Valeurs seuils pour le rapport coût-efficacité en soins de santé - KCE
Valeurs seuils pour le rapport coût-efficacité en soins de santé - KCE
Valeurs seuils pour le rapport coût-efficacité en soins de santé - KCE
Create successful ePaper yourself
Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.
<strong>Va<strong>le</strong>urs</strong> <strong>seuils</strong> <strong>pour</strong> <strong>le</strong> <strong>rapport</strong><br />
<strong>coût</strong>-<strong>efficacité</strong> <strong>en</strong> <strong>soins</strong><br />
<strong>de</strong> <strong>santé</strong><br />
<strong>KCE</strong> reports 100B<br />
Fe<strong>de</strong>raal K<strong>en</strong>nisc<strong>en</strong>trum voor <strong>de</strong> Gezondheidszorg<br />
C<strong>en</strong>tre fédéral d’expertise <strong>de</strong>s <strong>soins</strong> <strong>de</strong> <strong>santé</strong><br />
2008
Le C<strong>en</strong>tre fédéral d’expertise <strong>de</strong>s <strong>soins</strong> <strong>de</strong> <strong>santé</strong><br />
Prés<strong>en</strong>tation : Le C<strong>en</strong>tre fédéral d’expertise <strong>de</strong>s <strong>soins</strong> <strong>de</strong> <strong>santé</strong> est un parastatal, créé <strong>le</strong> 24<br />
décembre 2002 par la loi-programme (artic<strong>le</strong>s 262 à 266), sous tutel<strong>le</strong> du<br />
Ministre <strong>de</strong> la Santé publique et <strong>de</strong>s Affaires socia<strong>le</strong>s, qui est chargé <strong>de</strong> réaliser<br />
<strong>de</strong>s étu<strong>de</strong>s éclairant la décision politique dans <strong>le</strong> domaine <strong>de</strong>s <strong>soins</strong> <strong>de</strong> <strong>santé</strong> et<br />
<strong>de</strong> l’assurance maladie.<br />
Conseil d’administration<br />
Membres effectifs : Gil<strong>le</strong>t Pierre (Présid<strong>en</strong>t), Cuypers Dirk (Vice-Présid<strong>en</strong>t), Avontroodt Yolan<strong>de</strong>,<br />
De Cock Jo (Vice-Présid<strong>en</strong>t), De Meyere Frank, De Rid<strong>de</strong>r H<strong>en</strong>ri, Gil<strong>le</strong>t Jean-<br />
Bernard, Godin Jean-Noël, Goy<strong>en</strong>s Floris, Maes Jef, Mert<strong>en</strong>s Pascal, Mert<strong>en</strong>s<br />
Raf, Mo<strong>en</strong>s Marc, Perl François, Van Mass<strong>en</strong>hove Frank, Van<strong>de</strong>rmeer<strong>en</strong><br />
Philippe, Verertbrugg<strong>en</strong> Patrick, Vermey<strong>en</strong> Karel.<br />
Membres suppléants : Annemans Liev<strong>en</strong>, Bertels Jan, Collin B<strong>en</strong>oît, Cuypers Rita, Decoster<br />
Christiaan, Dercq Jean-Paul, Désir Daniel, Laasman Jean-Marc, Lemye Roland,<br />
Morel Amanda, Palsterman Paul, Ponce Annick, Remac<strong>le</strong> Anne, Schroot<strong>en</strong><br />
R<strong>en</strong>aat, Van<strong>de</strong>rstapp<strong>en</strong> Anne..<br />
Commissaire du gouvernem<strong>en</strong>t : Roger Yves<br />
Direction<br />
Directeur général a.i. : Jean-Pierre Closon<br />
Directeur général adjoint a.i. : Gert Peeters<br />
Contact<br />
C<strong>en</strong>tre fédéral d’expertise <strong>de</strong>s <strong>soins</strong> <strong>de</strong> <strong>santé</strong> (<strong>KCE</strong>).<br />
C<strong>en</strong>tre Administrative Doorbuilding<br />
Av<strong>en</strong>ue Jardin Botanique 55<br />
B-1000 Bruxel<strong>le</strong>s<br />
Belgium<br />
Tel: +32 [0]2 287 33 88<br />
Fax: +32 [0]2 287 33 85<br />
Email : info@kce.fgov.be<br />
Web : http://www.kce.fgov.be
<strong>Va<strong>le</strong>urs</strong> <strong>seuils</strong> <strong>pour</strong> <strong>le</strong> <strong>rapport</strong><br />
<strong>coût</strong>-<strong>efficacité</strong> <strong>en</strong> <strong>soins</strong> <strong>de</strong><br />
<strong>santé</strong><br />
<strong>KCE</strong> reports 100B<br />
IRINA CLEEMPUT, MATTIAS NEYT, NANCY THIRY,<br />
CHRIS DE LAET, MARK LEYS<br />
Fe<strong>de</strong>raal K<strong>en</strong>nisc<strong>en</strong>trum voor <strong>de</strong> Gezondheidszorg<br />
C<strong>en</strong>tre fédéral d’expertise <strong>de</strong>s <strong>soins</strong> <strong>de</strong> <strong>santé</strong><br />
2008
<strong>KCE</strong> REPORTS 100B<br />
Titre : <strong>Va<strong>le</strong>urs</strong> <strong>seuils</strong> <strong>pour</strong> <strong>le</strong> <strong>rapport</strong> <strong>coût</strong>-<strong>efficacité</strong> <strong>en</strong> <strong>soins</strong> <strong>de</strong> <strong>santé</strong><br />
Auteurs : Irina C<strong>le</strong>emput, Mattias Neyt, Nancy Thiry, Chris De Laet, Mark Leys<br />
Experts Externes : Liev<strong>en</strong> Annemans (UG<strong>en</strong>t, VUB), Jean-Luc Frère (SPF Budget et Contrô<strong>le</strong><br />
<strong>de</strong> Gestion/FOD Budget <strong>en</strong> Beheerscontro<strong>le</strong>), Marc Koopmanschap<br />
(Erasmus Medisch C<strong>en</strong>trum, Rotterdam, Ne<strong>de</strong>rland), Chantal Neirynck<br />
(Union nationa<strong>le</strong> <strong>de</strong>s mutualités libres), Catherine Van <strong>de</strong>r Auwera (SPF<br />
Budget et Contrô<strong>le</strong> <strong>de</strong> Gestion/FOD Budget <strong>en</strong> Beheerscontro<strong>le</strong>)<br />
Validateurs : Louis Niess<strong>en</strong> (Johns Hopkins School of Public Health, Maryland, USA),<br />
James Raftery (NCCHTA, Southampton, UK), Philippe Van Wil<strong>de</strong>r<br />
(RIZIV/INAMI/NIHDI, Brussels)<br />
Conflit d’intérêt : None <strong>de</strong>clared<br />
Disclaimer: Les experts externes ont collaboré au <strong>rapport</strong> sci<strong>en</strong>tifique qui a <strong>en</strong>suite<br />
été soumis aux validateurs. La validation du <strong>rapport</strong> résulte d’un<br />
cons<strong>en</strong>sus ou d’un vote majoritaire <strong>en</strong>tre <strong>le</strong>s validateurs. Le <strong>KCE</strong> reste<br />
seul responsab<strong>le</strong> <strong>de</strong>s erreurs ou omissions qui <strong>pour</strong>rai<strong>en</strong>t subsister <strong>de</strong><br />
même que <strong>de</strong>s recommandations faites aux autorités publiques.<br />
Layout : Wim Van Moer, Verhulst Ine<br />
Bruxel<strong>le</strong>s, 14 janvier 2009<br />
Etu<strong>de</strong> nr 2008-38<br />
Domaine: Health Technology Assessm<strong>en</strong>t (HTA)<br />
MeSH: Cost-B<strong>en</strong>efit Analysis ; Decision Making ; Health Care Rationing / economics ; Quality-<br />
Adjusted Life Years ; Health Care Costs<br />
NLM classification : WA 525<br />
Langage : français, anglais<br />
Format : Adobe® PDF (A4)<br />
Dépot légal: D/2008/10.273/95<br />
Comm<strong>en</strong>t citer ce docum<strong>en</strong>t?<br />
C<strong>le</strong>emput I, Neyt M, Thiry N, De Laet C, Leys M. <strong>Va<strong>le</strong>urs</strong> <strong>seuils</strong> <strong>pour</strong> <strong>le</strong> <strong>rapport</strong> <strong>coût</strong>-<strong>efficacité</strong> <strong>en</strong><br />
<strong>soins</strong> <strong>de</strong> <strong>santé</strong>. Health Technology Assessm<strong>en</strong>t (HTA). Bruxel<strong>le</strong>s: C<strong>en</strong>tre fédéral d'expertise <strong>de</strong>s <strong>soins</strong><br />
<strong>de</strong> <strong>santé</strong> (<strong>KCE</strong>); 2008. <strong>KCE</strong> reports 100B (D/2008/10.273/95)
<strong>KCE</strong> Reports 100B Utilisations <strong>de</strong>s va<strong>le</strong>urs <strong>seuils</strong> i<br />
AVANT-PROPOS<br />
Le <strong>KCE</strong> (C<strong>en</strong>tre Fédéral d'Expertise <strong>de</strong>s Soins <strong>de</strong> Santé) <strong>en</strong> est actuel<strong>le</strong>m<strong>en</strong>t à sa<br />
cinquième année d'élaboration <strong>de</strong> <strong>rapport</strong>s d'évaluation "Health Technology<br />
Assessm<strong>en</strong>ts" (HTA). Ces <strong>rapport</strong>s HTA recè<strong>le</strong>nt souv<strong>en</strong>t une évaluation économique.<br />
Dans <strong>le</strong>s cas où une interv<strong>en</strong>tion débouche sur <strong>de</strong>s résultats plus favorab<strong>le</strong>s <strong>pour</strong> la<br />
<strong>santé</strong> mais à un <strong>coût</strong> supérieur <strong>en</strong> comparaison avec la meil<strong>le</strong>ure interv<strong>en</strong>tion<br />
alternative visant la même affection, un <strong>rapport</strong> <strong>coût</strong>-<strong>efficacité</strong> différ<strong>en</strong>tiel (Increm<strong>en</strong>tal<br />
Cost-Effectiv<strong>en</strong>ess Ratio, ICER) est calculé. Cet ICER reflète <strong>le</strong> sur<strong>coût</strong> par unité<br />
supplém<strong>en</strong>taire d'effet sur la <strong>santé</strong>.<br />
Toutefois, la manière dont <strong>le</strong>s déci<strong>de</strong>urs belges trait<strong>en</strong>t ce type d'information reste<br />
floue. Le prés<strong>en</strong>t <strong>rapport</strong> constitue une introduction au contexte méthodologique et à<br />
la pertin<strong>en</strong>ce <strong>de</strong>s ICERs et <strong>de</strong>s va<strong>le</strong>urs <strong>seuils</strong> <strong>de</strong> l’ ICER ainsi qu'aux débats qu'ils<br />
<strong>en</strong>g<strong>en</strong>dr<strong>en</strong>t. Le <strong>rapport</strong> estime éga<strong>le</strong>m<strong>en</strong>t <strong>le</strong> niveau actuel <strong>de</strong>s connaissances et <strong>de</strong><br />
l'utilisation <strong>de</strong>s ICERs par <strong>le</strong>s déci<strong>de</strong>urs belges <strong>en</strong> <strong>soins</strong> <strong>de</strong> <strong>santé</strong>. Nous t<strong>en</strong>ons à saluer<br />
et à exprimer notre appréciation <strong>pour</strong> la collaboration et <strong>le</strong>s contributions <strong>de</strong>s<br />
membres <strong>de</strong> la Commission <strong>de</strong> Remboursem<strong>en</strong>t <strong>de</strong>s Médicam<strong>en</strong>ts (CRM) et du Conseil<br />
Technique <strong>de</strong>s Implants (CTI).<br />
Le prés<strong>en</strong>t docum<strong>en</strong>t est un <strong>rapport</strong> méthodologique comparab<strong>le</strong> aux<br />
recommandations <strong>pour</strong> <strong>le</strong>s évaluations pharmacoéconomiques <strong>en</strong> Belgique (<strong>KCE</strong> Report<br />
78C). Depuis <strong>le</strong>ur publication <strong>en</strong> avril 2008, l'INAMI incite <strong>le</strong>s <strong>en</strong>treprises<br />
pharmaceutiques à suivre ces recommandations lors <strong>de</strong> la soumission d'une évaluation<br />
pharmacoéconomique dans <strong>le</strong> cadre d'une <strong>de</strong>man<strong>de</strong> <strong>de</strong> remboursem<strong>en</strong>t <strong>pour</strong> un<br />
médicam<strong>en</strong>t. Avec <strong>le</strong> prés<strong>en</strong>t <strong>rapport</strong>, nous espérons, d'une part, informer <strong>le</strong>s déci<strong>de</strong>urs<br />
ainsi que <strong>le</strong>s autres parties pr<strong>en</strong>antes à propos <strong>de</strong> l'utilisation et <strong>de</strong> la pertin<strong>en</strong>ce <strong>de</strong>s<br />
évaluations économiques aux fins <strong>de</strong>s décisions <strong>en</strong> matière <strong>de</strong> politique <strong>de</strong> <strong>soins</strong> <strong>de</strong><br />
<strong>santé</strong>. D'autre part, nous souhaitons améliorer la compréh<strong>en</strong>sion <strong>de</strong> concepts<br />
économiques comp<strong>le</strong>xes et <strong>de</strong>s résultats d'étu<strong>de</strong>s.<br />
Gert Peeters Jean-Pierre Closon<br />
Directeur général adjoint a.i. Directeur général a.i.
ii Utilisations <strong>de</strong>s va<strong>le</strong>urs <strong>seuils</strong> <strong>KCE</strong> reports 100B<br />
CONTEXTE<br />
OBJECTIFS<br />
METHODES<br />
Sommaire<br />
Les évaluations <strong>de</strong>s technologies <strong>de</strong> <strong>santé</strong> (Health Technology Assessm<strong>en</strong>ts - HTA)<br />
comport<strong>en</strong>t souv<strong>en</strong>t une analyse <strong>coût</strong>-<strong>efficacité</strong> (ACE). L'objectif d'une ACE est<br />
d'informer <strong>le</strong>s déci<strong>de</strong>urs politiques à propos <strong>de</strong> la r<strong>en</strong>tabilité d'une interv<strong>en</strong>tion. La<br />
question <strong>de</strong> la r<strong>en</strong>tabilité se pose parce que <strong>de</strong>s choix sont inévitab<strong>le</strong>s lorsque <strong>le</strong>s<br />
ressources sont limitées. Toutefois, <strong>le</strong>s déci<strong>de</strong>urs politiques <strong>en</strong> <strong>soins</strong> <strong>de</strong> <strong>santé</strong> ont<br />
souv<strong>en</strong>t du mal à interpréter correctem<strong>en</strong>t <strong>le</strong>s résultats <strong>de</strong>s ACE et à savoir comm<strong>en</strong>t<br />
<strong>le</strong>s utiliser dans la prise <strong>de</strong> décisions. Pour cette raison, <strong>le</strong> Conseil d'Administration du<br />
C<strong>en</strong>tre Fédéral d'Expertise <strong>de</strong>s Soins <strong>de</strong> Santé (<strong>KCE</strong>) a <strong>de</strong>mandé une analyse<br />
approfondie <strong>de</strong>s concepts fondam<strong>en</strong>taux <strong>de</strong>s évaluations économiques <strong>en</strong> <strong>soins</strong> <strong>de</strong> <strong>santé</strong><br />
ainsi que <strong>de</strong> l'interprétation <strong>de</strong>s <strong>rapport</strong>s <strong>coût</strong>-<strong>efficacité</strong> dans un docum<strong>en</strong>t <strong>de</strong>stiné à un<br />
public non formé à l’économie <strong>de</strong> la <strong>santé</strong>.<br />
Les objectifs du <strong>rapport</strong> sont <strong>le</strong>s suivants :<br />
• Prés<strong>en</strong>ter <strong>le</strong>s concepts fondam<strong>en</strong>taux d'une évaluation économique: analyse<br />
<strong>coût</strong>-<strong>efficacité</strong> (ACE), <strong>rapport</strong> <strong>coût</strong>-<strong>efficacité</strong> différ<strong>en</strong>tiel (ICER) et va<strong>le</strong>urs<br />
<strong>seuils</strong> <strong>de</strong> l’ICER.<br />
• Expliquer <strong>le</strong>s fon<strong>de</strong>m<strong>en</strong>ts théoriques et la pertin<strong>en</strong>ce <strong>de</strong>s va<strong>le</strong>urs <strong>seuils</strong> <strong>de</strong><br />
l’ICER fréquemm<strong>en</strong>t <strong>rapport</strong>és <strong>en</strong> économie <strong>de</strong> la <strong>santé</strong>.<br />
• Décrire comm<strong>en</strong>t <strong>le</strong>s va<strong>le</strong>urs <strong>seuils</strong> <strong>de</strong> l’ICER sont utilisées et <strong>pour</strong>rai<strong>en</strong>t<br />
l'être dans <strong>le</strong>s processus décisionnels <strong>en</strong> <strong>soins</strong> <strong>de</strong> <strong>santé</strong>.<br />
• Faire un tour d'horizon <strong>de</strong> la manière dont d'autres pays utilis<strong>en</strong>t <strong>le</strong>s va<strong>le</strong>urs<br />
<strong>seuils</strong> <strong>de</strong> l’ICER.<br />
• Explorer la façon dont <strong>de</strong>ux <strong>en</strong>tités consultatives belges trait<strong>en</strong>t actuel<strong>le</strong>m<strong>en</strong>t<br />
<strong>le</strong>s problèmes <strong>de</strong> <strong>coût</strong>-<strong>efficacité</strong> dans <strong>le</strong>s processus décisionnels.<br />
Aucune connaissance préalab<strong>le</strong> <strong>en</strong> évaluation économique n'est requise <strong>pour</strong> lire <strong>le</strong><br />
prés<strong>en</strong>t <strong>rapport</strong>.<br />
Nous avons procédé à une revue sommaire <strong>de</strong> la littérature fondée sur une recherche<br />
incrém<strong>en</strong>ta<strong>le</strong> <strong>de</strong> la littérature reflétant <strong>le</strong>s différ<strong>en</strong>tes perspectives sur <strong>le</strong>s ICERs et <strong>le</strong>s<br />
va<strong>le</strong>urs <strong>seuils</strong> <strong>de</strong> l’ICER. La comparaison internationa<strong>le</strong> s'est fondée sur une revue qui a<br />
démarré par la liste <strong>de</strong>s recommandations pharmacoéconomiques publiées par l'ISPOR<br />
(International Society for Pharmacoeconomics ans Outcomes Research). Dans une<br />
étu<strong>de</strong> <strong>de</strong> terrain, nous avons réalisé <strong>de</strong>ux <strong>en</strong>treti<strong>en</strong>s col<strong>le</strong>ctifs : l'un avec <strong>de</strong>s membres<br />
du Bureau <strong>de</strong> la Commission <strong>de</strong> Remboursem<strong>en</strong>t <strong>de</strong>s Médicam<strong>en</strong>ts (CRM) et l'autre<br />
avec <strong>de</strong>s membres du Conseil Technique <strong>de</strong>s Implants (CTI).
<strong>KCE</strong> Reports 100B Utilisations <strong>de</strong>s va<strong>le</strong>urs <strong>seuils</strong> iii<br />
EVALUATION ÉCONOMIQUE EN SOINS DE SANTE<br />
L'évaluation économique <strong>de</strong>s interv<strong>en</strong>tions <strong>en</strong> <strong>santé</strong> consiste <strong>en</strong> une analyse<br />
comparative <strong>de</strong> diverses possibilités d'action <strong>en</strong> termes <strong>de</strong> <strong>coût</strong>s et <strong>de</strong> conséqu<strong>en</strong>ces.<br />
Les techniques d'évaluation économique <strong>le</strong>s plus fréquemm<strong>en</strong>t utilisées <strong>en</strong> <strong>soins</strong> <strong>de</strong><br />
<strong>santé</strong> sont <strong>le</strong>s analyses <strong>coût</strong>-<strong>efficacité</strong> et <strong>le</strong>s analyses <strong>coût</strong>-utilité. Les techniques<br />
diffèr<strong>en</strong>t au niveau <strong>de</strong> la manière dont sont exprimés <strong>le</strong>s résultats mais <strong>le</strong>ur finalité est la<br />
même, à savoir déterminer comm<strong>en</strong>t obt<strong>en</strong>ir <strong>le</strong>s meil<strong>le</strong>urs résultats <strong>pour</strong> la <strong>santé</strong> <strong>de</strong><br />
l’<strong>en</strong>semb<strong>le</strong> <strong>de</strong> la société avec un budget limité. Une mesure du résultat-<strong>santé</strong><br />
couramm<strong>en</strong>t utilisée dans <strong>le</strong>s analyses <strong>coût</strong>-<strong>efficacité</strong> est <strong>le</strong> LYG (Life-Years Gained),<br />
c’est à dire <strong>le</strong> nombre d'années <strong>de</strong> vie gagnées grâce à une interv<strong>en</strong>tion. Dans <strong>le</strong>s<br />
analyses <strong>coût</strong>-utilité, <strong>le</strong>s résultats-<strong>santé</strong> sont fréquemm<strong>en</strong>t exprimés <strong>en</strong> QALYs<br />
(Quality-Adjusted Life Years), une mesure où <strong>le</strong>s années <strong>de</strong> vies gagnées grâce à une<br />
interv<strong>en</strong>tion (LYG) sont pondérées par la qualité <strong>de</strong> vie, liée à la <strong>santé</strong>, durant ces<br />
années. Dans <strong>le</strong> prés<strong>en</strong>t <strong>rapport</strong>, nous utilisons ACE comme terme générique couvrant<br />
<strong>le</strong>s <strong>de</strong>ux techniques d'évaluation économique.<br />
LE RAPPORT COUT-EFFICACITE DIFFERENTIEL<br />
Le <strong>rapport</strong> <strong>coût</strong>-<strong>efficacité</strong> différ<strong>en</strong>tiel (ICER) est <strong>le</strong> <strong>rapport</strong> <strong>en</strong>tre la différ<strong>en</strong>ce estimée<br />
<strong>en</strong>tre <strong>le</strong> <strong>coût</strong> <strong>de</strong> <strong>de</strong>ux interv<strong>en</strong>tions et la différ<strong>en</strong>ce estimée <strong>de</strong> ces <strong>de</strong>ux interv<strong>en</strong>tions<br />
<strong>en</strong> termes <strong>de</strong> résultats. Le ICER représ<strong>en</strong>te <strong>le</strong> sur<strong>coût</strong> estimé par unité <strong>de</strong> <strong>santé</strong><br />
supplém<strong>en</strong>taire générée par cette interv<strong>en</strong>tion, comparé à son alternative la plus<br />
r<strong>en</strong>tab<strong>le</strong> <strong>pour</strong> la même affection. Le ICER est utilisé ess<strong>en</strong>tiel<strong>le</strong>m<strong>en</strong>t <strong>pour</strong> contribuer à<br />
une prise <strong>de</strong> décision éclairée lorsqu’il s’agit <strong>de</strong> choisir <strong>en</strong>tre <strong>de</strong>s interv<strong>en</strong>tions qui sont<br />
à la fois plus <strong>coût</strong>euses et d'une <strong>efficacité</strong> réel<strong>le</strong> supérieure par <strong>rapport</strong> à <strong>le</strong>ur<br />
comparateur.<br />
Il faut toutefois t<strong>en</strong>ir compte <strong>de</strong> plusieurs considérations méthodologiques relatives aux<br />
ICERs.<br />
• Il n'est pas toujours possib<strong>le</strong> <strong>de</strong> procé<strong>de</strong>r à une comparaison probante <strong>en</strong>tre<br />
<strong>le</strong>s ICERs <strong>de</strong> différ<strong>en</strong>tes interv<strong>en</strong>tions, a fortiori si <strong>le</strong>s unités <strong>de</strong> mesure <strong>de</strong>s<br />
résultats ne sont pas id<strong>en</strong>tiques (par exemp<strong>le</strong>, <strong>de</strong>s LYG versus QALY gagnés)<br />
ou si <strong>le</strong>s métho<strong>de</strong>s utilisées <strong>pour</strong> calcu<strong>le</strong>r <strong>le</strong>s ICERs sont différ<strong>en</strong>tes.<br />
• Les projections <strong>de</strong> <strong>coût</strong>s et <strong>de</strong> conséqu<strong>en</strong>ces incrém<strong>en</strong>tiels étant par<br />
définition incertaines, l’ICER l'est éga<strong>le</strong>m<strong>en</strong>t. Pour <strong>le</strong>s déci<strong>de</strong>urs, la mesure <strong>de</strong><br />
l’incertitu<strong>de</strong> liée à l’estimation <strong>de</strong> l’ICER constitue une information<br />
importante qu’ils <strong>de</strong>vrai<strong>en</strong>t pouvoir pr<strong>en</strong>dre <strong>en</strong> compte dans <strong>le</strong> processus <strong>de</strong><br />
décision.<br />
• En raison <strong>de</strong> l'exist<strong>en</strong>ce d'une préfér<strong>en</strong>ce temporel<strong>le</strong>, <strong>le</strong>s <strong>coût</strong>s et bénéfices<br />
futurs doiv<strong>en</strong>t être actualisés. En d'autres termes, <strong>le</strong>s <strong>coût</strong>s et conséqu<strong>en</strong>ces<br />
futurs doiv<strong>en</strong>t être réduits <strong>pour</strong> t<strong>en</strong>ir compte du fait que <strong>le</strong>s individus<br />
accord<strong>en</strong>t davantage d'importance aux effets sur la <strong>santé</strong> et à l’arg<strong>en</strong>t actuels,<br />
par <strong>rapport</strong> aux effets sur la <strong>santé</strong> et à l’arg<strong>en</strong>t futurs. Le choix du taux<br />
d’actualisation peut avoir un impact important sur l’ICER estimé. Le débat<br />
<strong>pour</strong> savoir si <strong>le</strong>s effets sur la <strong>santé</strong> doiv<strong>en</strong>t être actualisés (ou non) au même<br />
taux que <strong>le</strong>s <strong>coût</strong>s est toujours ouvert.<br />
• Généra<strong>le</strong>m<strong>en</strong>t, <strong>le</strong>s ICERs <strong>en</strong>glob<strong>en</strong>t <strong>de</strong>s <strong>coût</strong>s et <strong>de</strong>s résultats <strong>pour</strong> la <strong>santé</strong> à<br />
court terme et à long terme. Ceci signifie que la prise <strong>de</strong> décisions fondées<br />
sur <strong>le</strong>s ICERs a <strong>de</strong>s implications <strong>pour</strong> <strong>le</strong>s ressources utilisées et la <strong>santé</strong> tant<br />
prés<strong>en</strong>tes que futures. L'incertitu<strong>de</strong> à propos <strong>de</strong> l'av<strong>en</strong>ir r<strong>en</strong>d toujours<br />
quelque peu incertaine la justesse <strong>de</strong> la décision actuel<strong>le</strong>. Cette incertitu<strong>de</strong><br />
doit être pondérée <strong>de</strong> manière adéquate dans <strong>le</strong> processus décisionnel.
iv Utilisations <strong>de</strong>s va<strong>le</strong>urs <strong>seuils</strong> <strong>KCE</strong> reports 100B<br />
LA VALEUR SEUIL DE L’ICER<br />
En soi, l’ICER ne permet pas aux déci<strong>de</strong>urs politiques <strong>de</strong> tirer <strong>de</strong>s conclusions sur<br />
l’effici<strong>en</strong>ce d'une interv<strong>en</strong>tion. De tel<strong>le</strong>s conclusions exig<strong>en</strong>t une comparaison avec une<br />
va<strong>le</strong>ur <strong>de</strong> référ<strong>en</strong>ce <strong>pour</strong> l’ICER, au-<strong>de</strong>là <strong>de</strong> laquel<strong>le</strong> une interv<strong>en</strong>tion ne serait pas<br />
considérée comme r<strong>en</strong>tab<strong>le</strong> (parce que <strong>le</strong> sur<strong>coût</strong> <strong>pour</strong> une unité <strong>de</strong> <strong>santé</strong><br />
supplém<strong>en</strong>taire est perçu comme trop é<strong>le</strong>vé) et <strong>en</strong> <strong>de</strong>çà <strong>de</strong> laquel<strong>le</strong> l’interv<strong>en</strong>tion serait<br />
considérée comme r<strong>en</strong>tab<strong>le</strong>. Dans la théorie économique néoclassique « Welfariste », il<br />
est possib<strong>le</strong> <strong>de</strong> démontrer que, sous la contrainte d'un budget fixe, on peut fixer une<br />
va<strong>le</strong>ur seuil <strong>de</strong> l’ICER au-<strong>de</strong>ssus <strong>de</strong> laquel<strong>le</strong> <strong>le</strong>s interv<strong>en</strong>tions n'amélior<strong>en</strong>t pas l'effici<strong>en</strong>ce<br />
(i.e. une maximisation <strong>de</strong> la <strong>santé</strong> tota<strong>le</strong> avec <strong>le</strong>s ressources disponib<strong>le</strong>s) et au-<strong>de</strong>ssous<br />
<strong>de</strong> laquel<strong>le</strong> el<strong>le</strong>s l’amélior<strong>en</strong>t. La va<strong>le</strong>ur seuil <strong>de</strong> l’ICER correspond à l’ICER <strong>de</strong> la<br />
<strong>de</strong>rnière interv<strong>en</strong>tion d'un classem<strong>en</strong>t (« <strong>le</strong>ague tab<strong>le</strong> ») qui resterait financée (dans sa<br />
totalité ou <strong>en</strong> partie) par un budget donné fixe. Le classem<strong>en</strong>t est ordonnancé <strong>de</strong><br />
manière tel<strong>le</strong> que la première interv<strong>en</strong>tion prés<strong>en</strong>te l’ICER <strong>le</strong> plus faib<strong>le</strong>, et la <strong>de</strong>rnière<br />
interv<strong>en</strong>tion du classem<strong>en</strong>t l’ICER <strong>le</strong> plus é<strong>le</strong>vé.<br />
La va<strong>le</strong>ur seuil <strong>de</strong> l’ICER ainsi obt<strong>en</strong>ue se fon<strong>de</strong> sur différ<strong>en</strong>ts postulats :<br />
• Le budget <strong>de</strong> <strong>soins</strong> <strong>de</strong> <strong>santé</strong> est fixe, autrem<strong>en</strong>t dit, il ne peut être dépassé.<br />
• Le seul but <strong>de</strong>s décisions <strong>en</strong> <strong>soins</strong> <strong>de</strong> <strong>santé</strong> est <strong>de</strong> maximaliser <strong>le</strong>s avantages<br />
<strong>pour</strong> la <strong>santé</strong> <strong>de</strong> la population, <strong>en</strong> termes <strong>de</strong> QALYs ou <strong>de</strong> LYGs.<br />
• On dispose d'une information complète sur <strong>le</strong>s ICERs <strong>de</strong> toutes <strong>le</strong>s<br />
interv<strong>en</strong>tions.<br />
• Les programmes sont parfaitem<strong>en</strong>t divisib<strong>le</strong>s, autrem<strong>en</strong>t dit, ils peuv<strong>en</strong>t être<br />
réduits jusqu'à chaque niveau souhaité.<br />
• Les programmes propos<strong>en</strong>t <strong>de</strong>s r<strong>en</strong><strong>de</strong>m<strong>en</strong>ts d'échel<strong>le</strong> constants, ce qui<br />
signifie que tout élargissem<strong>en</strong>t d'un programme génère une élévation<br />
proportionnel<strong>le</strong> id<strong>en</strong>tique <strong>de</strong>s <strong>coût</strong>s et <strong>de</strong>s effets. En d'autres termes, réduire<br />
ou élargir un programme (dans la même population cib<strong>le</strong>) n'a aucune<br />
influ<strong>en</strong>ce sur l’ICER.<br />
• Les programmes <strong>de</strong> <strong>santé</strong> sont indép<strong>en</strong>dants <strong>le</strong>s uns <strong>de</strong>s autres, <strong>de</strong> sorte<br />
qu'un changem<strong>en</strong>t dans un programme n'a pas d'impact sur <strong>le</strong>s autres.<br />
Cette va<strong>le</strong>ur seuil <strong>de</strong> l’ICER est <strong>le</strong> résultat d'un modè<strong>le</strong> <strong>de</strong> maximisation <strong>de</strong> la <strong>santé</strong> qui<br />
s'applique à un contexte spécifique (budget, modè<strong>le</strong> d'organisation <strong>de</strong>s <strong>soins</strong> <strong>de</strong> <strong>santé</strong>,<br />
type d’assurance maladie, …), à un mom<strong>en</strong>t précis dans <strong>le</strong> temps et à <strong>de</strong>s conditions<br />
spécifiques. En conséqu<strong>en</strong>ce, la va<strong>le</strong>ur seuil <strong>de</strong> l’ICER n'est pas statique, mais évolue au<br />
fil du temps, puisqu'el<strong>le</strong> est tributaire <strong>de</strong>s modifications budgétaires, <strong>de</strong>s interv<strong>en</strong>tions<br />
financées et <strong>de</strong> la productivité <strong>de</strong>s <strong>soins</strong> <strong>de</strong> <strong>santé</strong>. Un contexte à budget fixe exige une<br />
va<strong>le</strong>ur seuil <strong>de</strong> l’ICER variab<strong>le</strong>. Par contre, une va<strong>le</strong>ur seuil <strong>de</strong> l’ICER fixe exigerait un<br />
budget <strong>de</strong> <strong>soins</strong> <strong>de</strong> <strong>santé</strong> f<strong>le</strong>xib<strong>le</strong>.<br />
DE LA THÉORIE A LA PRATIQUE<br />
Les postulats théoriques <strong>de</strong> l'approche va<strong>le</strong>ur seuil <strong>de</strong> l’ICER sont irréalistes. Primo, la<br />
va<strong>le</strong>ur seuil <strong>de</strong> l’ICER théorique ne peut pas être id<strong>en</strong>tifiée dans la pratique <strong>en</strong> raison<br />
d'un manque d'informations. Secundo, même si la va<strong>le</strong>ur seuil <strong>de</strong> l’ICER pouvait être<br />
id<strong>en</strong>tifiée, el<strong>le</strong> ne <strong>pour</strong>rait pas être appliquée parce que certaines conditions théoriques<br />
ne sont pas remplies : <strong>le</strong>s budgets ne sont pas nécessairem<strong>en</strong>t fixes, la maximisation <strong>de</strong><br />
la <strong>santé</strong> n'est pas <strong>le</strong> seul souci <strong>de</strong>s déci<strong>de</strong>urs politiques <strong>en</strong> matière <strong>de</strong> <strong>santé</strong>, <strong>de</strong>s<br />
problèmes d'équité se pos<strong>en</strong>t toujours lorsque <strong>de</strong>s ressources doiv<strong>en</strong>t être affectées,<br />
<strong>le</strong>s programmes <strong>de</strong> <strong>santé</strong> peuv<strong>en</strong>t ne pas produire <strong>de</strong> r<strong>en</strong><strong>de</strong>m<strong>en</strong>ts d'échel<strong>le</strong> constants et<br />
<strong>le</strong>s programmes peuv<strong>en</strong>t ne pas être parfaitem<strong>en</strong>t divisib<strong>le</strong>s. Certaines <strong>de</strong> ces conditions<br />
peuv<strong>en</strong>t être considérées plus importantes que d'autres.<br />
Dans un système mixte public privé, dans <strong>le</strong>quel <strong>le</strong>s pati<strong>en</strong>ts pai<strong>en</strong>t une quote-part,<br />
l'application <strong>de</strong> l'approche va<strong>le</strong>ur seuil <strong>de</strong> l’ICER n'est pas exempte <strong>de</strong> jugem<strong>en</strong>ts <strong>de</strong><br />
va<strong>le</strong>ur: la va<strong>le</strong>ur seuil <strong>de</strong> l’ICER propose <strong>le</strong>s interv<strong>en</strong>tions qui va<strong>le</strong>nt la peine d'être mises<br />
<strong>en</strong> œuvre, mais n'ai<strong>de</strong> pas à déterminer <strong>le</strong> niveau <strong>de</strong> remboursem<strong>en</strong>t optimal
<strong>KCE</strong> Reports 100B Utilisations <strong>de</strong>s va<strong>le</strong>urs <strong>seuils</strong> v<br />
(à savoir la part du <strong>coût</strong> total qui sera financée par <strong>le</strong> budget <strong>de</strong> <strong>soins</strong> <strong>de</strong> <strong>santé</strong> du<br />
gouvernem<strong>en</strong>t). L’utilisation malavisée d’une seu<strong>le</strong> va<strong>le</strong>ur seuil <strong>de</strong> l’ICER risque dès lors<br />
d’<strong>en</strong>traîner un niveau inopportun <strong>de</strong> dép<strong>en</strong>ses personnel<strong>le</strong>s <strong>pour</strong> certaines populations<br />
<strong>de</strong> pati<strong>en</strong>ts et ainsi occasionner une diminution <strong>de</strong> l’accessibilité financière aux <strong>soins</strong> <strong>de</strong><br />
<strong>santé</strong>.<br />
L'utilisation d'une va<strong>le</strong>ur seuil <strong>de</strong> l’ICER explicite peut induire <strong>de</strong>s manipulations dans <strong>le</strong>s<br />
évaluations économiques prés<strong>en</strong>tées visant à rester <strong>en</strong> <strong>de</strong>ssous du seuil <strong>de</strong> façon à<br />
obt<strong>en</strong>ir <strong>le</strong> remboursem<strong>en</strong>t d’un produit. Cela sera notamm<strong>en</strong>t <strong>le</strong> cas si aucune<br />
méthodologie standard n'est recommandée aux fins <strong>de</strong> l'évaluation économique. En tout<br />
état <strong>de</strong> cause, même avec une méthodologie standard, la manipulation reste possib<strong>le</strong><br />
dans une certaine mesure (par exemp<strong>le</strong>, <strong>en</strong> procédant à une utilisation sé<strong>le</strong>ctive <strong>de</strong>s<br />
données d'<strong>en</strong>trées dans un modè<strong>le</strong> économique ou <strong>en</strong> "optimisant" <strong>le</strong> prix d'une<br />
interv<strong>en</strong>tion). De surcroît, la va<strong>le</strong>ur seuil <strong>de</strong> l’ICER risque <strong>de</strong> <strong>de</strong>v<strong>en</strong>ir une légitimation <strong>en</strong><br />
soi, alors que, dans <strong>le</strong>urs décisions, <strong>le</strong>s déci<strong>de</strong>urs politiques désir<strong>en</strong>t peut-être, dans<br />
certains cas, accor<strong>de</strong>r une pondération plus importante à d'autres critères qu'à la<br />
maximisation <strong>de</strong> la <strong>santé</strong>.<br />
SOLUTIONS DE RECHANGE AUX ICERS ET A LA VALEUR SEUIL DE<br />
L’ICER<br />
Dans la littérature, <strong>de</strong>s solutions <strong>de</strong> rechange à la va<strong>le</strong>ur seuil <strong>de</strong> l’ICER définie selon la<br />
théorie économique néoclassique « welfariste » ont été proposées. Cel<strong>le</strong>s-ci diffèr<strong>en</strong>t<br />
au niveau <strong>de</strong> l'importance du poids qu'el<strong>le</strong>s accord<strong>en</strong>t à la notion <strong>de</strong> va<strong>le</strong>ur seuil <strong>de</strong><br />
l’ICER <strong>en</strong> tant que principe directeur dans l'affectation <strong>de</strong>s ressources.<br />
• Plutôt que <strong>de</strong> définir la va<strong>le</strong>ur seuil <strong>de</strong> l’ICER comme <strong>le</strong> <strong>coût</strong> par unité <strong>de</strong><br />
<strong>santé</strong> <strong>de</strong> l'interv<strong>en</strong>tion la moins r<strong>en</strong>tab<strong>le</strong> qui reste financée, la va<strong>le</strong>ur seuil <strong>de</strong><br />
l’ICER <strong>pour</strong>rait être définie comme la disposition à payer (Willingness To<br />
Pay, WTP) <strong>de</strong> la société <strong>pour</strong> un QALY (ou une LYG). Définir la va<strong>le</strong>ur seuil<br />
<strong>de</strong> l’ICER <strong>de</strong> cette manière exige un budget f<strong>le</strong>xib<strong>le</strong>, puisque cela signifie <strong>en</strong><br />
principe que toute interv<strong>en</strong>tion dont l’ICER est inférieur à la disposition à<br />
payer (WTP) sociéta<strong>le</strong> <strong>pour</strong> un QALY <strong>de</strong>vrait être financée. L'approche <strong>de</strong> la<br />
WTP sociéta<strong>le</strong> est attrayante, car el<strong>le</strong> se fon<strong>de</strong> explicitem<strong>en</strong>t sur la va<strong>le</strong>ur<br />
sociéta<strong>le</strong> <strong>de</strong> la <strong>santé</strong>. Toutefois, il semb<strong>le</strong> diffici<strong>le</strong>, voire impossib<strong>le</strong>, <strong>de</strong><br />
quantifier cette WTP sociéta<strong>le</strong> <strong>pour</strong> un QALY (ou LYG) générique. Il est <strong>en</strong><br />
effet malaisé d'imaginer la va<strong>le</strong>ur d'un QALY (ou d'une LYG) <strong>en</strong> <strong>de</strong>hors <strong>de</strong><br />
tout contexte concret. La solution alternative <strong>de</strong> redéfinir la WTP sociéta<strong>le</strong><br />
<strong>pour</strong> un QALY donné au cas par cas contourne <strong>le</strong> problème <strong>de</strong> quantification<br />
<strong>de</strong> la WTP sociéta<strong>le</strong> <strong>pour</strong> un QALY générique et indép<strong>en</strong>dant d'un contexte,<br />
mais risque <strong>de</strong> déboucher sur <strong>de</strong>s exig<strong>en</strong>ces budgétaires int<strong>en</strong>ab<strong>le</strong>s. Pr<strong>en</strong>dre<br />
<strong>en</strong> compte <strong>le</strong>s décisions prises par <strong>le</strong> passé constitue une autre option <strong>pour</strong><br />
id<strong>en</strong>tifier la disposition sociéta<strong>le</strong> à payer <strong>pour</strong> un QALY. Cela étant, <strong>de</strong>s<br />
va<strong>le</strong>urs <strong>seuils</strong> <strong>de</strong> l’ICER empiriques ou une fourchette <strong>de</strong> va<strong>le</strong>urs <strong>seuils</strong> <strong>de</strong><br />
l’ICER observées dans <strong>le</strong>s décisions passées <strong>de</strong>vrai<strong>en</strong>t toujours être<br />
interprétées dans <strong>le</strong>ur contexte budgétaire, sociétal et politique. Les<br />
décisions étant rarem<strong>en</strong>t – voire jamais – inspirées uniquem<strong>en</strong>t par <strong>de</strong>s<br />
considérations économiques, <strong>le</strong>s décisions prises par <strong>le</strong> passé ne fourniront<br />
jamais une véritab<strong>le</strong> estimation <strong>de</strong> la disposition sociéta<strong>le</strong> à payer <strong>pour</strong> un<br />
QALY (ou une LYG), mais <strong>pour</strong>rai<strong>en</strong>t à tout <strong>le</strong> moins représ<strong>en</strong>ter une<br />
fourchette <strong>de</strong> va<strong>le</strong>urs qui <strong>pour</strong>rai<strong>en</strong>t être acceptab<strong>le</strong>s.<br />
• Dans <strong>le</strong> cadre du processus décisionnel, l’ICER <strong>pour</strong>rait être pondéré<br />
implicitem<strong>en</strong>t ou explicitem<strong>en</strong>t par <strong>rapport</strong> à d'autres élém<strong>en</strong>ts. Une tel<strong>le</strong><br />
approche exige une mesure et/ou une objectivisation <strong>de</strong> chacun <strong>de</strong>s élém<strong>en</strong>ts<br />
jugés pertin<strong>en</strong>ts <strong>pour</strong> <strong>le</strong> processus <strong>de</strong> prise <strong>de</strong> décision. Dans ce contexte,<br />
comparer un ICER avec une va<strong>le</strong>ur seuil, <strong>pour</strong>rait constituer une manière<br />
d'inclure <strong>de</strong>s considérations relatives à l'effici<strong>en</strong>ce dans <strong>le</strong> processus<br />
décisionnel. Le poids réel <strong>de</strong> chaque élém<strong>en</strong>t <strong>de</strong> décision peut être r<strong>en</strong>du<br />
explicite ou rester implicite. Toutefois, cette <strong>de</strong>rnière option, à savoir rester<br />
dans l'implicite <strong>pour</strong> <strong>le</strong>s pondérations relatives, réduit la transpar<strong>en</strong>ce du<br />
processus décisionnel.
vi Utilisations <strong>de</strong>s va<strong>le</strong>urs <strong>seuils</strong> <strong>KCE</strong> reports 100B<br />
Une définition explicite <strong>de</strong>s pondérations réduit <strong>le</strong> risque <strong>de</strong> discussions à<br />
répétition sur l'importance relative <strong>de</strong> chacun <strong>de</strong>s critères <strong>de</strong> décision. Cela<br />
étant, il sera malaisé <strong>de</strong> déterminer <strong>le</strong>s pondérations réel<strong>le</strong>s. Un débat restera<br />
nécessaire, car chaque décision est affectée par <strong>de</strong>s circonstances<br />
particulières et <strong>de</strong>s conditions loca<strong>le</strong>s.<br />
• Plutôt que <strong>de</strong> définir une va<strong>le</strong>ur seuil <strong>de</strong> l’ICER , on <strong>pour</strong>rait <strong>en</strong> déterminer<br />
une <strong>pour</strong> <strong>le</strong> <strong>rapport</strong> <strong>coût</strong>-<strong>efficacité</strong> moy<strong>en</strong>. Le Produit Intérieur Brut (PIB)<br />
moy<strong>en</strong> par habitant <strong>pour</strong>rait représ<strong>en</strong>ter une tel<strong>le</strong> va<strong>le</strong>ur seuil, qui reflèterait<br />
la "juste part" <strong>de</strong>s citoy<strong>en</strong>s dans la richesse d'une nation. Le problème<br />
inhér<strong>en</strong>t à cette approche est qu'el<strong>le</strong> <strong>pour</strong>rait aboutir à une situation dans<br />
laquel<strong>le</strong> la totalité du PIB (voire davantage) <strong>de</strong>vrait être consacrée aux <strong>soins</strong><br />
<strong>de</strong> <strong>santé</strong>.<br />
• L'approche <strong>coût</strong> d'opportunité abandonne l'idée d'un ICER <strong>pour</strong> ori<strong>en</strong>ter <strong>le</strong>s<br />
décisions. Cette approche souti<strong>en</strong>t que <strong>le</strong>s bénéfices <strong>en</strong> <strong>santé</strong> perdus d'autres<br />
interv<strong>en</strong>tions qui <strong>de</strong>vront être abandonnées <strong>pour</strong> financer la nouvel<strong>le</strong><br />
<strong>de</strong>vrai<strong>en</strong>t être r<strong>en</strong>dus explicites et comparés directem<strong>en</strong>t avec <strong>le</strong>s bénéfices<br />
<strong>en</strong> <strong>santé</strong> <strong>de</strong> la nouvel<strong>le</strong> interv<strong>en</strong>tion. La mise <strong>en</strong> œuvre <strong>de</strong> cette approche au<br />
niveau national peut être compliquée, a fortiori si <strong>le</strong>s budgets ne sont pas<br />
strictem<strong>en</strong>t fixés, mais <strong>pour</strong>rait néanmoins être applicab<strong>le</strong> à un niveau local,<br />
notamm<strong>en</strong>t dans un établissem<strong>en</strong>t hospitalier particulier.<br />
• L'approche <strong>coût</strong>-conséqu<strong>en</strong>ce plai<strong>de</strong> <strong>en</strong> faveur d'une prés<strong>en</strong>tation désagrégée<br />
<strong>de</strong> tous <strong>le</strong>s élém<strong>en</strong>ts économiquem<strong>en</strong>t pertin<strong>en</strong>ts : <strong>le</strong>s inputs <strong>pour</strong> la<br />
modélisation ainsi que <strong>le</strong>s outputs qui sont <strong>en</strong>globés dans l’ICER estimé. Une<br />
tel<strong>le</strong> approche permet au déci<strong>de</strong>ur <strong>de</strong> pondérer <strong>de</strong> manière explicite <strong>le</strong>s<br />
élém<strong>en</strong>ts économiques par <strong>rapport</strong> à d'autres paramètres.<br />
Chacune <strong>de</strong> ces approches recherche <strong>de</strong>s façons <strong>de</strong> r<strong>en</strong>dre <strong>le</strong>s considérations<br />
économiques explicites dans <strong>le</strong> processus décisionnel <strong>en</strong> <strong>soins</strong> <strong>de</strong> <strong>santé</strong>. Plutôt que <strong>de</strong><br />
s'exclure mutuel<strong>le</strong>m<strong>en</strong>t, ces métho<strong>de</strong>s peuv<strong>en</strong>t être considérées comme<br />
complém<strong>en</strong>taires <strong>pour</strong> atteindre la transpar<strong>en</strong>ce dans la prise <strong>de</strong> décisions.<br />
LE RÔLE DES ÉVALUATIONS ÉCONOMIQUES<br />
DANS LE PROCESSUS DE PRISE DE DÉCISIONS EN<br />
SOINS DE SANTE<br />
La prise <strong>de</strong> décisions est un processus nettem<strong>en</strong>t plus comp<strong>le</strong>xe qu'une évaluation<br />
éclairée et rationnel<strong>le</strong> <strong>de</strong>s problèmes, une pondération <strong>de</strong>s solutions <strong>de</strong> rechange et la<br />
formulation <strong>de</strong> meil<strong>le</strong>ures solutions. La rationalité est par ess<strong>en</strong>ce limitée, <strong>le</strong>s décisions<br />
sont incrém<strong>en</strong>tiel<strong>le</strong>s et <strong>le</strong> processus décisionnel est politique par nature. Les décisions<br />
sont <strong>le</strong> plus souv<strong>en</strong>t prises non pas sur la base <strong>de</strong>s seu<strong>le</strong>s informations sci<strong>en</strong>tifiques ou<br />
techniques, mais <strong>en</strong> se fondant sur un mélange <strong>de</strong> sources d'information. Dans <strong>le</strong> mon<strong>de</strong><br />
réel <strong>de</strong> la prise <strong>de</strong> décisions, <strong>le</strong>s évaluations économiques ne suffis<strong>en</strong>t pas, à el<strong>le</strong>s seu<strong>le</strong>s,<br />
à informer <strong>le</strong>s déci<strong>de</strong>urs politiques. L'<strong>efficacité</strong> réel<strong>le</strong> et la r<strong>en</strong>tabilité ne sont que <strong>de</strong>ux<br />
considérations parmi <strong>de</strong> nombreuses autres <strong>pour</strong> poser <strong>de</strong>s choix politiques. Ni la<br />
théorie ni <strong>le</strong>s preuves empiriques ne souti<strong>en</strong>n<strong>en</strong>t l'att<strong>en</strong>te selon laquel<strong>le</strong> <strong>le</strong>s va<strong>le</strong>urs<br />
<strong>seuils</strong> <strong>de</strong> l’ICER vont évoluer <strong>en</strong> tant que critère <strong>de</strong> décision unique. Les résultats d'une<br />
ACE peuv<strong>en</strong>t être utilisés <strong>en</strong> tant qu'input dans un processus décisionnel délibératif<br />
fondé sur <strong>le</strong>s preuves qui pr<strong>en</strong>d <strong>en</strong> compte <strong>le</strong>s points <strong>de</strong> vue et <strong>le</strong>s va<strong>le</strong>urs <strong>de</strong> multip<strong>le</strong>s<br />
parties pr<strong>en</strong>antes.<br />
En tout état <strong>de</strong> cause, <strong>de</strong>s efforts sont déployés <strong>pour</strong> "rationaliser" la prise <strong>de</strong> décision<br />
<strong>en</strong> <strong>soins</strong> <strong>de</strong> <strong>santé</strong> <strong>en</strong> proposant <strong>de</strong>s informations et <strong>de</strong>s connaissances rassemblées dans<br />
<strong>le</strong> respect d'une méthodologie soli<strong>de</strong>.<br />
Les preuves issues du savoir et <strong>de</strong>s recherches peuv<strong>en</strong>t être utilisées <strong>de</strong> trois façons<br />
différ<strong>en</strong>tes par <strong>le</strong>s déci<strong>de</strong>urs : directem<strong>en</strong>t, sé<strong>le</strong>ctivem<strong>en</strong>t ou comme source<br />
d'information. Le processus <strong>de</strong> décision éclairée se distingue toutefois par divers<br />
élém<strong>en</strong>ts inhibiteurs et facilitateurs.
