General and mental health status following colorectal cancer treatment and its association with mortality among a racially diverse population-based cohort.

Abstract

45 Background: Patient-reported outcomes (PROs) are recognized as strong predictors of cancer prognosis, outcomes, and care. However, racial/ethnic minorities with colorectal cancer (CRC) tend to report poorer general health status (GHS) and mental health status (MHS) compared to non-Hispanic whites. The objectives of this study were to determine: (1) if there are racial/ethnic differences in GHS and MHS within 36 months of CRC diagnosis and (2) if poorer GHS and MHS in recently diagnosed CRC patients are associated with mortality. Methods: We used the population-based Surveillance, Epidemiology, and End Results (SEER)-Consumer Assessment of Healthcare Providers and Systems (CAHPS) dataset to analyze Medicare beneficiaries aged ≥65 years who were diagnosed with CRC between 1998 and 2011, received surgical resection for their tumor, and completed a CAHPS survey within 6-36 months post-diagnosis. CAHPS surveys captured patient-reported GHS and MHS on a five-point Likert scale ranging from “poor” to “excellent.” We used stepwise multivariable logistic regression to examine associations between patient race/ethnicity and fair or poor health status, adjusting for clinical and sociodemographic factors. Additionally, a multivariable Cox proportional hazards regression was used to determine the risks of mortality associated with fair or poor GHS and MHS. Results: Of 1,867 patients with CRC, 79.5% were non-Hispanic white (NHW), 6.4% were non-Hispanic black (NHB), 7.5% were Hispanic, and 6.6% were non-Hispanic Asian (NHA). In Model 1 of our stepwise logistic regression, NHB patients had higher unadjusted odds for fair or poor GHS (OR 1.56, 95% CI 1.06-2.28) compared to NHW patients while Hispanic patients had higher unadjusted odds for both fair or poor GHS (1.48, 1.04-2.11) and MHS (1.92, 1.23-3.01). In Model 2, this relationship persisted after adjusting for clinical factors, with NHB patients being more likely to report fair or poor GHS (1.62, 1.10-2.40) and Hispanic patients being more likely to report fair or poor GHS (1.49, 1.04-2.13) and MHS (1.92, 1.22-3.00). In Model 3, after adjusting for both clinical and sociodemographic factors, the association between race/ethnicity and fair or poor GHS ( p = 0.53) and MHS ( p = 0.23) no longer remained. Reporting fair or poor GHS and MHS was associated with a greater risk of mortality among all CRC patients (HR 1.52, 95% CI 1.31-1.76 and 1.62, 1.34-1.99, respectively). Conclusions: Our study illustrates that racial/ethnic differences in PROs are largely driven by sociodemographic factors as opposed to clinical factors. As fair or poor GHS and MHS shortly after diagnosis reflect a higher risk of mortality in CRC patients, efforts to understand unmet biopsychosocial concerns may help further elucidate racial differences in CRC survival that may be otherwise overlooked in standard clinical practice.