Differences in Medicare Expenditures Between Appalachian and Nationally Representative Cohorts of Elderly Women With Breast Cancer: An Application of Decomposition Technique

Abstract

Background: Differences in Medicare expenditures during the initial phase of cancer care among rural and medically underserved elderly women with breast cancer (BC) and those from a nationally representative cohort have not been reported. The objective of this study was to determine Medicare expenditures during the initial phase of care among women in West Virginia (WV) who were Medicare beneficiaries with BC and compare them with national estimates. The magnitude of differences in these expenditures was also determined by using a linear decomposition technique. Methods: A retrospective observational study was conducted using the WV Cancer Registry-Medicare database and the SEER-Medicare database. Our study cohorts consisted of elderly women aged ≥66 years diagnosed with incident BC in 2003 to 2006. Medicare expenditures during the initial year after BC diagnosis were derived from all of the Medicare files. Generalized linear regressions were performed to model expenditures, after controlling for predisposing factors, enabling resources, need, healthcare use, and external healthcare environmental factors. Blinder-Oaxaca decomposition was conducted to examine the proportion of the differences in the average expenditures explained by independent variables included in the model. Results: Average Medicare expenditures for the WV Medicare cohort during the initial phase of BC care were $25,626 compared with $29,502 for the SEER-Medicare cohort; a difference of $3,876. In the multivariate regression, this difference decreased to $708 and remained significant. Only 16% of the differences in the average expenditures between the cohorts were explained by the independent variables included in the model. Enabling resources (6.86%), healthcare use (7.55%), and external healthcare environmental factors (3.33%) constituted most of the explained portion of the differences in the average expenditures. Conclusions: The difference in average Medicare expenditures between the elderly beneficiaries with BC from a rural state (WV) and their national counterparts narrowed but remained significantly lower after multivariate adjustment. The explained portion of this difference was mainly driven by enabling and healthcare use factors, whereas 84% of this difference remained unexplained.