<strong>KCE</strong> Reports 100B Utilisations <strong>de</strong>s va<strong>le</strong>urs <strong>seuils</strong> vii<br />
Les inhibiteurs à l'utilisation <strong>de</strong>s évaluations économiques dans la formulation <strong>de</strong>s<br />
politiques se résum<strong>en</strong>t à l'accessibilité <strong>de</strong>s preuves extraites <strong>de</strong>s recherches <strong>de</strong> même<br />
qu'à l'acceptabilité <strong>de</strong> cel<strong>le</strong>s-ci. La faculté <strong>de</strong> compr<strong>en</strong>dre <strong>le</strong>s analyses économiques, <strong>le</strong>s<br />
comportem<strong>en</strong>ts face aux évaluations économiques (y compris <strong>le</strong>s préoccupations<br />
relatives à la base <strong>de</strong> ces analyses et à <strong>le</strong>ur utilisation), <strong>le</strong> champ d'application <strong>de</strong>s<br />
questions <strong>de</strong> la recherche économique <strong>de</strong> même que la portée <strong>de</strong>s questions politiques<br />
<strong>en</strong>trav<strong>en</strong>t l'utilisation <strong>de</strong>s ACE dans <strong>le</strong> processus décisionnel.<br />
Un axe <strong>de</strong> la recherche se conc<strong>en</strong>tre sur <strong>le</strong> « courtage du savoir » dont quatre modè<strong>le</strong>s<br />
ont été id<strong>en</strong>tifiés : un modè<strong>le</strong> <strong>de</strong> type "information push", un modè<strong>le</strong> <strong>de</strong> type<br />
"information pull", un modè<strong>le</strong> d'échange et un modè<strong>le</strong> intégré.<br />
En raison <strong>de</strong>s difficultés pratiques liées à la prise <strong>de</strong> décisions, dans une perspective <strong>de</strong><br />
justice socia<strong>le</strong> et <strong>de</strong> démocratie, on <strong>pour</strong>rait à tout <strong>le</strong> moins s'att<strong>en</strong>dre à ce que <strong>le</strong><br />
processus décisionnel soit transpar<strong>en</strong>t et à une responsabilisation <strong>de</strong>s déci<strong>de</strong>urs par<br />
<strong>rapport</strong> à <strong>le</strong>urs décisions.<br />
L'UTILISATION DES VALEURS SEUILS DE L’ICER DANS D'AUTRES<br />
PAYS<br />
Aucun <strong>de</strong>s dix pays étudiés dans <strong>le</strong> prés<strong>en</strong>t <strong>rapport</strong> n'utilise une va<strong>le</strong>ur seuil <strong>de</strong> l’ICER<br />
unique. Actuel<strong>le</strong>m<strong>en</strong>t, <strong>le</strong> Royaume-Uni utilise une fourchette <strong>de</strong> <strong>seuils</strong> allant <strong>de</strong> 20 000 £<br />
à 30 000 £ par QALY gagné, mais <strong>le</strong> débat y est toujours <strong>en</strong> cours à propos <strong>de</strong><br />
l'utilisation et du niveau <strong>de</strong>s va<strong>le</strong>urs <strong>seuils</strong> <strong>de</strong> l’ICER .<br />
Certains pays ont essayé <strong>de</strong> dériver une va<strong>le</strong>ur seuil <strong>de</strong> l’ICER implicite <strong>en</strong> se basant sur<br />
<strong>de</strong>s décisions passées <strong>en</strong> matière d'affectation <strong>de</strong>s ressources. L'Australie a estimé une<br />
va<strong>le</strong>ur seuil <strong>de</strong> 69 900 AU$/QALY, la Nouvel<strong>le</strong>-Zélan<strong>de</strong> <strong>de</strong> 20 000 NZ$/QALY et <strong>le</strong><br />
Canada a défini une fourchette d'acceptation allant du <strong>coût</strong> d’une interv<strong>en</strong>tion<br />
dominante (moins chère et plus efficace) jusqu'à 80 000 CAN$/QALY, avec une<br />
fourchette <strong>de</strong> rejet allant <strong>de</strong> 31 000 à 137 000 CAN$/QALY.<br />
Des va<strong>le</strong>urs uniques ou fourchettes <strong>de</strong> <strong>seuils</strong> proposées par <strong>de</strong>s personnes ou <strong>de</strong>s<br />
institutions ont été trouvées aux États-Unis (50 000 $/QALY), aux Pays-Bas (80 000<br />
€/QALY) et au Canada (20 000 - 100 000 CAN$/QALY).<br />
Même dans l'unique pays (Royaume-Uni), possédant une fourchette <strong>de</strong> <strong>seuils</strong> explicite, la<br />
prise <strong>de</strong> décision ne se fon<strong>de</strong> pas uniquem<strong>en</strong>t sur <strong>de</strong>s considérations <strong>de</strong> <strong>coût</strong>-<strong>efficacité</strong>.<br />
L'équité est généra<strong>le</strong>m<strong>en</strong>t considérée comme un critère <strong>de</strong> décision subsidiaire. Il y a<br />
peu, dans <strong>le</strong> but d'attribuer <strong>de</strong>s "pondérations d'équité" aux QALYs gagnés dans <strong>le</strong>s<br />
ACE, on a procédé au Royaume-Uni à une quantification <strong>de</strong>s va<strong>le</strong>urs sociéta<strong>le</strong>s <strong>pour</strong> <strong>le</strong>s<br />
gains <strong>de</strong> <strong>santé</strong> <strong>en</strong> fonction <strong>de</strong>s populations bénéficiaires <strong>de</strong> ceux-ci.<br />
Face à <strong>de</strong>s ICERs é<strong>le</strong>vés, d'autres élém<strong>en</strong>ts d'appréciation peuv<strong>en</strong>t pr<strong>en</strong>dre une<br />
importance plus gran<strong>de</strong>. Dans la plupart <strong>de</strong>s pays, il ressort que <strong>le</strong>s interv<strong>en</strong>tions à ICER<br />
bas sont davantage susceptib<strong>le</strong>s d'être acceptées que <strong>le</strong>s interv<strong>en</strong>tions à ICER é<strong>le</strong>vé.<br />
Une analyse <strong>de</strong> cette relation <strong>pour</strong> la Belgique sortait du champ d'application <strong>de</strong> la<br />
prés<strong>en</strong>te étu<strong>de</strong>.<br />
L'UTILISATION DES VALEURS SEUILS DE L’ICER EN BELGIQUE<br />
En dépit <strong>de</strong>s efforts déployés <strong>pour</strong> rationaliser <strong>le</strong> processus décisionnel et étayer <strong>le</strong>s<br />
<strong>de</strong>man<strong>de</strong>s <strong>de</strong> remboursem<strong>en</strong>t avec <strong>de</strong>s preuves sci<strong>en</strong>tifiques, la prise <strong>de</strong> décision reste<br />
<strong>en</strong> Belgique un processus <strong>de</strong> délibération interactif. L'<strong>efficacité</strong> clinique réel<strong>le</strong> est <strong>le</strong><br />
critère sci<strong>en</strong>tifique <strong>le</strong> plus important utilisé tant par la Commission <strong>de</strong> Remboursem<strong>en</strong>t<br />
<strong>de</strong>s Médicam<strong>en</strong>ts que par <strong>le</strong> Conseil Technique <strong>de</strong>s Implants. Il arrive que <strong>le</strong> <strong>rapport</strong><br />
<strong>coût</strong>-<strong>efficacité</strong> soit pris <strong>en</strong> considération par la CRM mais plus rarem<strong>en</strong>t par <strong>le</strong> CTI. Les<br />
<strong>de</strong>ux <strong>en</strong>tités considèr<strong>en</strong>t que l'impact budgétaire est plus important que l’ICER.
viii Utilisations <strong>de</strong>s va<strong>le</strong>urs <strong>seuils</strong> <strong>KCE</strong> reports 100B<br />
CONCLUSION<br />
En tant que mesure d'évaluation <strong>de</strong> la capacité <strong>de</strong>s interv<strong>en</strong>tions à améliorer l’effici<strong>en</strong>ce<br />
pot<strong>en</strong>tiel<strong>le</strong> <strong>de</strong>s <strong>soins</strong> <strong>de</strong> <strong>santé</strong>, l’ICER prés<strong>en</strong>te <strong>de</strong>s faib<strong>le</strong>sses. La va<strong>le</strong>ur seuil <strong>de</strong> l’ICER<br />
par <strong>rapport</strong> à laquel<strong>le</strong> <strong>le</strong>s ICERs <strong>de</strong>s interv<strong>en</strong>tions <strong>de</strong>vrai<strong>en</strong>t être comparés est inconnue<br />
et soumise à <strong>de</strong>s variations temporel<strong>le</strong>s. Toutefois, ceci ne constitue pas un argum<strong>en</strong>t<br />
contre l'utilisation <strong>de</strong> considérations économiques dans la prise <strong>de</strong> décision <strong>en</strong> <strong>soins</strong> <strong>de</strong><br />
<strong>santé</strong>. Faire abstraction <strong>de</strong>s considérations économiques n'est pas éthique, car dép<strong>en</strong>ser<br />
<strong>de</strong>s ressources <strong>pour</strong> un programme <strong>de</strong> <strong>santé</strong> réduit l'<strong>en</strong>veloppe disponib<strong>le</strong> <strong>pour</strong> d'autres<br />
programmes <strong>de</strong> <strong>santé</strong>.<br />
RECOMMENDATIONS<br />
• L'effici<strong>en</strong>ce <strong>de</strong>vrait systématiquem<strong>en</strong>t faire partie <strong>de</strong>s critères <strong>de</strong> décision. La<br />
passer sous si<strong>le</strong>nce est contraire à l'éthique. Les dossiers prés<strong>en</strong>tés aux<br />
déci<strong>de</strong>urs <strong>de</strong>vrai<strong>en</strong>t donc systématiquem<strong>en</strong>t comporter <strong>de</strong>s analyses<br />
économiques.<br />
• Les modè<strong>le</strong>s économiques <strong>de</strong>vrai<strong>en</strong>t être <strong>rapport</strong>és <strong>de</strong> manière transpar<strong>en</strong>te.<br />
Toutes <strong>le</strong>s informations utilisées dans <strong>le</strong> modè<strong>le</strong> doiv<strong>en</strong>t être prés<strong>en</strong>tées <strong>de</strong><br />
manière tel<strong>le</strong> qu’el<strong>le</strong>s permett<strong>en</strong>t aux déci<strong>de</strong>urs <strong>de</strong> vérifier <strong>le</strong>s postulats,<br />
d’évaluer l’incertitu<strong>de</strong> et <strong>de</strong> pondérer l’importance <strong>de</strong> ces postulats et <strong>de</strong><br />
cette incertitu<strong>de</strong> <strong>pour</strong> la prise <strong>de</strong> décision.<br />
• Les résultats <strong>de</strong>s évaluations économiques <strong>de</strong>vrai<strong>en</strong>t être prés<strong>en</strong>tés sous<br />
forme désagrégée. Une prés<strong>en</strong>tation qui implique un éclatem<strong>en</strong>t <strong>de</strong> l’ICER,<br />
mais éga<strong>le</strong>m<strong>en</strong>t la prés<strong>en</strong>tation d'autres paramètres <strong>de</strong> résultats<br />
économiquem<strong>en</strong>t pertin<strong>en</strong>ts pouvant être dérivés <strong>de</strong> l'évaluation économique<br />
mais qui ne sont pas nécessairem<strong>en</strong>t visib<strong>le</strong>s dans l'estimation du ICER.<br />
• Outre la prés<strong>en</strong>tation désagrégée d'élém<strong>en</strong>ts revêtant une importance<br />
économique, <strong>le</strong>s ICERs <strong>de</strong>vrai<strong>en</strong>t continuer à être prés<strong>en</strong>tés et calculés selon<br />
<strong>de</strong>s directives méthodologiques normalisées.<br />
• Il convi<strong>en</strong>t <strong>de</strong> <strong>pour</strong>suivre l'utilisation <strong>de</strong> la recherche sci<strong>en</strong>tifique dans <strong>le</strong><br />
processus décisionnel d'affectation <strong>de</strong>s ressources <strong>en</strong> <strong>soins</strong> <strong>de</strong> <strong>santé</strong>. Cette<br />
approche permettra aux déci<strong>de</strong>urs politiques d'étayer <strong>le</strong>urs argum<strong>en</strong>taires.<br />
• Les déci<strong>de</strong>urs <strong>de</strong>vrai<strong>en</strong>t être plus transpar<strong>en</strong>ts <strong>en</strong> ce qui concerne <strong>le</strong>s critères<br />
<strong>de</strong> décision et l'importance relative <strong>de</strong> chacun <strong>de</strong> ces critères dans chaque<br />
décision.
<strong>KCE</strong> reports 100 ICER Thresholds 1<br />
Sci<strong>en</strong>tific summary<br />
Tab<strong>le</strong> of cont<strong>en</strong>ts<br />
GLOSSARY ................................................................................................................................. 3<br />
ABBREVIATIONS ...................................................................................................................... 5<br />
1 GENERAL INTRODUCTION ........................................................................................ 6<br />
1.1 BACKGROUND........................................................................................................................................... 6<br />
1.2 PROBLEM STATEMENT............................................................................................................................. 6<br />
1.3 SCOPE OF THIS REPORT.......................................................................................................................... 7<br />
1.4 OBJECTIVES................................................................................................................................................... 7<br />
1.5 METHODOLOGY........................................................................................................................................ 7<br />
1.6 STRUCTURE OF THIS REPORT .............................................................................................................. 8<br />
2 ICERS AND ICER THRESHOLD VALUES................................................................... 9<br />
2.1 GENERAL PRINCIPLES OF ECONOMIC EVALUATION IN HEALTH CARE ............................. 9<br />
2.2 THE INCREMENTAL COST-EFFECTIVENESS RATIO (ICER)........................................................10<br />
2.3 THE COST-EFFECTIVENESS PLANE ....................................................................................................11<br />
2.4 METHODOLOGICAL ISSUES OF THE ICER......................................................................................12<br />
2.4.1 Comparability of ICERs for differ<strong>en</strong>t interv<strong>en</strong>tions................................................................12<br />
2.4.2 Uncertainty around the ICER......................................................................................................13<br />
2.4.3 Wh<strong>en</strong> do we incur costs and wh<strong>en</strong> do we reap the b<strong>en</strong>efits? .............................................14<br />
2.5 THE ICER THRESHOLD VALUE IN A FIXED BUDGET SETTING...............................................15<br />
2.5.1 Basic assumptions...........................................................................................................................15<br />
2.5.2 Id<strong>en</strong>tifying the ICER threshold value..........................................................................................16<br />
2.5.3 Characteristics of the ICER threshold value in a fixed budget setting................................17<br />
2.5.4 Interpretation of the ICER threshold value in a fixed budget setting .................................17<br />
2.6 METHODOLOGICAL ISSUES OF THE ICER THRESHOLD VALUE ............................................18<br />
2.6.1 Uncertainty around the ICER and the ICER threshold value ...............................................18<br />
2.6.2 Comparison with an appropriate comparator.........................................................................19<br />
2.6.3 Measurem<strong>en</strong>t units in nominator and d<strong>en</strong>ominator ...............................................................20<br />
2.7 HOW WELL ARE THE THEORETICAL ASSUMPTIONS FOR THE ICER THRESHOLD<br />
VALUE FULFILLED IN REAL LIFE?.........................................................................................................20<br />
2.7.1 Fixed budget....................................................................................................................................21<br />
2.7.2 Comp<strong>le</strong>te information on costs and effects of all health interv<strong>en</strong>tions..............................23<br />
2.7.3 Perfect divisibility and constant returns to sca<strong>le</strong>.....................................................................24<br />
2.7.4 Health programmes are in<strong>de</strong>p<strong>en</strong>d<strong>en</strong>t from one another ......................................................25<br />
2.7.5 Health maximisation as the so<strong>le</strong> goal of health policy makers .............................................25<br />
2.7.6 Additional caveats ..........................................................................................................................27<br />
2.8 ALTERNATIVES TO ICERS AND ICER THRESHOLD VALUES ....................................................28<br />
2.8.1 The ICER threshold value as a ref<strong>le</strong>ction of societal willingness to pay.............................29<br />
2.8.2 Comparison with past <strong>de</strong>cisions .................................................................................................30<br />
2.8.3 Weighing the ICER against other <strong>de</strong>cision criteria in the <strong>de</strong>cision making process........31<br />
2.8.4 The average GDP per capita as a threshold value for the average cost-effectiv<strong>en</strong>ess ratio<br />
32<br />
2.8.5 The opportunity costs approach.................................................................................................33<br />
2.8.6 Cost-consequ<strong>en</strong>ces analysis.........................................................................................................34<br />
3 THE ROLE OF ECONOMIC EVALUATIONS IN HEALTH CARE DECISION<br />
MAKING..........................................................................................................................36<br />
3.1 DECISION MAKING PROCESSES .........................................................................................................36<br />
3.2 INFORMED POLICY DECISION MAKING.........................................................................................37
2 ICER Thresholds <strong>KCE</strong> Reports 100<br />
3.3 EMPIRICAL EVIDENCE ON THE USE OF ECONOMIC EVALUATIONS IN HEALTH CARE<br />
DECISION MAKING .................................................................................................................................39<br />
3.4 COST-EFFECTIVENESS ANALYSIS, ICER THRESHOLD VALUES AND DECISION MAKING<br />
........................................................................................................................................................................40<br />
3.5 THE USE OF ICER THRESHOLD VALUES IN OTHER COUNTRIES ..........................................43<br />
3.5.1 Methodology ...................................................................................................................................43<br />
3.5.2 England and Wa<strong>le</strong>s .........................................................................................................................43<br />
3.5.3 Canada..............................................................................................................................................44<br />
3.5.4 The Netherlands.............................................................................................................................45<br />
3.5.5 USA ...................................................................................................................................................46<br />
3.5.6 Australia ...........................................................................................................................................47<br />
3.5.7 New Zealand...................................................................................................................................47<br />
3.5.8 Finland...............................................................................................................................................48<br />
3.5.9 Swed<strong>en</strong> .............................................................................................................................................48<br />
3.5.10 Norway.............................................................................................................................................49<br />
3.5.11 D<strong>en</strong>mark ..........................................................................................................................................49<br />
3.6 THE USE OF ICER THRESHOLD VALUES IN BELGIUM.................................................................50<br />
3.6.1 Background on DRC and TCI .....................................................................................................50<br />
3.6.2 Aims and methods of the field study..........................................................................................51<br />
3.6.3 Results of the field study...............................................................................................................52<br />
4 GENERAL DISCUSSION.............................................................................................. 56<br />
4.1 ECONOMIC EVALUATION AND ICERS............................................................................................56<br />
4.2 WAYS TO INTRODUCE EFFICIENCY CONSIDERATIONS IN HEALTH CARE DECISION<br />
MAKING.......................................................................................................................................................56<br />
4.3 HEALTH CARE DECISION MAKING CONTEXTS ..........................................................................58<br />
4.4 SUGGESTIONS FOR FURTHER RESEARCH ......................................................................................58<br />
5 CONCLUSION .............................................................................................................. 60<br />
6 RECOMMENDATIONS ................................................................................................ 61<br />
7 REFERENCES................................................................................................................. 62
<strong>KCE</strong> reports 100 ICER Thresholds 3<br />
GLOSSARY<br />
Beveridge-type health care system Taxation-fun<strong>de</strong>d public health service system (named after<br />
William Beveridge, 20 th c<strong>en</strong>tury British economist and politician).<br />
Bismarck-type health care system Health care system where public and private provi<strong>de</strong>rs are<br />
reimbursed by compulsory health insurance funds (named after<br />
Otto von Bismarck, 19 th c<strong>en</strong>tury German chancellor).<br />
Confid<strong>en</strong>ce interval (CI) Statistical concept. Interval likely to inclu<strong>de</strong> the estimated<br />
parameter with a giv<strong>en</strong> confid<strong>en</strong>ce <strong>le</strong>vel, for examp<strong>le</strong> 95% CI.<br />
Results are pres<strong>en</strong>ted as a point estimate surroun<strong>de</strong>d by its<br />
confid<strong>en</strong>ce interval.<br />
Cost-b<strong>en</strong>efit analysis Type of economic evaluation in which all costs incurred and<br />
resulting b<strong>en</strong>efits of an interv<strong>en</strong>tion are expressed in monetary<br />
units (e.g. €) and a net monetary gain/loss or cost-b<strong>en</strong>efit ratio is<br />
computed.<br />
Cost-consequ<strong>en</strong>ce analysis A variant of cost-effectiv<strong>en</strong>ess analysis in which the compon<strong>en</strong>ts<br />
of increm<strong>en</strong>tal costs and consequ<strong>en</strong>ces (health outcomes) of<br />
alternative programmes are computed and listed, without<br />
Cost-effectiv<strong>en</strong>ess acceptability<br />
curve<br />
aggregation into a cost-effectiv<strong>en</strong>ess ratio or cost-utility ratio.<br />
Curve repres<strong>en</strong>ting the probability of an interv<strong>en</strong>tion being costeffective<br />
(Y-axis), giv<strong>en</strong> differ<strong>en</strong>t values for the ICER threshold<br />
value (X-axis). The curve ref<strong>le</strong>cts the uncertainty around the<br />
ICER estimate.<br />
Cost-effectiv<strong>en</strong>ess analysis Method of comparing alternative treatm<strong>en</strong>ts in which the costs<br />
and consequ<strong>en</strong>ces of the treatm<strong>en</strong>ts vary. The outcomes of<br />
alternative treatm<strong>en</strong>ts are measured in the same non-monetary<br />
(natural) unit (e.g. life years gained, ev<strong>en</strong>ts avoi<strong>de</strong>d, …).<br />
Cost-minimisation analysis Method of comparing the costs of alternative health<br />
interv<strong>en</strong>tions that are assumed to have an equiva<strong>le</strong>nt effect on<br />
health outcomes.<br />
Cost-utility analysis Special form of cost-effectiv<strong>en</strong>ess analysis in which the costs per<br />
unit of ‘utility’ are calculated. The term is also frequ<strong>en</strong>tly used<br />
for economic evaluations that take the impact of an interv<strong>en</strong>tion<br />
on health-related quality of life into account, irrespective of<br />
whether the outcome measure can be regar<strong>de</strong>d as a true utility<br />
measure in its theoretical economic s<strong>en</strong>se. The most commonly<br />
used outcome measure in cost-utility analyses is the qualityadjusted<br />
life year (QALY).<br />
Credibility interval Confid<strong>en</strong>ce interval around a cost-effectiv<strong>en</strong>ess ratio resulting<br />
from an economic mo<strong>de</strong>l. In contrast to statistical confid<strong>en</strong>ce<br />
intervals, the values within a credibility interval are not actually<br />
observed but result from a mathematical mo<strong>de</strong>l, making<br />
assumptions about the relationships and distributions of input<br />
variab<strong>le</strong>s.<br />
Discounting Economic concept to hand<strong>le</strong> time-prefer<strong>en</strong>ce, using a method of<br />
calculation by which costs and b<strong>en</strong>efits occurring at differ<strong>en</strong>t<br />
mom<strong>en</strong>ts in time can be compared. Discounting converts the<br />
value of future costs and b<strong>en</strong>efits into their pres<strong>en</strong>t value to<br />
account for positive time prefer<strong>en</strong>ces for b<strong>en</strong>efits (prefer<strong>en</strong>ce<br />
for curr<strong>en</strong>t b<strong>en</strong>efits as compared to future b<strong>en</strong>efits) and negative<br />
time prefer<strong>en</strong>ces for costs (prefer<strong>en</strong>ce for future costs as<br />
compared to curr<strong>en</strong>t costs).<br />
Economic evaluation Comparative analysis of alternative courses of action in terms of<br />
Effectiv<strong>en</strong>ess<br />
(effectiviteit/doeltreff<strong>en</strong>dheid;<br />
<strong>efficacité</strong> réel<strong>le</strong>/<strong>efficacité</strong> pratique<br />
Efficacy (efficaciteit/werkzaamheid;<br />
<strong>efficacité</strong> théorique/<strong>efficacité</strong><br />
expérim<strong>en</strong>ta<strong>le</strong>/<strong>efficacité</strong><br />
both their costs and consequ<strong>en</strong>ces.<br />
The ext<strong>en</strong>t to which health interv<strong>en</strong>tions achieve health<br />
improvem<strong>en</strong>ts in real-life settings.<br />
The ext<strong>en</strong>t to which health interv<strong>en</strong>tions achieve health<br />
improvem<strong>en</strong>ts un<strong>de</strong>r i<strong>de</strong>al control<strong>le</strong>d conditions (as for examp<strong>le</strong><br />
in randomised control<strong>le</strong>d trials)
4 ICER Thresholds <strong>KCE</strong> Reports 100<br />
pot<strong>en</strong>tiel<strong>le</strong>)<br />
Effici<strong>en</strong>cy<br />
(efficiëntie/doelmatigheid;<br />
effici<strong>en</strong>ce)<br />
In economic theory <strong>de</strong>fined as the condition in which no<br />
productive resources are wasted in the manufacture of a certain<br />
product; i.e. where output is produced at minimum cost or the<br />
<strong>le</strong>vel of output is maximised at a giv<strong>en</strong> cost (i.e. cannot be<br />
increased). In health care, effici<strong>en</strong>cy implies that choices should<br />
be ma<strong>de</strong> so as to <strong>de</strong>rive the maximum total health b<strong>en</strong>efit from<br />
the availab<strong>le</strong> resources. ‘Allocative effici<strong>en</strong>cy’ occurs wh<strong>en</strong> the<br />
outcomes achieved with the availab<strong>le</strong> resources match the<br />
priorities of society.<br />
Health-related quality of life A multidim<strong>en</strong>sional construct measuring the physical, social and<br />
emotional aspects that are re<strong>le</strong>vant and important to a pati<strong>en</strong>t’s<br />
well-being.<br />
Health maximisation Maximisation of re<strong>le</strong>vant health outcomes. In health economics<br />
oft<strong>en</strong> maximisation of the number of LYG or the number of<br />
QALYs gained.<br />
Health outcome Result of health interv<strong>en</strong>tion for the health of a pati<strong>en</strong>t or a<br />
population.<br />
ICER threshold value B<strong>en</strong>chmark for ICERs (increm<strong>en</strong>tal cost-effectiv<strong>en</strong>ess ratios) to<br />
assess an interv<strong>en</strong>tion’s cost-effectiv<strong>en</strong>ess. Interv<strong>en</strong>tions with an<br />
ICER below the ICER threshold value are consi<strong>de</strong>red costeffective,<br />
interv<strong>en</strong>tions with an ICER above the ICER threshold<br />
value are not cost-effective.<br />
Increm<strong>en</strong>tal analysis Analysis of additional costs and additional health outcomes<br />
Increm<strong>en</strong>tal cost-effectiv<strong>en</strong>ess<br />
ratio (ICER)<br />
associated with differ<strong>en</strong>t treatm<strong>en</strong>ts.<br />
Ratio of additional costs and additional health outcomes<br />
associated with differ<strong>en</strong>t treatm<strong>en</strong>ts: (C 2 – C 1)/(E 2 – E 1), where<br />
C 2 and C 1 repres<strong>en</strong>t the costs of interv<strong>en</strong>tion 2 and 1<br />
respectively and E 2 and E 1 repres<strong>en</strong>t the health outcomes<br />
(effects) of interv<strong>en</strong>tion 2 and 1 respectively.<br />
League tab<strong>le</strong> Tab<strong>le</strong> ranking health interv<strong>en</strong>tions according to their increm<strong>en</strong>tal<br />
cost-effectiv<strong>en</strong>ess ratio with the purpose to gui<strong>de</strong> resourceallocation<br />
<strong>de</strong>cisions.<br />
Opportunity costs The costs of resources consumed expressed as the value of the<br />
next best alternative for using these resources.<br />
Private insurance based health System where health care is fun<strong>de</strong>d from premiums paid to<br />
care system<br />
private insurance companies.<br />
Quality-adjusted life year Measure for health outcomes that inclu<strong>de</strong>s both quality and<br />
quantity of life a pati<strong>en</strong>t is expected to have. Quality-adjusted life<br />
years are calculated by estimating the total life years gained from<br />
a treatm<strong>en</strong>t and weighting each time period within these life<br />
years gained with a quality-of-life score betwe<strong>en</strong> 0 (<strong>de</strong>ad) to 1<br />
(perfect health) that ref<strong>le</strong>cts the health-related quality of life in<br />
that period.<br />
S<strong>en</strong>sitivity analysis Technique used in economic evaluation to allow for uncertainty<br />
by testing whether plausib<strong>le</strong> changes in the values of the main<br />
variab<strong>le</strong>s affect the results of the analysis.<br />
Societal Willingness to Pay (WTP) Societal willingness to pay refers to the maximum amount<br />
society is willing to pay for a unit of health gain (e.g. QALY or<br />
life-year gained). It ref<strong>le</strong>cts what society is willing to sacrifice in<br />
terms of other goods or services for a unit of health gain.<br />
Uncertainty A state in which the true value of a parameter or the structure<br />
of a process is unknown.<br />
Utility A measure of the prefer<strong>en</strong>ce for, or <strong>de</strong>sirability of, a specific<br />
<strong>le</strong>vel of health status or specific health outcomes.
<strong>KCE</strong> reports 100 ICER Thresholds 5<br />
ABBREVIATIONS<br />
CADTH Canadian Ag<strong>en</strong>cy for Drugs and Technologies in Health<br />
CBA Cost-B<strong>en</strong>efit Analysis<br />
CEA Cost-Effectiv<strong>en</strong>ess Analysis<br />
CEDAC Canadian Expert Drug Advisory Committee<br />
CTG/CRM Drug Reimbursem<strong>en</strong>t Committee (Commissie Tegemoetkoming<br />
G<strong>en</strong>eesmid<strong>de</strong><strong>le</strong>n/Commission <strong>de</strong> Remboursem<strong>en</strong>t <strong>de</strong>s Médicam<strong>en</strong>ts)<br />
(=DRC) (Belgium)<br />
CUA Cost-Utility Analysis<br />
CVZ Dutch Health care Insurance Board (Col<strong>le</strong>ge voor zorgverzekering<strong>en</strong>)<br />
DACEHTA Danish C<strong>en</strong>tre for Health Technology Assessm<strong>en</strong>t<br />
DRC Drug Reimbursem<strong>en</strong>t Committee (=CTG/CRM) (Belgium)<br />
FOD/SPF Fe<strong>de</strong>ral Public Service (Fe<strong>de</strong>ra<strong>le</strong> Overheidsdi<strong>en</strong>st / Service Public Fédéral)<br />
(=FPS) (Belgium)<br />
FPS Fe<strong>de</strong>ral Public Service (=FOD/SPF) (Belgium)<br />
GDP Gross Domestic Product<br />
HPV Human Papillomavirus<br />
HRQoL Health-Related Quality of Life<br />
ICER Increm<strong>en</strong>tal Cost-Effectiv<strong>en</strong>ess Ratio<br />
<strong>KCE</strong> Belgian Health Care Know<strong>le</strong>dge C<strong>en</strong>tre (Fe<strong>de</strong>raal K<strong>en</strong>nisc<strong>en</strong>trum voor <strong>de</strong><br />
Gezondheidszorg / C<strong>en</strong>tre fédéral d’expertise <strong>de</strong>s <strong>soins</strong> <strong>de</strong> <strong>santé</strong><br />
(Belgium)<br />
MCDA Multi-Criteria Decision Analysis<br />
MRI Magnetic Resonance Imaging<br />
NHS National Health Service<br />
NICE National Institute for Health and Clinical Excel<strong>le</strong>nce (UK)<br />
NIHDI National Institute for Health and Disability Insurance (=RIZIV/INAMI)<br />
(Belgium)<br />
PBAC Pharmaceutical B<strong>en</strong>efits Advisory Committee (Australia)<br />
PET Positron Emission Tomography<br />
PHARMAC Pharmaceutical Managem<strong>en</strong>t Ag<strong>en</strong>cy (New Zealand)<br />
QALY Quality-Adjusted Life Year<br />
RIZIV/INAMI National Institute for Health and Disability Insurance (Rijksinstituut voor<br />
Ziekte- <strong>en</strong> Invaliditeitsverzekering/National d’Assurance Maladie-<br />
Invalidité) (=NIHDI) (Belgium)<br />
TCI Technical Council for Implants (=TRI/CTI) (Belgium)<br />
TRI/CTI Technical Council for Implants (Technische Raad voor<br />
Implantat<strong>en</strong>/Conseil Technique <strong>de</strong>s Implants) (=TCI) (Belgium)<br />
WTP Willingness-To-Pay
6 ICER Thresholds <strong>KCE</strong> Reports 100<br />
1 GENERAL INTRODUCTION<br />
1.1 BACKGROUND<br />
One of the research domains of the Belgian Health Care Know<strong>le</strong>dge C<strong>en</strong>tre (<strong>KCE</strong>) is<br />
Health Technology Assessm<strong>en</strong>t (HTA). HTA aims to inform health care <strong>de</strong>cision<br />
makers about the (most likely) clinical, economic, organisational and ethical implications<br />
of imp<strong>le</strong>m<strong>en</strong>ting and financing health interv<strong>en</strong>tions. In comparison with existing health<br />
interv<strong>en</strong>tions, new interv<strong>en</strong>tions may reduce mortality or morbidity, improve healthrelated<br />
quality of life (HRQoL), <strong>de</strong>liver similar care for <strong>le</strong>ss money or <strong>en</strong>hance the<br />
organisation of health care. Unfortunately, although many of these technologies have<br />
their own intrinsic value, society cannot afford all of them simultaneously. Therefore,<br />
the assessm<strong>en</strong>t of the ‘value for money’ of health interv<strong>en</strong>tions is an important part of<br />
any HTA. It usually takes the form of a cost-effectiv<strong>en</strong>ess or cost-utility analysis.<br />
This report is writt<strong>en</strong> at the request of the Board of the Belgian Health Care<br />
Know<strong>le</strong>dge C<strong>en</strong>tre (<strong>KCE</strong>). Some of the <strong>KCE</strong> HTA-reports pres<strong>en</strong>t cost-effectiv<strong>en</strong>ess<br />
analyses and increm<strong>en</strong>tal cost-effectiv<strong>en</strong>ess ratios (ICERs). For peop<strong>le</strong> who are not<br />
trained in health economics it may not be c<strong>le</strong>ar, however, what ICERs repres<strong>en</strong>t and<br />
what their contribution can be for health care policy making. Therefore a <strong>de</strong>mand was<br />
formulated to explain to lay peop<strong>le</strong> the concept of ICERs and their pot<strong>en</strong>tial use in daily<br />
practice.<br />
1.2 PROBLEM STATEMENT<br />
Economic evaluation assesses the relative ‘value for money’ of health interv<strong>en</strong>tions. 1<br />
Health interv<strong>en</strong>tions are to be interpreted in a broad s<strong>en</strong>se. They inclu<strong>de</strong> prev<strong>en</strong>tive<br />
health programmes, curative health care, rehabilitation services and palliative care and<br />
within these health interv<strong>en</strong>tions, use of drugs, medical <strong>de</strong>vices, behavioural therapy<br />
etc. 2-4<br />
The basic goal of economic evaluations is to inform health care policy makers about the<br />
best way to allocate limited resources in health care in or<strong>de</strong>r to maximize health gains.<br />
They help <strong>de</strong>cision makers to evaluate whether funding a particular (high cost)<br />
technology is worthwhi<strong>le</strong>.<br />
A commonly used measure for the assessm<strong>en</strong>t of an interv<strong>en</strong>tion’s relative value for<br />
money is the increm<strong>en</strong>tal cost-effectiv<strong>en</strong>ess ratio (ICER). The ICER repres<strong>en</strong>ts the<br />
additional cost per extra unit of ‘effect’ g<strong>en</strong>erated by an interv<strong>en</strong>tion compared with an<br />
appropriate comparator. This comparator is an alternative interv<strong>en</strong>tion for the same<br />
condition or no interv<strong>en</strong>tion wh<strong>en</strong> none is curr<strong>en</strong>tly availab<strong>le</strong>. Whi<strong>le</strong> costs are normally<br />
expressed as monetary values (for a giv<strong>en</strong> year), effects can be expressed in various<br />
units, such as life years gained (LYG), quality-adjusted life years (QALY) gained, or<br />
natural units (e.g. number of infections avoi<strong>de</strong>d). 2-4<br />
If an interv<strong>en</strong>tion offers better outcomes at a lower cost than its comparator, i.e. it is<br />
more effective and <strong>le</strong>ss costly, it is straightforward to conclu<strong>de</strong> that it offers better<br />
value for money (see 2.3). More oft<strong>en</strong>, however, an interv<strong>en</strong>tion offers better outcomes<br />
at a higher cost. For the evaluation of these cases, economic evaluation can be helpful to<br />
policymakers, especially wh<strong>en</strong> <strong>de</strong>ciding about the reimbursem<strong>en</strong>t of health interv<strong>en</strong>tions<br />
because reimbursem<strong>en</strong>t has implications for the effici<strong>en</strong>cy of the allocation of the scarce<br />
health care resources.<br />
However, it is not so straightforward to give a meaning to ICERs or to use them in a<br />
<strong>de</strong>cision making context. For examp<strong>le</strong>, is €60 000 per QALY gained reasonab<strong>le</strong> to<br />
<strong>de</strong>ci<strong>de</strong> for reimbursem<strong>en</strong>t or is it too high? Where do we draw the line of acceptability?<br />
Without ways to <strong>de</strong>al with these questions, economic evaluations and their resulting<br />
ICERs will be of limited value to health care <strong>de</strong>cision makers.
<strong>KCE</strong> reports 100 ICER Thresholds 7<br />
1.3 SCOPE OF THIS REPORT<br />
This report is writt<strong>en</strong> to support non-economically trained peop<strong>le</strong> involved in health<br />
care <strong>de</strong>cision making. Therefore, the aim of this report is primarily didactic and it does<br />
not pret<strong>en</strong>d to be a fully elaborated sci<strong>en</strong>tific (theoretical and methodological) study.<br />
The c<strong>en</strong>tral questions in this report are: “Wh<strong>en</strong> can interv<strong>en</strong>tions be consi<strong>de</strong>red costeffective"<br />
and “Is there a threshold for the ICER above which interv<strong>en</strong>tions can no longer be<br />
consi<strong>de</strong>red cost-effective?”.<br />
This report offers information to health care policy makers about economic evaluations<br />
in health care, their results and their re<strong>le</strong>vance for health care policy. It gives an<br />
introduction on how the results of economic evaluations should be interpreted,<br />
whether a threshold value for the ICER can be <strong>de</strong>fined above which an interv<strong>en</strong>tion<br />
cannot be consi<strong>de</strong>red cost-effective and how the ICER can be used in health care policy<br />
<strong>de</strong>cision making. Recomm<strong>en</strong>dations are formulated for health care policy makers on<br />
how to <strong>de</strong>al with results of economic evaluations within <strong>de</strong>cision making contexts and<br />
also for researchers on how to make the results of economic evaluations more useful<br />
for policy makers.<br />
1.4 OBJECTIVES<br />
This report <strong>de</strong>scribes the basic concepts of cost-effectiv<strong>en</strong>ess analysis (CEA),<br />
increm<strong>en</strong>tal cost-effectiv<strong>en</strong>ess ratios (ICERs) and ICER threshold values. It aims to<br />
explain how they are obtained and tries to explore how they are or could be used in<br />
health care <strong>de</strong>cision making contexts.<br />
More specifically, the report addresses six questions:<br />
• What is an ICER and an ICER threshold value and where does it come<br />
from, i.e. what is its theoretical background? (sections 2.1 to 2.6)<br />
• What is the external validity of the theoretical assumptions for an ICER<br />
threshold value? (section 2.7)<br />
• What are possib<strong>le</strong> alternatives for the ICER threshold value approach?<br />
(section 2.8)<br />
• What is the (pot<strong>en</strong>tial) ro<strong>le</strong> of an ICER threshold value in health care<br />
<strong>de</strong>cision making contexts? (sections 3.1 to 3.4)<br />
• Are ICER threshold values used in other countries and how are they<br />
used? (section 3.5)<br />
• How do specific Belgian advisory councils <strong>de</strong>al with the issue of ‘value for<br />
money’ in health care? (section 3.6)<br />
For interested rea<strong>de</strong>rs, refer<strong>en</strong>ces for further reading are provi<strong>de</strong>d. 2-5<br />
1.5 METHODOLOGY<br />
Giv<strong>en</strong> the objectives of this report, we did not perform a systematic review of the<br />
literature. We ma<strong>de</strong> a narrative review of the literature on ICERs and ICER threshold<br />
values using an increm<strong>en</strong>tal search strategy: starting from re<strong>le</strong>vant refer<strong>en</strong>ces id<strong>en</strong>tified<br />
through an explorative search in Medline, (key-words used were ‘cost-effectiv<strong>en</strong>ess’,<br />
‘ICER’ and ‘threshold’) and applying the snowball princip<strong>le</strong> to id<strong>en</strong>tify additional re<strong>le</strong>vant<br />
refer<strong>en</strong>ces. We consi<strong>de</strong>red differ<strong>en</strong>t viewpoints on CEA and ICERs, trying not to<br />
exclu<strong>de</strong> or prefer<strong>en</strong>tially inclu<strong>de</strong> any specific perspective.<br />
The methods used for the international comparison of the use of ICER threshold values<br />
and for the explorative field study in the Belgian <strong>de</strong>cision making context are <strong>de</strong>tai<strong>le</strong>d in<br />
the re<strong>le</strong>vant sections.
8 ICER Thresholds <strong>KCE</strong> Reports 100<br />
1.6 STRUCTURE OF THIS REPORT<br />
First, we briefly <strong>de</strong>scribe the basic concepts of economic evaluation and the theoretical<br />
foundations for ICER threshold values (question 1). Next, we <strong>de</strong>scribe to what ext<strong>en</strong>t<br />
the theoretical assumptions of the ICER threshold value approach are valid (question 2)<br />
and which alternatives are suggested in literature for ICERs and ICER threshold values<br />
along with their str<strong>en</strong>gths and weaknesses (question 3).<br />
In a subsequ<strong>en</strong>t chapter, we <strong>de</strong>scribe <strong>de</strong>cision making processes and the pot<strong>en</strong>tial place<br />
of economic consi<strong>de</strong>rations within these (question 4). The application of ICER<br />
thresholds in other countries is <strong>de</strong>scribed (question 5) and we conclu<strong>de</strong> the chapter<br />
with a <strong>de</strong>scription of the results of an explorative field study within the Belgian <strong>de</strong>cision<br />
making context and the place of ICERs and ICER threshold values in two committees<br />
that advice the Ministry of Social Affairs about the reimbursem<strong>en</strong>t of health<br />
interv<strong>en</strong>tions (question 6). The field study explores to what ext<strong>en</strong>t CEA and ICERs are<br />
curr<strong>en</strong>tly known and used in Belgian health care reimbursem<strong>en</strong>t <strong>de</strong>cision making.<br />
The report conclu<strong>de</strong>s with a g<strong>en</strong>eral discussion on the possib<strong>le</strong> ro<strong>le</strong> of economic<br />
evaluation in health care <strong>de</strong>cision making and the conclusions from this sci<strong>en</strong>tific<br />
overview. Finally, we formulate some recomm<strong>en</strong>dations for Belgian health care policy<br />
makers with respect to the pot<strong>en</strong>tial use of ICERs and ICER threshold values in policy<br />
<strong>de</strong>cisions.
<strong>KCE</strong> reports 100 ICER Thresholds 9<br />
2 ICERS AND ICER THRESHOLD VALUES<br />
2.1 GENERAL PRINCIPLES OF ECONOMIC EVALUATION IN<br />
HEALTH CARE<br />
Economic evaluation is <strong>de</strong>fined as the comparative analysis of alternative courses of<br />
action in terms of both their costs and consequ<strong>en</strong>ces. 2 In economic evaluation of health<br />
care interv<strong>en</strong>tions, ‘consequ<strong>en</strong>ces’ are most oft<strong>en</strong> interpreted as ‘health effects’ or<br />
‘health outcomes’. Both terms will be used interchangeably in this report.<br />
Only evaluations that compare two or more alternatives and consi<strong>de</strong>r both costs and<br />
consequ<strong>en</strong>ces are consi<strong>de</strong>red full economic evaluations (see Figure 1). 2<br />
Figure 1: Overview of partial and full economic evaluations<br />
Are both costs (inputs) and consequ<strong>en</strong>ces (outputs) of alternatives examined?<br />
No<br />
Yes<br />
Outputs only Inputs only<br />
No Outcome <strong>de</strong>scription Cost <strong>de</strong>scription Cost-outcome <strong>de</strong>scription<br />
Cost-utility analysis (CUA),<br />
Efficacy or effectiv<strong>en</strong>ess<br />
Yes Cost comparison<br />
evaluation<br />
Cost-b<strong>en</strong>efit analysis (CBA),<br />
Cost-effectiv<strong>en</strong>ess analysis (CEA),<br />
Cost-minimisation analysis (CMA)<br />
Comparison of<br />
2 alternatives?<br />
Partial evaluation<br />
Adapted from Drummond et al. (2005) 2<br />
Full economic evaluation<br />
Full economic evaluations are classified according to the way in which the health effects<br />
are expressed. Health effects can be expressed in physical units (cost-effectiv<strong>en</strong>ess<br />
analysis, CEA), in terms of utility values (cost-utility analysis, CUA) or in monetary<br />
terms (cost-b<strong>en</strong>efit analysis, CBA). A commonly used outcome measure in CEA is the<br />
‘number of life years gained’ (LYG) by the interv<strong>en</strong>tion un<strong>de</strong>r study. In CUA, the number<br />
of quality-adjusted life years gained (QALYs) is oft<strong>en</strong> used as an outcome parameter,<br />
where LYG are ‘weighted’ for a quantified measure of health-related quality of life<br />
(HRQoL) in those LYG. QALYs are oft<strong>en</strong> regar<strong>de</strong>d as just another measure of<br />
effectiv<strong>en</strong>ess rather than as a utility measure in its strict utilitarian s<strong>en</strong>se. 6, 7 Therefore<br />
and for the ease of reading, we use the term cost-effectiv<strong>en</strong>ess analysis (CEA) for both<br />
cost-per-LYG and cost-per-QALY gained analyses throughout this report. CBA, where<br />
both costs and effects are expressed in monetary terms, is becoming increasingly<br />
unpopular as a technique for health economic evaluations because of the difficulties<br />
experi<strong>en</strong>ced in valuing health outcomes in monetary terms. a 3 Therefore, we will<br />
conc<strong>en</strong>trate on CEA in this report. Cost-minimisation analysis is a specific case of CEA<br />
where the health outcomes of the interv<strong>en</strong>tion and its comparator are assumed to be<br />
equiva<strong>le</strong>nt and where the aim th<strong>en</strong> becomes to obtain those outcomes at the lowest<br />
cost. 2-4<br />
CEA is used to assess effici<strong>en</strong>cy in the production of <strong>de</strong>sirab<strong>le</strong> health outcomes. More<br />
specifically, it aims to help id<strong>en</strong>tify how the highest number of LYG or QALYs can be<br />
achieved by allocating limited resources betwe<strong>en</strong> all possib<strong>le</strong> health interv<strong>en</strong>tions. 2-4<br />
a In contrast to CEA or CUA, where one specific measure of health effect is chos<strong>en</strong> for the evaluation (e.g.<br />
LYG or QALYs), cost-b<strong>en</strong>efit analysis in princip<strong>le</strong> allows the consi<strong>de</strong>ration of non-health effects of an<br />
interv<strong>en</strong>tion as well. It is therefore a broa<strong>de</strong>r form of economic evaluation than CEA or CUA. I.e. if all<br />
effects of an interv<strong>en</strong>tion could be expressed in monetary units, there is no reason to restrict the analysis<br />
to health effects only. Consequ<strong>en</strong>tly, broa<strong>de</strong>r comparisons would become possib<strong>le</strong> with CBA, ev<strong>en</strong> with<br />
interv<strong>en</strong>tions outsi<strong>de</strong> the health care sector. An acceptab<strong>le</strong> monetary valuation of all effects remains,<br />
however, difficult.
10 ICER Thresholds <strong>KCE</strong> Reports 100<br />
The assumption of CEA in its neo-classical welfarist form is that health care <strong>de</strong>cision<br />
makers’ primary goal is to maximise health within giv<strong>en</strong> budget constraints, making<br />
abstraction of other pot<strong>en</strong>tial concerns <strong>de</strong>cision makers may have in real life, such as<br />
equity, political and macro-economic consi<strong>de</strong>rations.<br />
Its aim is to show how resources can be allocated to meet the goal of health outcomes<br />
maximisation, where health outcomes are strictly <strong>de</strong>fined in terms of LYG or QALYs<br />
gained. There is still <strong>de</strong>bate about whether this kind of CEA is useful for <strong>de</strong>cision<br />
makers, i.e. as an imperfect aid to <strong>de</strong>cision making, 2 or whether CEA should try to<br />
incorporate, in some way or another, the other goals health care policy makers might<br />
have, 8 for instance by weighting the health outcomes of specific populations more<br />
heavily to ref<strong>le</strong>ct social prefer<strong>en</strong>ces for health outcomes allocation, as in the so-cal<strong>le</strong>d<br />
extra-welfarist approach. Although theoretically appealing, there is yet no cons<strong>en</strong>sus on<br />
how other goals besi<strong>de</strong>s health maximisation should be incorporated in CEA or what<br />
other goals should be inclu<strong>de</strong>d. 4, 9-11 CEA, as commonly performed in practice now, still<br />
approaches the resource allocation <strong>de</strong>cision prob<strong>le</strong>m from the economic effici<strong>en</strong>cy<br />
point of view, not explicitly consi<strong>de</strong>ring other health policy goals such as equity in its<br />
framework. Therefore, we focus on this basic approach in this chapter. Chapter 3<br />
elaborates on other possib<strong>le</strong> approaches.<br />
Key points<br />
• Economic evaluation is the comparative analysis of alternative courses of<br />
action in terms of both their costs and consequ<strong>en</strong>ces.<br />
• Cost-effectiv<strong>en</strong>ess (CEA) and cost-utility analysis (CUA) are the most<br />
frequ<strong>en</strong>tly used techniques for economic evaluation in health care. In this<br />
report CEA is used as the g<strong>en</strong>eric term to cover both techniques.<br />
• CEA aims to inform health policy makers about the best way to allocate<br />
limited health care resources in or<strong>de</strong>r to obtain maximal health<br />
outcomes in terms of LYG or QALYs gained.<br />
2.2 THE INCREMENTAL COST-EFFECTIVENESS RATIO (ICER)<br />
The ICER is the ratio of the differ<strong>en</strong>ce in costs (C) and the differ<strong>en</strong>ce in outcomes (E,<br />
effects) betwe<strong>en</strong> an interv<strong>en</strong>tion and its comparator. Expressed in a formula:<br />
C2 − C1<br />
ICER =<br />
E2 − E1<br />
where C2 (E2) is the cost (effect) of the interv<strong>en</strong>tion and C1 (E1) is the cost (effect) of the<br />
comparator. The costs of the interv<strong>en</strong>tion and the comparator inclu<strong>de</strong> not only the<br />
costs of a specific drug, <strong>de</strong>vice or act but the cost of the <strong>en</strong>tire treatm<strong>en</strong>t path followed<br />
by the pati<strong>en</strong>ts un<strong>de</strong>rgoing the treatm<strong>en</strong>t, including the costs of follow-up and<br />
treatm<strong>en</strong>t of pot<strong>en</strong>tial complications and/or si<strong>de</strong>-effects (lifetime perspective). Also on<br />
the effect-si<strong>de</strong>, effects are not limited to the immediate effect of a product, <strong>de</strong>vice or<br />
act but both the positive and negative health effects associated with the <strong>en</strong>tire<br />
treatm<strong>en</strong>t path and the pot<strong>en</strong>tial complications and si<strong>de</strong>-effects. The comparator should<br />
be a cost-effective alternative interv<strong>en</strong>tion for the same condition. b2, 3<br />
The ICER can be expressed as a cost per LYG or as a cost per QALY gained. It is used<br />
for making <strong>de</strong>cisions about interv<strong>en</strong>tions that are both more costly and more effective<br />
than their comparator (or <strong>le</strong>ss effective but cheaper).<br />
b Comparing with an interv<strong>en</strong>tion that is not cost-effective will ev<strong>en</strong>tually <strong>le</strong>ad to an estimate of the ICER<br />
that is unab<strong>le</strong> to inform policy makers about the best way to allocate scarce resources to obtain the<br />
highest health b<strong>en</strong>efits (see 2.6.2). It is h<strong>en</strong>ce assumed that it is possib<strong>le</strong> to establish the cost-effectiv<strong>en</strong>ess<br />
of the comparators before the comparison is ma<strong>de</strong>. In practice, it is oft<strong>en</strong> assumed that ‘curr<strong>en</strong>t practice’<br />
is the most appropriate cost-effective comparator.
