Donepezil, rivastigmine, galantamine and memantine for ...
Donepezil, rivastigmine, galantamine and memantine for ...
Donepezil, rivastigmine, galantamine and memantine for ...
Create successful ePaper yourself
Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.
120<br />
Economic analysis<br />
The model structure has been in<strong>for</strong>med by the<br />
systematic review of the literature on the costeffectiveness<br />
of pharmaceuticals <strong>for</strong> AD <strong>and</strong> on the<br />
available methods <strong>for</strong> modelling disease<br />
progression in AD (see discussion above) <strong>and</strong><br />
discussion with physicians involved in the<br />
treatment of AD. The model structure used is<br />
based on the AHEAD model presented by Caro<br />
<strong>and</strong> colleagues 131 (discussed above), which has<br />
been used in the published literature to estimate<br />
the cost-effectiveness of <strong>galantamine</strong>. Figure 19<br />
describes the simple model structure. The<br />
approach applies the predictive risk equation from<br />
the AHEAD model to determine a monthly hazard<br />
<strong>for</strong> progression of disease to a point where<br />
patients require FTC. This hazard rate has been<br />
used in a disease progression model, together with<br />
mortality data <strong>for</strong> AD patients, to model over time<br />
the experiences of a cohort of AD patients. The<br />
cohort analysis predicts the proportion of patients<br />
that will be located in the three possible health<br />
states (pre-FTC, FTC <strong>and</strong> death) at each monthly<br />
cycle (i.e. a time horizon of 60 months). Costs <strong>and</strong><br />
benefits have been allocated to each of these<br />
health states, <strong>and</strong> the analysis compares disease<br />
progression <strong>and</strong> subsequent costs <strong>and</strong> benefits<br />
over time to estimate the cost-effectiveness of the<br />
intervention.<br />
The AHEAD model presented by Caro <strong>and</strong><br />
colleagues 131 has been described above. The<br />
model is based on the progression of disease to a<br />
Drug therapy<br />
(improved ADAS-cog)<br />
COHORT OF<br />
AD PATIENTS<br />
Costs outcomes<br />
cost-effectiveness?<br />
Usual care<br />
(baseline ADAS-cog)<br />
Pre-FTC Pre-FTC<br />
Death FTC FTC Death<br />
FIGURE 19 Diagrammatic representation of SHTAC model/approach<br />
point where patients require FTC; where they have<br />
a requirement <strong>for</strong> a significant amount (<strong>for</strong> the<br />
greater part of the day) of paid care <strong>and</strong><br />
supervision each day, regardless of the location of<br />
care (institution or community setting) or who<br />
provides the care. Given the concerns in the<br />
literature over the use of cognition alone to model<br />
disease progression over time, <strong>and</strong> based on<br />
discussions with physicians, it was felt that of the<br />
methods available, the use (in our modelling<br />
approach) of the AHEAD methodology <strong>and</strong> the<br />
endpoint FTC was preferable to the use of transit<br />
probabilities from trial data <strong>for</strong> health states<br />
described solely by cognitive scores (i.e. MMSE<br />
scores). The statistical techniques used by Stern<br />
<strong>and</strong> colleagues 132 <strong>and</strong> thereafter Caro <strong>and</strong><br />
colleagues 131 to determine the predictive risk<br />
equations used in the AHEAD model are subject<br />
to a number of concerns, but the method could be<br />
used to illustrate the potential progression of AD<br />
over time to in<strong>for</strong>m on the current consideration<br />
of the cost-effectiveness of interventions.<br />
The predictive risk equation <strong>for</strong> FTC (used in the<br />
SHTAC model) has a two-stage process to<br />
calculate a monthly hazard (risk) of patients<br />
entering FTC from the starting health state of pre-<br />
FTC. A Cox proportional hazards model is used to<br />
calculate a risk index, <strong>and</strong> this has coefficients <strong>for</strong><br />
the presence of EPS, the presence of psychotic<br />
symptoms, a young age at disease onset (i.e.<br />