Suicide risk among patients with cancer in the United States, 2000-2016.

Abstract

12119 Background: Individuals diagnosed with cancer have elevated suicide risk in the US, although little is known about risk associated with state of residence, health insurance coverage, or time since diagnosis by cancer types. This study used a recent national dataset to examine a wide range of patients’ sociodemographic and clinical factors that may be associated with suicide risks. Methods: We identified patients diagnosed with cancer from 43 population-based state cancer registries in 2000-2016 with follow-up through Dec 31, 2016. Standardized Mortality Ratios (SMR) and 95% confidence intervals (95CI) were calculated by state of residence, attained age group, sex, and race/ethnicity to compare suicide risks in the cohort vs. the general US population. Hazard Ratios (HR) and 95CI from multivariable Cox proportional hazard models were derived to identify cancer-specific risk factors of suicide among the cohort, controlling for competing risks from other causes of death. Results: Among 16,771,397 patients, 7,972,782 (47.5%) died during the study period, and 20,792 (0.3%) from suicide. The overall SMR for suicide was 1.26 (95CI = 1.24-1.28), decreasing from 1.67 (95CI 1.47-1.88) in 2000 to 1.16 (95CI 1.11-1.21) in 2016. Patients from Alaska, Colorado and Idaho, those aged 65-69 years (SMR = 1.44, 95CI = 1.39-1.50), Hispanic patients (SMR = 1.48, 95CI = 1.38-1.58), those uninsured (SMR = 1.66, 95CI = 1.53-1.80) or insured with Medicaid (SMR = 1.72, 95CI = 1.61-1.84) or ≤64 years of age with Medicare (SMR = 1.94, 95CI = 1.80-2.07) had the highest suicide risks compared to the general population. Moreover, the highest suicide risk occurred within two years of diagnosis (SMR [95CI] = 7.19 [6.97-7.41], 5.60 [5.35-5.84] and 4.18 [4.03-4.33] for ≤5 months, 6-11 months, and 12-23 months after cancer diagnosis, respectively). In the first two years following diagnosis, the risk of suicide was higher in patients diagnosed with distant-stage than early-stage diseases (HR = 1.29, 95CI = 1.21-1.37), and in patients with more cancer types with poor prognoses and high symptom burdens, such as cancers of oral cavity & pharynx, esophagus, stomach, brain, lung and pancreas (HRs ranged 1.23-2.10 vs. colorectal cancer, all P≤0.001). After two years, patients diagnosed with cancers subject to long-term quality of life impairment, such as cancers of oral cavity & pharynx, female breast, bladder, and leukemia (HRs ranged 1.17-1.54 vs. colorectal cancer, all P≤0.01), had higher suicide risks. Conclusions: Suicide risk among patients diagnosed with cancer decreased during the past two decades but remained elevated compared to the general population. Different geographic, racial/ethnic, socioeconomic, and clinical factors, some of which are modifiable, contribute to increased suicide risk among patients diagnosed with cancer. Tailored social and psych-oncological interventions are warranted for suicide prevention in this vulnerable population.