Background: Care fragmentation is associated with a number of clinical and socio-demographic features which have been thought to increase hospital utilization as well as morbidity and mortality. Patients who receive fragmented care at the hospital level are often underinsured and uninsured. It is unclear to what degree lack of insurance and subsequently lack of continuous preventive care drives fragmentation and high hospital use. In this study we aim to determine if Medicaid expansion was effective in improving insurance rates in this challenging population and improve hospital continuity and utilization.
Methods: We identified high-utilizing patients within the Health Care Utilization Project’s State Inpatient Dataset from 6 states: 3 that expanded Medicaid (IA, VT, NY) and 3 that did not (GA, FL, and UT) from 2011 – 2015. Super-utilizers were identified if they had 4 or more hospitalizations in 1 year. We determined the efficacy of medicaid expansion in reducing the rate of uninsured using differences-in-differences (DiD) analysis using non-expansion states as a control following the first quarter of 2014. Then we then used DiD to test the association between Medicaid expansion and the number of different hospitals visited, number of encounters, total length of stay, diagnosis rates, and charges of high utilizing patients.
Results: Medicaid expansion was associated with a reduction in the number of uninsured encounters (OR 0.24, 95% CI 0.23 – 0.25, p <0.001) and specifically better for patients with fragmented care (p<0.001 for interaction). Medicaid expansion was specifically successful in reducing the rate of uninsured of patients with care fragmentation (p <0.001 for interaction). Medicaid expansion was associated with a lower degree of fragmentation as measured by number of different hospitals visited in 1 year (coef -0.025, p<0.001), lower number of encounters (coef -0.03, p<0.001), lower total charges (coef -0.249, p<0.001), but higher total length of stay, (coef 0.106, p<0.001). Hospital reported mortality and chronic diagnosis rates did not change.
Conclusions: Medicaid expansion was effective in reducing the rate of uninsured patients with care fragmentation, particularly in patients that lack continuity. We also found reductions in the degree of fragmentation, and lower hospitalization rates; however, this did not translate to a reduction in the total number of days hospitalized nor mortality. We conclude that while improving insurance coverage improved continuity to a small degree, more work is needed to specifically address drivers of fragmentation and hospital utilization.