Does recruitment method make a difference? Effects on protocol retention and treatment outcome in elderly depressed patients.

The specific aim of this study was to contrast effects of recruitment method (solicited, referred) on demographic, psychosocial, medical, and treatment outcome measures in an ongoing clinical trial of maintenance therapies in late-life depression. Data from 125 elderly patients (56 solicited via media campaign, 69 clinically referred) with recurrent, unipolar major depression were available for analysis. Several statistical contrast procedures, including group t tests, chi 2 tests, survival analysis, and logistic regression, were used to assess differences in patient profiles related to method of recruitment. Referred patients included a higher proportion of African Americans and had a lower level of education, fewer economic resources, and higher chronic medical burden. Solicited patients had been in the index episode longer than the referred patients at the time of protocol entry and were 3.4 times more likely to have experienced a "provoking agent" (severe life event or chronic difficulty) during the 6 months that preceded the onset of depressive symptoms. In contrast to these demographic and illness history differences, there were no differences in treatment response rates or time to response related to recruitment method. Solicited patients had an overall treatment response rate of 71% versus 62% in the referred group. Median time to response was 14.3 weeks in the solicited group and 13.6 weeks in the referred group. These results suggest that the inclusion of solicited patients in geriatric depression clinical trials does not bias short-term treatment outcome.