This project tests how payment increases affect physician supply of care for Medicaid patients. The study uses a natural experiment in which Medicaid reimbursement for primary care services increased by an average of 73 percent, as a result of a provision in the Affordable Care Act that increased Medicaid primary care fees to Medicare levels. This policy change affected physicians differently across states, based on their Medicaid-to-Medicare ratios. I will use this policy experiment to test whether physicians respond to payment increases by increasing how much care they provide to Medicaid patients - as observed in a database of claims and electronic health record data compiled by athenahealth, Inc. The Affordable Care Act greatly increased access to insurance, partially through state Medicaid expansions. Expanded coverage has generated concern about access to care, particularly as physicians have historically been less likely to accept Medicaid patients. The increase in Medicaid reimbursement for primary care was designed to incentivize physicians to treat more Medicaid patients. My proposed research will directly test the efficacy of this policy and inform an ongoing debate about whether increasing Medicaid reimbursement is an effective strategy for ensuring access. If physicians respond to a payment increase by changing their supply of medical care for Medicaid patients, this change may take a few possible forms. First, physicians who currently treat Medicaid patients may treat more.
Aim 1 will test for this, using a modified difference-in-differences (DD) framework and outcome measures that quantify Medicaid patients as a share of a physician's panel and a physician's total scheduled appointment time. The DD framework identifies the effects of payment increases by comparing physician supply of medical before and after the Medicaid primary care payment increase, and across physicians that experienced different price shocks as a result of the pre-period fee schedule generosity in their state. Physician and time fixed effects control for baseline differences in outcomes among physicians and time trends. A payment increase may also incentivize physicians who had previously excluded Medicaid patients to begin accepting them.
Aim 2 will test for this using a binary outcome measure equal to one if the physician accepted any Medicaid patients during that time period. The estimating equation is similar to Aim 1, but will use logistic regression, due to the binary nature of the dependent variable. Finally, physicia response to a payment increase may differ by new versus established patients.
Aim 3 focuses on changes in access to physician services for new patients with Medicaid. To assess whether increased payment resulted in greater access for new patients, I will test whether there was an expansion in the percent of patients in a physician's panel who had not previously been seen, as identified by a collection of procedure codes reserved for office visits with new, rather than established, patients.
Understanding the effect of a payment increase on physician supply of care for Medicaid patients is crucial for policymakers, as more than 10 million individuals have recently gained insurance coverage through state Medicaid expansions. Medicaid has historically been considered an ungenerous payer with concerns about access to care. Knowing whether payment increases encourage physicians to participate in Medicaid, devote more time to existing Medicaid patients, or accept new Medicaid patient's represents a key contribution to health policy research and future evidence-based policy-making.