<strong>KCE</strong> reports 100 ICER Thresholds 11<br />
The ICER of an interv<strong>en</strong>tion could be compared to a certain ICER threshold value or to<br />
ICERs of other interv<strong>en</strong>tions for other conditions. The lower the ICER, the more<br />
additional health can be obtained with the same additional value of resource inputs, and<br />
thus the more cost-effective an interv<strong>en</strong>tion is consi<strong>de</strong>red. 2-4<br />
Key points<br />
• The increm<strong>en</strong>tal cost-effectiv<strong>en</strong>ess ratio (ICER) is the ratio of the<br />
estimated differ<strong>en</strong>ce betwe<strong>en</strong> the costs of two interv<strong>en</strong>tions and the<br />
estimated differ<strong>en</strong>ce betwe<strong>en</strong> the outcomes of these two interv<strong>en</strong>tions.<br />
• The ICER repres<strong>en</strong>ts the estimated additional cost per extra unit of<br />
health b<strong>en</strong>efit g<strong>en</strong>erated by an interv<strong>en</strong>tion compared with an<br />
appropriate comparator. The appropriate comparator is the most costeffective<br />
alternative for the same health condition.<br />
• The ICER int<strong>en</strong>ds to support informed <strong>de</strong>cision making about<br />
interv<strong>en</strong>tions that are both more costly and more effective than their<br />
comparator (or inversely <strong>le</strong>ss effective but cheaper).<br />
2.3 THE COST-EFFECTIVENESS PLANE<br />
The results of a full economic evaluation can be repres<strong>en</strong>ted graphically on a costeffectiv<strong>en</strong>ess<br />
plane (Figure 2). 2<br />
Figure 2: The cost-effectiv<strong>en</strong>ess plane<br />
Increm<strong>en</strong>tal cost (€)<br />
North-West or<br />
Fourth quadrant<br />
€50 000<br />
€40 000<br />
€30 000<br />
€20 000<br />
€10 000<br />
-5 -4 -3 -2 -1 1 2 3 4 5<br />
South-West or<br />
Third quadrant<br />
-€10 000<br />
-€20 000<br />
-€30 000<br />
-€40 000<br />
-€50 000<br />
Increm<strong>en</strong>tal effect (LYG or QALYs)<br />
LYG: life years gained, QALY: quality-adjusted life years<br />
A<br />
North-East or<br />
First quadrant<br />
South-East or<br />
Second quadrant<br />
The origin of the plane repres<strong>en</strong>ts the costs and effects of the comparator, against<br />
which the costs and effects of the interv<strong>en</strong>tion are compared. The values on the axes<br />
are increm<strong>en</strong>tal values. The horizontal axis repres<strong>en</strong>ts the differ<strong>en</strong>ce in health effects<br />
betwe<strong>en</strong> the interv<strong>en</strong>tion and the comparator, expressed as either LYG or QALYs<br />
gained. The vertical axis repres<strong>en</strong>ts the cost differ<strong>en</strong>ce betwe<strong>en</strong> the interv<strong>en</strong>tion and<br />
the comparator. The comparator is either no interv<strong>en</strong>tion (or curr<strong>en</strong>t situation) or a<br />
re<strong>le</strong>vant cost-effective alternative interv<strong>en</strong>tion for the same condition.<br />
The cost-effectiv<strong>en</strong>ess plane is subdivi<strong>de</strong>d in four parts, cal<strong>le</strong>d quadrants.<br />
In the second and fourth quadrant it is straightforward which interv<strong>en</strong>tion offers the<br />
highest value for money, i.e. the interv<strong>en</strong>tion with the lowest costs and the largest<br />
health effects.
12 ICER Thresholds <strong>KCE</strong> Reports 100<br />
• In the second quadrant, the new treatm<strong>en</strong>t is better and cheaper than the<br />
alternative; in economic literature this new treatm<strong>en</strong>t is th<strong>en</strong> cal<strong>le</strong>d<br />
‘dominant’.<br />
• In the fourth quadrant the new treatm<strong>en</strong>t is worse and more exp<strong>en</strong>sive, it<br />
is ‘being dominated’.<br />
The conclusion becomes more difficult if the interv<strong>en</strong>tion is situated in the first or the<br />
third quadrant.<br />
• In the first quadrant, the interv<strong>en</strong>tion is better but also more exp<strong>en</strong>sive.<br />
In this case an ICER can be calculated.<br />
• In the third quadrant, interv<strong>en</strong>tions are cheaper but <strong>le</strong>ss effective than<br />
their comparator. These can in theory be treated in the same way as<br />
interv<strong>en</strong>tions in the first quadrant 12 , although in practice they are very<br />
differ<strong>en</strong>t. Interv<strong>en</strong>tions in the third quadrant would save money, at the<br />
exp<strong>en</strong>se of worse health. It is oft<strong>en</strong> argued, however, that in real life<br />
proof of better clinical effectiv<strong>en</strong>ess is a necessary (but not suffici<strong>en</strong>t)<br />
condition for reimbursem<strong>en</strong>t. 13 In other words, before ev<strong>en</strong> consi<strong>de</strong>ring<br />
an interv<strong>en</strong>tion for reimbursem<strong>en</strong>t, policy makers will first look at<br />
whether the interv<strong>en</strong>tion offers better health outcomes. If this is not the<br />
case, costs or savings are oft<strong>en</strong> not ev<strong>en</strong> consi<strong>de</strong>red.<br />
• The third quadrant may also repres<strong>en</strong>t situations of disinvestm<strong>en</strong>t:<br />
reversing a reimbursem<strong>en</strong>t <strong>de</strong>cision might pot<strong>en</strong>tially result in large<br />
savings at the exp<strong>en</strong>se of a limited loss in health. Again, in real life these<br />
consi<strong>de</strong>rations are rarely ma<strong>de</strong>.<br />
The value of the ICER is equal to the slope of the line through the origin and the<br />
interv<strong>en</strong>tion’s cost-effectiv<strong>en</strong>ess pair (e.g. point A). In the examp<strong>le</strong> in Figure 2, the ICER<br />
is €30 000 per QALY gained (or LYG <strong>de</strong>p<strong>en</strong>ding on the units on the horizontal axis)<br />
compared to its comparator.<br />
Key points<br />
• A cost-effectiv<strong>en</strong>ess plane visualises the cost differ<strong>en</strong>ce (Y-axis) and effect<br />
differ<strong>en</strong>ce (X-axis) betwe<strong>en</strong> an interv<strong>en</strong>tion and its comparator.<br />
• If an interv<strong>en</strong>tion is more costly and more effective than its comparator<br />
(or <strong>le</strong>ss costly and <strong>le</strong>ss effective), the slope of the line through the origin<br />
and the point corresponding to the increm<strong>en</strong>tal cost and effect on the<br />
plane is equal to the value of the ICER.<br />
2.4 METHODOLOGICAL ISSUES OF THE ICER<br />
2.4.1 Comparability of ICERs for differ<strong>en</strong>t interv<strong>en</strong>tions<br />
Differ<strong>en</strong>t consi<strong>de</strong>rations have to be ma<strong>de</strong> wh<strong>en</strong> using and comparing ICERs.<br />
First, ICERs of differ<strong>en</strong>t interv<strong>en</strong>tions can only be compared if their numerator and<br />
d<strong>en</strong>ominator are expressed in the same units. Increm<strong>en</strong>tal costs are g<strong>en</strong>erally expressed<br />
in monetary units (for a giv<strong>en</strong> country and year) but increm<strong>en</strong>tal effects can be<br />
expressed in differ<strong>en</strong>t units, for instance as LYG or as QALYs.<br />
Second, the methodology used for calculating increm<strong>en</strong>tal costs and increm<strong>en</strong>tal effects<br />
is important for the comparability of ICERs across interv<strong>en</strong>tions. Methodological issues<br />
may <strong>de</strong>crease the comparability of ICERs across interv<strong>en</strong>tions and their suitability for<br />
health care <strong>de</strong>cision making. Consist<strong>en</strong>cy and transpar<strong>en</strong>cy of economic evaluations is<br />
crucial for their credibility and usefulness for health care <strong>de</strong>cision making.<br />
Therefore, several HTA ag<strong>en</strong>cies have elaborated methodological gui<strong>de</strong>lines to help<br />
those who conduct economic evaluations to calculate ICERs consist<strong>en</strong>tly. 14-18<br />
One of the important e<strong>le</strong>m<strong>en</strong>ts in all these gui<strong>de</strong>lines is the perspective of the economic<br />
analysis: differ<strong>en</strong>t perspectives <strong>le</strong>ad to differ<strong>en</strong>t values for the ICER, due to differ<strong>en</strong>ces<br />
in the costs inclu<strong>de</strong>d in the analysis.
<strong>KCE</strong> reports 100 ICER Thresholds 13<br />
For examp<strong>le</strong>, productivity losses are costs from the societal perspective but not from<br />
the health care payers’ perspective. Including or excluding these costs may have an<br />
important impact on the ICER estimate.<br />
Another important methodological issue relates to the measure for health gains. Both<br />
LYG and QALYs have their weaknesses as measures for health gains.<br />
• Using LYG as the so<strong>le</strong> outcome measure of interv<strong>en</strong>tions could create a<br />
<strong>de</strong>cision bias against interv<strong>en</strong>tions that only impact upon quality of life. 2<br />
• QALYs, on the other hand, are oft<strong>en</strong> still fraught with measurem<strong>en</strong>t<br />
prob<strong>le</strong>ms and are oft<strong>en</strong> not comparab<strong>le</strong> betwe<strong>en</strong> studies due to the<br />
variety in measurem<strong>en</strong>t techniques for HRQoL. Differ<strong>en</strong>t measurem<strong>en</strong>t<br />
techniques give differ<strong>en</strong>t results (e.g. Griebsch et al. 19 , Scuffham et al. 20 ,<br />
Read et al. 21 , Hornberger et al. 22 , and Marra et al. 23 ). As there is no ‘gold<br />
standard’ for measuring HRQoL, it is difficult to <strong>de</strong>termine which<br />
measurem<strong>en</strong>t technique gives the most appropriate results for the<br />
purposes of the evaluation. As long as differ<strong>en</strong>t measurem<strong>en</strong>t techniques<br />
for HRQoL are being used in CEAs, ICERs expressed in terms of costper-QALY<br />
gained will be difficult to compare across interv<strong>en</strong>tions.<br />
2.4.2 Uncertainty around the ICER<br />
The calculated increm<strong>en</strong>tal costs and effects that are used to <strong>de</strong>termine the ICER are<br />
both estimates, and estimates are by <strong>de</strong>finition uncertain. The uncertainty of the<br />
numerator and d<strong>en</strong>ominator of the ICER translates into uncertainty around the ICER<br />
estimate.<br />
Whi<strong>le</strong> the <strong>de</strong>gree of uncertainty may differ betwe<strong>en</strong> estimates, the uncertainty should<br />
not be ignored. 24, 25 Uncertainty may, for examp<strong>le</strong>, relate to the expected effectiv<strong>en</strong>ess<br />
of the treatm<strong>en</strong>t in a specific pati<strong>en</strong>t population, the proportion of pati<strong>en</strong>ts complying<br />
with the treatm<strong>en</strong>t, the costs associated with the organisation of the treatm<strong>en</strong>t in<br />
routine care, etc... Very oft<strong>en</strong>, assumptions have to be ma<strong>de</strong> about these parameters,<br />
assumptions that are translated into data distributions around a c<strong>en</strong>tral estimate.<br />
The uncertainty around the ICER estimate can be expressed as a credibility interval,<br />
comparab<strong>le</strong> to a confid<strong>en</strong>ce interval for empirical data, or graphically on the costeffectiv<strong>en</strong>ess<br />
plane as a scatter plot, repres<strong>en</strong>ting the individual values resulting from<br />
probabilistic s<strong>en</strong>sitivity analysis (Figure 3). 26<br />
Figure 3: Cost-effectiv<strong>en</strong>ess plane and pres<strong>en</strong>tation of uncertainty around<br />
the estimate of the cost-effectiv<strong>en</strong>ess ratio<br />
In this examp<strong>le</strong>, the Mean ICER is €30 803 (95% credibility interval: 19 433 – 46 747)
14 ICER Thresholds <strong>KCE</strong> Reports 100<br />
2.4.3 Wh<strong>en</strong> do we incur costs and wh<strong>en</strong> do we reap the b<strong>en</strong>efits?<br />
The future is uncertain and this uncertainty has to be <strong>de</strong>alt with. Costs and outcomes of<br />
health interv<strong>en</strong>tions usually do not occur at the same time. This raises the issue of<br />
valuing future outcomes and costs and choosing the appropriate time horizon for<br />
economic evaluations.<br />
2, 4, 27, 28<br />
The timing issue is <strong>de</strong>alt with by means of discounting future costs and outcomes.<br />
Through discounting the value of future costs and outcomes is reduced to account for<br />
the fact that peop<strong>le</strong> value future health outcomes and costs <strong>le</strong>ss than immediate health<br />
outcomes and costs. The choice of the relative discount rate for costs compared to<br />
outcomes is important, as a discount rate that is lower for outcomes than for costs<br />
<strong>le</strong>ads to relatively lower ICERs for programmes with outcomes in a far-away future as<br />
compared to a situation where both costs and outcomes would be discounted at the<br />
same rate. 29-31 The <strong>de</strong>bate on whether an equal or differ<strong>en</strong>tial discount rate should be<br />
used for costs and health outcomes is ongoing. 32-36<br />
A crucial question is “what is a reasonab<strong>le</strong> time horizon for health economic<br />
evaluations”, knowing that these evaluations will be used for curr<strong>en</strong>t <strong>de</strong>cision making.<br />
By taking <strong>de</strong>cisions now on the basis of economic evaluations with a long time horizon,<br />
<strong>de</strong>cisions are actually tak<strong>en</strong> about costs and effects occurring now and in the future.<br />
Uncertainty exists, amongst others, about the emerg<strong>en</strong>ce of new interv<strong>en</strong>tions, the<br />
future population eligib<strong>le</strong> for a specific treatm<strong>en</strong>t and the long-term effectiv<strong>en</strong>ess of an<br />
interv<strong>en</strong>tion. 37 As a consequ<strong>en</strong>ce, the expected b<strong>en</strong>efits of an interv<strong>en</strong>tion for which<br />
<strong>de</strong>cisions are tak<strong>en</strong> now might never occur. Some of these uncertainties can be solved<br />
25, 38<br />
by further research whi<strong>le</strong> others cannot.<br />
However, this does not imply that it would be better to take <strong>de</strong>cisions only within a<br />
short term perspective, as this would preclu<strong>de</strong> the application of prev<strong>en</strong>tive<br />
interv<strong>en</strong>tions and could pot<strong>en</strong>tially jeopardize future g<strong>en</strong>erations.<br />
Vaccinations are a typical examp<strong>le</strong> of interv<strong>en</strong>tions with an important upfront cost and<br />
with outcomes much further in the future, outcomes that can oft<strong>en</strong> only be<br />
<strong>de</strong>monstrated after actual imp<strong>le</strong>m<strong>en</strong>tation of a large vaccination programme. In other<br />
words, the expected ICER might never be realised and maybe resources would have<br />
be<strong>en</strong> better sp<strong>en</strong>t elsewhere.<br />
Decision makers are forced to make a tra<strong>de</strong>-off betwe<strong>en</strong> waiting until more evid<strong>en</strong>ce<br />
becomes availab<strong>le</strong> and the risks inher<strong>en</strong>t to taking a <strong>de</strong>cision in a situation of<br />
uncertainty. 37, 39-41 Remaining uncertainty can be addressed in ext<strong>en</strong>sive s<strong>en</strong>sitivity<br />
24, 25, 38, 42<br />
analyses that <strong>le</strong>ad to an uncertainty range around the ICER point estimate.<br />
Key points<br />
• Variability across economic evaluations in terms of outcome measures<br />
(LYG or QALYs) and methods used to calculate the ICER reduce the<br />
comparability of ICERs across interv<strong>en</strong>tions.<br />
• The ICER of an interv<strong>en</strong>tion is by <strong>de</strong>finition an estimate and therefore<br />
uncertain.<br />
• Future costs and b<strong>en</strong>efits should be discounted. The <strong>de</strong>bate on whether<br />
equal discount rates should be used for costs and health outcomes or<br />
lower discount rates for outcomes than for costs is still ongoing, but the<br />
rates used can have an important effect on the ICER estimate and should<br />
be fully appreciated wh<strong>en</strong> comparing ICERs.<br />
• Future costs and b<strong>en</strong>efits are oft<strong>en</strong> uncertain. This uncertainty should be<br />
a<strong>de</strong>quately consi<strong>de</strong>red in the <strong>de</strong>cision making process.
<strong>KCE</strong> reports 100 ICER Thresholds 15<br />
2.5 THE ICER THRESHOLD VALUE IN A FIXED BUDGET<br />
SETTING<br />
As a stand-alone value the ICER does not offer information about whether an<br />
interv<strong>en</strong>tion is worth its costs. Health care policy makers still have to <strong>de</strong>ci<strong>de</strong> whether<br />
the value of the ICER is acceptab<strong>le</strong> or not.<br />
According to neoclassical welfare economic theoryc, an ICER threshold value can be<br />
<strong>de</strong>fined below which an interv<strong>en</strong>tion is cost-effective (increases effici<strong>en</strong>cy) and above<br />
which it is not. 2 This is subject to specific conditions (see 2.5.1). In this paragraph, the<br />
basis and the meaning of this ICER threshold value and the conditions to which it is<br />
subject are explained.<br />
2.5.1 Basic assumptions<br />
To be ab<strong>le</strong> to <strong>de</strong>fine the ICER threshold value, the following basic assumptions have to<br />
be fulfil<strong>le</strong>d: 45-47<br />
• the health care budget is fixedd • the health care policy makers’ so<strong>le</strong> objective is to maximise health giv<strong>en</strong><br />
this fixed budgete • full information exists on the costs and effects of all availab<strong>le</strong> health<br />
interv<strong>en</strong>tions,<br />
• health programmes are perfectly divisib<strong>le</strong>, meaning that it is possib<strong>le</strong> to<br />
realise only part of a programme<br />
• health programmes are in<strong>de</strong>p<strong>en</strong>d<strong>en</strong>t from one another<br />
• health programmes have constant returns to sca<strong>le</strong>, meaning that reducing<br />
a programme does not change its ICER.<br />
A fixed health care budget is in this context not a budget that remains constant over<br />
time or grows at a constant rate. ‘Fixed budget’ means that the budget cannot be<br />
increased or oversp<strong>en</strong>t within a giv<strong>en</strong> year. f<br />
Perfect divisibility of a health programme would imply that the programme can be<br />
imp<strong>le</strong>m<strong>en</strong>ted or downgra<strong>de</strong>d to whatever ext<strong>en</strong>t. Basically, this refers to programmes<br />
without fixed costs (see 2.7.3).<br />
c The differ<strong>en</strong>ce betwe<strong>en</strong> welfarims and extra-welfarism is conceptually and methodologically comp<strong>le</strong>x.<br />
Welfarism asserts that social welfare is a function of individual welfare (approached as utility) obtained<br />
only from the consumption of goods and services.43 Extra-welfarism argues that the superiority of one<br />
social state (allocation of resources) over another may also <strong>de</strong>p<strong>en</strong>d on the non-utility aspects of each<br />
state.43 For examp<strong>le</strong>, whi<strong>le</strong> in the welfarist approach the aim is to maximise the total number of QALYs,<br />
extra-welfarism also inclu<strong>de</strong>s consi<strong>de</strong>rations that are not inclu<strong>de</strong>d in the QALY, such as the allocation of<br />
QALYs across pati<strong>en</strong>t groups or severity of illnesses. These additional consi<strong>de</strong>rations may justify an<br />
allocation of resources that is sub-optimal according to the welfarist approach. Relative societal values of<br />
health gains (QALYs) have rec<strong>en</strong>tly be<strong>en</strong> studied empirically in the UK.44 This fits with the extra-welfarist<br />
approach.<br />
d A fixed budget is not specifically required for the welfarist approach. Also in a variab<strong>le</strong> budget context the<br />
welfarist approach can be applied, but th<strong>en</strong> the meaning of the ICER threshold value is differ<strong>en</strong>t from the<br />
one pres<strong>en</strong>ted in this section (see 2.8.1 for the meaning of the welfaristic ICER threshold value in a<br />
variab<strong>le</strong> budget context).<br />
e This refers to the welfarist approach.<br />
f In some systems, the health care budget will be strictly fixed, i.e. it cannot be increased and resources<br />
from other sectors cannot be applied to fill pot<strong>en</strong>tial gaps. Such a system prevails in countries such as the<br />
UK and New Zealand although it should be noted that budgets are never comp<strong>le</strong>tely fixed. In other<br />
systems, the budget is fixed in princip<strong>le</strong> but can be used in a f<strong>le</strong>xib<strong>le</strong> way. For examp<strong>le</strong>, in Belgium the<br />
health care budget is fixed but due to the prospective financing of some health services (e.g. GP<br />
consultations) the budget can exceed the pre-<strong>de</strong>fined budget.
16 ICER Thresholds <strong>KCE</strong> Reports 100<br />
2.5.2 Id<strong>en</strong>tifying the ICER threshold value<br />
If these assumptions are fulfil<strong>le</strong>d, it is possib<strong>le</strong> to construct an ICER <strong>le</strong>ague tab<strong>le</strong>, where<br />
interv<strong>en</strong>tions are ranked from lowest to highest ICER. The health-maximising allocation<br />
of a fixed budget is obtained by financing the interv<strong>en</strong>tions with the lowest ICERs first<br />
and th<strong>en</strong> moving down in the <strong>le</strong>ague tab<strong>le</strong> until the budget is exhausted. 45 The ICER of<br />
the last interv<strong>en</strong>tion still financed from the budget th<strong>en</strong> repres<strong>en</strong>ts the cost of the<br />
marginal (i.e. the last) QALY gained (or LYG) from the budget. This ICER th<strong>en</strong><br />
repres<strong>en</strong>ts the threshold value for new ICERs: if a new interv<strong>en</strong>tion can produce an<br />
additional QALY (LYG) at a lower increm<strong>en</strong>tal cost (i.e. has a lower ICER) than the last<br />
interv<strong>en</strong>tion already financed from the budget (the ICER threshold value), it is<br />
economically more effici<strong>en</strong>t to produce that additional QALY (LYG) instead of the<br />
curr<strong>en</strong>t marginal QALY (LYG). If the cost of an additional QALY (LYG) from the new<br />
interv<strong>en</strong>tion is higher than the cost of the marginal QALY (LYG) curr<strong>en</strong>tly fun<strong>de</strong>d, it is<br />
not worth replacing the existing marginal interv<strong>en</strong>tion with the new one, as it would<br />
reduce total health. 45<br />
Therefore, according to the ICER threshold approach, 2-4<br />
• interv<strong>en</strong>tion A is not cost-effective if ICERA > ICER threshold value;<br />
• interv<strong>en</strong>tion A is cost-effective if ICERA < ICER threshold value.<br />
An examp<strong>le</strong>: consi<strong>de</strong>r three health interv<strong>en</strong>tions in an exemplary health system. The<br />
increm<strong>en</strong>tal costs and effects of these interv<strong>en</strong>tions, each time relative to their re<strong>le</strong>vant<br />
comparator are pres<strong>en</strong>ted in Tab<strong>le</strong> 1. Each interv<strong>en</strong>tion treats a differ<strong>en</strong>t disease.<br />
Interv<strong>en</strong>tion A offers a treatm<strong>en</strong>t to 10 pati<strong>en</strong>ts, interv<strong>en</strong>tion B to 15 pati<strong>en</strong>ts and C to<br />
8 pati<strong>en</strong>ts. Based on the increm<strong>en</strong>tal costs and effects an ICER can be calculated for<br />
each interv<strong>en</strong>tion. The ICERs are th<strong>en</strong> ranked from low to high (<strong>le</strong>ague tab<strong>le</strong>). The<br />
higher the ICER, the <strong>le</strong>ss cost-effective an interv<strong>en</strong>tion is consi<strong>de</strong>red. Finally, the budget<br />
impact of the interv<strong>en</strong>tions in the curr<strong>en</strong>t year is indicated. With a giv<strong>en</strong> budget of<br />
€1 100 000, for instance, health interv<strong>en</strong>tions A, B and part of C can be financed. In the<br />
threshold approach, financing part of a programme is an option, as explained in section<br />
2.5.1. In real life, however, this might not always be the case. We elaborate on this in<br />
section 2.7.3.<br />
Consi<strong>de</strong>r now a new interv<strong>en</strong>tion D with an increm<strong>en</strong>tal cost of €800 000 and a budget<br />
impact of €700 000. For a <strong>de</strong>cision about the reimbursem<strong>en</strong>t of the interv<strong>en</strong>tion within<br />
the limits of a fixed budget, D’s ICER has to be compared with C’s ICER. If D’s ICER is<br />
lower than 20 000€/QALY, say 18 000€/QALY it is worthwhi<strong>le</strong> to imp<strong>le</strong>m<strong>en</strong>t part of D<br />
(up to the point where €100 000 is sp<strong>en</strong>t on D) and sp<strong>en</strong>d <strong>le</strong>ss resources on (disinvest<br />
in) interv<strong>en</strong>tion C. g<br />
Tab<strong>le</strong> 1: Resource allocation based on ICERs in the i<strong>de</strong>al world<br />
∆C ∆E ∆C/∆E<br />
Total<br />
increm<strong>en</strong>tal<br />
effectiv<strong>en</strong>ess<br />
Total<br />
increm<strong>en</strong>tal<br />
cost<br />
Budget<br />
impact in<br />
curr<strong>en</strong>t year<br />
A 100.000 10 10.000 100 1.000.000 200.000<br />
B 200.000 12 16.667 180 3.000.000 800.000<br />
C 100.000 5 20.000 40 800.000 150.000<br />
For every new interv<strong>en</strong>tion that is consi<strong>de</strong>red to be financed by the same fixed budget,<br />
the ICER of that interv<strong>en</strong>tion should be compared with the ICER of the last interv<strong>en</strong>tion<br />
in the <strong>le</strong>ague tab<strong>le</strong> still financed. 48 If the ICER of the new interv<strong>en</strong>tion is higher than the<br />
ICER of the last financed interv<strong>en</strong>tion (the threshold value), the new programme should<br />
not be accepted. If its ICER is lower than the threshold value, financing of this new<br />
interv<strong>en</strong>tion would increase the total number of QALYs gained and h<strong>en</strong>ce a health<br />
maximising <strong>de</strong>cision ru<strong>le</strong> would <strong>de</strong>mand the inclusion of the new health programme.<br />
g Note that un<strong>de</strong>r these conditions it is not cost-effective to finance the <strong>en</strong>tire programme, as the ICER of<br />
D is still higher than the ICER of B. Financing the <strong>en</strong>tire programme would require disinvestm<strong>en</strong>t in B, but<br />
giv<strong>en</strong> the lower ICER of B this would not be effici<strong>en</strong>t.
<strong>KCE</strong> reports 100 ICER Thresholds 17<br />
The <strong>de</strong>cision implies a reduction in the budget sp<strong>en</strong>t on the interv<strong>en</strong>tion with the<br />
highest ICER. If not, un<strong>de</strong>r a fixed budget, the new interv<strong>en</strong>tion cannot be paid for.<br />
2.5.3 Characteristics of the ICER threshold value in a fixed budget setting<br />
The ICER threshold value has the following characteristics:<br />
1. The threshold changes as the composition of the health programmes fun<strong>de</strong>d<br />
changes. Because the threshold value is equal to the ICER of the last<br />
programme se<strong>le</strong>cted before the budget is exhausted, the threshold changes<br />
each time a new programme is inclu<strong>de</strong>d in the package of fun<strong>de</strong>d<br />
programmes. 47, 48 In the previous examp<strong>le</strong>, the new ICER threshold value<br />
becomes 18 000€/QALY, being the ICER of D, the marginally financed<br />
programme.<br />
2. The ICER threshold value <strong>de</strong>p<strong>en</strong>ds on the availab<strong>le</strong> budget. The higher the<br />
49, 50<br />
budget, ceteris paribus, the higher the threshold will be.<br />
3. The ICER threshold value <strong>de</strong>p<strong>en</strong>ds on the productivity in the health care sector.<br />
If the productivity increases, meaning that more LYG or more QALYs can be<br />
g<strong>en</strong>erated with the same amount of resources, and the budget does not change,<br />
the threshold value will <strong>de</strong>crease.<br />
These characteristics suggest that the ICER threshold value is not a static value but<br />
changes over time due to changes in budgets, innovations, productivity, etc. 48 In<br />
addition, they imply that the ICER threshold value from one country is not necessarily<br />
applicab<strong>le</strong> to other countries, as other countries have other budgets, other practices,<br />
other productivity, other health programmes already financed etc. 46<br />
Other reasons explain why ICER threshold values are not easily transferab<strong>le</strong> betwe<strong>en</strong><br />
countries: the standard methodology used for calculating ICERs may differ, financing<br />
systems may differ, budgets may be more or <strong>le</strong>ss fixed etc.<br />
2.5.4 Interpretation of the ICER threshold value in a fixed budget setting<br />
The theoretical ICER threshold value as pres<strong>en</strong>ted above does not take into account<br />
societal willingness to pay for a QALY or for a LYG and it is neither an absolute<br />
34, 48<br />
criterion for evaluating the cost-effectiv<strong>en</strong>ess of health interv<strong>en</strong>tions in real life.<br />
Rather, it is the result of an economic theoretical mo<strong>de</strong>l for maximising health gains<br />
from a giv<strong>en</strong> fixed budget that applies to a specific context, at a specific mom<strong>en</strong>t in time<br />
and un<strong>de</strong>r specific conditions.<br />
The ICER threshold value repres<strong>en</strong>ts the highest amount of money for a QALY (or<br />
LYG) society still can pay at a specific mom<strong>en</strong>t in time, giv<strong>en</strong> its fixed health care budget<br />
and its health maximising goal. With every new interv<strong>en</strong>tion that <strong>en</strong>ters the package of<br />
reimbursed interv<strong>en</strong>tions, the ICER threshold value has to be revisited. Therefore, the<br />
ICER threshold value cannot be interpreted as a value that stands for a long time. It will<br />
only apply for as long as the comparisons with new pot<strong>en</strong>tial candidates for financing do<br />
not <strong>le</strong>ad to the inclusion of new interv<strong>en</strong>tions in the fun<strong>de</strong>d package.
18 ICER Thresholds <strong>KCE</strong> Reports 100<br />
Key points<br />
The ICER threshold value is a theoretical construct for maximising health<br />
within the constraints of a fixed budget.<br />
This theoretical ICER threshold value assumes:<br />
• A fixed health care budget, <strong>de</strong>fined as a budget that cannot be excee<strong>de</strong>d.<br />
• The one and only aim of health care <strong>de</strong>cisions is to maximise health<br />
b<strong>en</strong>efits in terms of QALYs or in terms of LYG.<br />
• Comp<strong>le</strong>te information on the ICERs of all interv<strong>en</strong>tions is availab<strong>le</strong>.<br />
• Perfect divisibility of health programmes.<br />
• Constant returns to sca<strong>le</strong>; i.e. reducing or ext<strong>en</strong>ding a programme (in the<br />
same target population) does not influ<strong>en</strong>ce its ICER.<br />
• Health programmes are in<strong>de</strong>p<strong>en</strong>d<strong>en</strong>t from each other.<br />
The ICER threshold value is the ICER of the last interv<strong>en</strong>tion in a <strong>le</strong>ague<br />
tab<strong>le</strong> that is still (fully or ev<strong>en</strong> partially) financed from the fixed budget.<br />
The ICER threshold value is not a static value but changes over time, subject<br />
to changes in the budget, the interv<strong>en</strong>tions fun<strong>de</strong>d and the productivity of<br />
health care.<br />
2.6 METHODOLOGICAL ISSUES OF THE ICER THRESHOLD<br />
VALUE<br />
2.6.1 Uncertainty around the ICER and the ICER threshold value<br />
As discussed in paragraph 2.4.2 an ICER is an estimate and the exact value is uncertain.<br />
This implies that in comparisons with an ICER threshold value, the only possib<strong>le</strong><br />
conclusion is that there is a probability that the ICER falls below the threshold value.<br />
This probability can be quantified based on the results of probabilistic s<strong>en</strong>sitivity<br />
analyses.<br />
The probabilistic s<strong>en</strong>sitivity analysis calculates a distribution around the ICER. The costeffectiv<strong>en</strong>ess<br />
acceptability curve th<strong>en</strong> ref<strong>le</strong>cts the proportion of the distribution of the<br />
ICER below the threshold value for all possib<strong>le</strong> ICER threshold values. 51 Graphically, it<br />
would be the proportion of dots from Figure 3 falling below the ICER threshold value;<br />
i.e. to the right of the line through the origin with a slope equal to the ICER threshold<br />
value. By varying the ICER threshold value, the proportion of dots falling below the<br />
ICER threshold line also varies. As the proportion of dots falling below the ICER<br />
threshold line corresponds to the probability that the interv<strong>en</strong>tion is consi<strong>de</strong>red costeffective<br />
for the pre-<strong>de</strong>fined ICER threshold value, the probability that an interv<strong>en</strong>tion is<br />
consi<strong>de</strong>red cost-effective changes if the ICER threshold changes. The graphical<br />
pres<strong>en</strong>tation of this probability is cal<strong>le</strong>d a ‘cost-effectiv<strong>en</strong>ess acceptability curve’ and is<br />
2, 3<br />
shown in Figure 4.
<strong>KCE</strong> reports 100 ICER Thresholds 19<br />
Figure 4: Cost-effectiv<strong>en</strong>ess acceptability curve<br />
Probability interv<strong>en</strong>tion being cost<br />
effective<br />
1<br />
0,9<br />
0,8<br />
0,7<br />
0,6<br />
0,5<br />
0,4<br />
0,3<br />
0,2<br />
0,1<br />
0<br />
€ 0 €20 000 €40 000 €60 000 €80 000 €100 000<br />
Value of the threshold ratio (cost per QALY gained)<br />
However, not only the ICER of interv<strong>en</strong>tions is uncertain, also the ICER threshold value<br />
-being the ICER of the last interv<strong>en</strong>tion financed from the budget- is subject to<br />
uncertainty. Therefore, the ICER threshold value will not be a sing<strong>le</strong> value but rather<br />
again a variab<strong>le</strong> with a distribution. 47 For practical use, the ICER threshold value could<br />
be <strong>de</strong>fined as a range with limits <strong>de</strong>fined by the upper- and lower limits of the 95%<br />
confid<strong>en</strong>ce or credibility interval around the ICER of the marginally fun<strong>de</strong>d health<br />
programme. h The cost-effectiv<strong>en</strong>ess acceptability curve, however, does not account for<br />
the uncertainty around the ICER threshold value. 47<br />
Giv<strong>en</strong> the uncertainty about the precise value of the ICER threshold and its <strong>de</strong>finition in<br />
terms of an interval, the kind of conclusions drawn from these curves can no longer be<br />
that “there is a probability of Y% that the ICER is below the threshold value” but rather<br />
that “the probability that the ICER falls below the ICER threshold value is betwe<strong>en</strong> X%<br />
and Z%.” The range for the probabilities is <strong>de</strong>fined by the applied range for the ICER<br />
threshold value.<br />
In conclusion, the uncertainty around the ICER threshold value adds to the uncertainty<br />
around the ICER estimate, thereby increasing the uncertainty about an interv<strong>en</strong>tion’s<br />
cost-effectiv<strong>en</strong>ess.<br />
2.6.2 Comparison with an appropriate comparator<br />
The theoretical ICER threshold value can only be <strong>de</strong>fined if for each interv<strong>en</strong>tion in the<br />
<strong>le</strong>ague tab<strong>le</strong> the ICER is calculated relative to a cost-effective comparator or relative to<br />
doing nothing. 3 If the comparator is an alternative interv<strong>en</strong>tion (and h<strong>en</strong>ce not ‘doing<br />
nothing’) it should be an interv<strong>en</strong>tion that is curr<strong>en</strong>tly financed because it is consi<strong>de</strong>red<br />
worthwhi<strong>le</strong> giv<strong>en</strong> the fixed budget constraint and the health maximisation objective. i In<br />
other words, the <strong>le</strong>ague tab<strong>le</strong> approach assumes that all health interv<strong>en</strong>tions curr<strong>en</strong>tly<br />
financed from the healthcare budget fit within the health maximisation rationa<strong>le</strong> and are<br />
financed only because they are cost-effective. H<strong>en</strong>ce, if a new interv<strong>en</strong>tion emerges as<br />
an alternative to an existing and already fun<strong>de</strong>d interv<strong>en</strong>tion, the existing interv<strong>en</strong>tion is<br />
an appropriate comparator.<br />
h In case of economic mo<strong>de</strong>lling, the term “credibility interval” is used rather than “confid<strong>en</strong>ce interval” to<br />
make the distinction betwe<strong>en</strong> variability in directly observed values versus variability in values resulting<br />
from an economic mo<strong>de</strong>l.<br />
i Comparison with an appropriate alternative treatm<strong>en</strong>t is recomm<strong>en</strong><strong>de</strong>d in most gui<strong>de</strong>lines for economic<br />
evaluation. The WHO’s “Gui<strong>de</strong>lines on g<strong>en</strong>eralized cost-effectiv<strong>en</strong>ess analysis”, however, recomm<strong>en</strong>d the<br />
evaluation of an interv<strong>en</strong>tion’s cost-effectiv<strong>en</strong>ess relative to “doing nothing” (i.e. relative to the natural<br />
history of disease) as a standard approach.52 The WHO has a very specific mandate and has therefore<br />
specific reasons for e<strong>le</strong>cting this approach. For a full discussion on the g<strong>en</strong>eralized cost-effectiv<strong>en</strong>ess<br />
analysis, see WHO (2003).53
20 ICER Thresholds <strong>KCE</strong> Reports 100<br />
If there is no alternative that is already fun<strong>de</strong>d for the new interv<strong>en</strong>tion, it should be<br />
compared with ‘doing nothing’, because this means that no cost-effective comparator is<br />
yet availab<strong>le</strong> for this new interv<strong>en</strong>tion.<br />
If interv<strong>en</strong>tions that are not cost-effective are curr<strong>en</strong>tly fun<strong>de</strong>d, CEA using these<br />
interv<strong>en</strong>tions as comparator might <strong>le</strong>ad to ICERs that look attractive compared to the<br />
ICER threshold value. However, as the comparator should not have be<strong>en</strong> fun<strong>de</strong>d in the<br />
first place according to the economic effici<strong>en</strong>cy argum<strong>en</strong>t, application of the ICER<br />
threshold value ru<strong>le</strong> for the new interv<strong>en</strong>tion would not <strong>le</strong>ad to maximal health<br />
outcomes. The approach h<strong>en</strong>ce imposes the strong requirem<strong>en</strong>t of being ab<strong>le</strong> to<br />
establish the cost-effectiv<strong>en</strong>ess of the comparator before proceeding to the costeffectiv<strong>en</strong>ess<br />
analysis.<br />
2.6.3 Measurem<strong>en</strong>t units in nominator and d<strong>en</strong>ominator<br />
Because of the c<strong>le</strong>ar differ<strong>en</strong>ce betwe<strong>en</strong> measures of ‘health gain’ (e.g. LYG or QALYs),<br />
any threshold value should be consi<strong>de</strong>red taking explicitly into account the curr<strong>en</strong>cy<br />
used in the numerator and outcome parameter used in the d<strong>en</strong>ominator. Although, this<br />
remark may seem obvious, it has be<strong>en</strong> observed that oft<strong>en</strong> ‘round’ numbers are<br />
preferred, which are easily copied without paying much att<strong>en</strong>tion to the units<br />
accompanying the number and applied threshold values do not seem to change over<br />
42, 54-56<br />
time <strong>de</strong>spite changing economic <strong>en</strong>vironm<strong>en</strong>t and inflation.<br />
Key points<br />
• The ICER threshold value is the result of a health maximisation mo<strong>de</strong>l<br />
that applies to a specific context (fixed budget, country), at a specific<br />
mom<strong>en</strong>t in time and un<strong>de</strong>r specific conditions.<br />
• The ICER threshold value is subject to uncertainty and variability.<br />
Therefore, the ICER threshold value is not a sing<strong>le</strong> value but a range of<br />
values. This is important for the kind of conclusions that can be drawn<br />
from cost-effectiv<strong>en</strong>ess analyses.<br />
• The <strong>le</strong>ague tab<strong>le</strong> approach used to <strong>de</strong>fine the ICER threshold value<br />
assumes that each interv<strong>en</strong>tion’s ICER is calculated compared to a costeffective<br />
alternative or to doing nothing if no cost-effective alternative<br />
exists.<br />
• The units in which the costs and health effects are expressed are<br />
important for the interpretation of the ICER threshold value: an ICER<br />
threshold value of €30 000/QALY is differ<strong>en</strong>t from an ICER threshold<br />
value of £30 000/LYG.<br />
2.7 HOW WELL ARE THE THEORETICAL ASSUMPTIONS FOR<br />
THE ICER THRESHOLD VALUE FULFILLED IN REAL LIFE?<br />
As explained previously in 2.5.1, the ICER threshold value can be <strong>de</strong>fined as the ICER of<br />
the last interv<strong>en</strong>tion still financed from a fixed budget, but only if a series of basic<br />
45, 57<br />
assumptions are fulfil<strong>le</strong>d.<br />
The conditions for the appropriate id<strong>en</strong>tification of the ICER threshold value are highly<br />
theoretical and seldom met in practice. 50 The ext<strong>en</strong>t to which the conditions are met<br />
<strong>de</strong>p<strong>en</strong>ds partly on the characteristics of the health care reimbursem<strong>en</strong>t system.<br />
Globally, three main reimbursem<strong>en</strong>t mo<strong>de</strong>ls can be id<strong>en</strong>tified: the Beveridge mo<strong>de</strong>l<br />
(national health provi<strong>de</strong>r paid directly from taxes), the Bismarck mo<strong>de</strong>l (compreh<strong>en</strong>sive<br />
social security based system but mainly paid from contributions of employees) and the<br />
private health insurance mo<strong>de</strong>l. 58-60<br />
A Beveridge system is characterised by a c<strong>en</strong>trally organised National Health Service<br />
(NHS), where health care is provi<strong>de</strong>d mainly by public provi<strong>de</strong>rs. 59 Examp<strong>le</strong>s of such<br />
systems are found in the UK, Italy and Spain. In Beveridge systems the state is the<br />
c<strong>en</strong>tral actor.
<strong>KCE</strong> reports 100 ICER Thresholds 21<br />
The Ministry of Health prepares the annual health care budget which is consi<strong>de</strong>red to<br />
be fixed for that year. Health care is primarily financed from public resources obtained<br />
through g<strong>en</strong>eral taxation. As such, the health care budget competes with other sp<strong>en</strong>ding<br />
priorities as a consequ<strong>en</strong>ce of which the health care budget cannot easily be excee<strong>de</strong>d. j<br />
Beveridge systems are, overall, characterised by many public provi<strong>de</strong>rs and relatively<br />
few private provi<strong>de</strong>rs. 60<br />
A Bismarck system, as prevailing in Belgium, France and Germany, is a social security<br />
based system where social insurance is compreh<strong>en</strong>sive and mandatory. 59 Resources<br />
availab<strong>le</strong> for social security sp<strong>en</strong>ding come from social security contributions, mainly<br />
from salaried employees. There is g<strong>en</strong>erally a strong influ<strong>en</strong>ce of stakehol<strong>de</strong>rs. For<br />
instance, reimbursem<strong>en</strong>t of health care procedures is oft<strong>en</strong> negotiated betwe<strong>en</strong> health<br />
care provi<strong>de</strong>rs, insurers and governm<strong>en</strong>t. There is a mixture of private and public<br />
provi<strong>de</strong>rs and the health care budget is consi<strong>de</strong>red somewhat more f<strong>le</strong>xib<strong>le</strong>. 60<br />
In a private insurance system, health care is paid out of premiums paid to private<br />
insurance companies. The obvious examp<strong>le</strong> of this system is the US, where this system<br />
is combined with a few limited social care fallback systems such as Medicare and<br />
Medicaid.<br />
2.7.1 Fixed budget<br />
The ICER threshold value approach as <strong>de</strong>scribed in 2.5.2 is applicab<strong>le</strong> in situations<br />
where the health care budget is strictly fixed (whi<strong>le</strong> other conditions also apply, see<br />
2.5.1).<br />
A fixed health care budget requires a variab<strong>le</strong> ICER threshold value. Un<strong>de</strong>r a fixed<br />
budget constraint, an ICER threshold value (with an appropriate range around it<br />
repres<strong>en</strong>ting uncertainty) against which other ICERs should be compared to maximise<br />
health outcomes can be <strong>de</strong>fined at a specific mom<strong>en</strong>t in time. But the ICER threshold<br />
value cannot be fixed over time in a fixed budget situation, It has to be revised every<br />
48, 61<br />
time a positive reimbursem<strong>en</strong>t <strong>de</strong>cision about a new interv<strong>en</strong>tion is tak<strong>en</strong>.<br />
A f<strong>le</strong>xib<strong>le</strong> health care budget does allow the use of a fixed ICER threshold value to a<br />
certain ext<strong>en</strong>t. The budget will th<strong>en</strong> have to expand every time a new interv<strong>en</strong>tion with<br />
a lower ICER than the threshold ICER value becomes availab<strong>le</strong>. 62-64 k However, the<br />
meaning and h<strong>en</strong>ce the interpretation of the ICER threshold value would in that case be<br />
comp<strong>le</strong>tely differ<strong>en</strong>t (see 2.8.1). It is th<strong>en</strong> no longer the health-maximising threshold<br />
criterion for a fixed budget.<br />
As explained previously the health care budget is more fixed in an NHS (Beveridge)<br />
system than in a social security (Bismarck) system. In the UK, for instance, the budget of<br />
the NHS is mainly exog<strong>en</strong>ously <strong>de</strong>termined by Parliam<strong>en</strong>t. 61<br />
It is a fixed budget that should cover most or all health care exp<strong>en</strong>ditures of the<br />
citiz<strong>en</strong>s. l As a consequ<strong>en</strong>ce, the cost of an interv<strong>en</strong>tion is equal to the resources nee<strong>de</strong>d<br />
from the health care budget.<br />
j In practice, the health care budget of a giv<strong>en</strong> year can be excee<strong>de</strong>d in a Beveridge system, for examp<strong>le</strong> if<br />
in a specific year more prescription drugs are used than initially expected. This will, however, be more<br />
difficult than in a Bismarck system where budgets are more oft<strong>en</strong> negotiated.<br />
k Expansion of the budget every time the ICER of an interv<strong>en</strong>tion is lower than the ICER threshold value is<br />
not t<strong>en</strong>ab<strong>le</strong> in any system, be it a fixed or f<strong>le</strong>xib<strong>le</strong> budget system. H<strong>en</strong>ce, regular adaptation of the ICER<br />
threshold value will always be necessary. In systems with more f<strong>le</strong>xib<strong>le</strong> budgets, the revision of the ICER<br />
threshold value might be <strong>le</strong>ss frequ<strong>en</strong>t than in systems with <strong>le</strong>ss f<strong>le</strong>xib<strong>le</strong> budgets (e.g. once a year or every<br />
two years, in the context of an evaluation of the health care package fun<strong>de</strong>d from public resources and<br />
their budgetary consequ<strong>en</strong>ces). But, the constant revision of the ICER threshold value still requires the<br />
satisfaction of the basic (and prob<strong>le</strong>matic) assumptions of the ‘conv<strong>en</strong>tional’ ICER threshold value<br />
approach: perfect divisibility of health programmes, constant returns to sca<strong>le</strong> and making abstraction of<br />
equity consi<strong>de</strong>rations across pati<strong>en</strong>t populations in case of unweighted QALYs (or LYG). It could be<br />
argued that perfect divisibility of programmes and constant returns to sca<strong>le</strong> are <strong>le</strong>ss important conditions<br />
for the ICER threshold value approach, but this only applies to systems with f<strong>le</strong>xib<strong>le</strong> budgets.<br />
l In practice, the health care budget of a giv<strong>en</strong> year can be excee<strong>de</strong>d in a Beveridge system, for examp<strong>le</strong> if<br />
in a specific year more prescription drugs are used than initially expected, but this will be more difficult<br />
than in a Bismarck system where budgets are much more negotiated.
22 ICER Thresholds <strong>KCE</strong> Reports 100<br />
In a social security system, where pati<strong>en</strong>ts oft<strong>en</strong> pay individual co-paym<strong>en</strong>ts for the<br />
services they consume, the total cost of an interv<strong>en</strong>tion is not equal to its total budget<br />
impact for the public health care payer. In the case of Belgium, the NIHDI as well as the<br />
‘Fe<strong>de</strong>ral Public Service (FPS) Health, Food Chain Safety and Environm<strong>en</strong>t’ pay for (part<br />
of) the health care services. Together they are the Belgian public (governm<strong>en</strong>tal) payers<br />
for health care. According to the Belgian pharmacoeconomic gui<strong>de</strong>lines costs are<br />
calculated from the perspective of the health care payer, i.e. the pati<strong>en</strong>t plus the public<br />
payer. 18 It inclu<strong>de</strong>s the impact on the governm<strong>en</strong>tal health care budget as well as on the<br />
pati<strong>en</strong>t’s health care exp<strong>en</strong>ditures. This is a reasonab<strong>le</strong> perspective if the objective of<br />
the economic evaluation is to allocate the health care budget effici<strong>en</strong>tly, but it implies<br />
that the health care ‘budget’ is <strong>le</strong>ss fixed as it inclu<strong>de</strong>s a pati<strong>en</strong>t’s co-paym<strong>en</strong>t<br />
compon<strong>en</strong>t. Calculating costs from either the governm<strong>en</strong>t or the pati<strong>en</strong>t’s perspective<br />
alone would not make s<strong>en</strong>se.<br />
Why not? If costs would be calculated from the governm<strong>en</strong>t’s perspective only,<br />
interv<strong>en</strong>tions with limited reimbursem<strong>en</strong>t (i.e. with a limited impact on the health care<br />
budget) will g<strong>en</strong>erally be more cost-effective than fully reimbursed interv<strong>en</strong>tions. The<br />
lower the reimbursem<strong>en</strong>t is, the lower the ICER will be. As a consequ<strong>en</strong>ce, if costs are<br />
calculated from the governm<strong>en</strong>t’s perspective only, interv<strong>en</strong>tions with a low<br />
reimbursem<strong>en</strong>t rate will <strong>le</strong>ad in the ICER <strong>le</strong>ague tab<strong>le</strong>. However, basing reimbursem<strong>en</strong>t<br />
<strong>de</strong>cisions on such a <strong>le</strong>ague tab<strong>le</strong> has ethical consequ<strong>en</strong>ces, as interv<strong>en</strong>tions that would<br />
require a higher reimbursem<strong>en</strong>t, e.g. because they would otherwise have a huge impact<br />
on pati<strong>en</strong>t’s income, would not be reimbursed according to the ICER threshold value<br />
approach. H<strong>en</strong>ce, taking the governm<strong>en</strong>t’s perspective only for the calculation of costs<br />
in the ICER is not useful for resource allocation <strong>de</strong>cisions in a system with pati<strong>en</strong>t copaym<strong>en</strong>ts.<br />
However, taking the broa<strong>de</strong>r perspective of both pati<strong>en</strong>t and governm<strong>en</strong>t as health care<br />
payers is also prob<strong>le</strong>matic for the application of the ICER threshold value approach for<br />
maximising health b<strong>en</strong>efits from a giv<strong>en</strong> fixed budget. Wh<strong>en</strong> the health care payers’<br />
perspective is tak<strong>en</strong>, the threshold value shows how the health care payers’ col<strong>le</strong>ctive<br />
budget should be allocated to obtain maximal health. Who pays, the pati<strong>en</strong>t or the<br />
governm<strong>en</strong>t, is consi<strong>de</strong>red irre<strong>le</strong>vant for the approach. In other words, the approach<br />
does not <strong>de</strong>termine the optimal <strong>le</strong>vel of reimbursem<strong>en</strong>t. This means that the ICER<br />
threshold value approach gives no indication about how to allocate the governm<strong>en</strong>t’s<br />
health care budget in or<strong>de</strong>r to obtain maximal health b<strong>en</strong>efits. From a distributive<br />
justice point of view, however, the <strong>le</strong>vel of reimbursem<strong>en</strong>t might not be irre<strong>le</strong>vant.<br />
Suppose two interv<strong>en</strong>tions cost the same from the health care payers’ perspective but<br />
the new interv<strong>en</strong>tion, that society wants to see reimbursed at 100% for whatever<br />
reason, replaces another interv<strong>en</strong>tion that is curr<strong>en</strong>tly reimbursed at 10%. Th<strong>en</strong> either<br />
the reimbursem<strong>en</strong>t of the new interv<strong>en</strong>tion requires disinvestm<strong>en</strong>t in more other<br />
interv<strong>en</strong>tions or the budget has to be adapted, or the <strong>de</strong>cision to reimburse the new<br />
interv<strong>en</strong>tion at 100% should be revised. But again, we either move away from the ‘fixed<br />
budget’ condition or we take <strong>de</strong>cisions with specific ethical consequ<strong>en</strong>ces.<br />
In conclusion, in a system with pati<strong>en</strong>t co-paym<strong>en</strong>ts, the health care budget is not as<br />
fixed as in a NHS system. This reasoning can be expan<strong>de</strong>d to all economic evaluations<br />
performed from a societal perspective. From a societal point of view, it is hard to see<br />
what the fixed health care budget would be. 47 The rationa<strong>le</strong> for taking a societal<br />
perspective in economic evaluation is that one should strive for an effici<strong>en</strong>t allocation of<br />
resources across all sectors, not only within the health sector. Allocation of the budget<br />
betwe<strong>en</strong> sectors is one e<strong>le</strong>m<strong>en</strong>t of effici<strong>en</strong>t resource use within a society. The ICER<br />
threshold value does not, however, help to <strong>de</strong>fine the appropriate budget for the health<br />
sector.<br />
Thus, the first condition for the ICER threshold value approach is not met in a social<br />
security based system or in a system where ICERs are calculated from a societal<br />
perspective.
