Insurance status and survival among female breast cancer patients in Florida

Introduction: Insurance type at the time of a cancer diagnosis can be a quality of care marker and/or a way to identify a population at increased risk of poor survival outcome. We examined associations between survival and insurance type in women diagnosed with breast cancer in Florida from 1996-2007.

Methods: Data from the Florida Cancer Data System (FCDS), the Agency for Health Care Administration (in- and out-patient treatment records), and the Census were linked (n=111,822). Survival time by insurance type was our primary endpoint, with adjustments for sociodemographic status, a neighborhood-based poverty measure, clinical characteristics, and co-morbidities. Insurance type was defined as private, defense/military/veteran, Indian Health Service, Medicaid, Medicare, other insurance, or uninsured. Cox regression models were performed to determine the hazard ratio (HR). Results: In the univariate model, compared to private insurance Medicaid (HR 2.42; 95% Confidence Interval=2.28-2.56), Medicare ([2.07]; 2.02-2.13), Indian Health Service ([2.03]; 1.61-2.55), and uninsured ([1.91]; 1.8-2.03) had increased risk of reduced survival. In the fully adjusted model, women with Indian Health Service insurance had the largest risk of reduced survival [1.72]; 1.33-2.23); those with Medicaid ([1.62]; 1.50-1.75), uninsured ([1.41]; 1.27-1.57), and Medicare ([1.10]; 1.06-1.15) were also at increased risk (all p<0.001). These risks were not significantly different for women with defense/military/veteran ([1.08]; 0.93-1.25). Conclusion: Insurance type at the time of diagnosis is predictive of survival in women with breast cancer. Studies are needed to identify strategies to reduce the excess mortality burden in women with Indian Health Service and Medicaid insurance.