To link data from a central cancer registry with Medicaid enrollment and claims files in order to assess cancer care in an economically disadvantaged population.
Over 500,000 cancer patients diagnosed between 2002 and 2006 reported to the New York State Cancer Registry were linked with New York State Medicaid enrollment and claims records.
A probabilistic linkage was performed between the two data sources. The resulting Medicaid and non‐Medicaid populations were compared in terms of demographics and stage at diagnosis.
Existing databases were used.
One‐quarter of cancer patients were enrolled in Medicaid at or near the time of cancer diagnosis. The Medicaid cohort was younger, more likely to be an ethnic minority, foreign born, never married, live in either an inner‐city or remote rural area, and have a higher stage at diagnosis.
The linked dataset will permit detailed analysis of cancer treatment and cancer treatment disparities among historically understudied groups. The linkage has also resulted in improvements in Cancer Registry quality through the identification of errors and missing values. The linkage did present technical challenges in the form of immense file sizes not easily adaptable to desktop computers.