<strong>KCE</strong> reports 100 ICER Thresholds 23<br />
2.7.2 Comp<strong>le</strong>te information on costs and effects of all health interv<strong>en</strong>tions<br />
The <strong>de</strong>termination of the ICER threshold value requires, among others, full information<br />
on the costs and consequ<strong>en</strong>ces of all health programmes. However, no sing<strong>le</strong> health<br />
care system, whether NHS, social security or private insurance based, has full<br />
information. As a consequ<strong>en</strong>ce, the real ICER of the marginal interv<strong>en</strong>tion covered by<br />
the budget is unknown. 4, 61, 64, 65 The health care policy maker may h<strong>en</strong>ce be continuously<br />
searching for an ICER threshold value rather than setting one. 61<br />
There are two ways for <strong>de</strong>aling with this prob<strong>le</strong>m of incomp<strong>le</strong>te information. The first is<br />
to make a rough estimate of the value of the threshold. 47 However, because very litt<strong>le</strong><br />
empirical evid<strong>en</strong>ce exists on the value of the ICER threshold, this approach is not<br />
evid<strong>en</strong>ce-based and will pot<strong>en</strong>tially -if the estimate is wrong- not <strong>le</strong>ad to maximal health<br />
from a giv<strong>en</strong> budget. If the estimated ICER threshold is higher than the ‘real’ threshold<br />
value in its theoretical s<strong>en</strong>se, i.e. higher than the ICER of the <strong>le</strong>ast cost-effective<br />
programme still fun<strong>de</strong>d, too many technologies will get a positive recomm<strong>en</strong>dation. 48 To<br />
fund these technologies, funds could have be<strong>en</strong> diverted from other healthcare services<br />
which provi<strong>de</strong>d better value for money. 66 As a result, maximal health gains are not<br />
reached for the giv<strong>en</strong> budget. Wh<strong>en</strong> the threshold value is un<strong>de</strong>restimated, some<br />
interv<strong>en</strong>tions (i.e. those with an ICER betwe<strong>en</strong> the threshold value that is too low and<br />
the real threshold value) that offer value for money are d<strong>en</strong>ied to society. The health<br />
budget is un<strong>de</strong>r-utilised and <strong>le</strong>ss health is gained from the availab<strong>le</strong> budget than could<br />
have be<strong>en</strong> gained.<br />
A second way for <strong>de</strong>aling with incomp<strong>le</strong>te information is to <strong>de</strong>fine the threshold as the<br />
ICER of the interv<strong>en</strong>tion that is most likely to be displaced by the new one. 48 In practice<br />
this would mean that the <strong>de</strong>cision maker should first consi<strong>de</strong>r where the resources for<br />
funding the new interv<strong>en</strong>tion should come from; i.e. the disinvestm<strong>en</strong>ts that will have to<br />
be ma<strong>de</strong> to finance the new interv<strong>en</strong>tion. I<strong>de</strong>ally, this should be the interv<strong>en</strong>tion with<br />
the highest ICER. The interv<strong>en</strong>tion with the highest ICER is, economically, the <strong>le</strong>ast<br />
effici<strong>en</strong>t and therefore the first candidate for disinvestm<strong>en</strong>t. If this interv<strong>en</strong>tion cannot<br />
be id<strong>en</strong>tified, the ICER of the interv<strong>en</strong>tion in which the disinvestm<strong>en</strong>t can be done<br />
should be used as the ICER threshold value against which the ICER of the new<br />
interv<strong>en</strong>tion is compared. Only if the ICER of the new interv<strong>en</strong>tion is lower than the<br />
ICER of the interv<strong>en</strong>tion that is being replaced, funding the interv<strong>en</strong>tion increases<br />
effici<strong>en</strong>cy in health care. If, in practice, the ICER of the interv<strong>en</strong>tion that will be replaced<br />
is at that time unknown, it should be calculated. Otherwise the <strong>de</strong>cision might be wrong<br />
from an effici<strong>en</strong>cy point of view.<br />
To illustrate this with an examp<strong>le</strong>, suppose a new interv<strong>en</strong>tion emerges for the<br />
treatm<strong>en</strong>t of Alzheimer disease and suppose that for the imp<strong>le</strong>m<strong>en</strong>tation of this<br />
interv<strong>en</strong>tion resources will be tak<strong>en</strong> away from a treatm<strong>en</strong>t for chronic low back pain.<br />
The ICER threshold value against which the ICER of the Alzheimer interv<strong>en</strong>tion should<br />
be compared is th<strong>en</strong> the ICER of the chronic low back pain treatm<strong>en</strong>t. This approach is<br />
useful only if the <strong>de</strong>cision maker takes the a priori position that financing of the<br />
Alzheimer disease treatm<strong>en</strong>t should come from disinvestm<strong>en</strong>ts in the chronic low back<br />
pain interv<strong>en</strong>tion. If afterwards another <strong>de</strong>cision with respect to disinvestm<strong>en</strong>t is tak<strong>en</strong>,<br />
the threshold was wrong and the investm<strong>en</strong>t <strong>de</strong>cision should be re-consi<strong>de</strong>red in the<br />
light of the ICER of the interv<strong>en</strong>tion that will actually be displaced. In a real-life <strong>de</strong>cision<br />
making context, however, this exercise would rarely be ma<strong>de</strong>. Decisions about the<br />
reimbursem<strong>en</strong>t of health interv<strong>en</strong>tions are mainly ma<strong>de</strong> on a case by case basis. 67
24 ICER Thresholds <strong>KCE</strong> Reports 100<br />
2.7.3 Perfect divisibility and constant returns to sca<strong>le</strong><br />
The condition of perfect divisibility and constant returns to sca<strong>le</strong> is highly theoretical<br />
46, 47, 68<br />
and never fulfil<strong>le</strong>d in real life.<br />
Perfect divisibility means that health care programmes can be ‘bought’ or downgra<strong>de</strong>d<br />
to whatever ext<strong>en</strong>t. This might not be the case, however, especially wh<strong>en</strong> a health<br />
programme requires high investm<strong>en</strong>t costs (e.g. an additional MRI or PET scanner, the<br />
building of additional premises). 69 For examp<strong>le</strong>, to provi<strong>de</strong> a diagnostic work-up<br />
involving PET to one pati<strong>en</strong>t, the full investm<strong>en</strong>t of a PET scanner has to be ma<strong>de</strong>. The<br />
cost per pati<strong>en</strong>t <strong>de</strong>creases as more pati<strong>en</strong>ts are inclu<strong>de</strong>d in the diagnostic work-up<br />
programme and there will be an optimal occupation rate for the PET scanner that will<br />
minimise the cost per pati<strong>en</strong>t and h<strong>en</strong>ce the ICER. The programme involving PET is<br />
therefore not perfectly divisib<strong>le</strong> without changing the ICER of the programme. Whi<strong>le</strong><br />
the ICER threshold approach assumes that all activities that are worth doing can be<br />
done in each and every volume, it is unlikely that this is the case in real life for all<br />
interv<strong>en</strong>tions. m 47 The examp<strong>le</strong> giv<strong>en</strong> might be an extreme case, but neverthe<strong>le</strong>ss<br />
illustrates the issue. Many health interv<strong>en</strong>tions do not require huge investm<strong>en</strong>ts, for<br />
examp<strong>le</strong> drug treatm<strong>en</strong>ts. The critique with respect to the abs<strong>en</strong>ce of perfect divisibility<br />
of sca<strong>le</strong> is not re<strong>le</strong>vant for these interv<strong>en</strong>tions. However, ev<strong>en</strong> in these cases it is<br />
unlikely that the imp<strong>le</strong>m<strong>en</strong>tation of only part of a health programme will not affect the<br />
ICER of that programme. This relates to the second assumption of constant returns to<br />
sca<strong>le</strong>.<br />
Constant returns to sca<strong>le</strong> means that the costs and health b<strong>en</strong>efits of a health<br />
programme are reduced (increased) proportionally to the reduction (increm<strong>en</strong>t) of the<br />
ext<strong>en</strong>t of the programme. 47 Consequ<strong>en</strong>tly, the ICER is assumed to remain constant<br />
wh<strong>en</strong> a health programme is reduced or increased. This assumption does not hold in<br />
cases where the fixed costs of a programme are high or where the increm<strong>en</strong>tal<br />
effectiv<strong>en</strong>ess does not change proportionally with the number of pati<strong>en</strong>ts treated. The<br />
higher the fixed costs, the <strong>le</strong>ss proportional the <strong>de</strong>crease in total cost with a <strong>de</strong>crease<br />
in the number of pati<strong>en</strong>ts treated and h<strong>en</strong>ce the higher the ICER becomes (if<br />
effectiv<strong>en</strong>ess is not affected, see earlier examp<strong>le</strong> of the PET scan). A reduction in the<br />
number of pati<strong>en</strong>ts treated might not proportionally <strong>de</strong>crease the effectiv<strong>en</strong>ess of the<br />
programme. For examp<strong>le</strong>, a school vaccination programme for HPV will be more<br />
effective in reducing the incid<strong>en</strong>ce of HPV than opportunistic vaccination because the<br />
effect of the vaccination exceeds the direct effect on the individual. 71 As a consequ<strong>en</strong>ce,<br />
the ICER of vaccination will <strong>de</strong>crease if more childr<strong>en</strong> are vaccinated because the<br />
effectiv<strong>en</strong>ess increases more than proportionally with the costs if the number of<br />
vaccinated childr<strong>en</strong> increases.<br />
Moreover, the ICERs may differ betwe<strong>en</strong> sub-groups of pati<strong>en</strong>ts. 46 Whi<strong>le</strong> the overall<br />
ICER of a health programme may be unfavourab<strong>le</strong> according to the threshold approach,<br />
it might well be the case that the ICER of the same programme is below the threshold<br />
value for a specific sub-group of pati<strong>en</strong>ts, e.g. because the programme is more effective<br />
for specified sub-populations. In that case, it would be effici<strong>en</strong>t to allocate resources to<br />
this specific sub-group but it would not be effici<strong>en</strong>t to imp<strong>le</strong>m<strong>en</strong>t the <strong>en</strong>tire health<br />
programme. Inability to make this distinction may <strong>le</strong>ad to ineffici<strong>en</strong>t resource allocation. n<br />
m Mathematical solutions have be<strong>en</strong> <strong>de</strong>veloped to address the prob<strong>le</strong>m of assumed constant returns to<br />
sca<strong>le</strong> and divisibility.70 The practical applicability of the techniques is yet limited because of the high data<br />
requirem<strong>en</strong>ts. The data nee<strong>de</strong>d are oft<strong>en</strong> not availab<strong>le</strong>.<br />
n Commercial companies may try to close an ‘all or nothing’ <strong>de</strong>al to cover the <strong>en</strong>tire population although<br />
the interv<strong>en</strong>tion is only cost-effective for a specific sub-group.72 The requested price is a weighted<br />
average of the prices that would r<strong>en</strong><strong>de</strong>r the interv<strong>en</strong>tion cost-effective in the differ<strong>en</strong>t sub-groups. In the<br />
sub-group where the interv<strong>en</strong>tion is most effective, the price can be relatively high if the interv<strong>en</strong>tion is<br />
to remain un<strong>de</strong>r the threshold value. In sub-groups where the interv<strong>en</strong>tion is <strong>le</strong>ss effective, the price will<br />
have to be lower to remain un<strong>de</strong>r the threshold value. At the ‘average’ price, the interv<strong>en</strong>tion will be<br />
cost-effective (below the threshold value) for some sub-groups but not for others. ‘All or nothing’ <strong>de</strong>als<br />
h<strong>en</strong>ce <strong>le</strong>ad to sub-optimal resource allocation, as in princip<strong>le</strong> the interv<strong>en</strong>tion should only be reimbursed<br />
for the sub-group in which the interv<strong>en</strong>tion is cost-effective at that price. Claxton et al. (2007) therefore<br />
argue that such <strong>de</strong>als should be rejected. They argue that the price should not be higher than the price<br />
that r<strong>en</strong><strong>de</strong>rs the interv<strong>en</strong>tion cost-effective in the sub-group with the highest ICER that is still consi<strong>de</strong>red
<strong>KCE</strong> reports 100 ICER Thresholds 25<br />
2.7.4 Health programmes are in<strong>de</strong>p<strong>en</strong>d<strong>en</strong>t from one another<br />
The assumption that health programmes are in<strong>de</strong>p<strong>en</strong>d<strong>en</strong>t from one another does not<br />
hold in real life. In economic evaluation, a health interv<strong>en</strong>tion is never looked at in<br />
isolation. For examp<strong>le</strong>, the costs associated with the implantation of a <strong>de</strong>vice (e.g. a<br />
coronary st<strong>en</strong>t) exceeds the pure cost of the <strong>de</strong>vice, as pati<strong>en</strong>ts will have to go to<br />
hospital, un<strong>de</strong>rgo diagnostic procedures before the <strong>de</strong>cision to implant the <strong>de</strong>vice is<br />
tak<strong>en</strong>, may have to follow an additional medication treatm<strong>en</strong>t after the interv<strong>en</strong>tion etc.<br />
H<strong>en</strong>ce, the ICER of the <strong>de</strong>vice cannot be reduced to the pure costs and effects of the<br />
<strong>de</strong>vice but also <strong>de</strong>p<strong>en</strong>ds on the costs and effects of the diagnostic procedures, the drug<br />
treatm<strong>en</strong>t etc.<br />
For the application of the ICER threshold value approach this might be prob<strong>le</strong>matic.<br />
Suppose in the previous examp<strong>le</strong> that the drug treatm<strong>en</strong>t (not prece<strong>de</strong>d by the <strong>de</strong>vice<br />
implant) is not reimbursed because its ICER is higher than the ICER threshold value. If<br />
the ICER of the interv<strong>en</strong>tion with the <strong>de</strong>vice, but including the drug treatm<strong>en</strong>t, is lower<br />
than the ICER threshold value, what should the <strong>de</strong>cision be? According to the ICER<br />
threshold value approach, the <strong>de</strong>vice should be reimbursed. The interv<strong>en</strong>tion with the<br />
<strong>de</strong>vice is only cost-effective, however, because it is followed by the drug treatm<strong>en</strong>t.<br />
Therefore, the reimbursem<strong>en</strong>t of the <strong>de</strong>vice cannot be <strong>de</strong>ci<strong>de</strong>d without reconsi<strong>de</strong>ration<br />
of the reimbursem<strong>en</strong>t of the drug treatm<strong>en</strong>t.<br />
This conclusion has implications for health care systems characterised by separate<br />
budgets for differ<strong>en</strong>t sub-sectors in the health care sector, e.g. for pharmaceuticals, for<br />
<strong>de</strong>vices and implants, for physician fees, etc. This is the case in many countries, including<br />
Belgium. Wh<strong>en</strong> pursuing effici<strong>en</strong>t resource allocation in the health care sector it is<br />
impossib<strong>le</strong> to stay within the rationa<strong>le</strong> of separate budgets. As <strong>de</strong>monstrated before,<br />
due to <strong>de</strong>p<strong>en</strong>d<strong>en</strong>cies betwe<strong>en</strong> interv<strong>en</strong>tions that are paid for out of differ<strong>en</strong>t health care<br />
sub-budgets, it does not make s<strong>en</strong>se to look at the interv<strong>en</strong>tions separately and<br />
consi<strong>de</strong>r only the impact on one particular sub-budget. Economic evaluation from the<br />
perspective of one of the sub-budgets only would not give an accurate i<strong>de</strong>a of the real<br />
impact of the interv<strong>en</strong>tions on the health care costs and effects and would h<strong>en</strong>ce not be<br />
useful for the evaluation of allocative effici<strong>en</strong>cy in health care.<br />
2.7.5 Health maximisation as the so<strong>le</strong> goal of health policy makers<br />
The ‘economic effici<strong>en</strong>cy in production’-argum<strong>en</strong>t for the use of ICER threshold values,<br />
or health maximisation (in terms of QALYs or LYG) as the primary aim of health care<br />
<strong>de</strong>cision making, might not a<strong>de</strong>quately ref<strong>le</strong>ct the reasons for <strong>de</strong>cisions about resource<br />
allocation in health care in real life. This applies to both NHS and social security-based<br />
health care systems. There is a large body of literature on distributional concerns in<br />
resource allocation based on CEA. They ess<strong>en</strong>tially provi<strong>de</strong> an argum<strong>en</strong>t for an extrawelfarist<br />
approach, where resource allocation <strong>de</strong>cisions take the relative societal value<br />
of health gains for differ<strong>en</strong>t population groups into account. 9-11, 20, 44, 57, 65, 74-92 Much of the<br />
discussion is related to the health outcome measures used in economic evaluation.<br />
QALYs, for instance, as other outcome measures, typically ignore societal prefer<strong>en</strong>ces<br />
for distributional aspects, such as prefer<strong>en</strong>ces related to the number of peop<strong>le</strong> receiving<br />
treatm<strong>en</strong>t (more pati<strong>en</strong>ts receiving QALYs versus fewer pati<strong>en</strong>ts) and prefer<strong>en</strong>ces<br />
related to the personal characteristics of the individuals receiving treatm<strong>en</strong>t (<strong>le</strong>vel of<br />
44, 76<br />
severity of the condition).<br />
In a NHS system where the budget is mainly fixed, health maximisation will of course<br />
not be pursued at whatever cost in terms of equity. Society has prefer<strong>en</strong>ces with<br />
respect to the allocation of health gains, that have to be tak<strong>en</strong> into account in the health<br />
care <strong>de</strong>cision making process. 44<br />
acceptab<strong>le</strong>.72 Other authors have argued that appropriation of the social surplus of an innovation to<br />
producers is c<strong>en</strong>tral to the dynamic effici<strong>en</strong>cy in health care (i.e. to <strong>en</strong>sure continuing effici<strong>en</strong>t R&D<br />
investm<strong>en</strong>ts) and is therefore justified.73
26 ICER Thresholds <strong>KCE</strong> Reports 100<br />
The same applies to a social security system, be it that the ICER threshold value<br />
approach would in this system not only impact on health inequality but also on income<br />
inequality. It is g<strong>en</strong>erally acknow<strong>le</strong>dged that <strong>de</strong>cision makers take other aspects into<br />
account besi<strong>de</strong>s the ICER of an interv<strong>en</strong>tion, to <strong>de</strong>ci<strong>de</strong> whether or not the interv<strong>en</strong>tion<br />
is worth its cost. Because these factors differ across interv<strong>en</strong>tions, a sing<strong>le</strong> threshold<br />
value for the ICER below which an interv<strong>en</strong>tion is consi<strong>de</strong>red value for money, is not<br />
consist<strong>en</strong>t with how society chooses to make <strong>de</strong>cisions (see also chapter 3).<br />
If it would be possib<strong>le</strong>, however, to inclu<strong>de</strong> these additional consi<strong>de</strong>rations in the ICER,<br />
e.g. by weighting the QALYs of populations the society wants to protect more heavily,<br />
the ICER threshold value approach might still be applicab<strong>le</strong>. 75 The ICER threshold value<br />
would in this case be <strong>de</strong>fined in terms of a cost-per-weighted QALY. The objective is no<br />
longer ‘health outcome maximisation’ but ‘weighted health outcome maximisation’. This<br />
weighted QALY approach has three pot<strong>en</strong>tial drawbacks. First, a prerequisite for this<br />
approach is transpar<strong>en</strong>cy of the composition of the weights for QALYs: which criteria<br />
are <strong>de</strong>terminant, what value is assigned to each <strong>de</strong>terminant and how are these values<br />
combined to obtain a unique weight? Obviously, this is not an easy requirem<strong>en</strong>t.<br />
A second pot<strong>en</strong>tial prob<strong>le</strong>m of including additional <strong>de</strong>cision criteria in the costeffectiv<strong>en</strong>ess<br />
ratio is the se<strong>le</strong>ction of the appropriate cost-effective comparator. If<br />
<strong>de</strong>cisions are not purely inspired by a pursuit of maximal health, it might happ<strong>en</strong> that<br />
interv<strong>en</strong>tions that are not cost-effective according to the theoretical ICER threshold<br />
value approach (without QALY weighting) are neverthe<strong>le</strong>ss reimbursed. For examp<strong>le</strong>,<br />
suppose that, <strong>de</strong>spite a high ICER, a specific interv<strong>en</strong>tion is reimbursed because there is<br />
no alternative treatm<strong>en</strong>t for treating a specific serious disease and pati<strong>en</strong>ts would<br />
otherwise be <strong>le</strong>ft untreated. Suppose that the conv<strong>en</strong>tional ICER of this programme is<br />
higher than the ICER threshold value that would imply maximal health outcomes but<br />
that the outcomes have be<strong>en</strong> giv<strong>en</strong> a higher weight in or<strong>de</strong>r to stay below the ICER<br />
threshold value. If after the <strong>de</strong>cision is tak<strong>en</strong> a new interv<strong>en</strong>tion for this pati<strong>en</strong>t<br />
population is <strong>de</strong>veloped, it might have a low ICER wh<strong>en</strong> the existing treatm<strong>en</strong>t is used<br />
as the comparator in the CEA. Its ICER might be below the ICER threshold value and<br />
h<strong>en</strong>ce it might be conclu<strong>de</strong>d that it is cost-effective. However, the existing reimbursed<br />
treatm<strong>en</strong>t might not be the re<strong>le</strong>vant comparator, as the reason for initial<br />
reimbursem<strong>en</strong>t (i.e. non-exist<strong>en</strong>ce of an alternative treatm<strong>en</strong>t for the pati<strong>en</strong>ts) no<br />
longer holds. Previous <strong>de</strong>cision might have to be revised in the light of new<br />
<strong>de</strong>velopm<strong>en</strong>ts. This <strong>le</strong>ads to the conclusion that including additional consi<strong>de</strong>rations in<br />
the ICER and including such ‘weighted’ ICER subsequ<strong>en</strong>tly in a <strong>le</strong>ague tab<strong>le</strong> might<br />
complicate their interpretation and their practical usability. Moreover, the risk for<br />
misuse or errors in the choice of the appropriate comparator increases.<br />
A third weakness of the weighted QALY approach, if used in combination with a<br />
threshold value, is the remaining requirem<strong>en</strong>ts of perfect divisibility of health<br />
programmes and constant returns to sca<strong>le</strong>, two requirem<strong>en</strong>ts that may not hold in real<br />
life (see 2.7.3).<br />
A few attempts have be<strong>en</strong> ma<strong>de</strong> to <strong>de</strong>rive an ICER threshold value from past health<br />
policy <strong>de</strong>cisions. 65, 93-95 The exercises showed that in<strong>de</strong>ed there is no sing<strong>le</strong> threshold<br />
value above which the <strong>de</strong>cision is always negative and below which it is always positive.<br />
Rather, a range of acceptab<strong>le</strong> ICERs has be<strong>en</strong> id<strong>en</strong>tified. This can mean differ<strong>en</strong>t things:<br />
(1) the <strong>de</strong>cision maker does not know the true ICER threshold value that would<br />
maximise health b<strong>en</strong>efits from a giv<strong>en</strong> budget; 61, 64, 96 (2) other consi<strong>de</strong>rations than health<br />
maximisation <strong>de</strong>termine the acceptability of an interv<strong>en</strong>tion with an ICER that is, strictly<br />
speaking, above the ICER threshold value 65, 67 , (3) differ<strong>en</strong>t methods are used to obtain<br />
the ICER estimates as a consequ<strong>en</strong>ce of which they are not always comparab<strong>le</strong>, (4) the<br />
<strong>le</strong>vel of uncertainty around the ICER estimates <strong>de</strong>termines their acceptability and (5)<br />
<strong>de</strong>cision makers do not ‘trust’ all ICER estimates to the same ext<strong>en</strong>t. The differ<strong>en</strong>t<br />
reasons probably all apply to some ext<strong>en</strong>t. 85 An empirically id<strong>en</strong>tified range of ICER<br />
threshold values should therefore be interpreted as the range of societal willingness to<br />
pay for an additional QALY or LYG at that time, in that specific budgetary and societal<br />
context and for those specific interv<strong>en</strong>tions rather than as an ICER threshold value in<br />
65, 96<br />
the purely theoretical meaning of an absolute criterion for health maximisation.
<strong>KCE</strong> reports 100 ICER Thresholds 27<br />
2.7.6 Additional caveats<br />
There is a serious risk of bias towards the ICER threshold value, if one is <strong>de</strong>fined. 97<br />
Once a threshold value is set, there is the danger that ICERs of new technologies will<br />
converge towards this threshold value by inducing commercial companies to adapt their<br />
prices in or<strong>de</strong>r to ‘satisfy’ the cost-effectiv<strong>en</strong>ess criterion (ICER
28 ICER Thresholds <strong>KCE</strong> Reports 100<br />
2.8 ALTERNATIVES TO ICERS AND ICER THRESHOLD<br />
VALUES<br />
As discussed in the previous chapter, the ICER threshold value approach is based on a<br />
number of assumptions that are highly theoretical. Because few practical solutions exist<br />
to overcome these issues, the id<strong>en</strong>tification and application of an ICER threshold value<br />
in its neo-classical welfarist meaning appears to be impossib<strong>le</strong>.<br />
Alternatives to the theoretical ICER threshold value have be<strong>en</strong> suggested in literature.<br />
They differ in the ext<strong>en</strong>t to which they follow the logic of the CEA, ICERs and ICER<br />
threshold values.<br />
Three lines of thought can be id<strong>en</strong>tified:<br />
1. those who suggest an alternative <strong>de</strong>finition for the ICER threshold value but<br />
stick to the princip<strong>le</strong> that an ICER threshold value should gui<strong>de</strong> health care<br />
policy <strong>de</strong>cisions<br />
2. those who abandon the i<strong>de</strong>a of an ICER threshold value but still use ICERs to<br />
support health care policy <strong>de</strong>cisions<br />
3. those who abandon the i<strong>de</strong>a of ICERs and CEA and suggest an alternative<br />
approach to bring economic consi<strong>de</strong>rations into the health-care <strong>de</strong>cision<br />
making process.<br />
This chapter gives a brief overview of the alternatives to the ICER threshold value<br />
un<strong>de</strong>r a fixed budget constraint <strong>de</strong>scribed in section 2.5. Whi<strong>le</strong> the alternatives are<br />
pres<strong>en</strong>ted here as stand-alone approaches, it should be appreciated that combinations<br />
of these approaches are possib<strong>le</strong> and are being examined. But, because of the particular<br />
scope and educational purpose of this report we ma<strong>de</strong> a c<strong>le</strong>ar distinction betwe<strong>en</strong> the<br />
approaches. The proposed alternatives are classified in Tab<strong>le</strong> 2 according to their <strong>le</strong>vel<br />
of acceptance of the ICER and a sing<strong>le</strong> ICER threshold value as a <strong>de</strong>cision criterion.<br />
Tab<strong>le</strong> 2: Classification of alternatives to the use of an ICER threshold value<br />
in its theoretical meaning according to their <strong>le</strong>vel of acceptance of ICERs<br />
and a sing<strong>le</strong> ICER threshold value as a <strong>de</strong>cision criterion in health care policy<br />
<strong>de</strong>cisions.<br />
Alternative Acceptance<br />
of ICER<br />
ICER threshold value<br />
as societal willingness<br />
to pay<br />
Comparison with past<br />
<strong>de</strong>cisions<br />
ICER as one e<strong>le</strong>m<strong>en</strong>t<br />
weighed against other<br />
e<strong>le</strong>m<strong>en</strong>ts in the<br />
<strong>de</strong>cision making<br />
process<br />
Average GDP per<br />
capita as a threshold<br />
value for average<br />
cost-effectiv<strong>en</strong>ess<br />
Opportunity cost<br />
approach<br />
Cost-consequ<strong>en</strong>ces<br />
analysis<br />
Acceptance of a<br />
sing<strong>le</strong> ICER threshold<br />
value<br />
YES Sing<strong>le</strong> or multip<strong>le</strong> ICER<br />
threshold values possib<strong>le</strong><br />
Paragraph Refer<strong>en</strong>ces<br />
2.8.1<br />
12, 56, 83<br />
YES YES 2.8.2 65, 93, 94<br />
YES Possib<strong>le</strong> but not<br />
nece©ssary<br />
CER, not<br />
ICER<br />
2.8.3<br />
YES 2.8.4 99<br />
8, 65, 67<br />
NO NO 2.8.5 63, 98<br />
NO NO 2.8.6 77
<strong>KCE</strong> reports 100 ICER Thresholds 29<br />
2.8.1 The ICER threshold value as a ref<strong>le</strong>ction of societal willingness to pay<br />
The concept of an ICER threshold value as <strong>de</strong>scribed in 2.5 is used to gui<strong>de</strong> <strong>de</strong>cision<br />
makers towards a health-maximising health care resource allocation giv<strong>en</strong> a fixed health<br />
care budget.<br />
If not <strong>de</strong>fined as the <strong>le</strong>ast cost-effective interv<strong>en</strong>tion still financed from a fixed health<br />
care budget, the ICER threshold value could be <strong>de</strong>fined as the maximum societal<br />
willingness to pay (WTP) for an additional QALY (or LYG). 12, 56, 83 The societal WTP for<br />
an additional QALY (or LYG) is <strong>de</strong>termined by the relative value of an extra QALY (or<br />
LYG) compared to the value of the b<strong>en</strong>efits g<strong>en</strong>erated in other sectors. o100<br />
The more b<strong>en</strong>efits from other sectors that the society is willing to give up for additional<br />
health, the higher the implied societal WTP for a QALY (or LYG) is.<br />
The societal WTP approach avoids the need for full information on the costs and health<br />
outcomes of all interv<strong>en</strong>tions, and would allow the evaluation one by one of every new<br />
interv<strong>en</strong>tion consi<strong>de</strong>red for funding. However, <strong>de</strong>fining the ICER threshold value like<br />
this has a number of implications and weaknesses, <strong>de</strong>p<strong>en</strong>ding on how it would be used.<br />
Two possibilities are consi<strong>de</strong>red:<br />
• either a g<strong>en</strong>eric ICER threshold value (WTP for a QALY) is applied to all<br />
new health programmes consi<strong>de</strong>red for funding, 12 or<br />
• the societal WTP for a QALY is reconsi<strong>de</strong>red for each new interv<strong>en</strong>tion<br />
consi<strong>de</strong>red for funding or for groups of interv<strong>en</strong>tions/conditions that are<br />
comparab<strong>le</strong> in terms of their characteristics that <strong>de</strong>termine societal<br />
WTP. 101<br />
Using the societal WTP for a QALY as the ICER threshold value is incompatib<strong>le</strong> with a<br />
fixed budget system. As argued in 2.7.1, fixed ICER threshold value requires a f<strong>le</strong>xib<strong>le</strong><br />
budget. p The measurem<strong>en</strong>t of the societal WTP for a (g<strong>en</strong>eric) QALY (or LYG) poses a<br />
number of methodological prob<strong>le</strong>ms and it is doubtful that a g<strong>en</strong>eric societal WTP value<br />
applicab<strong>le</strong> to all kinds of health programmes exists. In<strong>de</strong>ed, empirical studies suggest<br />
that the ICER threshold values oft<strong>en</strong> proposed in literature are lower than the actual<br />
WTP for a QALY, 56 whi<strong>le</strong> others find the opposite result. 83q The societal WTP for a<br />
QALY (or LYG) is always context-<strong>de</strong>p<strong>en</strong>d<strong>en</strong>t. It is hard to imagine the value of a life<br />
year, making abstraction of the person and his characteristics (curr<strong>en</strong>t health status, age,<br />
etc). In addition, appropriate measurem<strong>en</strong>t of WTP requires that respond<strong>en</strong>ts have to<br />
make tra<strong>de</strong>-offs and are aware that the value they place on a QALY (or LYG) has<br />
implications for the consumption of other goods and services (i.e. opportunity costs). If<br />
not, unrealistic and impractical values may be measured.<br />
An increasing amount of literature in health economics focuses on the incorporation of<br />
equity consi<strong>de</strong>rations in the ICER to overcome the prob<strong>le</strong>m of the previously <strong>de</strong>scribed<br />
approach that it does not take societal prefer<strong>en</strong>ces with respect to the distribution of<br />
health gains into acocount. 75, 81, 89, 90 This has be<strong>en</strong> addressed previously in section 2.7.5. r<br />
o The maximum societal willingness to pay for an additional QALY (or LYG) is the amount of “wealth”, in<br />
terms of b<strong>en</strong>efits from other sectors, society is willing to give up to obtain an additional QALY (or LYG).<br />
The health care budget is optimal from a societal point of view if the ICER of the <strong>le</strong>ast cost-effective<br />
interv<strong>en</strong>tion still financed from the health care budget is equal to the societal WTP for a QALY gained<br />
(or LYG). As long as society is willing to give up b<strong>en</strong>efits from other sectors to obtain additional b<strong>en</strong>efits<br />
in the health care sector (i.e. the value of the b<strong>en</strong>efits foregone in other sectors is lower than the value of<br />
the b<strong>en</strong>efits obtained in the health care sector), the budget should expand.<br />
p De facto this means that the maximum WTP for health gains will <strong>de</strong>termine the health care budget. The<br />
health care budget thus obtained is the optimal budget from a societal point of view because society<br />
would not be willing to tra<strong>de</strong> health for other b<strong>en</strong>efits in other sectors. Note that for an optimal budget<br />
from a societal point of view, the societal WTP approach gives the same results as the ICER threshold<br />
value approach if all other conditions are fulfil<strong>le</strong>d.<br />
q Note that the results of WTP studies <strong>de</strong>p<strong>en</strong>d heavily on the methods used to measure WTP. Differ<strong>en</strong>t<br />
methods yield differ<strong>en</strong>t results. As there is no gold-standard, it is difficult to assess the validity of the<br />
results.<br />
r This is an extra-welfarist approach.
30 ICER Thresholds <strong>KCE</strong> Reports 100<br />
A way to achieve the incorporation of equity consi<strong>de</strong>rations in the ICER metric is by<br />
weighting QALYs (or LYG). The weights assigned to the QALYs (or LYG) of specific<br />
population groups should ref<strong>le</strong>ct societal prefer<strong>en</strong>ces for the allocation of QALYs to<br />
these groups. H<strong>en</strong>ce, QALYs gained by pati<strong>en</strong>ts the society wishes to favour are valued<br />
higher than QALYs gained by pati<strong>en</strong>ts society does not want to favour. If a<strong>de</strong>quate<br />
weights could be <strong>de</strong>fined, ref<strong>le</strong>cting all societal equity concerns (both in terms of health<br />
and income distribution in case of a mixed public-private system, see 2.7.5), the ICER<br />
threshold value could be <strong>de</strong>fined as the societal WTP for a weighted QALY and the<br />
threshold value approach could again be applied, be it with the necessary caveats as<br />
pres<strong>en</strong>ted in 2.7.5. Its advantage would be increased transpar<strong>en</strong>cy of the <strong>de</strong>cision<br />
making process if the <strong>de</strong>termination of the weights could be transpar<strong>en</strong>t. The major<br />
prob<strong>le</strong>m is id<strong>en</strong>tifying all re<strong>le</strong>vant parameters from a societal point of view and<br />
measuring the weights. Differ<strong>en</strong>t empirical studies have <strong>de</strong>monstrated that peop<strong>le</strong><br />
76, 87, 92<br />
in<strong>de</strong>ed do find equity concerns important for resource allocation <strong>de</strong>cisions.<br />
Actual weighting of QALYs (or LYG) has, however, not yet be<strong>en</strong> imp<strong>le</strong>m<strong>en</strong>ted in<br />
routine CEA. It can moreover be argued that, ev<strong>en</strong> if it would be feasib<strong>le</strong> to a<strong>de</strong>quately<br />
capture and quantify equity (and other) concerns, certain dynamics in the <strong>de</strong>cision<br />
making process will remain implicit and variab<strong>le</strong> across health interv<strong>en</strong>tions, such as for<br />
instance the influ<strong>en</strong>ce of stakehol<strong>de</strong>rs.<br />
Rather than trying to <strong>de</strong>fine one g<strong>en</strong>eric ICER threshold value repres<strong>en</strong>ting the WTP of<br />
a (weighted) QALY (or LYG) in g<strong>en</strong>eral, the maximum WTP for a QALY (or LYG)<br />
could be ma<strong>de</strong> <strong>de</strong>p<strong>en</strong>d<strong>en</strong>t on specific characteristics of the interv<strong>en</strong>tion or the<br />
population (e.g. severity of disease, availability of an alternative treatm<strong>en</strong>t). For instance,<br />
WTP might be higher for interv<strong>en</strong>tions that reduce mortality risk than for interv<strong>en</strong>tions<br />
that improve quality of life, as already shown in empirical studies. 101 In its most extreme<br />
form, societal WTP could be re-assessed on a case-by-case basis for every individual<br />
interv<strong>en</strong>tion consi<strong>de</strong>red for funding. This approach requires a more f<strong>le</strong>xib<strong>le</strong> budget, as<br />
the budget will have to be adapted to the societal WTP for the health outcomes<br />
g<strong>en</strong>erated by each new interv<strong>en</strong>tion that is consi<strong>de</strong>red worthwhi<strong>le</strong>. Therefore, it is<br />
more difficult to apply this approach in an NHS based system. The advantage of re<strong>de</strong>fining<br />
a WTP for each interv<strong>en</strong>tion is that it allows taking all objectives of health care<br />
policy into account. Moreover, it does not require the additional theoretical<br />
assumptions of the ICER threshold value approach that have prov<strong>en</strong> to be prob<strong>le</strong>matic<br />
in real life. The price of this increased f<strong>le</strong>xibility is a pot<strong>en</strong>tial reduction in transpar<strong>en</strong>cy.<br />
The more room there is <strong>le</strong>ft for <strong>de</strong>viation from some kind of ‘ru<strong>le</strong> of thumb’, the <strong>le</strong>ss<br />
transpar<strong>en</strong>t <strong>de</strong>cisions become.<br />
2.8.2 Comparison with past <strong>de</strong>cisions<br />
One suggested way to id<strong>en</strong>tify the societal WTP for a QALY is to look at the ICERs of<br />
interv<strong>en</strong>tions for which a <strong>de</strong>cision has already be<strong>en</strong> ma<strong>de</strong> in the past. This <strong>le</strong>ads to a<br />
kind of restricted <strong>le</strong>ague tab<strong>le</strong> that could be used to <strong>de</strong>termine the relative position of<br />
interv<strong>en</strong>tions towards previously accepted or rejected interv<strong>en</strong>tions. However, as<br />
<strong>de</strong>cisions are rarely ma<strong>de</strong> on the basis of cost-effectiv<strong>en</strong>ess consi<strong>de</strong>rations alone, 98<br />
ICERs of interv<strong>en</strong>tions for which a positive or negative <strong>de</strong>cision has be<strong>en</strong> ma<strong>de</strong> in the<br />
past should always be consi<strong>de</strong>red along with all their argum<strong>en</strong>ts for the positive or<br />
negative recomm<strong>en</strong>dation if they are used for comparative purposes in curr<strong>en</strong>t <strong>de</strong>cision<br />
making processes. This evaluation might <strong>le</strong>ad to the conclusion that the <strong>de</strong>cision ma<strong>de</strong><br />
at that mom<strong>en</strong>t was actually not the optimal <strong>de</strong>cision and would maybe not have be<strong>en</strong><br />
ma<strong>de</strong> curr<strong>en</strong>tly. This complicates ev<strong>en</strong> further the feasibility of comparison with past<br />
<strong>de</strong>cisions. Moreover, comparison with ICERs calculated in the past is only warranted if<br />
the ICERs are obtained in the same way, i.e. using the same methodology, and un<strong>de</strong>r the<br />
same conditions, i.e. costs, existing technologies, experi<strong>en</strong>ce etc. Conditions change,<br />
however, as a consequ<strong>en</strong>ce of which this requirem<strong>en</strong>t is rarely fulfil<strong>le</strong>d.
<strong>KCE</strong> reports 100 ICER Thresholds 31<br />
2.8.3 Weighing the ICER against other <strong>de</strong>cision criteria in the <strong>de</strong>cision making<br />
process<br />
Policy <strong>de</strong>cisions about health care technologies are not tak<strong>en</strong> without consi<strong>de</strong>ration of<br />
e<strong>le</strong>m<strong>en</strong>ts and aspects beyond cost-effectiv<strong>en</strong>ess. There are differ<strong>en</strong>t ways to <strong>de</strong>al with<br />
this:<br />
1. either the additional e<strong>le</strong>m<strong>en</strong>ts are ma<strong>de</strong> explicit, measured or objectified and<br />
explicitly weighed in the <strong>de</strong>cision making process, 8, 57, 65, 102-105 or<br />
2. the additional e<strong>le</strong>m<strong>en</strong>ts are tak<strong>en</strong> implicitly into account in the <strong>de</strong>cision making<br />
process. 67<br />
The first approach assumes that all e<strong>le</strong>m<strong>en</strong>ts can be measured or objectified. The<br />
subsequ<strong>en</strong>t weighing of the e<strong>le</strong>m<strong>en</strong>ts in the <strong>de</strong>cision making process can be done in<br />
differ<strong>en</strong>t <strong>de</strong>grees of explicitness. One extreme is to remain implicit about the actual<br />
weight of each of the additional e<strong>le</strong>m<strong>en</strong>ts and <strong>le</strong>t the result <strong>de</strong>p<strong>en</strong>d on the discussions<br />
betwe<strong>en</strong> policy makers about the differ<strong>en</strong>t e<strong>le</strong>m<strong>en</strong>ts and their relative importance. 105<br />
Another extreme is to <strong>de</strong>termine the weights a priori, 103, 104 reducing the need for<br />
discussion betwe<strong>en</strong> health policy makers and/or stakehol<strong>de</strong>rs.<br />
Whatever the approach chos<strong>en</strong> for weighing the additional e<strong>le</strong>m<strong>en</strong>ts, the main objective<br />
is to increase transpar<strong>en</strong>cy in the e<strong>le</strong>m<strong>en</strong>ts that are consi<strong>de</strong>red in the <strong>de</strong>cision making<br />
process and at <strong>le</strong>ast make them explicit. s<br />
One of the e<strong>le</strong>m<strong>en</strong>ts could be the ICER and its relation to a pre<strong>de</strong>fined ICER threshold<br />
value. This is one way to consi<strong>de</strong>r the economic value of an interv<strong>en</strong>tion in the <strong>de</strong>cision<br />
making process. Other possibilities exist, however. For examp<strong>le</strong>, economic<br />
consi<strong>de</strong>rations can also be introduced in this approach by looking at the economic<br />
e<strong>le</strong>m<strong>en</strong>ts (cost, budget impact, g<strong>en</strong>eral health outcome) in a disaggregated form, 77 and<br />
weighing these separate e<strong>le</strong>m<strong>en</strong>ts explicitly in the <strong>de</strong>cision making process. We<br />
elaborate on this in section 2.8.6. In summary, being explicit about the <strong>de</strong>cision criteria<br />
offers ad<strong>de</strong>d value and does not <strong>de</strong>p<strong>en</strong>d on whether one accepts the i<strong>de</strong>a of an ICER<br />
threshold value or ICERs as such.<br />
In the case where the additional e<strong>le</strong>m<strong>en</strong>ts are tak<strong>en</strong> implicitly into account in the<br />
<strong>de</strong>cision making process, it is unlikely that a sing<strong>le</strong> threshold value can be id<strong>en</strong>tified that<br />
fits all policy <strong>de</strong>cisions about all health technologies. 64 This ultimately boils down to the<br />
<strong>le</strong>ss transpar<strong>en</strong>t situation <strong>de</strong>scribed earlier where the societal WTP for a QALY differs<br />
for every sing<strong>le</strong> interv<strong>en</strong>tion.<br />
With the ICER being one of the many consi<strong>de</strong>rations in health care policy making, the<br />
probability of rejecting an interv<strong>en</strong>tion increases as its ICER increases. If costeffectiv<strong>en</strong>ess<br />
consi<strong>de</strong>rations are tak<strong>en</strong> into account in the <strong>de</strong>cision making process, the<br />
g<strong>en</strong>eral i<strong>de</strong>a is that interv<strong>en</strong>tions with a relatively low ICER would be accepted more<br />
easily than interv<strong>en</strong>tions with a relatively high ICER. How much ‘more easily’ <strong>de</strong>p<strong>en</strong>ds<br />
on the other consi<strong>de</strong>rations, characteristics and societal concerns tak<strong>en</strong> into account<br />
wh<strong>en</strong> taking the <strong>de</strong>cision. 67<br />
The <strong>le</strong>vel of uncertainty around the ICER may also be an important <strong>de</strong>terminant in the<br />
<strong>de</strong>cision to reimburse an interv<strong>en</strong>tion. For instance, procedures with litt<strong>le</strong> evid<strong>en</strong>ce on<br />
effectiv<strong>en</strong>ess will typically be characterised by large credibility intervals around the<br />
ICER. It seems logical that, giv<strong>en</strong> the uncertainty about the effectiv<strong>en</strong>ess of a procedure,<br />
policy makers will be more reluctant to reimburse the procedure. In such a case, the<br />
risk of taking a ‘premature’ <strong>de</strong>cision might have to be assessed.<br />
s In this context it is worth noting that NICE makes a distinction betwe<strong>en</strong> “assessm<strong>en</strong>t” and “appraisal”.<br />
Assessm<strong>en</strong>t refers to the sci<strong>en</strong>tific review of the evid<strong>en</strong>ce about how well a group of similar treatm<strong>en</strong>ts<br />
work, and whether they offer value for money. The assessm<strong>en</strong>t report forms the basis for the appraisal.<br />
Appraisal refers to the formal assessm<strong>en</strong>t of the quality of research evid<strong>en</strong>ce and its re<strong>le</strong>vance to the<br />
clinical question or gui<strong>de</strong>line un<strong>de</strong>r consi<strong>de</strong>ration, according to pre<strong>de</strong>termined criteria. The Appraisal<br />
Committee <strong>de</strong>velops NICE’s guidance about using drugs or treatm<strong>en</strong>ts in the NHS (see<br />
http://www.nice.org.uk/website/glossary)
32 ICER Thresholds <strong>KCE</strong> Reports 100<br />
For examp<strong>le</strong>, is the <strong>de</strong>cision to wait for more evid<strong>en</strong>ce socially acceptab<strong>le</strong> or would this<br />
<strong>de</strong>cision be harmful.<br />
A ‘cost-effectiv<strong>en</strong>ess probability of rejection’ curve ref<strong>le</strong>cts the likelihood of an<br />
interv<strong>en</strong>tion being rejected <strong>de</strong>p<strong>en</strong>ding on the value of its ICER (Figure 5). 67 This is a<br />
theoretical curve as the probability of rejection may be differ<strong>en</strong>t for differ<strong>en</strong>t types of<br />
interv<strong>en</strong>tions. Moreover, as it is impossib<strong>le</strong> to quantify the precise impact of all other<br />
policy consi<strong>de</strong>rations and policy consi<strong>de</strong>rations will differ across interv<strong>en</strong>tions, it will in<br />
practice be impossib<strong>le</strong> to calculate the probability of rejection.<br />
Figure 5: The relation betwe<strong>en</strong> the ICER and probability of rejection<br />
Probability of rejection<br />
1<br />
20 40 60 80 100<br />
Cost per QALY (in € thousand)<br />
2.8.4 The average GDP per capita as a threshold value for the average costeffectiv<strong>en</strong>ess<br />
ratio<br />
According to Williams 99 it makes s<strong>en</strong>se to allow each citiz<strong>en</strong> the average Gross<br />
Domestic Product (GDP) per capita, as a ref<strong>le</strong>ction of each citiz<strong>en</strong>’s ‘fair share’ of a<br />
nation’s wealth. Therefore, the average per capita GDP might be used as a threshold<br />
value for the average cost-effectiv<strong>en</strong>ess ratio in the evaluation of an interv<strong>en</strong>tion’s costeffectiv<strong>en</strong>ess.<br />
The same reasoning was followed by the World Health Organisation’s<br />
Commission on Macroeconomics and Health, be it with ‘Disability Adjusted Life Years<br />
(DALYs) averted’ as the g<strong>en</strong>eric health outcome measure instead of QALYs or LYG. t<br />
The Commission specifies that averting one DALY for <strong>le</strong>ss than the average per capita<br />
income is very cost-effective, averting one DALY for <strong>le</strong>ss than three times the average<br />
per capita income is still cost-effective and averting one DALY for more than this<br />
amount is not cost-effective. The construct and meaning of DALYs is fundam<strong>en</strong>tally<br />
differ<strong>en</strong>t from that of QALYs. The similarity betwe<strong>en</strong> the WHO criteria and the<br />
criterion suggested by Williams in 2004 is therefore highly mis<strong>le</strong>ading.<br />
In<strong>de</strong>p<strong>en</strong>d<strong>en</strong>t from this specific caveat, the suggested ICER threshold value of one time<br />
the average GDP per capita is prob<strong>le</strong>matic in differ<strong>en</strong>t ways. First, the approach actually<br />
proposes a threshold value for the average cost-effectiv<strong>en</strong>ess ratio and not for the<br />
increm<strong>en</strong>tal cost-effectiv<strong>en</strong>ess ratio. It can, therefore, not be consi<strong>de</strong>red an appropriate<br />
threshold value for ICERs. Moreover, it is inappropriate to base <strong>de</strong>cisions on average<br />
cost-effectiv<strong>en</strong>ess ratios because this would mean that the cost and health effects of the<br />
alternative treatm<strong>en</strong>t are both zero. Ev<strong>en</strong> in cases where there is no alternative<br />
treatm<strong>en</strong>t and the relative comparator is ‘doing nothing’, this will not be the case.<br />
Resource allocation based on average cost-effectiv<strong>en</strong>ess ratios will therefore not<br />
maximise health with the giv<strong>en</strong> budget. Second, the approach implicitly assumes society<br />
is willing to <strong>de</strong>vote its <strong>en</strong>tire GDP to health care. 5<br />
t The number of DALYs of a disease ref<strong>le</strong>cts the number of healthy life-years lost in a population due to<br />
the disease. Four aspects of disease are tak<strong>en</strong> into account in the DALY measure: the number of pati<strong>en</strong>ts<br />
suffering from the disease, the severity of the disease, mortality and the age at <strong>de</strong>ath.
