Analyzing Insurance Status and Cancer Outcomes

New Jersey State Cancer Registry

Previous studies suggest that having no health insurance or inadequate insurance coverage can be a barrier to accessing cancer screening or receiving appropriate cancer treatment. Such barriers have been shown to impact cancer patient survival.

To improve the completeness of insurance status in the registry, we matched cancer cases without insurance information to Hospital Patient Discharge Uniform Billing files. Six cancer sites were examined: breast, colorectal, lung, melanoma, bladder, and non-Hodgkin lymphoma. Thirteen percent of the cases in the registry diagnosed between 1999 and 2004 had an unknown insurance payer. Matching reduced the percentage of cases with an unknown payer to about 7%.

We then used our insurance data to examine associations between health insurance status (insured, Medicaid, uninsured) with cancer patient survival in New Jersey for five common sites diagnosed among persons aged 18 to 64 (female breast, colorectal, lung, non-Hodgkin lymphoma, urinary bladder) obtained from the New Jersey State Cancer Registry. Patients were followed to acquire vital status through 2005. Cox proportional hazards models were fit separately for each cancer site to estimate the risk of cancer death for three insurance categories (Medicaid, private, uninsured), after adjusting for age, stage at diagnosis and race/ethnicity.

Results for the five cancers among uninsured patients had a significantly higher probability of cancer death compared to privately insured patients. Medicaid-insured patients also had a significantly higher probability of cancer death compared to privately insured patients for four of the primary sites: female breast, colorectal, lung, and non-Hodgkin lymphoma. These results confirm previous findings indicating that patients without private health insurance experienced worse cancer survival. Additional research is needed to investigate factors that affect health care access and quality of care.