Predictors of placebo response: a retrospective analysis.

Abstract

This retrospective evaluation involved 197 patients from independent clinical investigations of four antidepressant medications. Six variables were analyzed for their discriminative utility in predicting placebo response rates: gender; marital status; age; education; duration of illness; and severity of depressive symptomatology, as measured by Hamilton Rating Scale for Depression (HAM-D or HDRS) scores. Men were slightly more responsive to placebo than were women (29.8%, n = 94 vs. 24.3%, n = 103). Married patients demonstrated the highest probability of a positive placebo response (38.15%, n = 76), as compared with widowed/separated/divorced (21.9%, n = 73) or single (16.7%, n = 48) patients. A trend toward an increased probability of placebo response was detected for patients aged 60 and above (35.7%, n = 14). Educational level achieved and duration of illness were not predictive. Severity of illness proved most noteworthy, with the placebo response rate higher in the more mild patients (40.8%, n = 27 vs. 23.4%, n = 77).