<strong>KCE</strong> reports 100 ICER Thresholds 33<br />
The <strong>en</strong>tire GDP might not ev<strong>en</strong> be <strong>en</strong>ough if citiz<strong>en</strong>s require combinations of<br />
treatm<strong>en</strong>ts whose total average cost-per-QALY exceeds the average GDP per capita.<br />
Therefore, this approach is not feasib<strong>le</strong> and conflicts with the effici<strong>en</strong>cy evaluation<br />
objective of economic evaluation.<br />
2.8.5 The opportunity costs approach<br />
Gafni and Birch have argued that CEA and ICERs may not be very useful in real life<br />
<strong>de</strong>cision making contexts, ev<strong>en</strong> for maximizing health from a giv<strong>en</strong> budget, simply<br />
because the basic conditions for using ICERs for this purpose are not and can never be<br />
fulfil<strong>le</strong>d. 50, 68 Either ICERs would be interpreted as in the theoretical ICER threshold<br />
value approach, which is inappropriate giv<strong>en</strong> that the baseline conditions are not fulfil<strong>le</strong>d<br />
(see 2.7). Or, alternatively, the ICER threshold value would be <strong>de</strong>fined as a value for the<br />
societal WTP per QALY (or LYG), in which case the threshold approach would<br />
inevitably <strong>le</strong>ad to budget expansions. 98 There is evid<strong>en</strong>ce from Ontario (Canada),<br />
England and Australia that the adoption of the ICER threshold value approach has<br />
in<strong>de</strong>ed be<strong>en</strong> associated with substantial unplanned increases in healthcare exp<strong>en</strong>ditures<br />
62, 64<br />
without any evid<strong>en</strong>ce of any increase in total health b<strong>en</strong>efit.<br />
The suggested alternative to the use of ICERs and CEA in a <strong>de</strong>cision making context<br />
characterized by fixed budgets is to pres<strong>en</strong>t the real opportunity costs of the<br />
imp<strong>le</strong>m<strong>en</strong>tation of the programme un<strong>de</strong>r consi<strong>de</strong>ration. 63, 98 The opportunity costs of<br />
the programme are equal to the health b<strong>en</strong>efits foregone in other programmes that<br />
have to be downgra<strong>de</strong>d or abolished to finance the new one. It implies the notion of<br />
choice betwe<strong>en</strong> <strong>de</strong>sirab<strong>le</strong> but mutually exclusive outcomes. If the b<strong>en</strong>efits foregone<br />
from the cancel<strong>le</strong>d programme are higher than the b<strong>en</strong>efits g<strong>en</strong>erated by the new<br />
programme, the new programme should not be fun<strong>de</strong>d from the limited budget (un<strong>le</strong>ss<br />
there are other non health-economic argum<strong>en</strong>ts to fund it).<br />
As such, the additional resource requirem<strong>en</strong>ts are id<strong>en</strong>tified and the implications of<br />
cancelling other interv<strong>en</strong>tions are ma<strong>de</strong> explicit. 46 This increases the transpar<strong>en</strong>cy of the<br />
<strong>de</strong>cision making process.<br />
The imp<strong>le</strong>m<strong>en</strong>tation of this approach on a national <strong>le</strong>vel might be prob<strong>le</strong>matic for<br />
differ<strong>en</strong>t reasons. It is difficult to know precisely which activities will be displaced to be<br />
ab<strong>le</strong> to imp<strong>le</strong>m<strong>en</strong>t a new interv<strong>en</strong>tion. As a result, only accepting new technologies if<br />
the source of the resources is ma<strong>de</strong> explicit could paralyse the system. Other <strong>de</strong>cision<br />
prob<strong>le</strong>ms might appear. For examp<strong>le</strong>, what happ<strong>en</strong>s if in a <strong>de</strong>c<strong>en</strong>tralised reimbursem<strong>en</strong>t<br />
<strong>de</strong>cision system <strong>de</strong>cision makers think to find the resources for two differ<strong>en</strong>t<br />
interv<strong>en</strong>tions from disinvestm<strong>en</strong>t in the same third interv<strong>en</strong>tion? Furthermore,<br />
interv<strong>en</strong>tions with a large budget impact will probably be more prob<strong>le</strong>matic to<br />
imp<strong>le</strong>m<strong>en</strong>t than projects with a smal<strong>le</strong>r budget impact, ev<strong>en</strong> if they may be more costeffective,<br />
since they will need to id<strong>en</strong>tify relatively more projects to sacrifice.<br />
On a local or institutional <strong>le</strong>vel, their may be more possibilities to use the opportunity<br />
cost approach. For examp<strong>le</strong>, hospitals that have to <strong>de</strong>ci<strong>de</strong> on buying a new <strong>de</strong>vice or<br />
imp<strong>le</strong>m<strong>en</strong>ting a new health care programme, might consi<strong>de</strong>r the savings they will have<br />
to realize elsewhere in their organization to free resources for the new investm<strong>en</strong>ts.<br />
Despite the pot<strong>en</strong>tial practical prob<strong>le</strong>ms, the opportunity cost approach makes the<br />
important point that disinvestm<strong>en</strong>ts are always nee<strong>de</strong>d in a system with a fixed budget.<br />
First candidates for disinvestm<strong>en</strong>t should be interv<strong>en</strong>tions that have become obso<strong>le</strong>te<br />
or are no longer consi<strong>de</strong>red worth their costs. In a mixed public-private financing<br />
system, the opportunity cost approach might become ev<strong>en</strong> more comp<strong>le</strong>x, because<br />
every <strong>de</strong>cision to reduce public financing of an interv<strong>en</strong>tion (in or<strong>de</strong>r to contain costs<br />
from the perspective of the public payer) has a pot<strong>en</strong>tial impact on both health and<br />
income inequalities.
34 ICER Thresholds <strong>KCE</strong> Reports 100<br />
2.8.6 Cost-consequ<strong>en</strong>ces analysis<br />
Many authors have suggested that ICERs and CEA may have a limited meaning to health<br />
care policy makers. 77, 106-112 Coast (2004) 77 sees three reasons for this:<br />
1. health policy makers might not have the same objectives as presumed by<br />
economic evaluations<br />
2. the ICER might not provi<strong>de</strong> suffici<strong>en</strong>t information about the effici<strong>en</strong>cy of an<br />
interv<strong>en</strong>tion from the health care policy makers’ point of view<br />
3. the pres<strong>en</strong>tation of and methods used in economic evaluations might not be<br />
meaningful to <strong>de</strong>cision makers.<br />
The danger arising from this situation is that either economic evaluations become<br />
marginalised in the <strong>de</strong>cision making process, or, more worrying, that economic<br />
evaluations are used without careful thought about what is behind them and<br />
consequ<strong>en</strong>tly serve as a basis for <strong>de</strong>cisions that do not ref<strong>le</strong>ct society’s objectives.<br />
An alternative to ICERs and ICER thresholds is to pres<strong>en</strong>t the separate e<strong>le</strong>m<strong>en</strong>ts of<br />
economic evaluations that do make s<strong>en</strong>se to <strong>de</strong>cision makers in disaggregated form, 77<br />
such as costs per pati<strong>en</strong>t, costs for the <strong>en</strong>tire population, outcomes in terms of life<br />
years gained, impact on quality of life, disease severity etc.<br />
In health economics, this is cal<strong>le</strong>d cost-consequ<strong>en</strong>ces analysis. 2, 4 These e<strong>le</strong>m<strong>en</strong>ts can<br />
th<strong>en</strong> be weighed –implicitly or explicitly- by policy makers in the <strong>de</strong>cision making<br />
process. The major advantage of this approach is that -in contrast to the ICER- the<br />
separate e<strong>le</strong>m<strong>en</strong>ts make s<strong>en</strong>se to policy makers. The ICER, being a ratio, masks<br />
important aspects of an interv<strong>en</strong>tion. The absolute values of the numerator and<br />
d<strong>en</strong>ominator are lost wh<strong>en</strong> only looking at the value of the ICER, whi<strong>le</strong> these absolute<br />
values are important. For examp<strong>le</strong>, an interv<strong>en</strong>tion offering 0.001 additional QALY per<br />
pati<strong>en</strong>t for an additional cost of €80 has the same ICER as an interv<strong>en</strong>tion offering 10<br />
QALYs for an additional cost of €800 000 per pati<strong>en</strong>t. If for the latter interv<strong>en</strong>tion only<br />
3 pati<strong>en</strong>ts are eligib<strong>le</strong> and for the former 100 000, both interv<strong>en</strong>tions may neverthe<strong>le</strong>ss<br />
be perceived differ<strong>en</strong>tly by policy makers. Moreover, it allows more than the other<br />
alternatives, to take uncertainty in the economic e<strong>le</strong>m<strong>en</strong>ts into account. The<br />
disadvantage of the approach is that, in the abs<strong>en</strong>ce of an ICER, it is impossib<strong>le</strong> to assess<br />
the effici<strong>en</strong>cy of an interv<strong>en</strong>tion.<br />
Whi<strong>le</strong> effici<strong>en</strong>cy in resource allocation is c<strong>le</strong>arly not the only concern of health care<br />
policy makers, it cannot be d<strong>en</strong>ied that it is a concern. Ignoring cost-effectiv<strong>en</strong>ess is as<br />
unethical as using cost-effectiv<strong>en</strong>ess as the so<strong>le</strong> criterion for resource allocation<br />
113, 114<br />
<strong>de</strong>cisions.
<strong>KCE</strong> reports 100 ICER Thresholds 35<br />
Key points<br />
Alternatives to the theoretical ICER threshold value in a fixed budget setting<br />
have be<strong>en</strong> suggested, differing in the ext<strong>en</strong>t to which they support the<br />
notion of an ICER threshold value as a guiding resource allocation ru<strong>le</strong>:<br />
• Instead of <strong>de</strong>fining the ICER threshold value as the ICER of the <strong>le</strong>ast costeffective<br />
interv<strong>en</strong>tion still financed, the ICER threshold value could be<br />
<strong>de</strong>fined in terms of societal willingness to pay for a (weighted) QALY (or<br />
LYG). This requires a f<strong>le</strong>xib<strong>le</strong> health care budget.<br />
• A second suggested alternative is to <strong>de</strong>rive the ICER threshold value<br />
from past reimbursem<strong>en</strong>t <strong>de</strong>cisions.<br />
• A third suggested alternative is to consi<strong>de</strong>r the ICER as one e<strong>le</strong>m<strong>en</strong>t in<br />
the <strong>de</strong>cision making process that has to be weighed against other<br />
e<strong>le</strong>m<strong>en</strong>ts. This weighing can be explicit or implicit.<br />
• A fourth suggested alternative is to <strong>de</strong>fine a threshold value for the<br />
average cost-effectiv<strong>en</strong>ess ratio as the average GDP per capita, ref<strong>le</strong>cting<br />
the citiz<strong>en</strong>s’ ‘fair share’ of a nation’s wealth.<br />
• A fifth suggested alternative abandons the i<strong>de</strong>a of an ICER to gui<strong>de</strong><br />
<strong>de</strong>cisions and argues that the real opportunity costs of financing an<br />
interv<strong>en</strong>tion should be ma<strong>de</strong> explicit and compared to its b<strong>en</strong>efits in<br />
terms of better health outcomes.<br />
• A last suggested alternative is to pres<strong>en</strong>t all economic e<strong>le</strong>m<strong>en</strong>ts that are<br />
re<strong>le</strong>vant for <strong>de</strong>cision making in disaggregated form, in or<strong>de</strong>r to allow the<br />
<strong>de</strong>cision maker to weigh the economic e<strong>le</strong>m<strong>en</strong>ts against other e<strong>le</strong>m<strong>en</strong>ts.<br />
• Each alternative has its practical weaknesses: either there is a lack of data<br />
to imp<strong>le</strong>m<strong>en</strong>t the approach, or there are unresolved measurem<strong>en</strong>t<br />
prob<strong>le</strong>ms, or in practice they do not really increase the transpar<strong>en</strong>cy of<br />
<strong>de</strong>cisions. They all have in common that they look for ways for making<br />
economic consi<strong>de</strong>rations explicit in the health care <strong>de</strong>cision making<br />
context.
36 ICER Thresholds <strong>KCE</strong> Reports 100<br />
3 THE ROLE OF ECONOMIC EVALUATIONS<br />
IN HEALTH CARE DECISION MAKING<br />
3.1 DECISION MAKING PROCESSES<br />
Major efforts are being <strong>de</strong>voted in many countries to <strong>de</strong>velop tools and methods to<br />
create processes of “research-informed” <strong>de</strong>cision making. Rational and informed<br />
<strong>de</strong>cision making is implicitly put forward as part of the emerging evid<strong>en</strong>ce-based and<br />
health technology assessm<strong>en</strong>t movem<strong>en</strong>ts. The use of ICERs and ICER threshold values<br />
is part of this <strong>de</strong>velopm<strong>en</strong>t. However, a purely instrum<strong>en</strong>talist approach to <strong>de</strong>cision<br />
making in health care is to be avoi<strong>de</strong>d. Decision making in health care is not to be<br />
reduced to well-informed and technical, rational assessm<strong>en</strong>t of prob<strong>le</strong>ms upon which<br />
best solutions are formulated. 115<br />
The study of <strong>de</strong>cision making processes is not rec<strong>en</strong>t. Differ<strong>en</strong>t mo<strong>de</strong>ls have be<strong>en</strong><br />
id<strong>en</strong>tified and discussed in the history of social sci<strong>en</strong>ces. Some of the basics of <strong>de</strong>cision<br />
making theories will help to un<strong>de</strong>rstand why <strong>de</strong>cisions on health care technologies or<br />
interv<strong>en</strong>tions cannot be so<strong>le</strong>ly based on economic and clinical (technical-rational)<br />
consi<strong>de</strong>rations.<br />
• The Rational Decision Making mo<strong>de</strong>l focuses on reasoned (rational and<br />
logical) <strong>de</strong>cisions. It is based on an axiomatic approach that <strong>de</strong>cisions are<br />
(or should be) the result of rational weighing of alternatives before<br />
se<strong>le</strong>cting a choice. The rational mo<strong>de</strong>l assumes, stated in a simplified<br />
manner, that it is possib<strong>le</strong> to se<strong>le</strong>ct one sing<strong>le</strong> and best solution to a<br />
prob<strong>le</strong>m. It is based on the assumption that the prob<strong>le</strong>m is well known<br />
and id<strong>en</strong>tified, that c<strong>le</strong>ar assessm<strong>en</strong>t ru<strong>le</strong>s are availab<strong>le</strong> to judge possib<strong>le</strong><br />
solutions and that solutions for a prob<strong>le</strong>m are chos<strong>en</strong> based on these<br />
assessm<strong>en</strong>t ru<strong>le</strong>s. It also presupposes that all possib<strong>le</strong> options or<br />
approaches to solving the prob<strong>le</strong>m un<strong>de</strong>r study are id<strong>en</strong>tified and that (in<br />
political <strong>de</strong>cisions) the costs and b<strong>en</strong>efits of each option are assessed and<br />
compared and that the best interv<strong>en</strong>tion is se<strong>le</strong>cted. The un<strong>de</strong>rlying<br />
assumptions of this rational <strong>de</strong>cision making have be<strong>en</strong> criticised. It has<br />
be<strong>en</strong> argued that not all know<strong>le</strong>dge is readily availab<strong>le</strong> to make c<strong>le</strong>ar<br />
assessm<strong>en</strong>ts of a situation, be it on one solution or on possib<strong>le</strong><br />
alternatives; that cognitive capacities of <strong>de</strong>cision makers are limited; that<br />
prefer<strong>en</strong>ces are not always c<strong>le</strong>ar or that prefer<strong>en</strong>ces do not remain stab<strong>le</strong><br />
over time.<br />
• The “boun<strong>de</strong>d rationality mo<strong>de</strong>l”, an adapted form of the rational mo<strong>de</strong>l,<br />
tries to <strong>de</strong>al with these criticisms. This mo<strong>de</strong>l assumes that a certain<br />
ext<strong>en</strong>t of rationality in <strong>de</strong>cision making is possib<strong>le</strong>, be it that differ<strong>en</strong>t<br />
cognitive (e.g. capacity to <strong>de</strong>al with information) and circumstantial factors<br />
(availability and timeliness of information on the solutions and alternatives<br />
at the time of <strong>de</strong>cision making) inhibit total rational assessm<strong>en</strong>ts of a<br />
prob<strong>le</strong>m. The boun<strong>de</strong>d rationality mo<strong>de</strong>l assumes that “as rational as<br />
possib<strong>le</strong>” <strong>de</strong>cisions can be tak<strong>en</strong>. Variants of the boun<strong>de</strong>d rationality<br />
mo<strong>de</strong>l refer to “procedural rationality” in which <strong>de</strong>cisions are <strong>de</strong>veloped<br />
within the procedural constraints of the ag<strong>en</strong>cies or actors responsib<strong>le</strong><br />
for taking <strong>de</strong>cisions.<br />
Moreover, peop<strong>le</strong> rarely adhere to logical mo<strong>de</strong>ls of choice. Other <strong>de</strong>cision making<br />
theories, using a comp<strong>le</strong>tely differ<strong>en</strong>t perspective, have docum<strong>en</strong>ted that <strong>de</strong>cision<br />
making is primarily an interactive process rather than a rational and well thoughtthrough<br />
process of assessing a prob<strong>le</strong>m.<br />
• The “increm<strong>en</strong>talism” or “muddling through” <strong>de</strong>cision making mo<strong>de</strong>l<br />
argues that goals are set as politically feasib<strong>le</strong> goals in which peop<strong>le</strong> strive<br />
for acceptab<strong>le</strong> rather than theoretical best solutions. 116-118 Other, more<br />
psychological <strong>de</strong>cision making theories, have highlighted the importance of<br />
schemata in <strong>de</strong>termining how peop<strong>le</strong> interpret new information based on
<strong>KCE</strong> reports 100 ICER Thresholds 37<br />
their pre-existing beliefs and cultural values. “Attribution theories” have<br />
argued that peop<strong>le</strong> use heuristics, ru<strong>le</strong>s to test their vision on a prob<strong>le</strong>m<br />
(schemata) and facilitate the processing of information. The <strong>de</strong>cision<br />
making process produces <strong>de</strong>cisions only marginally differ<strong>en</strong>t from past<br />
practice as increm<strong>en</strong>tal <strong>de</strong>cision making <strong>de</strong>als with se<strong>le</strong>ctive issues as they<br />
arise. Prob<strong>le</strong>ms or issues are <strong>de</strong>alt with ad hoc, using whatever analysis is<br />
close at hand, without any compreh<strong>en</strong>sive review of all the associated<br />
issues. The increm<strong>en</strong>tal <strong>de</strong>cision making mo<strong>de</strong>l stresses that objectives<br />
are seldom explicitly specified, that remedial action (rather than rational<br />
analysis) is tak<strong>en</strong> wh<strong>en</strong> it becomes ess<strong>en</strong>tial, and more important<br />
<strong>de</strong>cisions are <strong>de</strong>p<strong>en</strong>d<strong>en</strong>t on the power strugg<strong>le</strong>s betwe<strong>en</strong> interest groups.<br />
• Related mo<strong>de</strong>ls have docum<strong>en</strong>ted that <strong>de</strong>cision making processes are<br />
political in nature. Decisions can be explained from a “political rationality”<br />
point of view, in which actors take <strong>de</strong>cisions as a result of power plays,<br />
bargaining, coalitions, public acceptance of <strong>de</strong>cisions, etc. The political<br />
mo<strong>de</strong>ls recognize the process of reconciling the interests of differ<strong>en</strong>t<br />
stakehol<strong>de</strong>rs within a variety of internal and external constraints, rather<br />
than a purely rational assessm<strong>en</strong>t of all ins and outs of a (health care)<br />
prob<strong>le</strong>m.<br />
We limited ourselves to a very rudim<strong>en</strong>tary sketch of the comp<strong>le</strong>x field of <strong>de</strong>cision<br />
making theories. This sketch allows arguing that <strong>de</strong>cision making in daily practice is<br />
<strong>de</strong>termined by a comp<strong>le</strong>xity of factors, and certainly not only a rational instrum<strong>en</strong>tal<br />
consi<strong>de</strong>ration and balancing of availab<strong>le</strong> alternatives. Decision making is in ess<strong>en</strong>ce an<br />
interactive process in which differ<strong>en</strong>t factors affect the outcome, the <strong>de</strong>cision.<br />
What we do see in the curr<strong>en</strong>t time frame of health care <strong>de</strong>cision making, is that major<br />
efforts are being ma<strong>de</strong> to “rationalise” <strong>de</strong>cision making as much as possib<strong>le</strong> offering<br />
information col<strong>le</strong>cted and analysed using sci<strong>en</strong>tific methodological princip<strong>le</strong>s (evid<strong>en</strong>ce,<br />
economic evaluation,…) where possib<strong>le</strong>. It is one of the necessary steps to make<br />
<strong>de</strong>cision making processes more transpar<strong>en</strong>t. The <strong>de</strong>velopm<strong>en</strong>t and use of sci<strong>en</strong>tific<br />
know<strong>le</strong>dge is however only one of the compon<strong>en</strong>ts of real-life <strong>de</strong>cision making.<br />
Key points<br />
• Decision making is a far more comp<strong>le</strong>x process than an informed rational<br />
assessm<strong>en</strong>t of prob<strong>le</strong>ms, weighing of alternatives and the formulation of<br />
best solutions.<br />
• Differ<strong>en</strong>t <strong>de</strong>cision making mo<strong>de</strong>ls <strong>le</strong>arn that rationality -if any- is<br />
boun<strong>de</strong>d, that <strong>de</strong>cisions are <strong>de</strong>veloped increm<strong>en</strong>tally and that <strong>de</strong>cision<br />
making processes are political in nature.<br />
• Efforts are ma<strong>de</strong> to ‘rationalise’ health care <strong>de</strong>cision making by offering<br />
information and know<strong>le</strong>dge gathered on a methodological sound basis,<br />
and to make <strong>de</strong>cision making processes more transpar<strong>en</strong>t.<br />
3.2 INFORMED POLICY DECISION MAKING<br />
A particular field of study related to <strong>de</strong>cision making evolves around supporting <strong>de</strong>cision<br />
makers with the necessary (sci<strong>en</strong>tific) information. The use of sci<strong>en</strong>tific information in<br />
<strong>de</strong>cision making processes is a very comp<strong>le</strong>x process. It has be<strong>en</strong> argued that for policy<br />
makers, any form of information that supports a <strong>de</strong>cision is oft<strong>en</strong> consi<strong>de</strong>red as re<strong>le</strong>vant<br />
know<strong>le</strong>dge. Decisions are thus in most of the cases not tak<strong>en</strong> on the basis of sci<strong>en</strong>tific<br />
or technical information only, but on a mix of information sources. 119-121<br />
Issues on the use of research utilization have be<strong>en</strong> discussed in literature on know<strong>le</strong>dge<br />
brokering and information dissemination betwe<strong>en</strong> researchers and <strong>de</strong>cision makers. We<br />
will not elaborate much, but quote insights from two systematic reviews.<br />
One systematic review discusses the facilitators of and barriers to the use of research<br />
evid<strong>en</strong>ce by policy makers. 122
38 ICER Thresholds <strong>KCE</strong> Reports 100<br />
• The most commonly m<strong>en</strong>tioned facilitators were: personal contact<br />
betwe<strong>en</strong> researchers and policymakers; timeliness and re<strong>le</strong>vance of the<br />
research; research that inclu<strong>de</strong>d a summary and c<strong>le</strong>ar recomm<strong>en</strong>dations;<br />
research that confirmed curr<strong>en</strong>t policy or <strong>en</strong>dorsed self-interest;<br />
community or cli<strong>en</strong>t <strong>de</strong>mand for the research; and research that inclu<strong>de</strong>d<br />
effectiv<strong>en</strong>ess data.<br />
• The most commonly m<strong>en</strong>tioned barriers were: abs<strong>en</strong>ce of personal<br />
contact betwe<strong>en</strong> researchers and policy makers; lack of timeliness or<br />
re<strong>le</strong>vance of research; mutual mistrust; power and budget strugg<strong>le</strong>s; poorquality<br />
research and political instability or high turnover of policymaking<br />
staff.<br />
The review also clarifies that "Use" of research can be viewed in three differ<strong>en</strong>t ways: 1)<br />
direct use (research directly affects the <strong>de</strong>cision); 2) se<strong>le</strong>ctive use (research is se<strong>le</strong>cted<br />
to <strong>le</strong>gitimate pre-<strong>de</strong>termined positions); and 3) <strong>en</strong>light<strong>en</strong>ing use (research helps to set<br />
new b<strong>en</strong>chmarks for what is possib<strong>le</strong> and <strong>de</strong>ep<strong>en</strong>s un<strong>de</strong>rstanding).<br />
Lavis et al. 123 distinguish four broad categories or mo<strong>de</strong>ls for <strong>de</strong>veloping evid<strong>en</strong>ce-based<br />
policy making. These mo<strong>de</strong>ls are pres<strong>en</strong>ted as a typology, but can be applied either<br />
alone or in combination.<br />
• Enhancing push factors (Mo<strong>de</strong>l A) - “Push” efforts are g<strong>en</strong>erally <strong>le</strong>d by<br />
researchers, or communications staff of research institutes and aim to<br />
increase awar<strong>en</strong>ess of research evid<strong>en</strong>ce among policy makers and civil<br />
society. Such efforts are well suited to situations where the pot<strong>en</strong>tial<br />
research users are unaware they should be consi<strong>de</strong>ring a particular<br />
message or in some cases would prefer to continue to disregard evid<strong>en</strong>ce.<br />
HTA and evid<strong>en</strong>ce based research ag<strong>en</strong>cies operate most of the time<br />
within this mo<strong>de</strong>l.<br />
• Enhancing pull factors (Mo<strong>de</strong>l B) - User-pull and/or <strong>de</strong>mand for research<br />
evid<strong>en</strong>ce are critical for research evid<strong>en</strong>ce uptake. It occurs wh<strong>en</strong> policy<br />
and <strong>de</strong>cision makers id<strong>en</strong>tify an information gap and request evid<strong>en</strong>ce or<br />
commission research to fill this gap. These kinds of <strong>de</strong>mands are oft<strong>en</strong><br />
hand<strong>le</strong>d by policy advisory units within Ministries of Health and/or other<br />
policy making or technical support units (sometimes also applied research<br />
by universities or research institutes)<br />
• Supporting exchange efforts (Mo<strong>de</strong>l C) - Exchange efforts occur wh<strong>en</strong><br />
producers and users of research work in partnerships and establish links.<br />
• Supporting imp<strong>le</strong>m<strong>en</strong>tation of an integrated mo<strong>de</strong>l (Mo<strong>de</strong>l D) - integrated<br />
mo<strong>de</strong>ls combine e<strong>le</strong>m<strong>en</strong>ts of the three mo<strong>de</strong>ls <strong>de</strong>scribed above.<br />
Key points<br />
• Decisions are in most of the cases not tak<strong>en</strong> on the basis of sci<strong>en</strong>tific or<br />
technical information only, but on a mix of information sources.<br />
• Several barriers and facilitators have be<strong>en</strong> id<strong>en</strong>tified for <strong>de</strong>veloping<br />
sci<strong>en</strong>tifically informed <strong>de</strong>cision making.<br />
• Know<strong>le</strong>dge and research evid<strong>en</strong>ce can be used by <strong>de</strong>cision makers in<br />
three differ<strong>en</strong>t ways: direct, se<strong>le</strong>ctive and <strong>en</strong>light<strong>en</strong>ing.<br />
• Know<strong>le</strong>dge brokering is an important topic to be further explored in<br />
or<strong>de</strong>r to support <strong>de</strong>cision makers.<br />
• The process of ‘know<strong>le</strong>dge brokering’ for policy makers can be<br />
summarized in four mo<strong>de</strong>ls: an information push mo<strong>de</strong>l, an information<br />
pull mo<strong>de</strong>l, an exchange mo<strong>de</strong>l and an integrated mo<strong>de</strong>l.
<strong>KCE</strong> reports 100 ICER Thresholds 39<br />
3.3 EMPIRICAL EVIDENCE ON THE USE OF ECONOMIC<br />
EVALUATIONS IN HEALTH CARE DECISION MAKING<br />
“…, the results of rigorous clinical trials and s<strong>en</strong>sitive mo<strong>de</strong>lling techniques tell us litt<strong>le</strong> about<br />
how data on clinical and cost effectiv<strong>en</strong>ess are interpreted at the <strong>le</strong>vel of national policy<br />
formulation”. 124<br />
It has be<strong>en</strong> repeatedly argued in this report that health care <strong>de</strong>cision making is assumed,<br />
in an i<strong>de</strong>al situation, to be focussing on an optimal allocation of availab<strong>le</strong> resources with<br />
the purpose of maximising health. Decision makers are expected to focus on<br />
interv<strong>en</strong>tions that provi<strong>de</strong> the most health gains for a giv<strong>en</strong> exp<strong>en</strong>diture of resources. 114<br />
Moreover, a systematic approach is expected to increase transpar<strong>en</strong>cy and consist<strong>en</strong>cy<br />
in the <strong>de</strong>cisions tak<strong>en</strong>. 125<br />
A particular branch of empirical research has be<strong>en</strong> studying the impact of economic<br />
evaluation studies on the policy making processes. The successful application of costeffectiv<strong>en</strong>ess<br />
princip<strong>le</strong>s has be<strong>en</strong> docum<strong>en</strong>ted as being a prob<strong>le</strong>m in differ<strong>en</strong>t health care<br />
systems. 114, 126-128 Several barriers to the use of the results of economic evaluations have<br />
be<strong>en</strong> observed. 129 The barriers have be<strong>en</strong> summarized as issues of accessibility of<br />
research evid<strong>en</strong>ce and (sci<strong>en</strong>tific, structural/institutional and ethical/political) acceptability<br />
of research evid<strong>en</strong>ce. 17<br />
A systematic review on the use of economic evaluations in the UK revea<strong>le</strong>d that a<br />
number of features of the <strong>de</strong>cision making process hamper the use of cost-effectiv<strong>en</strong>ess<br />
analysis, such as capacity to un<strong>de</strong>rstand economic analysis, attitu<strong>de</strong>s to economic<br />
evaluations including concerns on the basis of analysis and its use, the scope of the<br />
research questions and the scope of the policy question. 130 A survey in nine European<br />
countries docum<strong>en</strong>ts that <strong>de</strong>cision makers use differ<strong>en</strong>t sources of economic<br />
information, but that many <strong>de</strong>cision makers also believe that a lot of the information<br />
obtained can be biased through sponsorship. 112 Despite the wi<strong>de</strong>spread use of mo<strong>de</strong>lling<br />
and cost-effectiv<strong>en</strong>ess ratios for health care <strong>de</strong>cision support, there are concerns with<br />
regard to the quality of the mo<strong>de</strong>ls: 131 concerns exist about the transpar<strong>en</strong>cy and<br />
validity of the mo<strong>de</strong>ls, the lack of high <strong>le</strong>vel clinical data, possib<strong>le</strong> bias wh<strong>en</strong><br />
observational data are used and difficulties with extrapolation.<br />
In or<strong>de</strong>r to <strong>de</strong>al with these perceptions and increase the <strong>le</strong>gitimacy of research findings<br />
major efforts are being <strong>de</strong>voted to the <strong>de</strong>velopm<strong>en</strong>t of gui<strong>de</strong>lines on how to perform<br />
economic evaluation. Moreover, <strong>de</strong>cision makers do not fully un<strong>de</strong>rstand health<br />
economics outcomes statem<strong>en</strong>ts such as in particular increm<strong>en</strong>tal cost-effectiv<strong>en</strong>ess<br />
ratios, willingness to pay, QALYs etc. 77 or consi<strong>de</strong>r them to be irre<strong>le</strong>vant. 132<br />
Decision makers are convinced that although economic evaluations can be useful in<br />
princip<strong>le</strong>, in practice their usefulness is consi<strong>de</strong>red limited as the studies do not always<br />
apply to the particular <strong>de</strong>cision making context. 133-136 Economic evaluations seldom take<br />
contextual factors into consi<strong>de</strong>ration 137 although health care systems and health<br />
insurance regimes differ, and particular cultural, social, economic and political conditions<br />
are important background variab<strong>le</strong>s to un<strong>de</strong>rstand <strong>de</strong>cision making procedures. A<br />
rec<strong>en</strong>t comparison of drug reimbursem<strong>en</strong>t <strong>de</strong>cisions betwe<strong>en</strong> the UK, Australia and<br />
New Zealand conclu<strong>de</strong>d that differ<strong>en</strong>t factors might drive reimbursem<strong>en</strong>t <strong>de</strong>cisions in<br />
differ<strong>en</strong>t countries. 138 Drugs that have the pot<strong>en</strong>tial to save lives (e.g. <strong>le</strong>ukaemia) or<br />
al<strong>le</strong>viate particularly comp<strong>le</strong>x diseases (e.g. multip<strong>le</strong> sc<strong>le</strong>rosis) were reimbursed in all<br />
three countries. For other drugs, severity of the disease becomes important in the<br />
e<strong>le</strong>m<strong>en</strong>ts consi<strong>de</strong>red during the <strong>de</strong>cision making process. Perceptions of disease severity<br />
might differ betwe<strong>en</strong> countries. Raftery (2008) suggests that “the perception of ‘dread’<br />
diseases <strong>de</strong>p<strong>en</strong>ds on social factors, such as pati<strong>en</strong>t lobbying and public perceptions. Decisions<br />
on which drugs to fund, in the final analysis, <strong>de</strong>p<strong>en</strong>d on their political and social<br />
acceptability”. 138<br />
Timeliness of information is a particular issue. Cost-effectiv<strong>en</strong>ess analyses crucially<br />
<strong>de</strong>p<strong>en</strong>d on evid<strong>en</strong>ce of effectiv<strong>en</strong>ess and therefore always come later in the life cyc<strong>le</strong> of<br />
a technology. Healthcare <strong>de</strong>cisions, however, are frequ<strong>en</strong>tly nee<strong>de</strong>d in the early stages<br />
of a technology’s life cyc<strong>le</strong>.
40 ICER Thresholds <strong>KCE</strong> Reports 100<br />
As a consequ<strong>en</strong>ce, <strong>de</strong>cision makers are sometimes in a position of having to take<br />
<strong>de</strong>cisions without having a<strong>de</strong>quate cost-effectiv<strong>en</strong>ess data at their disposal. 139<br />
Moreover, clinical effectiv<strong>en</strong>ess and cost-effectiv<strong>en</strong>ess are only two of many<br />
consi<strong>de</strong>rations in making policy choices. Valuing differ<strong>en</strong>t types of outcomes is<br />
inher<strong>en</strong>tly value-lad<strong>en</strong>, where economic evid<strong>en</strong>ce needs to be combined with<br />
stakehol<strong>de</strong>r <strong>de</strong>liberation. 114<br />
Key points<br />
• Barriers in the use of economic evaluation studies have be<strong>en</strong> summarized<br />
as issues of accessibility of research evid<strong>en</strong>ce and acceptability of research<br />
evid<strong>en</strong>ce<br />
• The capacity to un<strong>de</strong>rstand economic analysis, attitu<strong>de</strong>s to economic<br />
evaluations (including concerns about the basis of the analyses and their<br />
use), the scope of the research questions and the scope of the policy<br />
question, hamper the use of cost-effectiv<strong>en</strong>ess analysis in <strong>de</strong>cision making<br />
• Effectiv<strong>en</strong>ess and cost-effectiv<strong>en</strong>ess are only two of many consi<strong>de</strong>rations<br />
in making policy choices. Economic evid<strong>en</strong>ce needs to be combined with<br />
stakehol<strong>de</strong>r <strong>de</strong>liberation.<br />
3.4 COST-EFFECTIVENESS ANALYSIS, ICER THRESHOLD<br />
VALUES AND DECISION MAKING<br />
Internationally there is an ongoing methodological <strong>de</strong>bate on what could be the ro<strong>le</strong> of<br />
CEA and ICERs in health care <strong>de</strong>cision making. International ag<strong>en</strong>cies such as the World<br />
Health Organization u and the World Bank v promote the use of CEA. Eich<strong>le</strong>r and<br />
col<strong>le</strong>agues 125 predict “CE thresholds will gradually become a reality, irrespective of whether<br />
local <strong>de</strong>cision makers welcome them or remain critical, because it is meaning<strong>le</strong>ss to perform<br />
CE-studies in the abs<strong>en</strong>ce of an acceptance threshold […] neither theory nor empiric evid<strong>en</strong>ce<br />
supports the expectation that CE thresholds will evolve as the so<strong>le</strong> <strong>de</strong>cision criterion” (p525)<br />
Although some scholars are convinced about the pervasiv<strong>en</strong>ess of the use of ICERs,<br />
there still is fundam<strong>en</strong>tal methodological <strong>de</strong>bate on the foundations for using an ICER<br />
threshold value in <strong>de</strong>cision making. The methodological issues have be<strong>en</strong> discussed<br />
ext<strong>en</strong>sively previously in this report. Some authors have docum<strong>en</strong>ted that curr<strong>en</strong>tly<br />
accepted thresholds are <strong>de</strong>termined rather arbitrarily, and that further methodological<br />
<strong>de</strong>bate is nee<strong>de</strong>d. 47, 48, 140 A large number of factors might be expected to g<strong>en</strong>erate<br />
variation in the cost-effectiv<strong>en</strong>ess of healthcare interv<strong>en</strong>tions across locations. 137<br />
Argum<strong>en</strong>ts have be<strong>en</strong> <strong>de</strong>veloped that differ<strong>en</strong>tial threshold values are nee<strong>de</strong>d for<br />
diverse disease and treatm<strong>en</strong>t characteristics (e.g. higher thresholds for life-saving<br />
treatm<strong>en</strong>ts), age, g<strong>en</strong><strong>de</strong>r and race factors, and argum<strong>en</strong>ts are being <strong>de</strong>veloped to<br />
<strong>de</strong>velop equity adjustm<strong>en</strong>t procedures to cost-effectiv<strong>en</strong>ess thresholds. 141<br />
The main message of the critiques is that in real world <strong>de</strong>cision making some of the<br />
theoretical assumptions of ICER threshold values do not hold (see 2.7) and<br />
consi<strong>de</strong>rations of cost-effectiv<strong>en</strong>ess are insuffici<strong>en</strong>t to inform <strong>de</strong>cision makers.<br />
Moreover there remains the issue of implicit and explicit threshold values. Many<br />
countries do not use explicit thresholds for coverage <strong>de</strong>cisions (see also 3.5), whi<strong>le</strong><br />
some countries use an implicit ICER threshold value, above which the <strong>de</strong>cision would<br />
usually be negative (e.g. Australia, New Zealand and Canada).<br />
u The “Making Choices in Health: WHO Gui<strong>de</strong> to Cost-Effectiv<strong>en</strong>ess Analysis” seeks to provi<strong>de</strong> analysts<br />
with a method of assessing whether the curr<strong>en</strong>t as well as proposed mix of interv<strong>en</strong>tions is effici<strong>en</strong>t. It<br />
also seeks to maximize the g<strong>en</strong>eralizability of results across settings.<br />
http://www.who.int/choice/<strong>en</strong>/in<strong>de</strong>x.html<br />
v The World <strong>de</strong>velopm<strong>en</strong>t report 1993 “Investing in health” proposed a universal method to set health<br />
priorities for all countries based on the c<strong>en</strong>tral i<strong>de</strong>a that priority in allocating means and resources should<br />
go to prob<strong>le</strong>ms that cause a large disease burd<strong>en</strong> and with cost-effective interv<strong>en</strong>tions that are availab<strong>le</strong>.
<strong>KCE</strong> reports 100 ICER Thresholds 41<br />
The most ext<strong>en</strong>sive discussion on the use of ICER threshold values by governm<strong>en</strong>t<br />
ag<strong>en</strong>cies can be found in the UK. In the UK, argum<strong>en</strong>ts have be<strong>en</strong> <strong>de</strong>veloped why it is<br />
improper to apply a specific threshold. 61 For a number of reasons, NICE formally rejects<br />
the use of an absolute ICER threshold value for judging the <strong>le</strong>vel of acceptability of a<br />
technology: 67<br />
“Firstly, there is no c<strong>le</strong>ar empirical basis for <strong>de</strong>ciding at what value a threshold should be set.<br />
Secondly, there may be circumstances, as discussed below, in which the Institute would want to<br />
ignore a threshold ev<strong>en</strong> if one could be <strong>de</strong>fined. Thirdly, to set a threshold would imply,<br />
unreasonably, that effici<strong>en</strong>cy (health maximisation) had an absolute priority over other<br />
objectives (particularly equity or fairness). Fourthly, many of the supply industries whose<br />
products are appraised by the Institute are monopolies or oligopolies with high R&D costs but<br />
low production costs. Consequ<strong>en</strong>tly, there are natural t<strong>en</strong>d<strong>en</strong>cies towards monopoly pricing and<br />
a threshold would provi<strong>de</strong> an inc<strong>en</strong>tive to set prices to achieve an ICER just below the threshold<br />
and discourage price competition” w. In the same discussion context, NICE adopted a formal<br />
standpoint on the use of sci<strong>en</strong>tific and social values x “Social value judgem<strong>en</strong>ts are equally<br />
necessary but are concerned with the societal values embodied, explicitly or implicitly, in the<br />
Institute’s advice. The need for judgem<strong>en</strong>ts of this kind is in<strong>de</strong>p<strong>en</strong>d<strong>en</strong>t of the sci<strong>en</strong>tific or<br />
empirical validity of the evid<strong>en</strong>ce, and is concerned with what should be consi<strong>de</strong>red to be<br />
appropriate for the NHS”<br />
Therefore, judgm<strong>en</strong>ts about whether ICERs can be consi<strong>de</strong>red ‘reasonab<strong>le</strong>’ are ma<strong>de</strong> by<br />
in<strong>de</strong>p<strong>en</strong>d<strong>en</strong>t members of NICE's advisory committees (particularly the appraisal y<br />
committee) and the gui<strong>de</strong>line <strong>de</strong>velopm<strong>en</strong>t groups. Moreover, <strong>de</strong>cision makers have to<br />
judge anyway whether an ICER repres<strong>en</strong>ts good value by following a ‘ru<strong>le</strong> of thumb’<br />
rather than looking formally at opportunity cost (see for more <strong>de</strong>tails in 3.5.2). Decision<br />
makers have a very imperfect i<strong>de</strong>a of the costs and b<strong>en</strong>efits of curr<strong>en</strong>t health care<br />
interv<strong>en</strong>tions which have not always be<strong>en</strong> systematically docum<strong>en</strong>ted. Therefore, it is<br />
not always c<strong>le</strong>ar whether existing interv<strong>en</strong>tions or alternatives should (continue to) be<br />
reimbursed.<br />
Based on an analysis of cost-effectiv<strong>en</strong>ess research in US public health policy, Grosse et<br />
al (2007, p. 382) 114 conclu<strong>de</strong> that “although CEA methods pose ethical chal<strong>le</strong>nges, excluding<br />
cost-effectiv<strong>en</strong>ess as a consi<strong>de</strong>ration is also ethically prob<strong>le</strong>matic. Ultimately cost is an issue of<br />
fairness as well as of effici<strong>en</strong>cy. CEA findings should be used as inputs in a <strong>de</strong>liberative evid<strong>en</strong>ce<br />
based <strong>de</strong>cision making process that consi<strong>de</strong>rs the viewpoints and values of multip<strong>le</strong><br />
stakehol<strong>de</strong>rs.”<br />
The use of ICER threshold values is paradoxical. On the one hand it appears to be an<br />
easy way to communicate about the comp<strong>le</strong>x issue of effici<strong>en</strong>t use of public means. On<br />
the other hand the methodological prob<strong>le</strong>ms associated with <strong>de</strong>fining the value of the<br />
ICER threshold are an argum<strong>en</strong>t for <strong>de</strong>cision makers to maintain the <strong>de</strong>liberation and<br />
negotiation process.<br />
Economic evaluation (CEA or ICERs) cannot provi<strong>de</strong> a blue-print solution for <strong>de</strong>cision<br />
making. At best, it supports the process of a more rationalised <strong>de</strong>cision making process.<br />
Multip<strong>le</strong> criteria have to be discussed for setting priorities in the allocation of<br />
constrained resources.<br />
The observation that in priority setting multip<strong>le</strong> criteria play a ro<strong>le</strong> and that <strong>de</strong>cisions<br />
are the result of comp<strong>le</strong>x processes has <strong>le</strong>d to the exploration of multi-criteria <strong>de</strong>cision<br />
analysis (MCDA) techniques. Baltuss<strong>en</strong> and Niess<strong>en</strong> 103, 142 argue that MCDA may be an<br />
important tool towards a more rational priority setting process in health care,<br />
promoting the use of quantitative rather than qualitative analysis.<br />
w http://www.gserve.nice.org.uk/niceMedia/Pdf/boardmeeting/brdmay04item6.pdf<br />
x http://www.gserve.nice.org.uk/niceMedia/Pdf/boardmeeting/brdmay04item6.pdf<br />
y NICE c<strong>le</strong>arly distinguishes ‘assessm<strong>en</strong>t’ from ‘appraisal’. Assessm<strong>en</strong>t refers to the review of the evid<strong>en</strong>ce<br />
about how well a group of similar treatm<strong>en</strong>ts work, and whether they offer value for money. The<br />
assessm<strong>en</strong>t report forms the basis for the appraisal. Appraisal refers to the formal assessm<strong>en</strong>t of the<br />
quality of research evid<strong>en</strong>ce and its re<strong>le</strong>vance to the clinical question or gui<strong>de</strong>line un<strong>de</strong>r consi<strong>de</strong>ration,<br />
according to pre<strong>de</strong>termined criteria. The Appraisal Committee <strong>de</strong>velops NICE’s guidance about using<br />
drugs or treatm<strong>en</strong>ts in the NHS (see http://www.nice.org.uk/website/glossary/).
42 ICER Thresholds <strong>KCE</strong> Reports 100<br />
But they also pay att<strong>en</strong>tion to the ro<strong>le</strong> of advisory panels in the <strong>de</strong>finition of the<br />
re<strong>le</strong>vant criteria and their relative importance for priority setting, and in making<br />
recomm<strong>en</strong>dations for reallocating resources on the basis of MCDA results.<br />
Daniels (oft<strong>en</strong> in collaboration with Sabin) has ma<strong>de</strong> some ethical ref<strong>le</strong>ctions on the<br />
<strong>de</strong>cision making process within health care groun<strong>de</strong>d in <strong>de</strong>mocratic and social justice<br />
theory. The basic princip<strong>le</strong> in Daniels’ ref<strong>le</strong>ction 143-148 is that a fair process is nee<strong>de</strong>d to<br />
establish <strong>le</strong>gitimacy for critical resource allocation <strong>de</strong>cisions. This process, labe<strong>le</strong>d as<br />
“accountability for reasonab<strong>le</strong>ness”, is based on <strong>de</strong>liberative processes (or <strong>de</strong>mocracy)<br />
at differ<strong>en</strong>t <strong>de</strong>cision making <strong>le</strong>vels. It puts forward four conditions for fair rationing that<br />
would <strong>en</strong>ab<strong>le</strong> to break op<strong>en</strong> the black box of health care <strong>de</strong>cision making:<br />
• Publicity: <strong>de</strong>cisions and the rationa<strong>le</strong>s for <strong>de</strong>cisions must be accessib<strong>le</strong>.<br />
This princip<strong>le</strong> implies that careful consi<strong>de</strong>ration should be giv<strong>en</strong> by<br />
<strong>de</strong>cision makers to the argum<strong>en</strong>ts for the choices they make. Publicity<br />
would <strong>le</strong>ad to transpar<strong>en</strong>cy and coher<strong>en</strong>ce in the giving of reasons.<br />
• Re<strong>le</strong>vance: The grounds for <strong>de</strong>cisions must be ones that peop<strong>le</strong>, who seek<br />
to cooperate with others on terms that are mutually justifiab<strong>le</strong>, can accept<br />
as re<strong>le</strong>vant to meet citiz<strong>en</strong>s or pati<strong>en</strong>ts needs fairly un<strong>de</strong>r resource<br />
constraints. This is particularly re<strong>le</strong>vant wh<strong>en</strong> claims are ma<strong>de</strong> that<br />
treatm<strong>en</strong>ts or technologies “cost too much” to be reimbursed, especially<br />
in situations where information is lacking to make these claims. In such<br />
cases it is of particular importance for the “<strong>le</strong>gitimacy” of the <strong>de</strong>mocratic<br />
process to be explicit about the procedures to take <strong>de</strong>cisions and <strong>de</strong>velop<br />
argum<strong>en</strong>ts un<strong>de</strong>rlying the (moral) reasons to take <strong>de</strong>cisions. Decision<br />
makers should be ma<strong>de</strong> “accountab<strong>le</strong> for the reasonab<strong>le</strong>ness”.<br />
• Appeals: there must be mechanisms to chal<strong>le</strong>nge and resolve limit-setting<br />
<strong>de</strong>cisions to revise and improve policies in the light of new evid<strong>en</strong>ce or<br />
argum<strong>en</strong>t.<br />
• Regulation: there must be some form of regulation to <strong>en</strong>sure that the<br />
previous conditions are met. These regulations could come through<br />
governm<strong>en</strong>tal regulation or through voluntary auto-regulation (but<br />
Daniels p<strong>le</strong>ads for an accreditation of this latter governance mo<strong>de</strong>l)<br />
Elaborating further on the notion of accountability for reasonab<strong>le</strong>ness, Gruskin and<br />
Daniels (2008) propose a human rights approach. 149 A human rights approach sets out a<br />
process that requires analyzing which rights and which populations would be positively<br />
or negatively affected by each interv<strong>en</strong>tion. Specific att<strong>en</strong>tion must be paid to who<br />
would b<strong>en</strong>efit most, and in what ways, from each interv<strong>en</strong>tion, and who would be <strong>le</strong>ft<br />
out.<br />
Key points<br />
• In real world <strong>de</strong>cision making, economic evaluations alone are not<br />
suffici<strong>en</strong>t to inform policy makers.<br />
• Neither theory nor empirical evid<strong>en</strong>ce supports the expectation that<br />
ICER threshold values will evolve as the so<strong>le</strong> <strong>de</strong>cision criterion. CEA<br />
findings should be used as inputs in a <strong>de</strong>liberative evid<strong>en</strong>ce based <strong>de</strong>cision<br />
making process that consi<strong>de</strong>rs the viewpoints and values of multip<strong>le</strong><br />
stakehol<strong>de</strong>rs.<br />
• Ensuring ‘accountability for reasonab<strong>le</strong>ness’ requires op<strong>en</strong>ing the black<br />
box of health care <strong>de</strong>cision making. Four conditions contribute to the<br />
‘accountability for reasonab<strong>le</strong>ness’: publicity, re<strong>le</strong>vance, appeals and<br />
regulation.
