To the Editor: In their study of suicide risk in affective disorders (1), J. Michael Bostwick, M.D., and V. Shane Pankratz, Ph.D., did not include data from a population-based health maintenance organization (HMO) sample reported by a colleague and me (2). They did cite our article to support a methodologic point but did not include the primary findings in their meta-analysis. This omission is surprising given the large size of our sample (35,000 patients), as well as its clinical diversity (psychiatric inpatients, psychiatric outpatients, and primary care patients).

Inclusion of data from our sample would have further strengthened the primary findings reported by Drs. Bostwick and Pankratz. In fact, their conclusions closely echoed those of our 1998 report: "Overall suicide risk among patients treated for depression is considerably lower than previous estimates based on specialty and inpatient samples. Risk is strongly related to treatment history—a likely indicator of illness severity."