Presentation Authors: Zachary Klaassen*, Augusta, GA, Christopher J. D. Wallis, Hanan Goldberg, Toronto, Canada, Thenappan Chandrasekar, Philadelphia, PA, Rashid K. Sayyid, Augusta, GA, Stephen B. Williams, Galveston, TX, Kelvin A. Moses, Nashville, TN, Martha K. Terris, Augusta, GA, Robert K. Nam, Paul Kurdyak, Girish S. Kulkarni, Toronto, Canada
Introduction: There is emerging evidence that oncology patients with pre-existing mental illness may have poorer survival compared to patients without psychiatric disease. Furthermore, cancer diagnosis may be associated with an increased risk of suicide. However, studies published thus far have failed to account for utilization of psychiatric resources, which may confound this relationship. The objective of this study was to (i) assess the impact of psychiatric utilization (PU) prior to cancer diagnosis on cancer-specific mortality (CSM), and (ii) to assess the effect of cancer diagnosis on suicide risk compared to the general population, accounting for pre-diagnosis PU.
Methods: All residents of Ontario, Canada diagnosed with either prostate, bladder or kidney cancer (1997-2014) were included. Each patient was assigned a psychiatric utilization gradient (PUG) score in the five years prior to cancer diagnosis: 0 (none), 1 (outpatient), 2 (emergency department), 3 (hospital admission). First, a multivariable cause-specific hazard model was used to assess the effect of PUG score on CSM. Second, non-cancer controls were matched 4:1 to cancer patients based on sociodemographic variables and a marginal cause-specific hazard model was used to assess the effect of cancer on the risk of suicidal death.
Results: 191,068 patients were included (137,699 prostate, 29,884 bladder, 23,485 kidney cancer): 109,154 (57.1%) with PUG score 0, 79,553 (41.6%) PUG score 1, 1,596 (0.84%) PUG score 2, and 765 (0.40%) PUG score 3. Increasing pre-diagnosis PU was associated with increased CSM: HR 1.78 (95%CI 1.47-2.14) among patients with PUG score 3 (vs 0) and HR 1.14 (95%CI 0.99-1.32) among those with PUG score 2. These patients with GU malignancies were then matched to 528,387 controls without any cancer diagnosis. Patients with GU cancer had a higher risk of dying of suicide compared to controls (HR 1.16, 95%CI 1.00-1.36). Specifically, among individuals with PUG score 0, those with cancer were significantly more likely to die of suicide compared to patients without cancer (HR 1.39, 95%CI 1.12-1.74).
Conclusions: Pre-cancer diagnosis PU is associated with worse CSM following diagnosis among patients with GU malignancies, with a graded effect. Additionally, the cancer diagnosis confers an increased risk of suicide, compared to the general population, even after accounting for pre-cancer diagnosis PU.
Source of Funding: CUOG-CUA-Astellas