<strong>KCE</strong> reports 100 ICER Thresholds 43<br />
3.5 THE USE OF ICER THRESHOLD VALUES IN OTHER<br />
COUNTRIES<br />
Decisions that influ<strong>en</strong>ce the diffusion and uptake of technologies can be influ<strong>en</strong>ced by<br />
many differ<strong>en</strong>t factors such as availab<strong>le</strong> (public) resources, reimbursem<strong>en</strong>t mechanisms,<br />
regulatory frameworks and cultural and social <strong>de</strong>terminants (e.g. attitu<strong>de</strong>s towards<br />
technological innovations). It is therefore reasonab<strong>le</strong> to expect that threshold values will<br />
not be id<strong>en</strong>tical in differ<strong>en</strong>t countries, 125 as budgets and prefer<strong>en</strong>ces (might) differ.<br />
Decision makers may use implicit or explicit threshold values. Explicit threshold values<br />
means that <strong>de</strong>cision makers have formally adopted and ma<strong>de</strong> public a threshold by<br />
which their <strong>de</strong>cisions on resource allocation will be bound. By contrast, implicit<br />
thresholds are not official or public, but may be inferred retrospectively by analysis of<br />
the <strong>de</strong>cision making pattern in a giv<strong>en</strong> health-care system. 125 In this section we examine<br />
to what ext<strong>en</strong>t explicit ICER threshold values are used in health care policy <strong>de</strong>cisions in<br />
a se<strong>le</strong>ction of countries.<br />
3.5.1 Methodology<br />
We searched for existing writt<strong>en</strong> material about the use of economic consi<strong>de</strong>rations in<br />
health policy. Writt<strong>en</strong> docum<strong>en</strong>ts, oft<strong>en</strong> grey literature retrieved through the Internet,<br />
were scrutinized to find clues about the exist<strong>en</strong>ce and the use of ICER threshold values<br />
in health policy. We started by consulting the ISPOR website z to see whether<br />
pharmacoeconomic gui<strong>de</strong>lines were published for the se<strong>le</strong>cted countries and which<br />
organization was the author of the gui<strong>de</strong>lines (HTA ag<strong>en</strong>cies or others). Those<br />
gui<strong>de</strong>lines and the website of the authors’ organization were scrutinized for the use of<br />
ICER threshold values. In a next step, the websites of the health <strong>de</strong>partm<strong>en</strong>ts of the<br />
national (or local) governm<strong>en</strong>ts and the websites of the national (or local) bodies<br />
responsib<strong>le</strong> for <strong>de</strong>cision making and reimbursem<strong>en</strong>t <strong>de</strong>cisions about pharmaceuticals<br />
were consulted for further re<strong>le</strong>vant information.<br />
A summary of the findings for each country inclu<strong>de</strong>d in our review is provi<strong>de</strong>d in Tab<strong>le</strong><br />
3.<br />
3.5.2 England and Wa<strong>le</strong>s aa<br />
NICE, the National Institute for Health and Clinical Excel<strong>le</strong>nce in the UK, set an explicit<br />
threshold value as from 2002. 150<br />
NICE’s “Gui<strong>de</strong> to the Methods of Technology Appraisal 2004”, 151 m<strong>en</strong>tioned two<br />
threshold values: £20 000 and £30 000 per QALY gained. In November 2007, NICE<br />
issued a new draft “gui<strong>de</strong> to the methods of technology appraisal” for consultation. 152<br />
The consultation process continued until 29 February 2008. The updated gui<strong>de</strong> was<br />
published in June 2008. 153<br />
With respect to the threshold values, the gui<strong>de</strong> states:<br />
“The Appraisal Committee does not use a precise ICER threshold above which a technology<br />
would automatically be <strong>de</strong>fined as not cost effective or below which it would. Giv<strong>en</strong> the fixed<br />
budget of the NHS, the appropriate threshold to be consi<strong>de</strong>red is that of the opportunity cost<br />
of programmes displaced by new, more costly technologies. Therefore, the Appraisal Committee<br />
judges cost effectiv<strong>en</strong>ess in relation to the cost effectiv<strong>en</strong>ess of interv<strong>en</strong>tions curr<strong>en</strong>tly fun<strong>de</strong>d by<br />
the NHS and those previously agreed by the Committee to be cost ineffective. Consi<strong>de</strong>ration of<br />
the cost effectiv<strong>en</strong>ess of a technology is a necessary, but is not the so<strong>le</strong>, basis for <strong>de</strong>cision<br />
making. Consequ<strong>en</strong>tly, the Institute consi<strong>de</strong>rs technologies in relation to a threshold range,<br />
betwe<strong>en</strong> which other factors have an increasing influ<strong>en</strong>ce upon the <strong>de</strong>cision to recomm<strong>en</strong>d a<br />
technology.”<br />
z (“Pharmacoeconomic Gui<strong>de</strong>lines around the World” http://www.ispor.org/PEgui<strong>de</strong>lines/in<strong>de</strong>x.asp/<br />
aa Website consulted, accessed autumn 2008: National Institute for Health and Clinical Excel<strong>le</strong>nce<br />
(http://www.nice.org.uk)
44 ICER Thresholds <strong>KCE</strong> Reports 100<br />
3.5.3 Canada cc<br />
In this paragraph, it is stated that no precise ICER threshold value is used for <strong>de</strong>cision<br />
making but that rather a threshold range is <strong>de</strong>fined, based on the ICERs of curr<strong>en</strong>tly<br />
fun<strong>de</strong>d health interv<strong>en</strong>tions.<br />
This means that, whi<strong>le</strong> the Appraisal Committee is c<strong>le</strong>ar about not wanting to use a<br />
sing<strong>le</strong> threshold value for funding <strong>de</strong>cisions, it neverthe<strong>le</strong>ss chooses to follow a certain<br />
guiding princip<strong>le</strong> with respect to <strong>de</strong>cisions about interv<strong>en</strong>tions in relation to their<br />
ICERs.<br />
The guiding princip<strong>le</strong>s are as follows:<br />
• For interv<strong>en</strong>tions with an ICER < £20 000/QALY gained, <strong>de</strong>cisions will<br />
primarily be gui<strong>de</strong>d by cost-effectiv<strong>en</strong>ess consi<strong>de</strong>rations. In princip<strong>le</strong>, the<br />
recomm<strong>en</strong>dation will be to provi<strong>de</strong> this interv<strong>en</strong>tion, un<strong>le</strong>ss there are<br />
major doubts about the plausibility of and/or certainty around the<br />
estimated ICER. Thus, account is tak<strong>en</strong> of the results of the s<strong>en</strong>sitivity<br />
analysis and pot<strong>en</strong>tial limitations to the g<strong>en</strong>eralizability of the findings<br />
regarding effectiv<strong>en</strong>ess.<br />
• For interv<strong>en</strong>tions with an ICER betwe<strong>en</strong> £20 000/QALY gained and<br />
£30 000/QALY gained, NICE takes account of the following factors:<br />
o The <strong>de</strong>gree of (un)certainty about the ICER.<br />
o Whether there are strong reasons to indicate that the assessm<strong>en</strong>t of<br />
the HRQoL has ina<strong>de</strong>quately captured, and may therefore<br />
misrepres<strong>en</strong>t, the health utility gained.<br />
o The innovative nature of the technology, specifically where the<br />
innovation adds b<strong>en</strong>efits of a substantial nature compared with<br />
availab<strong>le</strong> alternatives which may not have be<strong>en</strong> captured in the QALY<br />
measure.<br />
• For interv<strong>en</strong>tions with an ICER > £30 000/QALY gained the same factors<br />
will be tak<strong>en</strong> into account. A stronger case is nee<strong>de</strong>d on these factors to<br />
approve such interv<strong>en</strong>tions.<br />
These guiding princip<strong>le</strong>s suggest that ‘additional e<strong>le</strong>m<strong>en</strong>ts’, are only explicitly consi<strong>de</strong>red<br />
in the <strong>de</strong>cision making process by the Appraisal Committee if the ICER exceeds<br />
£20 000/QALY. For interv<strong>en</strong>tions with ICERs below £20 000/QALY the only additional<br />
e<strong>le</strong>m<strong>en</strong>t consi<strong>de</strong>red is the uncertainty around the estimate of the ICER.<br />
Rec<strong>en</strong>tly, a report has be<strong>en</strong> published on the relative societal value of health gains<br />
(QALYs) to differ<strong>en</strong>t population groups in the UK. 44 The study id<strong>en</strong>tified attributes that<br />
<strong>de</strong>termine societal prefer<strong>en</strong>ces for the allocation of health gains and estimated equity<br />
weights for QALYs based on data from 688 interviews in peop<strong>le</strong> from the g<strong>en</strong>eral<br />
public. bb<br />
In Canada, cost-effectiv<strong>en</strong>ess data are formally required for all new outpati<strong>en</strong>t<br />
medications since 1996. 94 Giv<strong>en</strong> this long-standing requirem<strong>en</strong>t for economic evid<strong>en</strong>ce<br />
in Canada, it might be expected that the ro<strong>le</strong> of ICERs and ICER threshold values is well<br />
established in that country.<br />
Despite the formal requirem<strong>en</strong>t for cost-effectiv<strong>en</strong>ess evid<strong>en</strong>ce from the Canadian<br />
Ag<strong>en</strong>cy for Drugs and Technologies in Health (CADTH), no information was found on<br />
how this economic evid<strong>en</strong>ce is used for <strong>de</strong>cision making, neither from the CADTH<br />
published gui<strong>de</strong>lines 154 , nor from the other websites consulted.<br />
bb Before the interviews in the g<strong>en</strong>eral public were performed, re<strong>le</strong>vant attributes for societal prefer<strong>en</strong>ces<br />
were id<strong>en</strong>tified through focus groups with 57 members of the g<strong>en</strong>eral public and from a survey of 172<br />
NHS employees.<br />
cc Website consulted, accessed autumn 2008: the Canadian Ag<strong>en</strong>cy for Drugs and Technologies in Health<br />
(http://www.cadth.ca), including the Common Drug Review that conducts rigorous reviews of the clinical<br />
and cost effectiv<strong>en</strong>ess of drugs, and provi<strong>de</strong>s formulary listing recomm<strong>en</strong>dations to the publicly fun<strong>de</strong>d<br />
drug plans in Canada (except Québec).
<strong>KCE</strong> reports 100 ICER Thresholds 45<br />
In the Canadian literature, Laupacis et al. 155 suggested in 1992 that evid<strong>en</strong>ce for<br />
adoption of a new interv<strong>en</strong>tion in Canada was strong with an ICER below<br />
CAN$20 000/QALY (CAN$ of the year 1990), mo<strong>de</strong>rate with an ICER betwe<strong>en</strong><br />
CAN$20 000/QALY gained and CAN$100 000/QALY gained, and weak if the ICER<br />
exceeds CAN$100 000/QALY gained. Laupacis et al. 155 acknow<strong>le</strong>dged however that<br />
these lower and upper boundaries were arbitrary.<br />
Furthermore, there is no formal evid<strong>en</strong>ce that any of these boundaries has be<strong>en</strong><br />
accepted or used by any Canadian <strong>de</strong>cision making institution. 94<br />
Rec<strong>en</strong>tly, Rocchi et al. 94 reviewed the published (September 2003 to March 2007) drug<br />
reimbursem<strong>en</strong>t recomm<strong>en</strong>dations g<strong>en</strong>erated by the advisory board of the Common<br />
Drug Review (CEDAC – Canadian Expert Drug Advisory Committee) in or<strong>de</strong>r to<br />
id<strong>en</strong>tify the ro<strong>le</strong> of economic evaluations and indicate whether an implicit threshold was<br />
used. Of the 62 fi<strong>le</strong>s reviewed, ICERs were consi<strong>de</strong>red in <strong>le</strong>ss than half of the cases<br />
(40%, 25 fi<strong>le</strong>s), including 12 negative recomm<strong>en</strong>dations and 13 positive<br />
recomm<strong>en</strong>dations. Medications with a positive recomm<strong>en</strong>dation ranged from dominant<br />
to CAN$80 000/QALY. Medications with a negative recomm<strong>en</strong>dation ranged from<br />
CAN$32 000/QALY gained to CAN$137 000/QALY gained. From this, Rocchi et al. 94<br />
conclu<strong>de</strong>d that these implicit thresholds did not act as a c<strong>le</strong>ar <strong>de</strong>marcation line, because<br />
the ICER range for medications with a positive recomm<strong>en</strong>dation overlapped with the<br />
ICER range for medications with a negative recomm<strong>en</strong>dation. They also suggested that<br />
the resulting inconsist<strong>en</strong>cy in which ICERs <strong>le</strong>ad to a positive recomm<strong>en</strong>dation may be<br />
due to the fact that other factors are consi<strong>de</strong>red in the context of a specific review.<br />
3.5.4 The Netherlands dd<br />
The Dutch Health care Insurance Board (CVZ, Col<strong>le</strong>ge voor Zorgverzekering<strong>en</strong>)<br />
examines the basic package of care to which all Dutch pati<strong>en</strong>ts have access. They<br />
provi<strong>de</strong> the Ministry of Health, Welfare and Sport (VWS, Volksgezondheid Welzijn <strong>en</strong><br />
Sport,) with advice about what care should be ad<strong>de</strong>d or removed. The Pharmaceutical<br />
Aid Committee is an expert committee within the CVZ that assists in assessing whether<br />
new medicines need to be inclu<strong>de</strong>d in the basic package and be reimbursed. To be<br />
consi<strong>de</strong>red for reimbursem<strong>en</strong>t, manufacturers are formally required to provi<strong>de</strong> costeffectiv<strong>en</strong>ess<br />
data of all new drugs for which they claim an ad<strong>de</strong>d-value. With this<br />
respect, CVZ has edited the “Dutch Gui<strong>de</strong>lines for Pharmacoeconomic Research” since<br />
1999, with an updated version published in April 2006. 16 Those gui<strong>de</strong>lines do not<br />
m<strong>en</strong>tion how the pharmacoeconomic information is used for making <strong>de</strong>cisions about<br />
the reimbursem<strong>en</strong>t of drugs in the Drug Reimbursem<strong>en</strong>t System.<br />
The Council for Public Health and Health Care (Raad voor <strong>de</strong> Volksgezondheid <strong>en</strong><br />
Zorg) is an in<strong>de</strong>p<strong>en</strong>d<strong>en</strong>t body advising the governm<strong>en</strong>t on public health and health care.<br />
At the <strong>en</strong>d of June 2006, the Council published the report “Zinnige <strong>en</strong> Duurzame Zorg”<br />
(“S<strong>en</strong>sib<strong>le</strong> and Sustainab<strong>le</strong> Care”) that addresses issues such as which criteria should be<br />
applied in or<strong>de</strong>r to id<strong>en</strong>tify priorities for the funding of care from col<strong>le</strong>ctive resources.<br />
The Council divi<strong>de</strong>s the process of <strong>de</strong>ciding which forms of care should or should not<br />
be fun<strong>de</strong>d from col<strong>le</strong>ctive resources into four phases:<br />
• Ag<strong>en</strong>da-setting (scoping): <strong>de</strong>fining the priorities for the <strong>de</strong>cision making<br />
process (urg<strong>en</strong>cy princip<strong>le</strong>)<br />
• Assessm<strong>en</strong>t (quantifiab<strong>le</strong> criteria): disease burd<strong>en</strong>, efficacy and costeffectiv<strong>en</strong>ess<br />
• Appraisal (non-quantifiab<strong>le</strong> criteria): community review of the outcome of<br />
the assessm<strong>en</strong>t phase, princip<strong>le</strong>s of fairness and solidarity<br />
• Imp<strong>le</strong>m<strong>en</strong>tation: <strong>en</strong>suring the forms of care that have be<strong>en</strong> id<strong>en</strong>tified as<br />
warranting funding from col<strong>le</strong>ctive resources are in<strong>de</strong>ed fun<strong>de</strong>d in this<br />
way, and that other forms of care are not<br />
dd Websites consulted, accessed autumn 2008: the Health Care Insurance Board (http://www.cvz.nl/),<br />
Ministry of Health, Welfare and Sport (http://www.minvws.nl/<strong>en</strong>/), the Council for Public Health and<br />
Health Care (http://www.rvz.net/)
46 ICER Thresholds <strong>KCE</strong> Reports 100<br />
3.5.5 USA ee<br />
The Council specifies that, for this <strong>de</strong>cision making process to function optimally, the<br />
Minister must <strong>de</strong>fine an acceptab<strong>le</strong> limit for some parameters, such as the disease<br />
burd<strong>en</strong> and the cost-effectiv<strong>en</strong>ess value. The Council believes that it is not <strong>en</strong>tit<strong>le</strong>d to<br />
<strong>de</strong>fine such threshold values and that a <strong>de</strong>mocratic discussion has to <strong>de</strong>termine the<br />
limit. In or<strong>de</strong>r to foster the discussion about this topic, the Council suggests an absolute<br />
maximum ICER threshold value of €80 000/QALY gained, provi<strong>de</strong>d that the disease<br />
severity in<strong>de</strong>x exceeds a specific threshold value. 156<br />
The in<strong>de</strong>x ref<strong>le</strong>cting disease severity is obtained following the gui<strong>de</strong>lines from the Dutch<br />
Health Insurance Board. 157 The measure ref<strong>le</strong>cts the health-related quality of life<br />
associated with a specific condition and is based on the number of QALYs lost due to<br />
the disease relative to the number of QALYs expected without the disease.<br />
Therefore, although ICERs are consi<strong>de</strong>red in the <strong>de</strong>cision making process of the<br />
Council, no explicit ICER threshold value has be<strong>en</strong> <strong>de</strong>fined so far in The Netherlands.<br />
Other factors also play an important ro<strong>le</strong> in this process.<br />
In the USA, the figure of US$50 000/QALY gained has frequ<strong>en</strong>tly be<strong>en</strong> quoted for many<br />
years as being the cost-effectiv<strong>en</strong>ess threshold value. 125, 140 Hirth et al. 158 report that this<br />
number was originally based on the supposed annual cost per QALY for the Medicare<br />
program for pati<strong>en</strong>ts with chronic r<strong>en</strong>al failure, but they further argue that this standard<br />
might have be<strong>en</strong> based on a consi<strong>de</strong>rab<strong>le</strong> un<strong>de</strong>restimation of the chronic r<strong>en</strong>al failure<br />
program’s true costs.<br />
Rec<strong>en</strong>tly, Braithwaite et al. 96 investigated whether the advocated $50 000/QALY ru<strong>le</strong> is<br />
consist<strong>en</strong>t with curr<strong>en</strong>t resource allocation <strong>de</strong>cisions in the US. They estimated a lower<br />
bound for the societal WTP per LYG by calculating the increm<strong>en</strong>tal b<strong>en</strong>efits of all<br />
medical advances since 1950 in terms of mortality reduction and the associated<br />
increm<strong>en</strong>tal costs. They simulated the costs and health outcomes in a US birth cohort<br />
without the medical advances and the health outcomes and costs with the medical<br />
advances. Major assumptions about the mortality reduction and costs attributab<strong>le</strong> to<br />
medical advances had to be ma<strong>de</strong>. Based on the simulation, they estimated the ICER for<br />
‘mo<strong>de</strong>rn’ health care. From the empirical observation that most individuals in the US<br />
favour expanding the health care budget, they inferred that society’s WTP for health<br />
care must exceed the ICER of mo<strong>de</strong>rn health care and therefore the ICER threshold<br />
value must be higher than the estimated lower bound.<br />
The estimate of the upper bound for the societal WTP for a LYG was based on<br />
observed peop<strong>le</strong>’s <strong>de</strong>cisions not to buy unsubsidized insurance ev<strong>en</strong> if they are not<br />
insured otherwise. The approach assumes that individual’s unwillingness to get insured<br />
(ev<strong>en</strong> wh<strong>en</strong> income is suffici<strong>en</strong>tly high) implies societal unwillingness to pay. The costs<br />
and b<strong>en</strong>efits associated with and without unsubsidized insurance are simulated and used<br />
to obtain an ICER for insurance. According to Braithwaite et al. the prefer<strong>en</strong>ce not to<br />
get insured may point towards an upper-bound estimate for the societal WTP. The base<br />
case analysis suggests $183 000/LYG and $264 000/LYG as plausib<strong>le</strong> lower and upper<br />
bounds for the ICER threshold value. Wh<strong>en</strong> both quantity and quality of life were<br />
consi<strong>de</strong>red, in their s<strong>en</strong>sitivity analysis, the lower and upper bounds became<br />
$109 000/QALY and $297 000/QALY respectively. Braithwaite et al. 96 conclu<strong>de</strong> that an<br />
ICER threshold value of $50 000/QALY is not consist<strong>en</strong>t with curr<strong>en</strong>t allocation<br />
<strong>de</strong>cisions in the US. As the plausib<strong>le</strong> lower and upper bounds for the ICER are<br />
substantially higher than $50 000/QALY, it is very unlikely that this ICER threshold value<br />
is consist<strong>en</strong>t with societal prefer<strong>en</strong>ces in the United States.<br />
Despite the exist<strong>en</strong>ce of such thresholds published in the US literature, so far, the<br />
C<strong>en</strong>ters for Medicare and Medicaid Services have avoi<strong>de</strong>d the explicit use of costeffectiv<strong>en</strong>ess<br />
criteria in their coverage <strong>de</strong>cisions and it is unc<strong>le</strong>ar to what <strong>de</strong>gree costeffectiv<strong>en</strong>ess<br />
is used to gui<strong>de</strong> coverage <strong>de</strong>cisions in the private sector. 140<br />
ee Website consulted, accessed autumn 2008: the C<strong>en</strong>ters for Medicare & Medicaid Services<br />
(http://www.cms.hhs.gov/), the Aca<strong>de</strong>my of Managed Care Pharmacy (http://www.amcp.org/)
<strong>KCE</strong> reports 100 ICER Thresholds 47<br />
3.5.6 Australia ff<br />
In their rec<strong>en</strong>t editorial, Weinstein 140 adds that because economic evaluations are not<br />
used in any systematic or consist<strong>en</strong>t way in the United States, it should not be<br />
surprising that there is no cons<strong>en</strong>sus as to the appropriate value of the cost per QALY<br />
that should gui<strong>de</strong> health care <strong>de</strong>cisions and policies. 140 Further, in the Format for<br />
Formulary Submissions (the suggested template for performing pharmacoeconomic<br />
evaluations in the US) published by the Aca<strong>de</strong>my of Managed Care Pharmacy in 2005,<br />
we found no information about the formal exist<strong>en</strong>ce of an explicit ICER threshold value.<br />
The ro<strong>le</strong> of the Pharmaceutical B<strong>en</strong>efits Advisory Committee (PBAC) in Australia is to<br />
recomm<strong>en</strong>d to the Minister for Health which drugs and medicinal preparations should<br />
be subsidised by the Australian Governm<strong>en</strong>t un<strong>de</strong>r the Pharmaceutical B<strong>en</strong>efits Scheme.<br />
It further advises the Minister and the Pharmaceutical B<strong>en</strong>efits Pricing Authority about<br />
the cost-effectiv<strong>en</strong>ess (‘value for money’) of a proposed drug compared with other<br />
drugs already listed in the Pharmaceutical B<strong>en</strong>efits Scheme for the same, or similar,<br />
indications.<br />
The PBAC has edited Gui<strong>de</strong>lines for Preparing Submissions to the Pharmaceutical<br />
B<strong>en</strong>efits Advisory Committee whose aim is to provi<strong>de</strong> practical information (including<br />
guidance for economic evaluations) to the pharmaceutical industry for making a<br />
submission to PBAC. 14 The gui<strong>de</strong>lines are also int<strong>en</strong><strong>de</strong>d to help PBAC assess<br />
submissions. Wh<strong>en</strong> making choices betwe<strong>en</strong> competing therapeutic modalities, the<br />
factors consi<strong>de</strong>red by PBAC are cost-effectiv<strong>en</strong>ess, but also other important factors<br />
which inclu<strong>de</strong> uncertainty, equity, ext<strong>en</strong>t of use and total costs. A threshold un<strong>de</strong>r<br />
which an ICER is consi<strong>de</strong>red attractive by the PBAC is not explicitly specified in the<br />
gui<strong>de</strong>lines.<br />
Although the PBAC does not appear to work with an explicit threshold value, H<strong>en</strong>ry et<br />
al. 95 report that observation of the <strong>de</strong>cisions of the PBAC betwe<strong>en</strong> 1994 and 2003 point<br />
to an appar<strong>en</strong>t threshold of AU$69 900/QALY gained above which reimbursem<strong>en</strong>t has<br />
be<strong>en</strong> found to be unlikely. There is however no evid<strong>en</strong>ce that this implicit threshold is<br />
effectively used to gui<strong>de</strong> PBAC’s <strong>de</strong>cisions.<br />
3.5.7 New Zealand gg<br />
The ro<strong>le</strong> of the Pharmaceutical Managem<strong>en</strong>t Ag<strong>en</strong>cy (PHARMAC) in New Zealand is to<br />
manage the pharmaceutical budget on behalf of the District Health Boards, and to<br />
<strong>de</strong>ci<strong>de</strong> which medicines are fun<strong>de</strong>d by the Governm<strong>en</strong>t. Rec<strong>en</strong>tly, PHARMAC published<br />
an updated version of The Prescription for Pharmacoeconomic Analysis (June 2007)<br />
which <strong>de</strong>scribes how PHARMAC un<strong>de</strong>rtakes its economic evaluations and how it<br />
interprets ICERs. 159<br />
As stated in their ‘Prescription for Pharmacoeconomic Analysis’-report, PHARMAC<br />
<strong>de</strong>ci<strong>de</strong>d not to <strong>de</strong>fine an explicit threshold value below which a pharmaceutical is<br />
consi<strong>de</strong>red ‘cost-effective’. 160 Their justification for this is as follows:<br />
• The main reason for this is that cost-effectiv<strong>en</strong>ess is only one <strong>de</strong>cision<br />
criterion used by PHARMAC. One proposal may be more cost-effective<br />
than another but rate poorly on other <strong>de</strong>cision criteria and therefore may<br />
not be fun<strong>de</strong>d (h<strong>en</strong>ce, on ‘successfulness grounds’, it will not be<br />
consi<strong>de</strong>red cost-effective).<br />
ff Websites consulted, accessed autumn 2008: Pharmaceutical B<strong>en</strong>efits Advisory Committee<br />
(http://www.health.gov.au/internet/main/publishing.nsf/Cont<strong>en</strong>t/Pharmaceutical%20B<strong>en</strong>efits%20Scheme%2<br />
0(PBS)-1), the Australian Governm<strong>en</strong>t, Departm<strong>en</strong>t of Health and Ageing.<br />
gg Website consulted, accessed autumn 2008: the Pharmaceutical Managem<strong>en</strong>t Ag<strong>en</strong>cy in New Zealand<br />
(http://www.pharmac.govt.nz/)
48 ICER Thresholds <strong>KCE</strong> Reports 100<br />
3.5.8 Finland hh<br />
3.5.9 Swed<strong>en</strong> ii<br />
• Another reason for not having a fixed ICER threshold value is that the<br />
sp<strong>en</strong>ding on community pharmaceuticals is required to be kept within a<br />
fixed budget within a giv<strong>en</strong> year. Giv<strong>en</strong> the binding nature of this<br />
constraint and all things being equal, what is and is not consi<strong>de</strong>red ‘costeffective’<br />
varies with the amount of funding availab<strong>le</strong> (not just in terms of<br />
the total budget each year, but the availab<strong>le</strong> budget at any point in time, as<br />
explained in 2.5.3: a fixed budget requires a variab<strong>le</strong> ICER threshold<br />
value).<br />
Pritchard et al. 93 speculated that PHARMAC’s <strong>de</strong>cisions are broadly consist<strong>en</strong>t with an<br />
implicit threshold of NZ$20 000/QALY (NZ$ of the year 2000). The authors did not,<br />
however, perform a systematic analysis of PHARMAC’s <strong>de</strong>cision reports.<br />
The Finnish ‘gui<strong>de</strong>lines for preparing a health economic evaluation’ are published as an<br />
annex to the Decree by the Ministry of Social Affairs and Health on applications for a<br />
reasonab<strong>le</strong> who<strong>le</strong>sa<strong>le</strong> price, on special reimbursem<strong>en</strong>t status for a medicinal product,<br />
and on the docum<strong>en</strong>tation to be attached to the application (<strong>de</strong>cree 1111/2005).<br />
Despite the formal requirem<strong>en</strong>t to provi<strong>de</strong> health economic evaluations for new<br />
chemical <strong>en</strong>tities in Finland, we could not id<strong>en</strong>tify the use of any explicit ICER threshold<br />
value, neither from the websites consulted (including the <strong>le</strong>gislation of the institutions),<br />
nor from the Finnish pharmacoeconomic gui<strong>de</strong>lines.<br />
In 2003, the Swedish Pharmaceutical B<strong>en</strong>efits Board published g<strong>en</strong>eral gui<strong>de</strong>lines for<br />
conducting economic evaluations. 161 The English version of those gui<strong>de</strong>lines does not<br />
contain information about the use of an ICER threshold value for <strong>de</strong>cision making.<br />
The main task of the Pharmaceutical B<strong>en</strong>efits Board in Swed<strong>en</strong> is to ascertain if a<br />
pharmaceutical or medical <strong>de</strong>vice is to be inclu<strong>de</strong>d in the pharmaceutical b<strong>en</strong>efits<br />
scheme and be reimbursed by society. The Pharmaceutical B<strong>en</strong>efits Board weighs three<br />
criteria (‘princip<strong>le</strong>s’) wh<strong>en</strong> making its <strong>de</strong>cisions:<br />
• The human value princip<strong>le</strong>; which un<strong>de</strong>rlines the respect for equality of all<br />
human beings and the integrity of every individual.<br />
• The need and solidarity princip<strong>le</strong>; which says that those in greatest need<br />
take preced<strong>en</strong>ce wh<strong>en</strong> it comes to reimbursing pharmaceuticals. In other<br />
words, peop<strong>le</strong> with more severe diseases are prioritised over peop<strong>le</strong> with<br />
<strong>le</strong>ss severe conditions.<br />
• The cost-effectiv<strong>en</strong>ess princip<strong>le</strong>; which states that the cost for using a<br />
medicine should be reasonab<strong>le</strong> from a medical, humanitarian and socia<strong>le</strong>conomic<br />
perspective.<br />
Cost-effectiv<strong>en</strong>ess thus appears to be a c<strong>en</strong>tral concern in the Swedish reimbursem<strong>en</strong>t<br />
system, 162 but a threshold value un<strong>de</strong>r which a treatm<strong>en</strong>t is consi<strong>de</strong>red cost-effective<br />
does not appear to be explicitly stated. Other factors, besi<strong>de</strong>s cost-effectiv<strong>en</strong>ess, are<br />
also weighed in the <strong>de</strong>cision making process of the Pharmaceutical B<strong>en</strong>efit Board.<br />
hh Websites consulted, accessed autumn 2008: the Finnish Office for Health Technology Assessm<strong>en</strong>t<br />
(http://finohta.stakes.fi), the Finnish Ministry of Social Affairs and Health (http://www.stm.fi), the<br />
Pharmaceuticals Pricing Board (the body responsib<strong>le</strong> for pricing <strong>de</strong>cision and operating un<strong>de</strong>r the control<br />
of the Ministry of Social Affairs and Health), the Social Insurance Institution (the body responsib<strong>le</strong> for the<br />
reimbursem<strong>en</strong>t of pharmaceuticals, http://www.kela.fi).<br />
ii Websites consulted, accessed autumn 2008: the Swedish Council on Technology Assessm<strong>en</strong>t in Health<br />
Care (http://www.sbu.se), the C<strong>en</strong>tre for Medical Technology Assessm<strong>en</strong>t (http://www.cmt.liu.se), the<br />
Swedish Pharmaceutical B<strong>en</strong>efits Board (http://www.lfn.se/), the National C<strong>en</strong>tre for Priority Setting in<br />
Health Care (http://e.lio.se/prioriteringsc<strong>en</strong>trum)
<strong>KCE</strong> reports 100 ICER Thresholds 49<br />
3.5.10 Norway jj<br />
The Norwegian Medicines Ag<strong>en</strong>cy approves medicines and monitors their use, and<br />
<strong>en</strong>sures effici<strong>en</strong>t, effective and well-docum<strong>en</strong>ted use of medicines. On its website, the<br />
Norwegian Medicines Ag<strong>en</strong>cy has published gui<strong>de</strong>lines for pharmacoeconomic analysis in<br />
connection with applications for reimbursem<strong>en</strong>t in 2005. 163 Those gui<strong>de</strong>lines do not<br />
inform on wh<strong>en</strong> an interv<strong>en</strong>tion is consi<strong>de</strong>red cost-effective.<br />
From the websites consulted, no explicit threshold value could be id<strong>en</strong>tified in Norway.<br />
3.5.11 D<strong>en</strong>mark kk<br />
The Health Technology Assessm<strong>en</strong>t Handbook published in 2007 reports the views of<br />
DACEHTA (the Danish C<strong>en</strong>tre for Health Technology Assessm<strong>en</strong>t) on wh<strong>en</strong> to<br />
consi<strong>de</strong>r a technology cost-effective. 164 They state that wh<strong>en</strong> two technologies do not<br />
dominate one another and an increm<strong>en</strong>tal cost-effectiv<strong>en</strong>ess ratio (ICER) is calculated<br />
for those two interv<strong>en</strong>tions, whether we should accept the new (or old) technology as<br />
being cost-effective ultimately <strong>de</strong>p<strong>en</strong>d upon the maximum price that the <strong>de</strong>cision maker<br />
is willing to pay for the extra effect. DACEHTA further reports that there are no<br />
g<strong>en</strong>erally applicab<strong>le</strong> limits concerning what can be consi<strong>de</strong>red a reasonab<strong>le</strong> QALY price.<br />
Tab<strong>le</strong> 3: Explicit, implicit and assumed ICER threshold values in other<br />
countries<br />
Country Authors ICER threshold<br />
Explicit ICER threshold range<br />
UK NICE151 £20 000 - £30 000 per QALY<br />
Implicit ICER threshold values or ranges based on past allocation <strong>de</strong>cisions<br />
Australia H<strong>en</strong>ry et al. and the AU$69 900 per QALY<br />
PBAC 95<br />
New Zealand Pritchard et al. and<br />
PHARMAC 93<br />
Canada Rocchi et al. and the<br />
CDR 94<br />
NZ$20 000 per QALY<br />
ICER threshold values or ranges proposed by individuals or institutions<br />
USA Weinstein 140 $50 000 per QALY<br />
Range of acceptance: dominant to CAN$80 000<br />
per QALY<br />
Range of rejection: CAN$31 000 to<br />
CAN$137 000 per QALY<br />
USA Braithwaite et al. 96 $109 000 - $297 000 per QALY<br />
The<br />
The Council for Public<br />
Netherlands Health and Health Care156 €80 000 per QALY<br />
Canada Laupacis et al. 155 CAN$20 000 to CAN$100 000 per QALY<br />
No ICER threshold values or ranges id<strong>en</strong>tified<br />
Finland, Swed<strong>en</strong>, Norway, D<strong>en</strong>mark<br />
CDR: Common Drug Review; NICE: National Institute for Health and Clinical Excel<strong>le</strong>nce;<br />
PBAC: Pharmaceutical B<strong>en</strong>efits Advisory Committee; PHARMAC: Pharmaceutical<br />
Managem<strong>en</strong>t Ag<strong>en</strong>cy.<br />
jj Websites consulted, accessed autumn 2008: The Norwegian Medicines Ag<strong>en</strong>cy<br />
(http://www.<strong>le</strong>gemid<strong>de</strong>lverket.no), the Norwegian Know<strong>le</strong>dge C<strong>en</strong>tre for the Health Services<br />
(http://www.nokc.no, this website is in Norwegian only), The Ministry of Health and Care Services<br />
(http://www.regjering<strong>en</strong>.no/), The Norwegian Labour and Welfare Organisation (responsib<strong>le</strong> for<br />
reimbursem<strong>en</strong>t and medical b<strong>en</strong>efits, http://www.nav.no).<br />
kk Websites consulted, accessed autumn 2008: the Danish C<strong>en</strong>tre for Evaluation and Health Technology<br />
Assessm<strong>en</strong>t (http://www.dacehta.dk), the Danish Institute for Health Services Research<br />
(http://www.dsi.dk), the Danish medicine ag<strong>en</strong>cy (http://www.dkma.dk/)
50 ICER Thresholds <strong>KCE</strong> Reports 100<br />
Key points<br />
• In the UK, a discussion is ongoing on the use of ICER threshold values.<br />
Curr<strong>en</strong>tly the UK uses an explicit threshold range of £20 000 to £30 000<br />
per QALY gained.<br />
• Implicit threshold values based on past allocation <strong>de</strong>cisions were<br />
published in Australia (AU$69 900 / QALY gained), New Zealand<br />
(NZ$20 000 / QALY gained) and Canada (range of acceptance: dominant<br />
to $80 000 per QALY gained, range of rejection: $31 000 to $137 000 per<br />
QALY gained).<br />
• Thresholds values or ranges proposed by individuals or institutions were<br />
found in the USA (50 000$/QALY), in the Netherlands (€80 000 per<br />
QALY gained) and in Canada (20 000 - 100 000$/QALY).<br />
• The lower and upper limit for the social WTP for a QALY in the US was<br />
estimated to be 109 000$/QALY and 297 000$/QALY respectively.<br />
• With the exception of the UK, no explicit ICER threshold value (or<br />
range) is used in the countries examined.<br />
• In all countries <strong>de</strong>cision making is not so<strong>le</strong>ly based on cost-effectiv<strong>en</strong>ess<br />
consi<strong>de</strong>rations. The technology is assessed based on the threshold range<br />
together with other criteria. In the pres<strong>en</strong>ce of high ICERs, those other<br />
criteria become more important.<br />
• In most countries it appears that interv<strong>en</strong>tions with a low ICER are more<br />
likely to become accepted than interv<strong>en</strong>tions with a high ICER.<br />
3.6 THE USE OF ICER THRESHOLD VALUES IN BELGIUM<br />
Additional to the international comparison a limited field study explored the use of<br />
cost-effectiv<strong>en</strong>ess evaluations in health care reimbursem<strong>en</strong>t <strong>de</strong>cisions in Belgium. We<br />
focused on two committees: the Drug Reimbursem<strong>en</strong>t Committee (DRC or<br />
CTG/CRM) and the Technical Committee for Implants (TCI or TRI/CTI), both having an<br />
important ro<strong>le</strong> in reimbursem<strong>en</strong>t <strong>de</strong>cisions.<br />
3.6.1 Background on DRC and TCI<br />
The <strong>de</strong>cision to reimburse pharmaceutical products and <strong>de</strong>vices in Belgium is tak<strong>en</strong> by<br />
the Minister of Social Affairs after consultation with the Minister of Budget, but the<br />
Minister is advised on these matters by the DRC and the TCI respectively. Both<br />
committees are organised within the NIHDI but their structure, working procedures<br />
and place in the <strong>de</strong>cision making process are differ<strong>en</strong>t.<br />
The DRC advises the Minister of Social Affairs directly. Although the advice is not<br />
strictly binding, the Minister can only <strong>de</strong>viate from the advice formulated by the DRC<br />
for social or budgetary reasons. The composition of the DRC and the procedures for<br />
formulating a reimbursem<strong>en</strong>t proposal for a pharmaceutical product are stipulated in<br />
two Royal Decrees, approved on 21/12/2001. 165, 166 The DRC is composed of<br />
repres<strong>en</strong>tatives of sickness funds (the mutualities), universities, medical doctors and<br />
pharmacists. Repres<strong>en</strong>tatives of the pharmaceutical industry, the Ministry of Economic<br />
Affairs, the Ministry of Budget, the Ministry of Social Affairs, the Ministry of Public<br />
Health and the NIHDI can att<strong>en</strong>d the meetings and participate in the discussions but<br />
have no voting rights. The DRC has to formulate an advice within strict time limits: the<br />
<strong>de</strong>cision to reimburse a pharmaceutical product has to be tak<strong>en</strong> within 180 days after<br />
the submission of a reimbursem<strong>en</strong>t request fi<strong>le</strong> by a pharmaceutical company. The<br />
advice of the DRC with respect to reimbursem<strong>en</strong>t has to reach the Minister of Social<br />
Affairs at day 150 at the latest.
<strong>KCE</strong> reports 100 ICER Thresholds 51<br />
According to the European Transpar<strong>en</strong>cy Directive 167 any <strong>de</strong>cision not to reimburse a<br />
pharmaceutical product must contain a statem<strong>en</strong>t of the reasons “based upon objective<br />
and verifiab<strong>le</strong> criteria, including, if appropriate, any expert opinions or recomm<strong>en</strong>dations on<br />
which the <strong>de</strong>cision is based”. ll<br />
The TCI does not provi<strong>de</strong> direct advice to the Minister of Social Affairs. The TCI is<br />
embed<strong>de</strong>d in the historically <strong>de</strong>veloped NIHDI structures for reimbursem<strong>en</strong>t <strong>de</strong>cisions,<br />
where an advice of the TCI is typically first discussed in a conv<strong>en</strong>tion or agreem<strong>en</strong>t<br />
commission (overe<strong>en</strong>komst<strong>en</strong>- <strong>en</strong> akkoord<strong>en</strong>commissie / commission <strong>de</strong> conv<strong>en</strong>tions ou<br />
d’accords) before it is s<strong>en</strong>t to the Insurance Committee (Verzekeringscomité/Comité <strong>de</strong><br />
l’assurance <strong>soins</strong> <strong>de</strong> <strong>santé</strong>) and the Commission for Budgetary Control (Commissie voor<br />
begrotingscontro<strong>le</strong> / Commission du contrô<strong>le</strong> budgétaire). Its composition and procedures<br />
are stipulated in the ‘Sickness and Invalidity Insurance law’ 169 and a Royal Decree<br />
stipulating the practical working procedures within the TCI. 170 The TCI is composed of<br />
repres<strong>en</strong>tatives of sickness funds, universities and hospital pharmacists. Repres<strong>en</strong>tatives<br />
of the Ministry of Social Affairs and the Ministry of Public Health can att<strong>en</strong>d the<br />
discussions but have no voting rights wh<strong>en</strong> it comes to formulating the final advice of<br />
the TCI. As from 2009, the TCI will be reformed and an “Implants and Medical Devices<br />
Reimbursem<strong>en</strong>t Committee” will be established. 171 The structure and working<br />
procedures will from th<strong>en</strong> be similar but not id<strong>en</strong>tical to those of the DRC. For<br />
examp<strong>le</strong>, the committee will also advise the Minister of Social Affairs directly but will<br />
not be subject to the same strict <strong>de</strong>adlines as the DRC.<br />
3.6.2 Aims and methods of the field study<br />
The purpose of the field research is to better un<strong>de</strong>rstand whether and how costeffectiv<strong>en</strong>ess<br />
issues (ICERs) are consi<strong>de</strong>red and discussed in those two advisory<br />
committees within the NIHDI. The purpose of this field study is purely explorative. We<br />
tried to better un<strong>de</strong>rstand how clinical effectiv<strong>en</strong>ess is balanced against costeffectiv<strong>en</strong>ess<br />
and other criteria. Furthermore -if re<strong>le</strong>vant- we tried to grasp whether<br />
specific explicit or implicit ICER threshold values are used and what the opinions of<br />
<strong>de</strong>cision makers in these committees are on the (pot<strong>en</strong>tial) advantages and<br />
disadvantages of CEA and ICER threshold values.<br />
We conducted two group-interviews with respectively members of DRC and TCI. The<br />
interview with members of the DRC was done with members of the “bureau”<br />
(presid<strong>en</strong>t, secretary and two staff members of the NIHDI). The interview with TCI<br />
took place as part of a formal meeting of the TCI: 11 persons participated in the<br />
interview (excluding the NIHDI administrative staff members who were pres<strong>en</strong>t but did<br />
not actively participate in the interview).<br />
To prepare for the group discussions, the researchers studied the formal procedures of<br />
each of the committees.<br />
Each interview was conducted by one mo<strong>de</strong>rator (not the same for the two<br />
committees), who used a checklist of topics to be discussed: Three researchers took<br />
notes. The interviews were reported in a writt<strong>en</strong> common raw data docum<strong>en</strong>t (no<br />
transcripts) <strong>de</strong>veloped by all of the researchers.<br />
A thematic cont<strong>en</strong>t analysis was done col<strong>le</strong>ctively by the researchers based on these<br />
notes.<br />
ll The criteria which are tak<strong>en</strong> into account by the DRC in <strong>de</strong>ciding whether or not to reimburse a product<br />
are inclu<strong>de</strong>d in the Royal Decree of 21/12/2001.168 They inclu<strong>de</strong> the therapeutic value (taking into<br />
account the efficacy, effectiv<strong>en</strong>ess, si<strong>de</strong> effects, applicability and user-fri<strong>en</strong>dliness of the product), the<br />
market price and the requested reimbursem<strong>en</strong>t price, the clinical effectiv<strong>en</strong>ess and likely impact of the<br />
product (taking into account therapeutic and social needs), the budget impact for the NIHDI and for<br />
Class I products (drugs for which the company claims ad<strong>de</strong>d therapeutic value compared to existing<br />
drugs) the cost-effectiv<strong>en</strong>ess of the product from the NIHDI perspective.
52 ICER Thresholds <strong>KCE</strong> Reports 100<br />
3.6.3 Results of the field study<br />
The results in this section ref<strong>le</strong>ct the researchers’ summary of what has be<strong>en</strong> said<br />
during the interviews based on the notes tak<strong>en</strong> during the interviews. This <strong>de</strong>scription<br />
does not necessarily ref<strong>le</strong>ct the official position of the full committees and should not be<br />
interpreted as such.<br />
3.6.3.1 Drug Reimbursem<strong>en</strong>t Committee<br />
As <strong>de</strong>scribed previously, the DRC has to work according to a very formal and timelimited<br />
procedure to assess a reimbursem<strong>en</strong>t request and advise the minister on the<br />
reimbursem<strong>en</strong>t of a pharmaceutical product. This formal procedure has a major impact<br />
on the information used and the preparation of the <strong>de</strong>cision making ag<strong>en</strong>da within the<br />
committee. Especially the particular time constraints have an influ<strong>en</strong>ce on the <strong>de</strong>cision<br />
making process. For Class I pharmaceutical productsmm the reimbursem<strong>en</strong>t request fi<strong>le</strong><br />
submitted by the pharmaceutical company must contain a pharmacoeconomic<br />
evaluation. Therefore, according to the bureau, cost-effectiv<strong>en</strong>ess of pharmaceuticals is<br />
an ess<strong>en</strong>tial issue in the <strong>de</strong>cision making process. Internal NIHDI experts scrutinize the<br />
reimbursem<strong>en</strong>t request: they can ask the company to provi<strong>de</strong> the e<strong>le</strong>ctronic economic<br />
mo<strong>de</strong>l to verify the mo<strong>de</strong>l in-<strong>de</strong>pth and search additional literature and verify the<br />
literature review. The experts prepare an evaluation fi<strong>le</strong> which is pres<strong>en</strong>ted to all DRC<br />
members and discussed during a meeting.<br />
It is repeatedly un<strong>de</strong>rlined that the formal procedures and time constraints put a lot of<br />
pressure on the handling of reimbursem<strong>en</strong>t dossiers. This simp<strong>le</strong> procedural fact makes<br />
the work of the committee members, and the NIHDI experts sometimes stressful. The<br />
expected pace of handling dossiers also impacts on the <strong>de</strong>cision making process. The<br />
preparatory work of the NIHDI experts is fundam<strong>en</strong>tal to the <strong>de</strong>cision making process<br />
itself: members heavily rely on this preparatory work (without necessarily always<br />
agreeing with or following the evaluation of the NIHDI experts).<br />
The curr<strong>en</strong>t organisation of the <strong>de</strong>cision making process is recognised to “rationalize”<br />
the <strong>de</strong>cision making process. Especially since the committee has to c<strong>le</strong>arly justify its<br />
<strong>de</strong>cision. Although the factual <strong>de</strong>cision making process is not free of emotional and<br />
other factors, rational argum<strong>en</strong>ts are se<strong>en</strong> as an ess<strong>en</strong>tial part. It was m<strong>en</strong>tioned during<br />
the interview that for instance media can increase the societal pressure on (members<br />
of) the committee. Moreover within the committee differ<strong>en</strong>t stakehol<strong>de</strong>rs are<br />
repres<strong>en</strong>ted, <strong>le</strong>ading to situations where members also try to <strong>de</strong>f<strong>en</strong>d specific interests.<br />
The formal preparatory stage requires a c<strong>le</strong>ar pre-assessm<strong>en</strong>t of cost-effectiv<strong>en</strong>ess<br />
issues of a pharmaceutical product. The interviewees recognise that it is not always<br />
<strong>en</strong>tirely c<strong>le</strong>ar whether high <strong>le</strong>vel clinical evid<strong>en</strong>ce is readily availab<strong>le</strong> based on the<br />
submitted dossiers: it oft<strong>en</strong> lacks information on “hard” outcomes. It is said that, wh<strong>en</strong><br />
evid<strong>en</strong>ce on hard outcomes is lacking, the NIHDI experts doing the pre-assessm<strong>en</strong>t are<br />
more inclined to question the validity of the economic evaluation. Sometimes it is ev<strong>en</strong><br />
perceived that ICERs are used in the dossiers to conceal the lack of c<strong>le</strong>ar clinical<br />
evid<strong>en</strong>ce.<br />
The availab<strong>le</strong> budget is judged as being a far more important criterion for taking<br />
<strong>de</strong>cisions on pharmaceuticals than ICERs. This does however not mean that this<br />
criterion is <strong>de</strong>cisive: the budget impact is approached in a f<strong>le</strong>xib<strong>le</strong> way. It is c<strong>le</strong>ar for the<br />
respond<strong>en</strong>ts that budget impact remains a fundam<strong>en</strong>tal criterion in the ultimate voting<br />
outcome. The number of pati<strong>en</strong>ts that can be served within the budget constraints is an<br />
additional consi<strong>de</strong>ration.<br />
The <strong>de</strong>cision making process and the criteria used are c<strong>le</strong>arly differ<strong>en</strong>t wh<strong>en</strong> the pati<strong>en</strong>t<br />
population are childr<strong>en</strong>. The committee t<strong>en</strong>ds to be more to<strong>le</strong>rant for childr<strong>en</strong>, and<br />
tries to un<strong>de</strong>rstand what a product would mean for the future of the child.<br />
mm Class I pharmaceutical products are so-cal<strong>le</strong>d innovative products for which the company claims an ad<strong>de</strong>d<br />
therapeutic value compared to existing drugs.
