Evaluating characteristics of patient selection and dropout rates

Patient selection and dropout rates can affect the results of a clinical tr ial. Long lists of exclusions in the selection of patients for clinical tri als reduce the possibility of examining treatment responses for heterogenei ty and make recruitment difficult. In many cases, a pool of 100 potential s ubjects may yield only 2 or 3 qualified participants, a Fact that raises th e issue of generalizability of results. Dropouts should be carefully define d in advance and can be used as dependent variables for the comparison of d ifferent treatments. This article discusses some of the sampling characteri stics (gender, age, diagnosis, inpatient/outpatient status, prior neurolept ic use, and symptom severity) and dropout rates in 5 recent comparative cli nical trials of atypical antipsychotics.