Prognosis and risk of suicide after cancer diagnosis.

Abstract

6634 Background: Suicide rates are elevated acutely after cancer diagnosis. We sought to create a unifying theory that explains variations in suicide risk across cancer sites, stages, and demographics. Based on the stress-diathesis model, we hypothesized that suicide risk correlates with cancer prognosis and that the impact of prognosis on suicide risk is greatest for populations with higher baseline risks of suicide. Methods: We identified all patients with newly diagnosed cancers from 2000-2019 in the Surveillance, Epidemiology, and End Results (SEER) 17 database, representing 27% of U.S. cancers diagnosed over the past 20 years. Multiple primary-standardized mortality ratios (SMR) were used to determine the relative risk of suicide within the first 6 months compared to the general U.S. population, adjusted for age, sex, race, and year of follow-up. We correlated suicide and 2-year overall survival rates for 20 cancer sites using a weighted linear regression model. Results: We identified 9,046 suicides among 6,811,940 persons diagnosed with cancer, 1,610 of which occurred within 6 months of diagnosis. There were 5.3 suicides per 10,000 person-years, compared with an expected rate of 1.7 (SMR 3.1 [95% CI 3.0-3.3]). Suicide risk correlated strongly with prognosis (increased risk of 9.5% per percent survival deficit at 2 years, R2 = 0.88, P < .0001). The impact of prognosis on suicide risk was greater for demographic groups with higher baseline risks of suicide. For men, the risk of suicide increased by 3 suicides per 10,0000 person-years (R2 = 0.84, P < .0001) versus 0.3 in women (R2 = 0.54, P < .0001). The impact of prognosis on suicide risk was also greater for persons age 60+ and for white (versus black) race. Conclusions: Poorer prognosis correlates with suicide risk in patients with cancer and has a greater effect on populations with higher baseline rates of suicide. Our model should provide a useful framework for triaging patients that need increased surveillance or urgent psychiatric referral.