Survival Disparities in Multiple Myeloma by Health Insurance Status among US Non-Elderly Adults: A SEER-Based Comparative Analysis.

BACKGROUND/AIM: The impacts of health insurance status on survival outcomes in multiple myeloma (MM) have not been addressed in depth. The present study was conducted to identify definite relationships of cancer-specific survival (CSS) and overall survival (OS) with health insurance status in MM patients.

METHODS

MM patients aged 18-64 years and with complete insurance records between January 1, 2007, and December 31, 2016, were identified from 18 Surveillance, Epidemiology, and End Results (SEER) Database registries. Health insurance condition was categorized as uninsured, any Medicaid, insured, and insured (no specifics). Relationships of health insurance condition with OS/CSS were identified through Kaplan-Meier, and uni-/multivariate Cox regressions using the hazard ratio and 95% confidence interval. Potential baseline confounding was adjusted using multiple propensity score (mPS).

RESULTS

Totally 17,981 patients were included, including 68.3% with private insurance and only 4.9% with uninsurance. Log-rank test uncovered significant difference between health insurance status and OS/CSS among MM patients. Patients with non-insurance or Medicaid coverage in comparison with private insurance tended to present poorer OS/CSS both in multivariate Cox regression and in mPS-adjusted model (non-insurance vs. private insurance [OS/CSS]: 1.33 [1.20-1.48]/1.13 [1.00-1.28] and 1.45 [1.25-1.69]/1.18 [1.04-1.33], respectively; Medicaid coverage vs. private insurance [OS/CSS]: 1.67 [1.56-1.78]/1.25 [1.16-1.36] and 1.76 [1.62-1.90]/1.23 [1.13-1.35], respectively).

CONCLUSIONS

Our observational study of exposure-outcome associations suggests that insufficient or no insurance is moderately linked with OS among MM patients aged 18-64 years. Wide insurance coverage and health-care availability may strengthen some disparate outcomes. In the future, prospective cohort research is needed to further clarify concrete risks with insurance type, owing to the lack of definite division of insurance data in SEER.