OBJECTIVE: To survey physicians to determine whether sociodemographic and practice-related variables are predictive of their attitudes toward suicide. DESIGN: Mailed questionnaire survey. SETTING: Monroe County, New York. PARTICIPANTS: A 50 percent random sample of practicing primary care physicians. Of 186 potential respondents, 114 (61.3%) physicians returned questionnaires. MEASUREMENT: In addition to age and gender, independent variables included medical specialty, practice composition defined in terms of patient age, and textbook knowledge of depression and suicide. The dependent variable, attitude toward suicide, was measured using four standard opinion survey questions and six items based on a case vignette. Dependent variables were rated on a Likert-type scale and subsequently dichotomized for use in logistic regression analysis. RESULTS: Approximately 61 percent of the respondents believe that suicide can be rational under certain circumstances, 51 percent believe that physicians should not assist suicides under any circumstances, and 31 percent support legalization of assisted suicide under certain circumstances. Regression analyses revealed that gender was a significant predictor for three of the ten attitude items. Medical specialty, clinical and epidemiological knowledge of depression and suicide, and physician age significantly predicted one attitude item. CONCLUSIONS: These results suggest that physicians' attitudes toward suicide are multidetermined. Although gender, age, medical specialty, and knowledge of depression and suicide predict certain attitudes, the logistic regression analyses were not highly successful in predicting respondents' attitudes. Given the power of physicians to command public attention and to influence public and legal opinion, it is vital that future research continue to explore the determinants of physicians' attitudes toward self-determined death.