<strong>KCE</strong> reports 100 ICER Thresholds 53<br />
ICERs are used t<strong>en</strong>tatively, as one of differ<strong>en</strong>t criteria. Moreover, the DRC does not<br />
use a formal ICER threshold value wh<strong>en</strong> assessing pharmaceutical products: one<br />
interviewee ev<strong>en</strong> opposes the use of an “absolute” threshold value in the <strong>de</strong>cision<br />
making process, as each pharmaceutical product has its own particularities. nn In answer<br />
to the question whether the committee makes a differ<strong>en</strong>ce betwe<strong>en</strong> outcome measures<br />
(LYG or QALY) used in the ICER, the answer was negative. oo<br />
The perceived therapeutic need and the perceived therapeutic ad<strong>de</strong>d value are the<br />
factual and informal <strong>de</strong>cision criteria used during the <strong>de</strong>liberation. Additional criteria<br />
such as “is it a true innovation” or an adaptation or alternative of already availab<strong>le</strong><br />
product, pp has an impact on the assessm<strong>en</strong>t of medical necessity.<br />
In case of a revision of a product (revision of reimbursem<strong>en</strong>t <strong>de</strong>cision after the product<br />
has be<strong>en</strong> on the market for some time), the committee expects in princip<strong>le</strong> more<br />
information (more real life data both on clinical and cost-effectiv<strong>en</strong>ess) about the<br />
product, since it is th<strong>en</strong> also more feasib<strong>le</strong> for the industry to pres<strong>en</strong>t more and better<br />
data.<br />
It is also recognised by the interviewees that the interpretation of cost-effectiv<strong>en</strong>ess<br />
studies within the <strong>de</strong>cision making process in the committee had to go hand in hand<br />
with a <strong>le</strong>arning process of the members. For examp<strong>le</strong>, it took a gradual process to make<br />
c<strong>le</strong>ar to the members that cost-effectiv<strong>en</strong>ess is not the same as cost-saving. Moreover,<br />
it is not c<strong>le</strong>ar whether all members of the committee truly un<strong>de</strong>rstand the<br />
methodological background of an ICER, but neverthe<strong>le</strong>ss the use of an ICER <strong>le</strong>ads to<br />
expressions as “€80 000 per QALY is high”.<br />
In other words, our respond<strong>en</strong>ts recognise that the <strong>de</strong>cision making process within the<br />
DRC, although being rationalized and substantiated with clinical and economic data,<br />
remains a <strong>de</strong>liberation process of peop<strong>le</strong>, in which differ<strong>en</strong>t formal and informal criteria<br />
are used.<br />
3.6.3.2 Technical Council for Implants<br />
Although the working and <strong>de</strong>cision making procedure of the TCI will be modified in<br />
2009, the curr<strong>en</strong>t process of assessing a technology within TCI is <strong>le</strong>ss formally <strong>de</strong>fined<br />
than in DRC. It is recognized that in the future the need for more objectified criteria in<br />
the <strong>de</strong>cision making process will be nee<strong>de</strong>d. Therefore, it has be<strong>en</strong> <strong>de</strong>ci<strong>de</strong>d to work<br />
more along the lines of the DRC procedures from 2009 onwards, with the<br />
establishm<strong>en</strong>t of the ‘Implants and Medical Devices Reimbursem<strong>en</strong>t Committee’.<br />
The committee makes a differ<strong>en</strong>ce betwe<strong>en</strong> so-cal<strong>le</strong>d ‘me too’ <strong>de</strong>mands or <strong>de</strong>mands for<br />
reimbursem<strong>en</strong>t of ‘new technologies’. Demands for the latter have to be substantiated<br />
by the industry with clinical studies, which are oft<strong>en</strong> not availab<strong>le</strong> for implants and<br />
medical <strong>de</strong>vices. Oft<strong>en</strong> experts from the field are invited and heard by the committee to<br />
substantiate the existing know<strong>le</strong>dge and the information on the clinical effectiv<strong>en</strong>ess of a<br />
new <strong>de</strong>vice.<br />
The members of the TCI say that they consi<strong>de</strong>r cost-effectiv<strong>en</strong>ess as a <strong>de</strong>cision<br />
criterion. Members of the committee report to be aware of the re<strong>le</strong>vance of costeffectiv<strong>en</strong>ess<br />
of implants, but simultaneously state that CEAs or publications are not<br />
assessed on a systematic basis for <strong>de</strong>cisions within the committee.<br />
nn Neverthe<strong>le</strong>ss, we found an examp<strong>le</strong> of a threshold value of 30 000 €/LYG in a motivation docum<strong>en</strong>t for a<br />
reimbursem<strong>en</strong>t <strong>de</strong>cision of the Minister of Social Affairs<br />
(http://www.riziv.fgov.be/inami_prd/SSP/CNS2/Pages/MinisterialDecisionDet.asp?qs_SpcCod=00642119&q<br />
s_EffDat=20071101&qs_MdId=5023).<br />
oo In the evaluation reports prepared by the evaluators of the NIHDI the distinction is neverthe<strong>le</strong>ss<br />
sometimes ma<strong>de</strong> and retained by the Minister of Social Affairs in his motivation for reimbursem<strong>en</strong>t (for<br />
examp<strong>le</strong><br />
http://www.riziv.fgov.be/inami_prd/SSP/CNS2/Pages/MinisterialDecisionDet.asp?qs_SpcCod=00581188&q<br />
s_EffDat=20070301&qs_MdId=6175)<br />
pp 1.5 to 3 years after the initial reimbursem<strong>en</strong>t request and a positive reimbursem<strong>en</strong>t <strong>de</strong>cision for a class I<br />
pharmaceutical product, companies have to submit a revision fi<strong>le</strong>. This fi<strong>le</strong> should contain evid<strong>en</strong>ce on the<br />
effectiv<strong>en</strong>ess and cost-effectiv<strong>en</strong>ess of the product in real life situations.
54 ICER Thresholds <strong>KCE</strong> Reports 100<br />
3.6.3.3 Summary<br />
As implants are oft<strong>en</strong> pres<strong>en</strong>ted for reimbursem<strong>en</strong>t in early stages of use, feasibility to<br />
find and use sci<strong>en</strong>tific (clinical and economic) information is se<strong>en</strong> as a major prob<strong>le</strong>m.<br />
Moreover the availab<strong>le</strong> studies are not always consi<strong>de</strong>red re<strong>le</strong>vant, especially if they<br />
come from large organisations or c<strong>en</strong>tres of excel<strong>le</strong>nce.<br />
The <strong>de</strong>liberation and <strong>de</strong>cision making process is to a large ext<strong>en</strong>t negotiated and expert<br />
opinion based. The ro<strong>le</strong> of the staff members of the NIHDI is <strong>le</strong>ss ext<strong>en</strong>sive than in the<br />
DRC. The preliminary work is done by the working groups, specialised in specific<br />
domains (e.g. cardiovascular implants, orthopaedic implants etc). The working groups<br />
prepare an advice, which is afterwards discussed in the p<strong>le</strong>nary TCI meetings. Advice<br />
can be typified as sometimes supported by availab<strong>le</strong> (economic) studies, rather than<br />
systematically based on CEA or economic evaluations. ICERs are not used in the<br />
<strong>de</strong>cision making process.<br />
Decision makers focus mainly on the availab<strong>le</strong> budget. These budgetary constraints form<br />
the framework within which <strong>de</strong>cisions on reimbursem<strong>en</strong>t are tak<strong>en</strong>. The procedural<br />
particularity of the <strong>de</strong>cision making process of the committee is that budgets have to be<br />
set and prepared almost one year in advance of the following working year. A budget<br />
has to be reserved for the following year, but estimations of the budget nee<strong>de</strong>d are not<br />
always accurate, sometimes <strong>le</strong>ading to specific prob<strong>le</strong>ms.<br />
In the answers of the committee members it became c<strong>le</strong>ar that cost issues (in terms of<br />
reimbursem<strong>en</strong>t) and cost saving issues are consi<strong>de</strong>red more than cost-effectiv<strong>en</strong>ess<br />
issues. Costs are not always estimated or calculated, they are approached rather<br />
intuitively and experi<strong>en</strong>ce-based. They stressed that it is oft<strong>en</strong> rather obvious to<br />
<strong>de</strong>monstrate the ad<strong>de</strong>d value of a product.<br />
Besi<strong>de</strong>s this economic e<strong>le</strong>m<strong>en</strong>t it is stressed that quality of life (not expressed as a<br />
QALY) is a re<strong>le</strong>vant criterion to steer the <strong>de</strong>cision making process.<br />
One of the members of the committee explicitly refers to the differ<strong>en</strong>ce betwe<strong>en</strong> a<br />
theoretical and a political ratio in reimbursem<strong>en</strong>t <strong>de</strong>cision making processes. The<br />
differ<strong>en</strong>ce implies that other criteria than clinical effectiv<strong>en</strong>ess or cost effectiv<strong>en</strong>ess have<br />
to be consi<strong>de</strong>red. Cost effectiv<strong>en</strong>ess analysis can be of value but cannot be consi<strong>de</strong>red<br />
as the so<strong>le</strong> criterion to base reimbursem<strong>en</strong>t <strong>de</strong>cisions upon for implants. Moreover, it is<br />
m<strong>en</strong>tioned that the committee also has to consi<strong>de</strong>r other interests and has to work in a<br />
context of societal and media pressure.<br />
The ways in which the two se<strong>le</strong>cted committees operate illustrate the growing<br />
awar<strong>en</strong>ess of the pot<strong>en</strong>tial re<strong>le</strong>vance of clinical evid<strong>en</strong>ce and economic evaluation<br />
studies. However, factors <strong>de</strong>scribed in <strong>de</strong>cision making literature are equally affecting<br />
the <strong>de</strong>cision making process. Efforts are ma<strong>de</strong> to “rationalise” the <strong>de</strong>cision making<br />
process and substantiate <strong>de</strong>mands for reimbursem<strong>en</strong>t with sci<strong>en</strong>tific evid<strong>en</strong>ce. It helps<br />
to make the <strong>de</strong>cision making criteria more transpar<strong>en</strong>t. But it also has to be stressed<br />
that the <strong>de</strong>cision making process remains an interactive <strong>de</strong>liberation process, which is<br />
certainly not to be reduced to the technocratic rational application of sci<strong>en</strong>tific (clinical<br />
and economic) findings: <strong>de</strong>cisions on reimbursem<strong>en</strong>t are negotiated and can only be<br />
un<strong>de</strong>rstood taking into account circumstantial factors.<br />
The DRC is c<strong>le</strong>arly going through a <strong>le</strong>arning curve in the use of cost-effectiv<strong>en</strong>ess<br />
know<strong>le</strong>dge in their <strong>de</strong>cision making process. Clinical effectiv<strong>en</strong>ess and cost effectiv<strong>en</strong>ess<br />
(including ICERs) are becoming criteria to be docum<strong>en</strong>ted by the firms and are actively<br />
consi<strong>de</strong>red, be it with the necessary critical attitu<strong>de</strong>. It is stressed that <strong>de</strong>cisions on<br />
reimbursem<strong>en</strong>t are affected by a lot more criteria and by the nature of the <strong>de</strong>cision<br />
making process.
<strong>KCE</strong> reports 100 ICER Thresholds 55<br />
The TCI has be<strong>en</strong> ref<strong>le</strong>cting and working on procedures to make their <strong>de</strong>cision making<br />
process more rational and procedural. The TCI will move towards a more formal (<strong>le</strong>gal)<br />
<strong>de</strong>scription of working practices such as the DRC’s, but curr<strong>en</strong>tly both committees’<br />
processes are still c<strong>le</strong>arly distinct. Curr<strong>en</strong>tly the <strong>de</strong>cision making process is not<br />
systematically substantiated yet by sci<strong>en</strong>tific clinical evid<strong>en</strong>ce and economic evaluation<br />
studies. The use of cost-effectiv<strong>en</strong>ess analyses has not yet perva<strong>de</strong>d the <strong>de</strong>cision making<br />
process.<br />
Key points<br />
• Although efforts are ma<strong>de</strong> to ‘rationalise’ the <strong>de</strong>cision making process<br />
and substantiate reimbursem<strong>en</strong>t requests with sci<strong>en</strong>tific evid<strong>en</strong>ce,<br />
<strong>de</strong>cision making processes in Belgium remain mainly an interactive<br />
<strong>de</strong>liberation process.<br />
• Clinical effectiv<strong>en</strong>ess is the most important sci<strong>en</strong>tific criterion used in the<br />
<strong>de</strong>cision making process of both the DRC and the TCI.<br />
• Cost-effectiv<strong>en</strong>ess is sometimes consi<strong>de</strong>red in the DRC but rarely in the<br />
TCI.<br />
• Budget impact is by both committees consi<strong>de</strong>red more important than<br />
the ICER.
56 ICER Thresholds <strong>KCE</strong> Reports 100<br />
4 GENERAL DISCUSSION<br />
4.1 ECONOMIC EVALUATION AND ICERS<br />
The aim of this report was to provi<strong>de</strong> an introduction to economic evaluation and<br />
ICERs for non-health economists and summarize the pot<strong>en</strong>tial str<strong>en</strong>gths and<br />
weaknesses of the use of ICERs and ICER threshold values in health care <strong>de</strong>cision<br />
making.<br />
One important caveat for using health economic evid<strong>en</strong>ce is the comparability of the<br />
methodologies used to obtain the ICER estimate. ICERs are useful for health care policy<br />
makers only in as far as they are comparab<strong>le</strong> betwe<strong>en</strong> interv<strong>en</strong>tions. It is thus<br />
consi<strong>de</strong>red of utmost importance to critically analyse the context for which ICERs have<br />
be<strong>en</strong> calculated. Moreover solutions have to be sought for the existing methodological<br />
variability. Guidance for economic evaluations in health care can reduce methodological<br />
variability. The Belgian pharmacoeconomic gui<strong>de</strong>lines 18 were <strong>de</strong>veloped as a response to<br />
this request for standardisation.<br />
However, guidance is a necessary but not a suffici<strong>en</strong>t condition. Economic evaluations<br />
and more specifically economic mo<strong>de</strong>lsqq oft<strong>en</strong> remain black boxes. Without offering<br />
policy makers the possibility to ‘play’ with an economic mo<strong>de</strong>l, it is unlikely that they<br />
will trust the ICERs resulting from them. The Belgian pharmacoeconomic gui<strong>de</strong>lines<br />
therefore <strong>de</strong>mand that the DRC can ask for the e<strong>le</strong>ctronic version of the<br />
pharmacoeconomic mo<strong>de</strong>l pres<strong>en</strong>ted by the pharmaceutical company in its drug<br />
reimbursem<strong>en</strong>t request, if so <strong>de</strong>sired.<br />
Furthermore, methodological standardization and control does not make the ICER a<br />
blue-print solution for policy making. Although it is g<strong>en</strong>erally accepted that economic<br />
effici<strong>en</strong>cy is important and should be consi<strong>de</strong>red in resource allocation <strong>de</strong>cisions, other<br />
e<strong>le</strong>m<strong>en</strong>ts than effici<strong>en</strong>cy are tak<strong>en</strong> into consi<strong>de</strong>ration in a <strong>de</strong>cision making process. Many<br />
countries use the ICER to inform <strong>de</strong>cision makers about interv<strong>en</strong>tions’ relative value for<br />
money but there is still <strong>de</strong>bate about whether an ICER and more specifically the ICER<br />
threshold value is the most appropriate way to introduce effici<strong>en</strong>cy consi<strong>de</strong>rations in<br />
the <strong>de</strong>cision making process.<br />
The advantages of an explicit ICER threshold value would be an improved transpar<strong>en</strong>cy<br />
and consist<strong>en</strong>cy of <strong>de</strong>cisions, at <strong>le</strong>ast if methodological issues can be <strong>de</strong>alt with in a<br />
satisfactory manner. The drawbacks of using an explicit ICER threshold value might be<br />
the creation of an excessively mechanical <strong>de</strong>cision making process, without<br />
consi<strong>de</strong>ration of other re<strong>le</strong>vant variab<strong>le</strong>s or a t<strong>en</strong>d<strong>en</strong>cy of companies to price up to the<br />
ICER threshold value or manipulate economic mo<strong>de</strong>ls to stay below the ICER threshold<br />
value. rr<br />
4.2 WAYS TO INTRODUCE EFFICIENCY CONSIDERATIONS IN<br />
HEALTH CARE DECISION MAKING<br />
Approaches for bringing effici<strong>en</strong>cy consi<strong>de</strong>rations into the health care <strong>de</strong>cision making<br />
process vary in the ext<strong>en</strong>t to which they accept the ICER and the ICER threshold value.<br />
Most of the approaches remain rather theoretical, due to the practical prob<strong>le</strong>ms<br />
associated with imp<strong>le</strong>m<strong>en</strong>ting them.<br />
The ICER threshold value is g<strong>en</strong>erally perceived as a fixed value against which the ICERs<br />
of interv<strong>en</strong>tions can be compared with to <strong>de</strong>ci<strong>de</strong> whether the new interv<strong>en</strong>tion is costeffective.<br />
Besi<strong>de</strong>s the methodological prob<strong>le</strong>ms m<strong>en</strong>tioned before, this perception also<br />
ignores the differ<strong>en</strong>ce betwe<strong>en</strong> fixed and f<strong>le</strong>xib<strong>le</strong> budget situations.<br />
qq Note that all economic evaluations of health interv<strong>en</strong>tions are based on mo<strong>de</strong>lling to some <strong>de</strong>gree.<br />
Mo<strong>de</strong>ls are used for differ<strong>en</strong>t reasons, e.g. ext<strong>en</strong>sion of time horizons, extrapolation of intermediate<br />
outcome parameters to final outcome parameters, simulation of effectiv<strong>en</strong>ess as compared to efficacy.18<br />
rr NICE has introduced the distinction betwe<strong>en</strong> ‘assessm<strong>en</strong>t’ and ‘appraisal’ to reduce this risk (see 3.4).
<strong>KCE</strong> reports 100 ICER Thresholds 57<br />
• A fixed ICER threshold value is incompatib<strong>le</strong> with a fixed health care<br />
budget. A fixed budget requires the revision of the ICER threshold value<br />
every time a positive reimbursem<strong>en</strong>t <strong>de</strong>cision is tak<strong>en</strong>.<br />
• In a f<strong>le</strong>xib<strong>le</strong> budget context, the ICER threshold value can be <strong>de</strong>fined as<br />
the maximum societal willingness to pay for a QALY (or LYG). At first<br />
sight this approach looks appealing because it seems to take social values<br />
better into account. The imp<strong>le</strong>m<strong>en</strong>tation of this approach requires<br />
however:<br />
o that the societal willingness to pay for a g<strong>en</strong>eric QALY can be<br />
measured. Measurem<strong>en</strong>t of societal WTP for a QALY (or LYG) is<br />
methodologically impossib<strong>le</strong> because the value of a QALY is always<br />
context-<strong>de</strong>p<strong>en</strong>d<strong>en</strong>t.<br />
o that the societal WTP a<strong>de</strong>quately ref<strong>le</strong>cts the value of the health care<br />
sector relative to other sectors. The use of societal WTP for a QALY<br />
(or LYG) as a fixed ICER threshold value will ultimately <strong>de</strong>termine (the<br />
expansion of) the health care budget. The maximum societal WTP for<br />
a QALY will have to be revised regularly to make sure that the health<br />
care budget keeps ref<strong>le</strong>cting the relative societal value of the health<br />
care sector.<br />
Because of these requirem<strong>en</strong>ts, no country uses the maximum societal WTP for a<br />
(g<strong>en</strong>eric) QALY (or LYG) as an ICER threshold value.<br />
Some researchers have tried to <strong>de</strong>rive the societal WTP from past <strong>de</strong>cisions. As<br />
explained in the report, this is an invalid approach, as <strong>de</strong>cisions are never ma<strong>de</strong><br />
in<strong>de</strong>p<strong>en</strong>d<strong>en</strong>t from additional consi<strong>de</strong>rations (e.g. equity, valuing pati<strong>en</strong>t or interv<strong>en</strong>tion<br />
characteristics).<br />
Three approaches id<strong>en</strong>tified in literature suggest to forget about the ICER threshold<br />
value and to focus on other variab<strong>le</strong>s: the ICER of interv<strong>en</strong>tions as such (in comparison<br />
with ICERs of other interv<strong>en</strong>tions but without refer<strong>en</strong>ce to an ICER threshold value),<br />
the opportunity costs of interv<strong>en</strong>tions or the disaggregated outcomes of the economic<br />
evaluation.<br />
• The first approach argues that the ICER is to be compared to the ICER of<br />
another interv<strong>en</strong>tion without refer<strong>en</strong>ce to an ICER threshold value. The<br />
in-betwe<strong>en</strong> comparison of ICERs allows policy makers to draw<br />
conclusions on the relative cost-effectiv<strong>en</strong>ess of interv<strong>en</strong>tions, but does<br />
not give a yes or no answer to the question of whether the interv<strong>en</strong>tion<br />
increases the health care sector’s effici<strong>en</strong>cy. Neither does it allow policy<br />
makers to draw conclusions about the interv<strong>en</strong>tion’s value for money, as<br />
this requires the inclusion of other consi<strong>de</strong>rations, such as equity, in the<br />
<strong>de</strong>cision making process. The advantage of the approach is that it<br />
contributes to ‘objectifying’ the economic effici<strong>en</strong>cy e<strong>le</strong>m<strong>en</strong>t in the<br />
<strong>de</strong>cision making process.<br />
• The opportunity cost approach states that ICERs are not the right vehic<strong>le</strong><br />
for making resource allocation <strong>de</strong>cisions. It requires making explicit where<br />
the resources for financing a new interv<strong>en</strong>tion in a fixed budget context<br />
must come from. This may be difficult in practice, especially in systems<br />
where <strong>de</strong>cisions are typically tak<strong>en</strong> within sub-budgets of the total health<br />
care budget and transfers betwe<strong>en</strong> sub-budgets are not necessarily<br />
consi<strong>de</strong>red (as is the case in Belgium).<br />
• The cost-consequ<strong>en</strong>ces approach refers to the dis<strong>en</strong>tangling of economic<br />
evaluations into concrete e<strong>le</strong>m<strong>en</strong>ts and (economic) outcomes, <strong>en</strong>abling an<br />
explicit weighting of the separate e<strong>le</strong>m<strong>en</strong>ts in the <strong>de</strong>cision making process.<br />
A combination of the differ<strong>en</strong>t approaches will probably offer better support to <strong>de</strong>cision<br />
makers to assess the economic effici<strong>en</strong>cy of interv<strong>en</strong>tions. For examp<strong>le</strong> pres<strong>en</strong>ting the<br />
ICER as well as the disaggregated results of an economic evaluation will allow future<br />
research in the ICER value above which an interv<strong>en</strong>tion does never obtain a positive<br />
<strong>de</strong>cision (irrespective of their ‘score’ on other criteria) but below which a <strong>de</strong>cision can<br />
still be negative
58 ICER Thresholds <strong>KCE</strong> Reports 100<br />
(= an implied ICER threshold, ref<strong>le</strong>cting the absolute maximum society is willing to pay<br />
for an additional QALY or LYG). It will moreover familiarize policy makers with the<br />
ICER ev<strong>en</strong> if they do not wish to give a high weight to the ICER in <strong>de</strong>cisions about<br />
specific technologies. This will ev<strong>en</strong>tually <strong>le</strong>ad to a refer<strong>en</strong>ce set in the minds of health<br />
care policy makers, against which they can value the ICER of new interv<strong>en</strong>tions. This is<br />
obviously a long-term and gradual process.<br />
4.3 HEALTH CARE DECISION MAKING CONTEXTS<br />
This report also briefly discussed the use of economic evaluations in health care<br />
<strong>de</strong>cision making. The most important insight from this overview is that <strong>de</strong>cision making<br />
processes cannot be reduced to a purely technocratic and rational assessm<strong>en</strong>t. From a<br />
social justice perspective on <strong>de</strong>cision making, there are good argum<strong>en</strong>ts to pursue to<br />
clarify on and make the argum<strong>en</strong>tative logic more transpar<strong>en</strong>t. More ref<strong>le</strong>ction and<br />
rationality in health care <strong>de</strong>cision making is certainly worth pursuing. The princip<strong>le</strong> that<br />
<strong>de</strong>cisions should be substantiated with well docum<strong>en</strong>ted, transpar<strong>en</strong>tly brought sci<strong>en</strong>tific<br />
and other know<strong>le</strong>dge is increasingly accepted.<br />
Health technology assessm<strong>en</strong>t is becoming a very useful methodology to support this<br />
ambition. Economic evaluation is part of any HTA and neg<strong>le</strong>cting economic argum<strong>en</strong>ts<br />
would be unethical. As resources cannot be consumed twice, choices are inevitab<strong>le</strong>.<br />
Consuming health care resources for one interv<strong>en</strong>tion implies d<strong>en</strong>ying these resources<br />
to another interv<strong>en</strong>tion. And precisely these choices need <strong>de</strong>liberation. Besi<strong>de</strong>s<br />
economic and clinical research based argum<strong>en</strong>ts, social justice consi<strong>de</strong>rations remain a<br />
core e<strong>le</strong>m<strong>en</strong>t in the <strong>de</strong>cision making process. The question on the allocation of limited<br />
resources to obtain optimal outcomes is therefore not a technical “neutral” issue but<br />
also an issue of societal values. These values <strong>de</strong>velop within political, social and<br />
economic contexts. The economic effici<strong>en</strong>cy argum<strong>en</strong>t will weigh differ<strong>en</strong>tly in <strong>de</strong>cision<br />
making processes. Economic (technical rational) criteria will be giv<strong>en</strong> another meaning<br />
e.g. according to the health care field (e.g. prev<strong>en</strong>tive, curative, long term care, <strong>en</strong>d-of<br />
life care) or the population addressed with the interv<strong>en</strong>tion (e.g. childr<strong>en</strong>). This is one<br />
of the reasons why <strong>de</strong>cision makers should not so<strong>le</strong>ly rely on seemingly simp<strong>le</strong> tools<br />
such as ICERs and ICER threshold values. In or<strong>de</strong>r to make more optimal use of<br />
economic analyses in health care <strong>de</strong>cision making, researchers and analysts should<br />
become more aware that <strong>de</strong>cisions on the use of health interv<strong>en</strong>tions are likely to be<br />
influ<strong>en</strong>ced by a range of social, financial and institutional factors. Taking better into<br />
account this know<strong>le</strong>dge would bring us closer to the core aims of HTA.<br />
4.4 SUGGESTIONS FOR FURTHER RESEARCH<br />
We need more research on the appropriat<strong>en</strong>ess of the theoretical foundations of the<br />
ICER and ICER threshold value for differ<strong>en</strong>t health care systems. In particular, the<br />
differ<strong>en</strong>ce betwe<strong>en</strong> social security-based systems and NHS-based systems is re<strong>le</strong>vant for<br />
at <strong>le</strong>ast two reasons: on the one hand the budgetary context (fixed or f<strong>le</strong>xib<strong>le</strong>), on the<br />
other hand the characteristics of the <strong>de</strong>cision making processes. The literature<br />
curr<strong>en</strong>tly relies on the assumption of a universally applicab<strong>le</strong> theory of CEA, but<br />
argum<strong>en</strong>ts can be ma<strong>de</strong> in favour of a more context-s<strong>en</strong>sitive analysis:<br />
• First, much of the literature on ICERs and ICER threshold values implicitly<br />
assumes a Beveridge-type health care mo<strong>de</strong>l or -if not- simply ignores the<br />
specificities of the health care system. Health care systems do for instance<br />
not all operate within a fixed budget approach (e.g. we argued why a fixed<br />
ICER threshold value is incompatib<strong>le</strong> with a fixed budget (a NHS-based<br />
system) and why it would be more, yet not comp<strong>le</strong>tely, compatib<strong>le</strong> with a<br />
social security system). We need further theoretical and methodological<br />
elaboration of CEA taking the health care system characteristics into<br />
account.
<strong>KCE</strong> reports 100 ICER Thresholds 59<br />
• Second, political <strong>de</strong>cision making structures and cultures diverge. The ro<strong>le</strong><br />
and place of stakehol<strong>de</strong>r <strong>de</strong>liberation processes and the importance giv<strong>en</strong><br />
to technocratic analysis is differ<strong>en</strong>t betwe<strong>en</strong> countries. Moreover, the<br />
societal context (e.g. values of the welfare state, political prefer<strong>en</strong>ces, …)<br />
pot<strong>en</strong>tially impacts on the criteria consi<strong>de</strong>red and on health care <strong>de</strong>cision<br />
making processes and outcomes. The differ<strong>en</strong>ces betwe<strong>en</strong> the systems<br />
and the place of economic evaluations in <strong>de</strong>cision making processes within<br />
these systems merit further exploration, in or<strong>de</strong>r to op<strong>en</strong> the “black box”<br />
of <strong>de</strong>cision making on health technologies.<br />
A first practical step should aim at a better un<strong>de</strong>rstanding of the <strong>de</strong>cision making criteria<br />
on health technologies. It is an ethical princip<strong>le</strong> that these criteria should be transpar<strong>en</strong>t<br />
and <strong>de</strong>batab<strong>le</strong> within a <strong>de</strong>mocratic welfare state. The criteria used wh<strong>en</strong> taking<br />
<strong>de</strong>cisions about a technology have to be id<strong>en</strong>tified in a sci<strong>en</strong>tific manner. This requires<br />
multidisciplinary sci<strong>en</strong>tific research (economics, medical sci<strong>en</strong>ce, ethics, and social<br />
sci<strong>en</strong>ce). A c<strong>le</strong>arer <strong>de</strong>finition of the criteria policy makers wish to take into account<br />
wh<strong>en</strong> taking <strong>de</strong>cisions on health interv<strong>en</strong>tions would allow the HTA community to<br />
increase the value of its assessm<strong>en</strong>ts for policy makers.<br />
Key points<br />
• There is a cons<strong>en</strong>sus that economic effici<strong>en</strong>cy should be one of the<br />
<strong>de</strong>cision making criteria. There are differ<strong>en</strong>t ways to introduce effici<strong>en</strong>cy<br />
consi<strong>de</strong>rations in the <strong>de</strong>cision making process.<br />
• Using ICERs in combination with an ICER threshold value is one<br />
possibility.<br />
• Methodological variability in economic evaluations reduces the value of<br />
ICERs for assessing interv<strong>en</strong>tions’ cost-effectiv<strong>en</strong>ess. Compliance with the<br />
gui<strong>de</strong>lines for economic evaluations and control of the economic<br />
evaluations is nee<strong>de</strong>d if ICERs are to inform health care policy makers.<br />
• The ICER threshold value is g<strong>en</strong>erally perceived as a fixed value against<br />
which the ICERs of other interv<strong>en</strong>tions can be compared to <strong>de</strong>ci<strong>de</strong><br />
whether the new interv<strong>en</strong>tion is cost-effective. In real life, however, the<br />
ICER threshold value can never be a fixed value over a long period of<br />
time.<br />
• Alternative approaches are the in-betwe<strong>en</strong> comparison of ICERs without<br />
refer<strong>en</strong>ce to an ICER threshold value, the opportunity cost approach and<br />
the cost-consequ<strong>en</strong>ces approach. Combinations of differ<strong>en</strong>t alternatives<br />
are probably the most valuab<strong>le</strong>.<br />
• Although the weight of economic consi<strong>de</strong>rations might differ betwe<strong>en</strong><br />
<strong>de</strong>cisions, ignoring economic effici<strong>en</strong>cy in health care <strong>de</strong>cision making is<br />
unethical.<br />
• Besi<strong>de</strong>s economic and clinical research based argum<strong>en</strong>ts, social justice<br />
consi<strong>de</strong>rations remain a core e<strong>le</strong>m<strong>en</strong>t in the <strong>de</strong>cision making process.<br />
The question on the allocation of limited resources to obtain optimal<br />
outcomes is therefore not a technical “neutral” issue but also an issue of<br />
societal values.<br />
• More research is nee<strong>de</strong>d on the applicability of theoretical foundations<br />
for CEA in differ<strong>en</strong>t health care systems and on the place of CEA in<br />
differ<strong>en</strong>t political <strong>de</strong>cision making structures and cultures.<br />
• Research is also nee<strong>de</strong>d on the <strong>de</strong>cision criteria <strong>de</strong>emed re<strong>le</strong>vant in the<br />
Belgian context.
60 ICER Thresholds <strong>KCE</strong> Reports 100<br />
5 CONCLUSION<br />
The aim of this report is to provi<strong>de</strong> a refer<strong>en</strong>ce docum<strong>en</strong>t for non-health economists<br />
on economic evaluation in health care, its basic concepts and its pot<strong>en</strong>tial value for<br />
health care policy making. The report explains why ICER threshold values, <strong>de</strong>fined in<br />
their neo-classical welfarist s<strong>en</strong>se and un<strong>de</strong>r a fixed budget constraint, have a theoretical<br />
basis that is, however, unt<strong>en</strong>ab<strong>le</strong> in daily practice because basic assumptions are not<br />
fulfil<strong>le</strong>d. This raises the question about whether we still need ICERs, since, according to<br />
theory, they should be compared with an ICER threshold value.<br />
ICERs can be valuab<strong>le</strong> in two ways:<br />
• <strong>de</strong>fine the ICER threshold value as the maximum societal WTP for a unit<br />
of health effect. This option requires a f<strong>le</strong>xib<strong>le</strong> budget and the<br />
measurem<strong>en</strong>t of the maximum societal WTP for a g<strong>en</strong>eric QALY.<br />
• <strong>de</strong>termine the acceptability of an ICER on a case-by-case basis by<br />
evaluating the societal WTP for a unit of health effect for each<br />
interv<strong>en</strong>tion separately. This option does not require the id<strong>en</strong>tification of<br />
an ICER threshold value but <strong>de</strong>rives interv<strong>en</strong>tions’ relative costeffectiv<strong>en</strong>ess<br />
by means of in-betwe<strong>en</strong> comparisons of ICERs. Other<br />
consi<strong>de</strong>rations are weighed against the effici<strong>en</strong>cy criteria once the relative<br />
position of the interv<strong>en</strong>tion’s ICER compared to other interv<strong>en</strong>tions’<br />
ICERs is <strong>de</strong>termined.<br />
Other options, not using the ICER, to inform health policy makers about the effici<strong>en</strong>cy<br />
of interv<strong>en</strong>tions are:<br />
• the opportunity cost approach<br />
• the cost-consequ<strong>en</strong>ces approach.<br />
C<strong>le</strong>arly, each of these approaches has its merits and weaknesses. The budgetary context<br />
is an important <strong>de</strong>terminant for the applicability of the alternatives but also<br />
methodological issues may impe<strong>de</strong> the application of an approach. Because it is unethical<br />
to ignore economic effici<strong>en</strong>cy in the <strong>de</strong>cision making process, a combination of<br />
approaches will probably offer the best result in terms of informing health care policy<br />
makers.<br />
No sing<strong>le</strong> country inclu<strong>de</strong>d in our review used a sing<strong>le</strong> ICER threshold value. Either an<br />
‘acceptab<strong>le</strong>’ range is <strong>de</strong>fined as in the UK, or no explicit ICER threshold values are used<br />
at all. In most countries, it appears that interv<strong>en</strong>tions with a low ICER are more likely to<br />
become accepted than interv<strong>en</strong>tions with a high ICER. In the pres<strong>en</strong>ce of high ICERs,<br />
other assessm<strong>en</strong>t e<strong>le</strong>m<strong>en</strong>ts may become more important.<br />
In Belgium <strong>de</strong>cision making remains mainly an interactive <strong>de</strong>liberation process, although<br />
efforts are ma<strong>de</strong> to ‘rationalise’ the <strong>de</strong>cision making and substantiate reimbursem<strong>en</strong>t<br />
requests with sci<strong>en</strong>tific evid<strong>en</strong>ce. In contrast to clinical effectiv<strong>en</strong>ess, cost-effectiv<strong>en</strong>ess<br />
is sometimes consi<strong>de</strong>red in the <strong>de</strong>cision making process by the DRC but rarely by the<br />
TCI.<br />
A key message we <strong>de</strong>rive from this work is the importance of transpar<strong>en</strong>cy about the<br />
criteria and social values that are weighed in a health policy making process. Therefore<br />
it is important that the information pres<strong>en</strong>ted to health care policy makers makes s<strong>en</strong>se<br />
to them, e.g. by pres<strong>en</strong>ting the information in disaggregated form in addition to<br />
‘composite’ ICERs.
<strong>KCE</strong> reports 100 ICER Thresholds 61<br />
6 RECOMMENDATIONS<br />
• Cost-effectiv<strong>en</strong>ess should be a criterion in the <strong>de</strong>cision making process, as<br />
ignoring economic effici<strong>en</strong>cy is unethical. Dossiers submitted to support<br />
policy makers should therefore always inclu<strong>de</strong> an economic evaluation.<br />
• Economic mo<strong>de</strong>ls should be reported in a transpar<strong>en</strong>t way, pres<strong>en</strong>ting all<br />
information used in the mo<strong>de</strong>l in a way that allows the policy makers to<br />
verify the assumptions, view the uncertainties and weigh the importance<br />
of the assumptions and uncertainties for the <strong>de</strong>cision. Transpar<strong>en</strong>cy and<br />
control of economic mo<strong>de</strong>ls is crucial to increase their credibility.<br />
• The results of economic evaluations should be pres<strong>en</strong>ted in disaggregated<br />
form. This inclu<strong>de</strong>s “unpacking” the ICER but also pres<strong>en</strong>ting other<br />
economically re<strong>le</strong>vant outcome parameters that can be <strong>de</strong>rived from the<br />
economic evaluation but that are not necessarily visib<strong>le</strong> in the ICER<br />
estimate.<br />
• Alongsi<strong>de</strong> the disaggregated pres<strong>en</strong>tation of economically important<br />
e<strong>le</strong>m<strong>en</strong>ts, also the ICER should continue to be pres<strong>en</strong>ted, calculated<br />
following standard methodological gui<strong>de</strong>lines.<br />
• Sci<strong>en</strong>tific research should continue to be used in the <strong>de</strong>cision making<br />
processes on the allocation of health care resources. It will allow policy<br />
makers to back up argum<strong>en</strong>ts in favour of or against a particular <strong>de</strong>cision<br />
by sci<strong>en</strong>tific evid<strong>en</strong>ce.<br />
• Decision makers should be more transpar<strong>en</strong>t in their <strong>de</strong>cision making<br />
criteria and the relative importance of the differ<strong>en</strong>t criteria in each<br />
<strong>de</strong>cision.
62 ICER Thresholds <strong>KCE</strong> Reports 100<br />
7 REFERENCES<br />
1. Weinstein MC, Stason WB. Foundations of cost-effectiv<strong>en</strong>ess analysis for health and medical<br />
practices. N Engl J Med. 1977;296(13):716-21.<br />
2. Drummond M, Sculpher M, Torrance G, O'Bri<strong>en</strong> B, Stoddart G. Methods for the Economic<br />
Evaluation of Health Care Programmes (Third Edition). Oxford University Press, editor.<br />
Oxford; 2005.<br />
3. Morris S, Devlin N, Parkin D. Economic Analysis in Health Care. 1 ed. West Sussex: John<br />
Wi<strong>le</strong>y & Sons, Ltd.; 2007.<br />
4. Gold M, Siegel J, Russell L, Weinstein M. Cost-effectiv<strong>en</strong>ess in health and medicine. Oxford:<br />
Oxford University Press; 1996.<br />
5. McCabe C, Claxton K, Culyer AJ. The NICE Cost-Effectiv<strong>en</strong>ess Threshold: What it is and<br />
What that Means. Pharmacoeconomics. 2008;26(9):733-44.<br />
6. Rutigliano MJ. Cost effectiv<strong>en</strong>ess analysis: a review. Neurosurgery. 1995;37(3):436-43;<br />
discussion 43-4.<br />
7. Siegel JE, Torrance GW, Russell LB, Luce BR, Weinstein MC, Gold MR. Gui<strong>de</strong>lines for<br />
pharmacoeconomic studies. Recomm<strong>en</strong>dations from the panel on cost effectiv<strong>en</strong>ess in health<br />
and medicine. Panel on cost Effectiv<strong>en</strong>ess in Health and Medicine. Pharmacoeconomics.<br />
1997;11(2):159-68.<br />
8. Brouwer WB, Koopmanschap MA. On the economic foundations of CEA. Ladies and<br />
g<strong>en</strong>t<strong>le</strong>m<strong>en</strong>, take your positions! J Health Econ. 2000;19(4):439-59.<br />
9. Wagstaff A. QALYs and the equity-effici<strong>en</strong>cy tra<strong>de</strong>-off. J Health Econ. 1991;10(1):21-41.<br />
10. Williams A. Interg<strong>en</strong>erational equity: an exploration of the 'fair innings' argum<strong>en</strong>t. Health Econ.<br />
1997;6(2):117-32.<br />
11. B<strong>le</strong>ichrodt H. Health utility indices and equity consi<strong>de</strong>rations. J Health Econ. 1997;16(1):65-91.<br />
12. Dowie J. Why cost-effectiv<strong>en</strong>ess should trump (clinical) effectiv<strong>en</strong>ess: the ethical economics of<br />
the South West quadrant. Health Econ. 2004;13(5):453-9.<br />
13. Donaldson C, Currie G, Mitton C. Cost effectiv<strong>en</strong>ess analysis in health care: contraindications.<br />
BMJ. 2002;325(7369):891-4.<br />
14. Australian Departm<strong>en</strong>t of Health and Aging. Gui<strong>de</strong>lines for preparing submissions to the<br />
Pharmaceutical B<strong>en</strong>efits Advisory Committee (version 4.2). In: Pharmaceutical B<strong>en</strong>efits<br />
Advisory Committee; 2007.<br />
15. Barton P, Bryan S, Robinson S. Mo<strong>de</strong>lling in the economic evaluation of health care: se<strong>le</strong>cting<br />
the appropriate approach. J Health Serv Res Policy. 2004;9(2):110-8.<br />
16. Col<strong>le</strong>ge voor Zorgverzekering<strong>en</strong> (CVZ). Gui<strong>de</strong>lines for pharmacoeconomic research, updated<br />
version. In. Diem<strong>en</strong>: CVZ; 2006.<br />
17. Bryan S, Williams I, McIver S. Seeing the NICE si<strong>de</strong> of cost-effectiv<strong>en</strong>ess analysis: a qualitative<br />
investigation of the use of CEA in NICE technology appraisals. Health Econ. 2007;16(2):179-93.<br />
18. C<strong>le</strong>emput I, Van Wil<strong>de</strong>r P, Vrij<strong>en</strong>s F, Huybrechts M, Ramaekers D. Gui<strong>de</strong>lines for<br />
Pharmacoeconomic Evaluations in Belgium. Health Technology assessm<strong>en</strong>t (HTA). Bruxel<strong>le</strong>s:<br />
Belgian Health Care Know<strong>le</strong>dge C<strong>en</strong>tre (<strong>KCE</strong>); 2008. <strong>KCE</strong> Reports 78C (D/2008/10.273/27)<br />
Availab<strong>le</strong> from: http://kce.fgov.be/in<strong>de</strong>x_<strong>en</strong>.aspx?SGREF=5213&CREF=11009<br />
19. Griebsch I, Coast J, Brown J. Quality-adjusted life-years lack quality in pediatric care: a critical<br />
review of published cost-utility studies in child health. Pediatrics. 2005;115(5):e600-14.<br />
20. Scuffham PA, Whitty JA, Mitchell A, Viney R. The use of QALY weights for QALY calculations:<br />
a review of industry submissions requesting listing on the Australian Pharmaceutical B<strong>en</strong>efits<br />
Scheme 2002-4. Pharmacoeconomics.26(4):297-310.<br />
21. Read JL, Quinn RJ, Berwick DM, Fineberg HV, Weinstein MC. Prefer<strong>en</strong>ces for health<br />
outcomes. Comparison of assessm<strong>en</strong>t methods. Med Decis Making.4(3):315-29.<br />
22. Hornberger JC, Re<strong>de</strong>lmeier DA, Peters<strong>en</strong> J. Variability among methods to assess pati<strong>en</strong>ts' wellbeing<br />
and consequ<strong>en</strong>t effect on a cost-effectiv<strong>en</strong>ess analysis. J Clin Epi<strong>de</strong>miol. 1992;45(5):505-<br />
12.<br />
23. Marra CA, Marion SA, Guh DP, Najafza<strong>de</strong>h M, Wolfe F, Esdai<strong>le</strong> JM, et al. Not all "qualityadjusted<br />
life years" are equal. J Clin Epi<strong>de</strong>miol. 2007;60(6):616-24.
<strong>KCE</strong> reports 100 ICER Thresholds 63<br />
24. Briggs A. Economics notes: handling uncertainty in economic evaluation. BMJ.<br />
25.<br />
1999;319(7202):120.<br />
Briggs A, Sculpher M, Buxton M. Uncertainty in the economic evaluation of health care<br />
technologies: the ro<strong>le</strong> of s<strong>en</strong>sitivity analysis. Health Econ. 1994;3(2):95-104.<br />
26. Groot Koerkamp B, Hunink MGM, Stijn<strong>en</strong> T, Hammitt JK, Kuntz KM, Weinstein MC.<br />
Limitations of acceptability curves for pres<strong>en</strong>ting uncertainty in cost-effectiv<strong>en</strong>ess analysis.[see<br />
comm<strong>en</strong>t]. Medical Decision Making. 2007;27(2):101-11.<br />
27. Torgerson DJ, Raftery J. Economic notes. Discounting. BMJ. 1999;319(7214):914-5.<br />
28. Krahn M, Gafni A. Discounting in the economic evaluation of health care interv<strong>en</strong>tions. Med<br />
Care. 1993;31(5):403-18.<br />
29. Ganiats TG. Prev<strong>en</strong>tion, policy, and paradox: what is the value of future health? Am J Prev Med.<br />
1997;13(1):12-7.<br />
30. Milne R. Valuing prev<strong>en</strong>tion: discounting health b<strong>en</strong>efits and costs in New Zealand. N Z Med J.<br />
2005;118(1214):U1443.<br />
31. Bonneux L, Birnie E. The discount rate in the economic evaluation of prev<strong>en</strong>tion: a thought<br />
experim<strong>en</strong>t. J Epi<strong>de</strong>miol Community Health. 2001;55(2):123-5.<br />
32. Brouwer W, van Hout B, Rutt<strong>en</strong> F. A fair approach to discounting future effects: taking a<br />
societal perspective. J Health Serv Res Policy. 2000;5(2):114-8.<br />
33. Brouwer WB, Niess<strong>en</strong> LW, Postma MJ, Rutt<strong>en</strong> FF. Need for differ<strong>en</strong>tial discounting of costs<br />
and health effects in cost effectiv<strong>en</strong>ess analyses. BMJ. 2005;331(7514):446-8.<br />
34. Claxton K, Sculpher M, Culyer A, McCabe C, Briggs A, Akehurst R, et al. Discounting and<br />
cost-effectiv<strong>en</strong>ess in NICE - stepping back to sort out a confusion. Health Econ. 2006;15(1):1-<br />
4.<br />
35. Gravel<strong>le</strong> H, Smith D. Discounting for health effects in cost-b<strong>en</strong>efit and cost-effectiv<strong>en</strong>ess<br />
analysis. Health Econ. 2001;10(7):587-99.<br />
36. Gravel<strong>le</strong> H, Brouwer W, Niess<strong>en</strong> L, Postma M, Rutt<strong>en</strong> F. Discounting in economic evaluations:<br />
stepping forward towards optimal <strong>de</strong>cision ru<strong>le</strong>s. Health Econ. 2007;16(3):307-17.<br />
37. Philips Z, Claxton K, Palmer S. The half-life of truth: what are appropriate time horizons for<br />
research <strong>de</strong>cisions? Med Decis Making. 2008;28(3):287-99.<br />
38. Walker D, Fox-Rushby J. Allowing for uncertainty in economic evaluations: qualitative<br />
s<strong>en</strong>sitivity analysis. Health Policy Plan. 2001;16(4):435-43.<br />
39. Eckermann S, Willan AR. Expected value of information and <strong>de</strong>cision making in HTA. Health<br />
Econ. 2007;16(2):195-209.<br />
40. F<strong>en</strong>wick E, Claxton K, Sculpher M. The value of imp<strong>le</strong>m<strong>en</strong>tation and the value of information:<br />
combined and unev<strong>en</strong> <strong>de</strong>velopm<strong>en</strong>t. Med Decis Making. 2008;28(1):21-32.<br />
41. Claxton K, Ginnelly L, Sculpher M, Philips Z, Palmer S. A pilot study on the use of <strong>de</strong>cision<br />
theory and value of information analysis as part of the NHS Health Technology Assessm<strong>en</strong>t<br />
programme. Health Technol Assess. 2004;8(31):1-103, iii.<br />
42. Briggs A, Gray AM. Handling uncertainty wh<strong>en</strong> performing economic evaluation of healthcare<br />
interv<strong>en</strong>tions. Health Technology Assessm<strong>en</strong>t. 1999;3(2).<br />
43. Culyer AJ, Newhouse JP. Handbook of Health Economics. New York: Elsevier; 2000.<br />
44. Dolan P, Edlin R, Tsuchiya A. The relative societal value of health gains to differ<strong>en</strong>t<br />
b<strong>en</strong>eficiaries. 2008. RM03/JH11 Availab<strong>le</strong> from:<br />
45.<br />
http://www.pcpoh.bham.ac.uk/publichealth/methodology/projects/RM03_JH11_PD.shtml<br />
Weinstein M, Zeckhauser R. Critical ratios and effici<strong>en</strong>t allocation. Journal of Public Economics.<br />
1973;2:147-57.<br />
46. Birch S, Gafni A. Information created to eva<strong>de</strong> reality (ICER): things we should not look to for<br />
answers. Pharmacoeconomics. 2006;24(11):1121-31.<br />
47. S<strong>en</strong>di P, Gafni A, Birch S. Opportunity costs and uncertainty in the economic evaluation of<br />
health care interv<strong>en</strong>tions. Health Econ. 2002;11(1):23-31.<br />
48. National Institute for Clinical Excel<strong>le</strong>nce. Briefing paper for the Methods Working Party on the<br />
Cost-Effectiv<strong>en</strong>ess Threshold. 2007.<br />
49. Birch S, Gafni A. Economics and the evaluation of health care programmes: g<strong>en</strong>eralisability of<br />
methods and implications for g<strong>en</strong>eralisability of results. Health Policy. 2003;64(2):207-19.
64 ICER Thresholds <strong>KCE</strong> Reports 100<br />
50. Birch S, Gafni A. Changing the prob<strong>le</strong>m to fit the solution: Johannesson and Weinstein's (mis)<br />
application of economics to real world prob<strong>le</strong>ms. J Health Econ. 1993;12(4):469-76.<br />
51. van Hout BA, Al MJ, Gordon GS, Rutt<strong>en</strong> F. Costs, effects and C/E-ratios alongsi<strong>de</strong> a clinical<br />
trial. Health Econ. 1994;3(5):309-19.<br />
52. Murray CJ, Evans DB, Acharya A, Baltuss<strong>en</strong> RM. Developm<strong>en</strong>t of WHO gui<strong>de</strong>lines on<br />
g<strong>en</strong>eralized cost-effectiv<strong>en</strong>ess analysis. Health Econ. 2000;9(3):235-51.<br />
53. World Health Organization. Making Choices in Health: WHO gui<strong>de</strong> to cost-effectiv<strong>en</strong>ess<br />
analysis. G<strong>en</strong>eva: WHO; 2003.<br />
54. Garber AM, Phelps CE. Economic foundations of cost-effectiv<strong>en</strong>ess analysis. J Health Econ.<br />
1997;16(1):1-31.<br />
55. Weinstein MC, editor. From cost-effectiv<strong>en</strong>ess ratios to resource allocation: where to draw<br />
the line? Cambridge: Cambridge University Press; 1995.<br />
56. Ubel PA, Hirth RA, Chernew ME, F<strong>en</strong>drick AM. What is the price of life and why doesn't it<br />
increase at the rate of inflation? Arch Intern Med. 2003;163(14):1637-41.<br />
57. Birch S, Donaldson C. Valuing the b<strong>en</strong>efits and costs of health care programmes: where's the<br />
'extra' in extra-welfarism? Soc Sci Med. 2003;56(5):1121-33.<br />
58. Dickson M, Hurst J, Jacobzone S. Survey of Pharmacoeconomic Assessm<strong>en</strong>t Activity in E<strong>le</strong>v<strong>en</strong><br />
Countries. Paris: Directorate for Employm<strong>en</strong>t, Labour and Social Affairs; 2003. OECD Health<br />
Working Papers 4 (4) Availab<strong>le</strong> from: http://www.oecd.org/dataoecd/27/25/2955828.pdf<br />
59. van <strong>de</strong>r Zee J, Kroneman MW. Bismarck or Beveridge: a beauty contest betwe<strong>en</strong> dinosaurs.<br />
BMC Health Serv Res.7:94.<br />
60. Lameire N, Joffe P, Wie<strong>de</strong>mann M. Healthcare systems--an international review: an overview.<br />
Nephrol Dial Transplant.14 Suppl 6:3-9.<br />
61. Culyer A, McCabe C, Briggs A, Claxton K, Buxton M, Akehurst R, et al. Searching for a<br />
threshold, not setting one: the ro<strong>le</strong> of the National Institute for Health and Clinical Excel<strong>le</strong>nce.<br />
J Health Serv Res Policy. 2007;12(1):56-8.<br />
62. Birch S, Gafni A. The biggest bang for the buck or bigger bucks for the bang: the fallacy of the<br />
cost-effectiv<strong>en</strong>ess threshold. J Health Serv Res Policy. 2006;11(1):46-51.<br />
63. Gafni A. Economic Evaluation of Health Care Interv<strong>en</strong>tions: The biggest bang for the buck or<br />
the bigger bucks for the bang? In.<br />
64. Gafni A, Birch S. Increm<strong>en</strong>tal cost-effectiv<strong>en</strong>ess ratios (ICERs): the si<strong>le</strong>nce of the lambda. Soc<br />
Sci Med. 2006;62(9):2091-100.<br />
65. Devlin N, Parkin D. Does NICE have a cost-effectiv<strong>en</strong>ess threshold and what other factors<br />
influ<strong>en</strong>ce its <strong>de</strong>cisions? A binary choice analysis. Health Econ. 2004;13(5):437-52.<br />
66. App<strong>le</strong>by J, Devlin N, Parkin D. NICE's cost effectiv<strong>en</strong>ess threshold. BMJ. 2007;335(7616):358-9.<br />
67. Rawlins MD, Culyer AJ. National Institute for Clinical Excel<strong>le</strong>nce and its value judgm<strong>en</strong>ts. BMJ.<br />
2004;329(7459):224-7.<br />
68. Birch S, Gafni A. Cost effectiv<strong>en</strong>ess/utility analyses. Do curr<strong>en</strong>t <strong>de</strong>cision ru<strong>le</strong>s <strong>le</strong>ad us to where<br />
we want to be? J Health Econ. 1992;11(3):279-96.<br />
69. Karlsson G, Johannesson M. Cost-effectiv<strong>en</strong>ess analysis and capital costs. Soc Sci Med.<br />
1998;46(9):1183-91.<br />
70. Stinnett AA, Paltiel AD. Mathematical programming for the effici<strong>en</strong>t allocation of health care<br />
resources. J Health Econ. 1996;15(5):641-53.<br />
71. Thiry N, Lambert ML, C<strong>le</strong>emput I, Huybrechts M, Neyt M, Hulstaert F, et al. HPV Vaccination<br />
for the Prev<strong>en</strong>tion of Cervical Cancer in Belgium: Health Technology Assessm<strong>en</strong>t. In: <strong>KCE</strong><br />
Reports. Brussels: Belgian Health Care Know<strong>le</strong>dge C<strong>en</strong>tre (<strong>KCE</strong>); 2007.<br />
72. Claxton K. Oft, Vbp: Qed? Health Econ. 2007;16(6):545-58.<br />
73. J<strong>en</strong>a AB, Philipson TJ. Cost-effectiv<strong>en</strong>ess analysis and innovation. J Health Econ.<br />
2008;27(5):1224-36.<br />
74. Ethical aspects of cost-utility analysis. In: 4th Forum of National Ethics Councils in EU<br />
Countries. Amsterdam; 2004.<br />
75. B<strong>le</strong>ichrodt H, Diecidue E, Quiggin J. Equity weights in the allocation of health care: the rank<strong>de</strong>p<strong>en</strong>d<strong>en</strong>t<br />
QALY mo<strong>de</strong>l. J Health Econ. 2004;23(1):157-71.
