Evidence-based Medicine Toolkit

72 Evidence-based Medicine ToolkitAre your patients similar to those of thestudy?Of course, your patients weren’t in the trial, so you need to applyyour clinical expertise to decide whether they are sufficiently similarfor the results to be applicable to them. Factors which wouldaffect this decision include:• The age range included in the trial (many trials exclude the oldergenerations); your group of patients may have a different riskprofile. For example, although many drugs have increasing adverseeffects in the ageing population which may not be takeninto account in the study, they may also have greater benefits.• Many of your patients will have co-morbidity which could affectdrug interactions and adverse events as well as benefits.• Will your patients be able to comply with treatment dosagesand duration? For example, compliance might decrease if yourpatient is taking other medications or if the treatment requiresmultiple doses daily rather than single ones.• If NNTs are similar for different treatments, then the NNHs forharmful side effects will become more important; lesser side effectsmay increase compliance (Bloom, 2001).The inclusion and exclusion criteria for the study may help as astarting point for your clinical judgment here. It is unlikely, however,that your patient will present an exact match with the study;Sackett et al (2000) have recommended framing this question inreverse:How different would your patient have to be for the results ofthe study to be of no help?How much of the study effect can you expectfor your patient(s)?To work out how much effect your patient can expect from the intervention,you first need an estimate of their risk of the outcome.This information might be available from a number of externalsources, such as cardiovascular risk tables in the British NationalFormulary, systematic reviews, Department of Health data or evenlocal audit data.The control group in the study may also provide a good startingpoint. However, you should use your clinical judgment to arrive atan individual’s risk, taking account of his or her individual clinicalcharacteristics.