The major purpose of this study was to examine the effect of depressed mood in older, medically ill, hospitalized patients on their preferences regarding life-sustaining treatments, physician-assisted suicide (PAS), and euthanasia and to determine the degree to which financial constraints affected their choices. Cross-sectional study. General medical hospital. One hundred fifty-eight medically hospitalized, nondemented patients age 60 or older, mean age 74.1 (range 60-94). The sample was divided, based on Center for Epidemiologic Studies-Depression (CES-D) scores, into a depressed group (n = 71) and a nondepressed control group. Subjects underwent a structured interview evaluating their life-sustaining treatment choices and whether they would accept or refuse PAS or euthanasia under a variety of hypothetical conditions. These choices were reevaluated with the introduction of financial impact. In addition, assessment included measures of depression, suicide, cognition, social support, functioning, and religiosity. Depression was found to be highly associated with acceptance of PAS and euthanasia in most hypothetical clinical scenarios in addition to patients' current condition. Compared with nondepressed people, depressed respondents were 13 times as likely to accept PAS when considering their current condition (95% confidence interval [CI] 1.68-110.98), and over twice as likely to accept PAS when facing a hypothetical terminal illness or coma. Depression alone was weakly associated with life-sustaining treatment choices but, when financial impact was introduced, significantly more depressed subjects refused treatment options they had previously desired than did nondepressed subjects. The presence of suicidal ideation, even passive ideation, was strongly predictive of life-sustaining treatment refusals and increased interest in PAS and euthanasia. Depression's effect on acceptance of PAS was confirmed by logistic regression, which also showed that religious coping was significantly correlated with less interest in PAS in two hypothetical scenarios. CONCLUSION. Depressed subjects and even subjects with subtle, passive suicidal ideation were markedly more interested in PAS and euthanasia than nondepressed subjects in hypothetical situations. Depressed subjects were also particularly vulnerable to rejecting treatments if financial consequences might have resulted.