Poster Paper:
The Spillover Effects of Medicare Advantage on Health Care Utilization and Spending Among Traditional Medicare Beneficiaries with Multiple Chronic Conditions
*Names in bold indicate Presenter
Objectives: We estimated the spillover effects of MA plans, as measured by their market penetration, on health care spending, ED visits, and hospital readmission for all TM beneficiaries, and among TM beneficiaries with multiple chronic conditions.
Data: We used multiple data from the Centers for Medicare and Medicaid Services in 2009-2017: Multiple Chronic Conditions files, Geographic Variation Public Use files, annual risk and ratebook files, and monthly MA enrollment data.
Methods: Our outcomes were county-level Medicare per capita standardized spending, ED visits per 1,000 beneficiaries, and hospital readmission percentage for TM beneficiaries aged 65 and older. Our key independent variable was county-level MA penetration rate. However, the literature has shown evidence of endogenous insurer entry based on profitability. To control for such endogeneity, we performed a two-stage least squares regression (2SLS) model. Following the methodology used in the previous literature, we used a simulated benchmark payment as an instrument. We ran the 2SLS model by numbers of chronic conditions: 1) zero to one, 2) two to three, 3) four to five, and 4) more than six.
Results: We found that greater MA penetration was significantly associated with lower Traditional Medicare per capita standardized spending, ED visits, and hospital readmission. This was more pronounced for TM beneficiaries with a higher number of chronic conditions. Specifically, we estimated that a one percentage point increase in MA penetration was significantly associated with $29, $157, $424, $1,310 reductions in health care spending for TM beneficiaries with zero to one, two to three, four to five, and more than six chronic conditions, respectively. We also estimated that a one percentage point increase in MA penetration was significantly associated with 0.84, 2.72, 4.52, 10.91 fewer ED visits per 1,000 beneficiaries for TM beneficiaries with zero to one, two to three, four to five, and more than six chronic conditions, respectively. Finally, we estimated that a one percentage point increase in MA penetration was significantly associated with a 0.07 percentage point increase in hospital readmission for TM beneficiaries with zero to one, whereas it was significantly associated with 0.13 and 0.17 percentage point decreases in hospital readmission for TM beneficiaries with four to five and more than six chronic conditions, respectively.
Conclusion: Our findings suggest that the spillover effects of MA penetration are particularly pronounced for TM beneficiaries with multiple chronic conditions. This implies that MA growth may be playing a role in improving the efficiency of health care delivery.