Insurance Transitions and Changes in Health Care Utilization: Evidence from All Payer Claims Databases

Changes in the source of health insurance, particularly transitions between Medicaid and private coverage, may affect health care utilization and spending through several channels. First, the set of providers who accept Medicaid may be limited compared to private insurance; the breadth of provider networks may also vary across different sources of private insurance. Second, transitions between Medicaid and private insurance are likely to be accompanied by changes in other circumstances, such as changes in employment, health status, or family structure. Third, changes in cost sharing may change utilization patterns. Due to the scarcity of panel data on utilization tracking large samples of individuals who switch between public and private sources of coverage, however, evidence on the effect of transitions in coverage on health care utilization remains limited. This is especially true with respect to studies from the current health policy environment created by the rollout of the Affordable Care Act's major coverage expansions in 2014.

We use data on insurance enrollment and health care utilization from the Arkansas, Colorado, and Rhode Island All Payer Claims Databases to document changes in utilization associated with transitions across sources of coverage. All-Payer Claims Databases hold considerable promise for studying transitions in coverage, but they have been utilized in relatively few such studies to date. We focus on individuals who have experienced insurance transitions between private insurance sources (employer-sponsored insurance, Exchange coverage, and other non-group coverage) and Medicaid, using event-study regression models to document patterns of utilization six months before and six months after transitions to and from Medicaid. We consider office visits, emergency department (ED) visits, inpatient stays, and prescription fills, with a particular emphasis on Opioid and Medication Assisted Treatment (MAT) prescriptions.

We find clear evidence that transitions between private coverage and Medicaid are associated with major changes in health care utilization. However, we also find that utilization of some types of care leading up to a switch in coverage are systematically different from those observed among individuals who do not experience a change in the source of coverage--both in the level of health care utilization and in the trends in utilization of certain services and medications. Findings from Colorado include the following:

• Utilization by individuals who will leave or join Medicaid is systematically different from individuals who have continuous Medicaid coverage even prior to their insurance transition.
• Those who will switch between private and Medicaid coverage are more likely to use office-based care and less likely to use ED care or prescription drugs.
• After joining Medicaid, enrollees change their utilization patterns. Particularly, among individiuals joining Medicaid from Employer-Sponsored Insurance or Exchange coverage, we observe decreases in offices visits and increases in ED visits, inpatient stays, and prescription drug fills.
• Among those who leave Medicaid, we also see changes in utilization. In contrast to Medicaid joiners, Medicaid leavers increase their use of office visits and decrease their use of ED visits, inpatient stays, and prescription drugs.