An analysis of disenrollment from Medicare managed care plans by Medicare beneficiaries with diabetes.

Journal Article (Journal Article)

RESEARCH OBJECTIVE: The purpose of this work is to determine whether high-cost high-risk Medicare patients with diabetes in managed care plans disenroll more quickly than lower-cost lower-risk Medicare patients with diabetes. If high-cost high-risk patients with diabetes do disenroll more quickly, Medicare managed care plans benefit financially from favorable disenrollment. STUDY DESIGN: Time in a health maintenance organization (HMO) was modeled using a duration model with the number of months in the HMO as the dependent variable, controlling for censoring. Data were drawn from a representative sample of Medicare patients with diabetes in the FFS sector in 1994. The panel was followed for 4 years, 1995-1998. The sample included all 6839 individuals who enrolled in a Medicare HMO for at least 1 month during the 48-month observation window. PRINCIPAL FINDINGS: We found a statistically significant negative association between the time in an HMO and pre-enrollment Part B expenditures (beta = -0.00001, t = -4.39) and any Part A expenditures (beta = -0.465, t = -1.98), and 2 of 4 diabetic complications (heart complications: beta = -0.0773, t = -4.61; vision complications beta = -0.2474, t = -1.94). Of the plan characteristics, only the drug benefit variable (beta = 0.151, t = 5.64) had a statistically significant coefficient. CONCLUSIONS: Overall, our results support the hypothesis that high-cost, high-risk individuals disenroll from Medicare HMOs sooner than lower-cost lower-risk individuals. However, this effect is mitigated by plans offering better prescription drug benefits. We did find some evidence that patients with diabetes with very high pre-enrollment Part A costs may remain longer in HMOs relative to patients with diabetes with lower Part A prior year expenditures.

Full Text

Duke Authors

Cited Authors

  • Atherly, A; Hebert, PL; Maciejewski, ML

Published Date

  • May 2005

Published In

Volume / Issue

  • 43 / 5

Start / End Page

  • 500 - 506

PubMed ID

  • 15838416

International Standard Serial Number (ISSN)

  • 0025-7079

Digital Object Identifier (DOI)

  • 10.1097/


  • eng

Conference Location

  • United States