<strong>KCE</strong> reports 100 ICER Thresholds 65<br />
76. Bryan S, Roberts T, Heginbotham C, McCallum A. QALY-maximisation and public prefer<strong>en</strong>ces:<br />
results from a g<strong>en</strong>eral population survey. Health Econ. 2002;11(8):679-93.<br />
77. Coast J. Is economic evaluation in touch with society's health values? BMJ.<br />
78.<br />
2004;329(7476):1233-6.<br />
Dakin HA, Devlin NJ, O<strong>de</strong>yemi IAO. "Yes", "No" or "Yes, but"? Multinomial mo<strong>de</strong>lling of NICE<br />
<strong>de</strong>cision making. Health Policy. 2006;77(3):352-67.<br />
79. Dolan P. The measurem<strong>en</strong>t of individual utility and social welfare. J Health Econ. 1998;17(1):39-<br />
52.<br />
80. Dolan P. Utilitarianism and the measurem<strong>en</strong>t and aggregation of quality--adjusted life years.<br />
Health Care Anal. 2001;9(1):65-76.<br />
81. Dolan P, Shaw R, Tsuchiya A, Williams A. QALY maximisation and peop<strong>le</strong>'s prefer<strong>en</strong>ces: a<br />
methodological review of the literature. Health Econ. 2005;14(2):197-208.<br />
82. Dolan PA, Ols<strong>en</strong> JA. Equity in health: the importance of differ<strong>en</strong>t health streams. J Health Econ.<br />
2001;20(5):823-34.<br />
83. King JT, Jr., Tsevat J, Lave JR, Roberts MS. Willingness to pay for a quality-adjusted life year:<br />
implications for societal health care resource allocation. Med Decis Making. 2005;25(6):667-77.<br />
84. Ols<strong>en</strong> JA. A note on eliciting distributive prefer<strong>en</strong>ces for health. J Health Econ. 2000;19(4):541-<br />
50.<br />
85. Polsky D. Does willingness to pay per quality-adjusted life year bring us closer to a useful<br />
<strong>de</strong>cision ru<strong>le</strong> for cost-effectiv<strong>en</strong>ess analysis? Med Decis Making. 2005;25(6):605-6.<br />
86. Schwappach DL. Resource allocation, social values and the QALY: a review of the <strong>de</strong>bate and<br />
empirical evid<strong>en</strong>ce. Health Expect. 2002;5(3):210-22.<br />
87. Schwappach DL. Does it matter who you are or what you gain? An experim<strong>en</strong>tal study of<br />
prefer<strong>en</strong>ces for resource allocation. Health Econ. 2003;12(4):255-67.<br />
88. Stolk EA, Pickee SJ, Am<strong>en</strong>t AH, Busschbach JJ. Equity in health care prioritisation: an empirical<br />
inquiry into social value. Health Policy. 2005;74(3):343-55.<br />
89. Tsuchiya A. QALYs and ageism: philosophical theories and age weighting. Health Econ.<br />
2000;9(1):57-68.<br />
90. Tsuchiya A, Dolan P. The QALY mo<strong>de</strong>l and individual prefer<strong>en</strong>ces for health states and health<br />
profi<strong>le</strong>s over time: a systematic review of the literature. Med Decis Making. 2005;25(4):460-7.<br />
91. Tsuchiya A, Dolan P. Do NHS clinicians and members of the public share the same views about<br />
reducing inequalities in health? Soc Sci Med. 2007;64(12):2499-503.<br />
92. Tsuchiya A, Dolan P. Equality of what in health? Distinguishing betwe<strong>en</strong> outcome egalitarianism<br />
and gain egalitarianism. Health Econ. 2008.<br />
93. Pritchard C. Overseas approaches to <strong>de</strong>cision making. In: Towse A, Pritchard C, Devlin N,<br />
editors. Cost-Effectiv<strong>en</strong>ess Thresholds. Economic and Ethical Issues. London: King’s Fund and<br />
Office of Health Economics; 2002.<br />
94. Rocchi A, M<strong>en</strong>on D, Verma S, Mil<strong>le</strong>r E. The Ro<strong>le</strong> of Economic Evid<strong>en</strong>ce in Canadian Oncology<br />
Reimbursem<strong>en</strong>t Decision making: To Lambda and Beyond. Value Health. 2007.<br />
95. H<strong>en</strong>ry DA, Hill SR, Harris A. Drug prices and value for money: the Australian Pharmaceutical<br />
B<strong>en</strong>efits Scheme. JAMA. 2005;294(20):2630-2.<br />
96. Braithwaite RS, Meltzer DO, King JT, Jr., Leslie D, Roberts MS. What does the value of<br />
mo<strong>de</strong>rn medicine say about the $50,000 per quality-adjusted life-year <strong>de</strong>cision ru<strong>le</strong>? Med Care.<br />
2008;46(4):349-56.<br />
97. Miners AH, Garau M, Fidan D, Fischer AJ. Comparing estimates of cost effectiv<strong>en</strong>ess submitted<br />
to the National Institute for Clinical Excel<strong>le</strong>nce (NICE) by differ<strong>en</strong>t organisations: retrospective<br />
study. BMJ. 2005;330(7482):65.<br />
98. Gafni A, Birch S, Nice, National Health S. NICE methodological gui<strong>de</strong>lines and <strong>de</strong>cision making<br />
in the National Health Service in England and Wa<strong>le</strong>s. Pharmacoeconomics. 2003;21(3):149-57.<br />
99. Williams A. What could be nicer than NICE? London: Office of Health Economics; 2004.<br />
100. Buxton MJ. How much are health-care systems prepared to pay to produce a QALY? Eur J<br />
Health Econ. 2005;6(4):285-7.<br />
101. Byrne MM, O'Mal<strong>le</strong>y K, Suarez-Almazor ME. Willingness to pay per quality-adjusted life year in<br />
a study of knee osteoarthritis. Med Decis Making. 2005;25(6):655-66.
66 ICER Thresholds <strong>KCE</strong> Reports 100<br />
102. Williams A. Is the QALY a technical solution to a political prob<strong>le</strong>m? Of course not! Int J Health<br />
Serv. 1991;21(2):365-9; discussion 71-2.<br />
103. Baltuss<strong>en</strong> R, t<strong>en</strong> Asbroek AH, Koolman X, Shrestha N, Bhattarai P, Niess<strong>en</strong> LW. Priority<br />
setting using multip<strong>le</strong> criteria: should a lung health programme be imp<strong>le</strong>m<strong>en</strong>ted in Nepal?<br />
Health Policy Plan. 2007;22(3):178-85.<br />
104. Mitton C, Donaldson C. Health care priority setting: princip<strong>le</strong>s, practice and chal<strong>le</strong>nges. Cost<br />
Eff Resour Alloc. 2004;2(1):3.<br />
105. Gibson J, Mitton C, Martin D, Donaldson C, Singer P. Ethics and economics: does programme<br />
budgeting and marginal analysis contribute to fair priority setting? J Health Serv Res Policy.<br />
2006;11(1):32-7.<br />
106. Nord E, Richardson J, Street A, Kuhse H, Singer P. Maximizing health b<strong>en</strong>efits vs egalitarianism:<br />
an Australian survey of health issues. Soc Sci Med. 1995;41(10):1429-37.<br />
107. Dolan P, Cookson R. A qualitative study of the ext<strong>en</strong>t to which health gain matters wh<strong>en</strong><br />
choosing betwe<strong>en</strong> groups of pati<strong>en</strong>ts. Health Policy. 2000;51(1):19-30.<br />
108. K<strong>le</strong>vit HD, Bates AC, Castanares T, Kirk EP, Sipes-Metz<strong>le</strong>r PR, Wopat R. Prioritization of<br />
health care services. A progress report by the Oregon Health Services Commission. Arch<br />
Intern Med. 1991;151(5):912-6.<br />
109. Al<strong>le</strong>n D, Lee RH, Lowson K. The use of QALYS (quality-adjusted life years) in health service<br />
planning. Int J Health Plann Manage. 1989;4(4):261-73.<br />
110. Eddy DM. Oregon's methods. Did cost-effectiv<strong>en</strong>ess analysis fail? JAMA. 1991;266(15):2135-41.<br />
111. Drummond M, Brown R, F<strong>en</strong>drick AM, Ful<strong>le</strong>rton P, Neumann P, Taylor R, et al. Use of<br />
pharmacoeconomics information--report of the ISPOR Task Force on use of<br />
pharmacoeconomic/health economic information in health-care <strong>de</strong>cision making. Value Health.<br />
2003;6(4):407-16.<br />
112. Hoffmann C, Graf von <strong>de</strong>r Schu<strong>le</strong>nburg JM. The influ<strong>en</strong>ce of economic evaluation studies on<br />
<strong>de</strong>cision making. A European survey. The EUROMET group. Health Policy. 2000;52(3):179-92.<br />
113. Williams A. QALYS and ethics: a health economist's perspective. Soc Sci Med.<br />
1996;43(12):1795-804.<br />
114. Grosse SD, Teutsch SM, Haddix AC. Lessons from cost-effectiv<strong>en</strong>ess research for United<br />
States public health policy. Annu Rev Public Health. 2007;28:365-91.<br />
115. Williams I, Bryan S, McIver S. How should cost-effectiv<strong>en</strong>ess analysis be used in health<br />
technology coverage <strong>de</strong>cisions? Evid<strong>en</strong>ce from the National Institute for Health and Clinical<br />
Excel<strong>le</strong>nce approach. J Health Serv Res Policy. 2007;12(2):73-9.<br />
116. Simon HA. Administrative Behaviour: A Study of Decision making Processes in Administrative<br />
Organizations. 3rd ed. New York: Free Press; 1976.<br />
117. March JG, Ols<strong>en</strong> JP. Organizational <strong>le</strong>arning and the ambiguity of the past. . In: March JG, Ols<strong>en</strong><br />
JP, editors. Ambiguity and Choice in Organizations. 2nd ed. Berg<strong>en</strong>, Norway:<br />
118.<br />
Universitatsforlaget; 1987.<br />
Lindblom CE. The Sci<strong>en</strong>ce of Muddling Through. Public Administration Review. 1959;19(2):79-<br />
88.<br />
119. Elliott H, Popay J. How are policy makers using evid<strong>en</strong>ce? Mo<strong>de</strong>ls of research utilisation and<br />
local NHS policy making. J Epi<strong>de</strong>miol Community Health. 2000;54(6):461-8.<br />
120. Davis P, Howd<strong>en</strong>-Chapman P. Translating research findings into health policy. Soc Sci Med.<br />
1996;43(5):865-72.<br />
121. Kothari A, Birch S, Char<strong>le</strong>s C. "Interaction" and research utilisation in health policies and<br />
programs: does it work? Health Policy. 2005;71(1):117-25.<br />
122. Innvaer S, Vist G, Trommald M, Oxman A. Health policy-makers' perceptions of their use of<br />
evid<strong>en</strong>ce: a systematic review. J Health Serv Res Policy. 2002;7(4):239-44.<br />
123. Lavis JN, Lomas J, Hamid M, Sewankambo NK. Assessing country-<strong>le</strong>vel efforts to link research<br />
to action. Bull World Health Organ. 2006;84(8):620-8.<br />
124. Mi<strong>le</strong>wa T, Barry C. Health Policy and the Politics of Evid<strong>en</strong>ce. Social Policy and Administration.<br />
2005;39(5):498-512.<br />
125. Eich<strong>le</strong>r HG, Kong SX, Gerth WC, Mavros P, Jonsson B. Use of cost-effectiv<strong>en</strong>ess analysis in<br />
health-care resource allocation <strong>de</strong>cision making: how are cost-effectiv<strong>en</strong>ess thresholds<br />
expected to emerge? Value Health. 2004;7(5):518-28.
<strong>KCE</strong> reports 100 ICER Thresholds 67<br />
126. Williams I, Bryan S. Un<strong>de</strong>rstanding the limited impact of economic evaluation in health care<br />
resource allocation: a conceptual framework. Health Policy. 2007;80(1):135-43.<br />
127. Williams IP, Bryan S. Cost-effectiv<strong>en</strong>ess analysis and formulary <strong>de</strong>cision making in England:<br />
findings from research. Soc Sci Med. 2007;65(10):2116-29.<br />
128. Duthie T, Trueman P, Chancellor J, Diez L. Research into the use of health economics in<br />
<strong>de</strong>cision making in the United Kingdom--Phase II. Is health economics 'for good or evil'? Health<br />
Policy. 1999;46(2):143-57.<br />
129. Drummond M, Cooke J, Wal<strong>le</strong>y T. Economic evaluation un<strong>de</strong>r managed competition: evid<strong>en</strong>ce<br />
from the U.K. Soc Sci Med. 1997;45(4):583-95.<br />
130. Williams I, McIver S, Moore D, Bryan S. The use of economic evaluations in NHS <strong>de</strong>cision<br />
making: a review and empirical investigation. Health Technol Assess. 2008;12(7):iii, ix-x, 1-175.<br />
131. von <strong>de</strong>r Schu<strong>le</strong>nburg J, Vauth C, Mitt<strong>en</strong>dorf T, Greiner W. Methods for <strong>de</strong>termining costb<strong>en</strong>efit<br />
ratios for pharmaceuticals in Germany. Eur J Health Econ. 2007;8 Suppl 1:S5-31.<br />
132. Thurston SJ, Craig D, Wilson P, Drummond MF. Increasing <strong>de</strong>cision makers' access to<br />
economic evaluations: alternative methods of communicating the information. Int J Technol<br />
Assess Health Care. 2008;24(2):151-7.<br />
133. Hoffmann C, Stoykova BA, Nixon J, Glanvil<strong>le</strong> JM, Misso K, Drummond MF. Do health-care<br />
<strong>de</strong>cision makers find economic evaluations useful? The findings of focus group research in UK<br />
health authorities. Value Health. 2002;5(2):71-8.<br />
134. Ross J. The use of economic evaluation in health care: Australian <strong>de</strong>cision makers' perceptions.<br />
Health Policy. 1995;31(2):103-10.<br />
135. Eddama O, Coast J. Use of economic evaluation in local health care <strong>de</strong>cision making in England:<br />
A qualitative investigation. Health Policy. 2008.<br />
136. Eddama O, Coast J. A systematic review of the use of economic evaluation in local <strong>de</strong>cision<br />
making. Health Policy. 2008;86(2-3):129-41.<br />
137. Sculpher MJ, Pang FS, Manca A, Drummond MF, Gol<strong>de</strong>r S, Urdahl H, et al. G<strong>en</strong>eralisability in<br />
economic evaluation studies in healthcare: a review and case studies. Health Technol Assess.<br />
2004;8(49):iii-iv, 1-192.<br />
138. Raftery JP. Paying for costly pharmaceuticals: regulation of new drugs in Australia, England and<br />
New Zealand. Med J Aust. 2008;188(1):26-8.<br />
139. Stoykova B, Drummond M, Barbieri M, K<strong>le</strong>ijn<strong>en</strong> J. The lag betwe<strong>en</strong> effectiv<strong>en</strong>ess and costeffectiv<strong>en</strong>ess<br />
evid<strong>en</strong>ce of new drugs. Implications for <strong>de</strong>cision making in health care. Eur J<br />
Health Econ. 2003;4(4):313-8.<br />
140. Weinstein MC. How much are Americans willing to pay for a quality-adjusted life year? Med<br />
Care. 2008;46(4):343-5.<br />
141. Stolk EA, van Donselaar G, Brouwer WB, Busschbach JJ. Reconciliation of economic concerns<br />
and health policy: illustration of an equity adjustm<strong>en</strong>t procedure using proportional shortfall.<br />
Pharmacoeconomics. 2004;22(17):1097-107.<br />
142. Baltuss<strong>en</strong> R, Niess<strong>en</strong> L. Priority setting of health interv<strong>en</strong>tions: the need for multi-criteria<br />
<strong>de</strong>cision analysis. Cost Eff Resour Alloc. 2006;4:14.<br />
143. Daniels N, Sabin J. Limits to health care: fair procedures, <strong>de</strong>mocratic <strong>de</strong>liberation, and the<br />
<strong>le</strong>gitimacy prob<strong>le</strong>m for insurers. Philos Public Aff. 1997;26(4):303-50.<br />
144. Daniels N. Accountability for reasonab<strong>le</strong>ness. Bmj. 2000;321(7272):1300-1.<br />
145. Daniels N, Sabin JE. Accountability for reasonab<strong>le</strong>ness: an update. Bmj. 2008;337:a1850.<br />
146. Daniels N, Teagard<strong>en</strong> JR, Sabin JE. An ethical template for pharmacy b<strong>en</strong>efits. Health Aff<br />
(Millwood). 2003;22(1):125-37.<br />
147. Sabin JE, Daniels N. Making insurance coverage for new technologies reasonab<strong>le</strong> and<br />
accountab<strong>le</strong>. Jama. 1998;279(9):703-4.<br />
148. Teagard<strong>en</strong> JR, Daniels N, Sabin JE. A proposed ethical framework for prescription drug b<strong>en</strong>efit<br />
allocation policy. J Am Pharm Assoc (Wash). 2003;43(1):69-74.<br />
149. Gruskin S, Daniels N. Process is the point: justice and human rights: priority setting and fair<br />
<strong>de</strong>liberative process. Am J Public Health. 2008;98(9):1573-7.
68 ICER Thresholds <strong>KCE</strong> Reports 100<br />
150. Towse A, Pritchard C. Does NICE have a threshold? An external view. In: Towse A, Pritchard<br />
C, Devlin N, editors. Cost-Effectiv<strong>en</strong>ess Thresholds. Economic and Ethical Issues. London:<br />
King’s Fund and Office of Health Economics; 2002.<br />
151. National Institute for Clinical Excel<strong>le</strong>nce. Gui<strong>de</strong> to the Methods of Technology Appraisal. April<br />
2004.<br />
152. NICE. Gui<strong>de</strong> to the Methods of Technology Appraisal. Draft for consultation. 2007 November<br />
2007.<br />
153. National Institute for Health and Clinical Excel<strong>le</strong>nce. Gui<strong>de</strong> to the Methods of Technology<br />
Appraisal. Draft for Consultation. London: NICE; 2007 November 2007. Availab<strong>le</strong> from:<br />
http://www.nice.org.uk/media/8AE/5C/TAMethodsGui<strong>de</strong>UpdateFINALFORCONSULTATION2<br />
81107.pdf<br />
154. Gui<strong>de</strong>lines for the economic evaluation of health technologies: Canada (3rd edition). In.<br />
Ottawa: Canadian Ag<strong>en</strong>cy for Drugs and Technologies in Health; 2006.<br />
155. Laupacis A, Fe<strong>en</strong>y D, Detsky AS, Tugwell PX. How attractive does a new technology have to<br />
be to warrant adoption and utilization? T<strong>en</strong>tative gui<strong>de</strong>lines for using clinical and economic<br />
evaluations. CMAJ. 1992;146(4):473-81.<br />
156. Raad voor <strong>de</strong> Volksgezondheid <strong>en</strong> Zorg. Zinnige <strong>en</strong> duurzame zorg. Zoetermeer: 2006.<br />
Availab<strong>le</strong> from: http://www.rvz.net/data/download/advies_Zinnige_zorg.pdf<br />
157. Stolk EA, Po<strong>le</strong>y M, Brouwer W, van Busschbach J. Proeftoetsing van het iMTA-mo<strong>de</strong>l.<br />
Id<strong>en</strong>tificatie van aando<strong>en</strong>ing<strong>en</strong> met minima<strong>le</strong> ziektelast <strong>en</strong> proeftoetsing van <strong>de</strong> voor ziektelast<br />
gecorrigeer<strong>de</strong> doelmatigheidstoets. In: (CVZ) CvZ, editor. Vervolgon<strong>de</strong>rzoek breedte<br />
g<strong>en</strong>eesmid<strong>de</strong><strong>le</strong>npakket. Amstelve<strong>en</strong>; 2002. Availab<strong>le</strong> from:<br />
158.<br />
http://www.xs4all.nl/~jannetvb/busschbach/manuscripts/2002%20Vervolgon<strong>de</strong>rzoek%20CvZ.pdf<br />
Hirth RA, Chernew ME, Mil<strong>le</strong>r E, F<strong>en</strong>drick AM, Weissert WG. Willingness to pay for a qualityadjusted<br />
life year: in search of a standard. Med Decis Making. 2000;20(3):332-42.<br />
159. Grocott R, Schoe<strong>le</strong>r R, Priest V, Hall C, Metcalfe S, Brougham M, et al. Prescription for<br />
Pharmacoeconomic Analysis. Methdos for cost-utility analysis. 2007.<br />
160. PHARMAC. Prescription for pharmacoeconomic analysis: methods for cost-utility analysis.<br />
Pharmaceutical Managem<strong>en</strong>t Ag<strong>en</strong>cy; 2007.<br />
161. The Pharmaceutical B<strong>en</strong>efits Board. G<strong>en</strong>eral gui<strong>de</strong>lines for economic evaluations from the<br />
Pharmaceutical B<strong>en</strong>efits Board. LFNAR; 2003. (2003:2) Availab<strong>le</strong> from:<br />
162.<br />
http://www.lfn.se/upload/English/ENG_lfnar2003-<strong>en</strong>g.pdf<br />
Carlsson P. Health technology assessm<strong>en</strong>t and priority setting for health policy in Swed<strong>en</strong>. Int J<br />
Technol Assess Health Care. 2004;20(1):44-54.<br />
163. Norwegian Medicines Ag<strong>en</strong>cy Oslo;c 2005. Norwegian gui<strong>de</strong>lines for pharmacoeconomic<br />
analysis in connection with applications for reimbursem<strong>en</strong>t. Availab<strong>le</strong> from:<br />
http://www.<strong>le</strong>gemid<strong>de</strong>lverket.no/templates/InterPage____25644.aspx?filterBy=CopyToIndustry<br />
164. Krist<strong>en</strong>s<strong>en</strong> FB, Sigmund He. Health Technology Assessm<strong>en</strong>t Handbook. 2nd edition ed.<br />
Cop<strong>en</strong>hag<strong>en</strong>: Danish C<strong>en</strong>tre for Health Technology Assessm<strong>en</strong>t, National Board of Health;<br />
2007.<br />
165. Royal Decree of 21 December 2001. Koninklijk Besluit tot vaststelling van <strong>de</strong> procedures,<br />
termijn<strong>en</strong> <strong>en</strong> voorwaard<strong>en</strong> inzake <strong>de</strong> tegemoetkoming van <strong>de</strong> verplichte verzekering voor<br />
g<strong>en</strong>eeskundige verzorging <strong>en</strong> uitkering<strong>en</strong> in <strong>de</strong> kost<strong>en</strong> van farmaceutische specialiteit<strong>en</strong>/Arrêté<br />
Royal fixant <strong>le</strong>s procédures, <strong>de</strong>lais et conditions <strong>en</strong> matière d’interv<strong>en</strong>tion <strong>de</strong> l’assurance<br />
obligatoire <strong>soins</strong> <strong>de</strong> <strong>santé</strong> et in<strong>de</strong>mnités dans <strong>le</strong> <strong>coût</strong> <strong>de</strong>s spécialités pharmaceutiques. Belgisch<br />
Staatsblad/Moniteur Belge 29/12/2001: .<br />
166. Royal Decree of 21 December 2001. Koninklijk besluit tot wijziging van het koninklijk besluit<br />
van 3 juli 1996 tot uitvoering van <strong>de</strong> wet betreff<strong>en</strong><strong>de</strong> <strong>de</strong> verplichte verzekering voor<br />
g<strong>en</strong>eeskundige verzorging <strong>en</strong> uitkering<strong>en</strong>, gecoördineerd op 14 juli 1994/ Arrêté Royal portant<br />
modification <strong>de</strong> l’arrêté Royal du 3 juil<strong>le</strong>t 1996 portant exécution <strong>de</strong> la loi relative à l’assurance<br />
obligatoire <strong>soins</strong> <strong>de</strong> <strong>santé</strong> et in<strong>de</strong>mnités, coordonnée <strong>le</strong> 14 juil<strong>le</strong>t 1994. . Belgisch<br />
Staatsblad/Moniteur Belge 29/12/2001.<br />
167. Council Directive 89/105/EEC of 21 December 1988 relating to the transpar<strong>en</strong>cy of measures<br />
regulating the pricing of medicinal products for human use and their inclusion in the scope of<br />
national health insurance systems. OJ No L 40, 1989. Availab<strong>le</strong> from:<br />
http://ec.europa.eu/<strong>en</strong>terprise/phabiocom/docs/dir_1989_105/dir_1989_105_<strong>en</strong>.pdf
<strong>KCE</strong> reports 100 ICER Thresholds 69<br />
168. Royal Decree of 21 December 2001 (published 29 December 2001): Koninklijk besluit van 21<br />
<strong>de</strong>cember 2001 tot vaststelling van <strong>de</strong> procedures, termijn<strong>en</strong> <strong>en</strong> voorwaard<strong>en</strong> inzake <strong>de</strong><br />
tegemoetkoming van <strong>de</strong> verplichte verzekering voor g<strong>en</strong>eeskundige verzorging <strong>en</strong> uitkering<strong>en</strong><br />
in <strong>de</strong> kost<strong>en</strong> van farmaceutische specialiteit<strong>en</strong>, B.S. 29 <strong>de</strong>cember 2001 // Arrêté royal <strong>de</strong> 21<br />
décembre 2001 fixant <strong>le</strong>s procédures, délais et conditions concernant l'interv<strong>en</strong>tion <strong>de</strong><br />
l'assurance obligatoire <strong>soins</strong> <strong>de</strong> <strong>santé</strong> et in<strong>de</strong>mnités dans <strong>le</strong> <strong>coût</strong> <strong>de</strong>s spécialités<br />
pharmaceutiques, M.B. <strong>le</strong> 29 décembre 2001., 2001.<br />
169. Law coordinated on 14 July 1994. Wet betreff<strong>en</strong><strong>de</strong> <strong>de</strong> verplichte verzekering voor<br />
g<strong>en</strong>eeskundige verzorging <strong>en</strong> uitkering<strong>en</strong>, gecoördineerd op 14 juli 1994. / Loi relative à<br />
l’assurance obligatoire <strong>soins</strong> <strong>de</strong> <strong>santé</strong> et in<strong>de</strong>mnités, coordonnée <strong>le</strong> 14 juil<strong>le</strong>t 1994. Belgisch<br />
Staatsblad/Moniteur Belge 27/08/1994.<br />
170. Royal Decree of 18 January 1999. Koninklijk besluit tot goedkeuring van het huishou<strong>de</strong>lijk<br />
reg<strong>le</strong>m<strong>en</strong>t van <strong>de</strong> Technische Raad voor implantat<strong>en</strong> ingesteld bij <strong>de</strong> Di<strong>en</strong>st voor<br />
g<strong>en</strong>eeskundige verzorging van het Rijksinstituut voor ziekte- <strong>en</strong> invaliditeitsverzekering. /<br />
Arrêté Royal portant approbation du règ<strong>le</strong>m<strong>en</strong>t d'ordre intérieur du Conseil technique <strong>de</strong>s<br />
implants institué auprès du Service <strong>de</strong>s <strong>soins</strong> <strong>de</strong> <strong>santé</strong> <strong>de</strong> l'Institut national d'assurance maladieinvalidité.<br />
Belgisch Staatsblad/Moniteur Belge 17/02/1999.<br />
171. Royal Decree of 20 February 2008. Koninklijk besluit tot wijziging, wat <strong>de</strong> Commissie<br />
Tegemoetkoming Implantat<strong>en</strong> <strong>en</strong> invasieve Medische Hulpmid<strong>de</strong><strong>le</strong>n betreft, van het koninklijk<br />
besluit van 3 juli 1996 tot uitvoering van <strong>de</strong> wet betreff<strong>en</strong><strong>de</strong> <strong>de</strong> verplichte verzerkering voor<br />
g<strong>en</strong>eeskundige verzorging <strong>en</strong> uitkering<strong>en</strong>, gecoördineerd op 14 juli 1994. / Arrêté Royal<br />
modifiant, <strong>en</strong> ce qui concerne la Commission <strong>de</strong> Remboursem<strong>en</strong>t <strong>de</strong>s Implants et Dispositifs<br />
médicaux invasifs, l’arrêté Royal du 3 juil<strong>le</strong>t 1996 portant exécution <strong>de</strong> la loi relative à<br />
l’assurance obligatoire <strong>soins</strong> <strong>de</strong> <strong>santé</strong> et in<strong>de</strong>mnités, coordonnée <strong>le</strong> 14 juil<strong>le</strong>t 1994. Belgisch<br />
Staatsblad/Moniteur Belge 20/02/2008.
This page is <strong>le</strong>ft int<strong>en</strong>tionally blank.
Dépôt légal : D/2009/10.273/95
<strong>KCE</strong> reports<br />
1. Efficacité et r<strong>en</strong>tabilité <strong>de</strong>s thérapies <strong>de</strong> sevrage tabagique. D/2004/10.273/2.<br />
2. Etu<strong>de</strong> relative aux <strong>coût</strong>s pot<strong>en</strong>tiels liés à une év<strong>en</strong>tuel<strong>le</strong> modification <strong>de</strong>s règ<strong>le</strong>s du droit <strong>de</strong> la<br />
responsabilité médica<strong>le</strong> (Phase 1). D/2004/10.273/4.<br />
3. Utilisation <strong>de</strong>s antibiotiques <strong>en</strong> milieu hospitalier dans <strong>le</strong> cas <strong>de</strong> la pyélonéphrite aiguë.<br />
D/2004/10.273/6.<br />
4. Leucoréduction. Une mesure <strong>en</strong>visageab<strong>le</strong> dans <strong>le</strong> cadre <strong>de</strong> la politique nationa<strong>le</strong> <strong>de</strong> sécurité<br />
<strong>de</strong>s transfusions sanguines. D/2004/10.273/8.<br />
5. Evaluation <strong>de</strong>s risques préopératoires. D/2004/10.273/10.<br />
6. Validation du <strong>rapport</strong> <strong>de</strong> la Commission d’exam<strong>en</strong> du sous financem<strong>en</strong>t <strong>de</strong>s hôpitaux.<br />
D/2004/10.273/12.<br />
7. Recommandation nationa<strong>le</strong> relative aux <strong>soins</strong> prénatals: Une base <strong>pour</strong> un itinéraire clinique <strong>de</strong><br />
suivi <strong>de</strong> grossesses. D/2004/10.273/14.<br />
8. Systèmes <strong>de</strong> financem<strong>en</strong>t <strong>de</strong>s médicam<strong>en</strong>ts hospitaliers: étu<strong>de</strong> <strong>de</strong>scriptive <strong>de</strong> certains pays<br />
europé<strong>en</strong>s et du Canada. D/2004/10.273/16.<br />
9. Feedback: évaluation <strong>de</strong> l'impact et <strong>de</strong>s barrières à l'implém<strong>en</strong>tation – Rapport <strong>de</strong> recherche:<br />
partie 1. D/2005/10.273/02.<br />
10. Le <strong>coût</strong> <strong>de</strong>s prothèses d<strong>en</strong>taires. D/2005/10.273/04.<br />
11. Dépistage du cancer du sein. D/2005/10.273/06.<br />
12. Etu<strong>de</strong> d’une métho<strong>de</strong> <strong>de</strong> financem<strong>en</strong>t alternative <strong>pour</strong> <strong>le</strong> sang et <strong>le</strong>s dérivés sanguins labi<strong>le</strong>s<br />
dans <strong>le</strong>s hôpitaux. D/2005/10.273/08.<br />
13. Traitem<strong>en</strong>t <strong>en</strong>dovasculaire <strong>de</strong> la sténose carotidi<strong>en</strong>ne. D/2005/10.273/10.<br />
14. Variations <strong>de</strong>s pratiques médica<strong>le</strong>s hospitalières <strong>en</strong> cas d’infarctus aigu du myocar<strong>de</strong> <strong>en</strong><br />
Belgique. D/2005/10.273/12<br />
15. Evolution <strong>de</strong>s dép<strong>en</strong>ses <strong>de</strong> <strong>santé</strong>. D/2005/10.273/14.<br />
16. Etu<strong>de</strong> relative aux <strong>coût</strong>s pot<strong>en</strong>tiels liés à une év<strong>en</strong>tuel<strong>le</strong> modification <strong>de</strong>s règ<strong>le</strong>s du droit <strong>de</strong> la<br />
responsabilité médica<strong>le</strong>. Phase II : développem<strong>en</strong>t d'un modè<strong>le</strong> actuariel et premières<br />
estimations. D/2005/10.273/16.<br />
17. Evaluation <strong>de</strong>s montants <strong>de</strong> référ<strong>en</strong>ce. D/2005/10.273/18.<br />
18. Utilisation <strong>de</strong>s itinéraires cliniques et gui<strong>de</strong>s <strong>de</strong> bonne pratique afin <strong>de</strong> déterminer <strong>de</strong> manière<br />
prospective <strong>le</strong>s honoraires <strong>de</strong>s mé<strong>de</strong>cins hospitaliers: plus faci<strong>le</strong> à dire qu'à faire..<br />
D/2005/10.273/20<br />
19. Evaluation <strong>de</strong> l'impact d'une contribution personnel<strong>le</strong> forfaitaire sur <strong>le</strong> recours au service<br />
d'urg<strong>en</strong>ces. D/2005/10.273/22.<br />
20. HTA Diagnostic Moléculaire <strong>en</strong> Belgique. D/2005/10.273/24, D/2005/10.273/26.<br />
21. HTA Matériel <strong>de</strong> Stomie <strong>en</strong> Belgique. D/2005/10.273.28.<br />
22. HTA Tomographie par Emission <strong>de</strong> Positrons <strong>en</strong> Belgique. D/2005/10.273/30.<br />
23. HTA Le traitem<strong>en</strong>t é<strong>le</strong>ctif <strong>en</strong>dovasculaire <strong>de</strong> l’anévrysme <strong>de</strong> l’aorte abdomina<strong>le</strong> (AAA).<br />
D/2005/10.273.33.<br />
24. L'emploi <strong>de</strong>s pepti<strong>de</strong>s natriurétiques dans l'approche diagnostique <strong>de</strong>s pati<strong>en</strong>ts prés<strong>en</strong>tant une<br />
suspicion <strong>de</strong> décomp<strong>en</strong>sation cardiaque. D/2005/10.273.35<br />
25. Endoscopie par capsu<strong>le</strong>. D2006/10.273.02.<br />
26. Aspects médico-légaux <strong>de</strong>s recommandations <strong>de</strong> bonne pratique médica<strong>le</strong>. D2006/10.273/06.<br />
27. Qualité et organisation <strong>de</strong>s <strong>soins</strong> du diabète <strong>de</strong> type 2. D2006/10.273/08.<br />
28. Recommandations provisoires <strong>pour</strong> <strong>le</strong>s évaluations pharmacoéconomiques <strong>en</strong> Belgique.<br />
D2006/10.273/11.<br />
29. Recommandations nationa<strong>le</strong>s Collège d’oncologie : A. cadre général <strong>pour</strong> un manuel<br />
d’oncologie B. base sci<strong>en</strong>tifique <strong>pour</strong> itinéraires cliniques <strong>de</strong> diagnostic et traitem<strong>en</strong>t, cancer<br />
colorectal et cancer du testicu<strong>le</strong>. D2006/10.273/13.<br />
30. Inv<strong>en</strong>taire <strong>de</strong>s bases <strong>de</strong> données <strong>de</strong> <strong>soins</strong> <strong>de</strong> <strong>santé</strong>. D2006/10.273/15.<br />
31. Health Technology Assessm<strong>en</strong>t : l’antigène prostatique spécifique (PSA) dans <strong>le</strong> dépistage du<br />
cancer <strong>de</strong> la prostate. D2006/10.273/18.<br />
32. Feedback: évaluation <strong>de</strong> l'impact et <strong>de</strong>s barrières à l'implém<strong>en</strong>tation - Rapport <strong>de</strong> recherche:<br />
partie II. D2006/10.273/20.<br />
33. Effets et <strong>coût</strong>s <strong>de</strong> la vaccination <strong>de</strong>s <strong>en</strong>fants Belges au moy<strong>en</strong> du vaccin conjugué<br />
antipneumococcique. D2006/10.273/22.<br />
34. Trastuzumab <strong>pour</strong> <strong>le</strong>s sta<strong>de</strong>s précoces du cancer du sein. D2006/10.273/24.
35. Etu<strong>de</strong> relative aux <strong>coût</strong>s pot<strong>en</strong>tiels liés à une év<strong>en</strong>tuel<strong>le</strong> modification <strong>de</strong>s règ<strong>le</strong>s du droit <strong>de</strong> la<br />
responsabilité médica<strong>le</strong> – Phase III : affinem<strong>en</strong>t <strong>de</strong>s estimations. D2006/10.273/27.<br />
36. Traitem<strong>en</strong>t pharmacologique et chirurgical <strong>de</strong> l'obésité. Prise <strong>en</strong> charge résid<strong>en</strong>tiel<strong>le</strong> <strong>de</strong>s<br />
<strong>en</strong>fants sévèrem<strong>en</strong>t obèses <strong>en</strong> Belgique. D/2006/10.273/29.<br />
37. Health Technology Assessm<strong>en</strong>t Imagerie par Résonance Magnétique. D/2006/10.273/33.<br />
38. Dépistage du cancer du col <strong>de</strong> l’utérus et recherche du Papillomavirus humain (HPV).<br />
D/2006/10.273/36<br />
39. Evaluation rapi<strong>de</strong> <strong>de</strong> technologies émerg<strong>en</strong>tes s'appliquant à la colonne vertébra<strong>le</strong> :<br />
remplacem<strong>en</strong>t <strong>de</strong> disque intervertébral et vertébro/cyphoplastie par ballonnet.<br />
D/2006/10.273/39.<br />
40. Etat fonctionnel du pati<strong>en</strong>t: un instrum<strong>en</strong>t pot<strong>en</strong>tiel <strong>pour</strong> <strong>le</strong> remboursem<strong>en</strong>t <strong>de</strong> la<br />
kinésithérapie <strong>en</strong> Belgique? D/2006/10.273/41.<br />
41. Indicateurs <strong>de</strong> qualité cliniques. D/2006/10.273/44.<br />
42. Etu<strong>de</strong> <strong>de</strong>s disparités <strong>de</strong> la chirurgie é<strong>le</strong>ctive <strong>en</strong> Belgique. D/2006/10.273/46.<br />
43. Mise à jour <strong>de</strong> recommandations <strong>de</strong> bonne pratique existantes. D/2006/10.273/49.<br />
44. Procédure d'évaluation <strong>de</strong>s dispositifs médicaux émergeants. D/2006/10.273/51.<br />
45. HTA Dépistage du Cancer Colorectal : état <strong>de</strong>s lieux sci<strong>en</strong>tifique et impact budgétaire <strong>pour</strong> la<br />
Belgique. D/2006/10.273/54.<br />
46. Health Technology Assessm<strong>en</strong>t. Polysomnographie et monitoring à domici<strong>le</strong> <strong>de</strong>s nourrissons<br />
<strong>en</strong> prév<strong>en</strong>tion <strong>de</strong> la mort subite. D/2006/10.273/60.<br />
47. L'utilisation <strong>de</strong>s médicam<strong>en</strong>ts dans <strong>le</strong>s maisons <strong>de</strong> repos et <strong>le</strong>s maisons <strong>de</strong> repos et <strong>de</strong> <strong>soins</strong><br />
Belges. D/2006/10.273/62<br />
48. Lombalgie chronique. D/2006/10.273/64.<br />
49. Médicam<strong>en</strong>ts antiviraux <strong>en</strong> cas <strong>de</strong> grippe saisonnière et pandémique. Revue <strong>de</strong> littérature et<br />
recommandations <strong>de</strong> bonne pratique. D/2006/10.273/66.<br />
50. Contributions personnel<strong>le</strong>s <strong>en</strong> matière <strong>de</strong> <strong>soins</strong> <strong>de</strong> <strong>santé</strong> <strong>en</strong> Belgique. L'impact <strong>de</strong>s supplém<strong>en</strong>ts.<br />
D/2006/10.273/69.<br />
51. Besoin <strong>de</strong> <strong>soins</strong> chroniques <strong>de</strong>s personnes âgées <strong>de</strong> 18 à 65 ans et atteintes <strong>de</strong> lésions<br />
cérébra<strong>le</strong>s acquises. D/2007/10.273/02.<br />
52. Rapid Assessm<strong>en</strong>t: Prév<strong>en</strong>tion cardiovasculaire primaire dans la pratique du mé<strong>de</strong>cin<br />
généraliste <strong>en</strong> Belgique. D/2007/10.273/04.<br />
53. Financem<strong>en</strong>t <strong>de</strong>s <strong>soins</strong> Infirmiers Hospitaliers. D/2007/10 273/06<br />
54. Vaccination <strong>de</strong>s nourrissons contre <strong>le</strong> rotavirus <strong>en</strong> Belgique. Analyse <strong>coût</strong>-<strong>efficacité</strong><br />
55. Va<strong>le</strong>ur <strong>en</strong> termes <strong>de</strong> données probantes <strong>de</strong>s informations écrites <strong>de</strong> l’industrie pharmaceutique<br />
<strong>de</strong>stinées aux mé<strong>de</strong>cins généralistes. D/2007/10.273/13<br />
56. Matériel orthopédique <strong>en</strong> Belgique: Health Technology Assessm<strong>en</strong>t. D/2007/10.273/15.<br />
57. Organisation et Financem<strong>en</strong>t <strong>de</strong> la Réadaptation Locomotrice et Neurologique <strong>en</strong> Belgique<br />
D/2007/10.273/19<br />
58. Le Défibrillateur Cardiaque Implantab<strong>le</strong>.: un <strong>rapport</strong> d’évaluation <strong>de</strong> technologie <strong>de</strong> <strong>santé</strong><br />
D/2007/10.273/22<br />
59. Analyse <strong>de</strong> biologie clinique <strong>en</strong> mé<strong>de</strong>cine général. D/2007/10.273/25<br />
60. Tests <strong>de</strong> la fonction pulmonaire chez l'adulte. D/2007/10.273/28<br />
61. Traitem<strong>en</strong>t <strong>de</strong> plaies par pression négative: une évaluation rapi<strong>de</strong>. D/2007/10.273/31<br />
62. Radiothérapie Conformationel<strong>le</strong> avec Modulation d’int<strong>en</strong>sité (IMRT). D/2007/10.273/33.<br />
63. Support sci<strong>en</strong>tifique du Collège d’Oncologie: un gui<strong>de</strong>line <strong>pour</strong> la prise <strong>en</strong> charge du cancer du<br />
sein. D/2007/10.273/36.<br />
64. Vaccination HPV <strong>pour</strong> la prév<strong>en</strong>tion du cancer du col <strong>de</strong> l’utérus <strong>en</strong> Belgique: Health<br />
Technology Assessm<strong>en</strong>t. D/2007/10.273/42.<br />
65. Organisation et financem<strong>en</strong>t du diagnostic génétique <strong>en</strong> Belgique. D/2007/10.273/45.<br />
66. Drug Eluting St<strong>en</strong>ts <strong>en</strong> Belgique: Health Technology Assessm<strong>en</strong>t. D/2007/10.273/48.<br />
67. Hadronthérapie. D/2007/10.273/51.<br />
68. In<strong>de</strong>mnisation <strong>de</strong>s dommages résultant <strong>de</strong> <strong>soins</strong> <strong>de</strong> <strong>santé</strong> - Phase IV : Clé <strong>de</strong> répartition <strong>en</strong>tre<br />
<strong>le</strong> Fonds et <strong>le</strong>s assureurs. D/2007/10.273/53.<br />
69. Assurance <strong>de</strong> Qualité <strong>pour</strong> <strong>le</strong> cancer du rectum – Phase 1: Recommandation <strong>de</strong> bonne<br />
pratique <strong>pour</strong> la prise <strong>en</strong> charge du cancer rectal D/2007/10.273/55<br />
70. Etu<strong>de</strong> comparative <strong>de</strong>s programmes d’accréditation hospitalière <strong>en</strong> Europe. D/2008/10.273/02<br />
71. Recommandation <strong>de</strong> bonne pratique clinique <strong>pour</strong> cinq tests ophtalmiques. D/2008/10.273/05<br />
72. L’offre <strong>de</strong> mé<strong>de</strong>cins <strong>en</strong> Belgique. Situation actuel<strong>le</strong> et défis. D/2008/10.273/08
73. Financem<strong>en</strong>t du programme <strong>de</strong> <strong>soins</strong> <strong>pour</strong> <strong>le</strong> pati<strong>en</strong>t gériatrique dans l’hôpital<br />
classique : Définition et évaluation du pati<strong>en</strong>t gériatrique, fonction <strong>de</strong> liaison et évaluation d’un<br />
instrum<strong>en</strong>t <strong>pour</strong> un financem<strong>en</strong>t approprié. D/2008/10.273/12<br />
74. Oxygénothérapie Hyperbare: Rapid Assessm<strong>en</strong>t. D/2008/10.273/14.<br />
75. Gui<strong>de</strong>line <strong>pour</strong> la prise <strong>en</strong> charge du cancer oesophagi<strong>en</strong> et gastrique: élém<strong>en</strong>ts sci<strong>en</strong>tifiques à<br />
<strong>de</strong>stination du Collège d’Oncologie. D/2008/10.273/17.<br />
76. Promotion <strong>de</strong> la qualité <strong>de</strong> la mé<strong>de</strong>cine généra<strong>le</strong> <strong>en</strong> Belgique: status quo ou quo vadis ?<br />
D/2008/10.273/19.<br />
77. Orthodontie chez <strong>le</strong>s <strong>en</strong>fants et ado<strong>le</strong>sc<strong>en</strong>ts D/2008/10.273/21<br />
78. Recommandations <strong>pour</strong> <strong>le</strong>s évaluations pharmacoéconomiques <strong>en</strong> Belgique. D/2008/10.273/24.<br />
79. Remboursem<strong>en</strong>t <strong>de</strong>s radioisotopes <strong>en</strong> Belgique. D/2008/10.273/27.<br />
80. Évaluation <strong>de</strong>s effets du maximum à facturer sur la consommation et l’accessibilité financière<br />
<strong>de</strong>s <strong>soins</strong> <strong>de</strong> <strong>santé</strong>. D/2008/10.273/36.<br />
81. Assurance <strong>de</strong> qualité <strong>pour</strong> <strong>le</strong> cancer rectal – phase 2: développem<strong>en</strong>t et test d’un <strong>en</strong>semb<strong>le</strong><br />
d’indicateurs <strong>de</strong> qualité. D/2008/10.273/39<br />
82. Angiographie coronaire par tomod<strong>en</strong>sitométrie 64-détecteurs chez <strong>le</strong>s pati<strong>en</strong>ts suspects <strong>de</strong><br />
maladie coronari<strong>en</strong>ne. D/2008/10.273/41<br />
83. Comparaison internationa<strong>le</strong> <strong>de</strong>s règ<strong>le</strong>s <strong>de</strong> remboursem<strong>en</strong>t et aspects légaux <strong>de</strong> la chirurgie<br />
plastique D/2008/10.273/44<br />
84. Les séjours psychiatriques <strong>de</strong> longue durée <strong>en</strong> lits T. D/2008/10.273/47<br />
85. Comparaison <strong>de</strong> <strong>de</strong>ux systèmes <strong>de</strong> financem<strong>en</strong>t <strong>de</strong>s <strong>soins</strong> <strong>de</strong> première ligne <strong>en</strong> Belgique.<br />
D/2008/10.273/50.<br />
86. Différ<strong>en</strong>ciation <strong>de</strong> fonctions dans <strong>le</strong>s <strong>soins</strong> infirmiers :possibilités et limites D/2008/10.273/53<br />
87. Consommation <strong>de</strong> kinésithérapie et <strong>de</strong> mé<strong>de</strong>cine physique et <strong>de</strong> réadaptation <strong>en</strong> Belgique.<br />
D/2008/10.273/55<br />
88. Syndrome <strong>de</strong> Fatigue Chronique : diagnostic, traitem<strong>en</strong>t et organisation <strong>de</strong>s <strong>soins</strong>.<br />
D/2008/10.273/59.<br />
89. Evaluation <strong>de</strong>s certains nouveaux traitem<strong>en</strong>ts du cancer <strong>de</strong> la prostate et <strong>de</strong> l’hypertrophie<br />
bénigne <strong>de</strong> la prostate. D/2008/10.273/62<br />
90. Mé<strong>de</strong>cine généra<strong>le</strong>: comm<strong>en</strong>t promouvoir l’attraction et la rét<strong>en</strong>tion dans la profession ?<br />
D/2008/10.273/64.<br />
91. Appareils auditifs <strong>en</strong> Belgique: health technology assessm<strong>en</strong>t. D/2008/10.273/68<br />
92. Les infections nosocomia<strong>le</strong>s <strong>en</strong> Belgique : Vo<strong>le</strong>t I, Etu<strong>de</strong> Nationa<strong>le</strong> <strong>de</strong> Préva<strong>le</strong>nce.<br />
D/2008/10.273/71.<br />
93. Détection <strong>de</strong>s événem<strong>en</strong>ts indésirab<strong>le</strong>s dans <strong>le</strong>s bases <strong>de</strong> données administratives.<br />
D/2008/10.273/74.<br />
94. Soins maternels int<strong>en</strong>sifs (Maternal Int<strong>en</strong>sive Care) <strong>en</strong> Belgique. D/2008/10.273/78.<br />
95. Implantation percutanée <strong>de</strong>s valvu<strong>le</strong>s cardiaques dans <strong>le</strong> cas <strong>de</strong> maladies valvulaires congénita<strong>le</strong>s<br />
et dégénératives: A rapid Health Technology Assessm<strong>en</strong>t. D/2007/10.273/80<br />
96. Construction d’un in<strong>de</strong>x médical <strong>pour</strong> <strong>le</strong>s contrats privés d’assurance maladie.<br />
D/2008/10.273/83<br />
97. C<strong>en</strong>tres <strong>de</strong> réadaptation ORL/PSY : groupes cib<strong>le</strong>s, preuves sci<strong>en</strong>tifiques et organisation <strong>de</strong>s<br />
<strong>soins</strong>. D/2009/10.273/85<br />
98. Évaluation <strong>de</strong> programmes <strong>de</strong> vaccination généraux et ciblés contre l’hépatite A <strong>en</strong> Belgique.<br />
D/2008/10.273/89<br />
99. Financem<strong>en</strong>t <strong>de</strong> l’hôpital <strong>de</strong> jour gériatrique. D/2008/10.273/91<br />
100. <strong>Va<strong>le</strong>urs</strong> <strong>seuils</strong> <strong>pour</strong> <strong>le</strong> <strong>rapport</strong> <strong>coût</strong>-<strong>efficacité</strong> <strong>en</strong> <strong>soins</strong> <strong>de</strong> <strong>santé</strong>. D/2008/10.273/95