To the Editor: We read with great interest the article by Lorrin M. Koran, M.D., and his colleagues (1) that examined the effects of 28 weeks of double-blind placebo-controlled medication maintenance after 52 weeks of single-blind sertraline treatment.

A randomized controlled trial, more than any other method, can have a powerful impact on patient care and is accepted by medicine as an objective scientific method and, if ideally performed, produces knowledge untainted by bias. However, it can be flawed in design and is not immune to bias. The relevance of such studies has been criticized on the grounds of selection bias and use of the placebo arm.

Although the selection of patients is known to be a powerful factor affecting the results of clinical trials, little is known about recruitment issues. Many patients who are screened for a clinical trial are ultimately not included in the study. In this study, the authors gave an account of all the patients who dropped out but failed to provide information about how many subjects were initially assessed, how many were excluded, and the reasons for exclusion. We do not have any idea how many subjects responded to the advertisements and what was the participation rate, which has implications for generalizability and future research. In this context, the Consolidated Standard of Reporting Trials (CONSORT) guidelines state that all patients assessed for a trial should be accounted for and that the report should be accompanied by a diagram that explains what happened to all of the patients involved in the trial (2). The authors failed to follow the CONSORT guidelines in this regard.

Placebo-control design raises questions of deception, the withholding of patient information, informed consent, the unblinding of such information, and the withholding of active treatment by randomly allocating trial medication. In this context, the Declaration of Helsinki demands that individual patients in a study be assured of the best proven diagnostic and therapeutic methods, even in the control group (3). This statement discards the use of a placebo group as a control group when a proven treatment exists. In this study, one group of patients received no treatment (placebo) for more than 28 weeks when there were a number of control group options that could have fulfilled ethical and scientific needs.