Insurance acceptance trends in psychiatry

Bishop and colleagues' 2014 JAMA Psychiatry paper analyzed a national sample from the CDC's National Ambulatory Medical Care Survey. They compared office-based psychiatrists with office-based physicians in every other specialty on a single question: whether the practice accepted new patients with a given payer. The results were consistent across every payer they looked at. Psychiatrists accepted private non-capitated insurance at 55.3 percent versus 88.7 percent for other specialties, Medicare at 54.8 percent versus 86.1 percent, and Medicaid at 43.1 percent versus 73.0 percent. Almost a decade of secret-shopper studies since then have found similar gaps in real-world appointment availability. See the Malowney psychiatric secret-shopper study (Psychiatric Services, 2015) for one widely-cited replication.

Three factors show up repeatedly in the workforce literature. First, reimbursement per visit for evaluation and management codes has not kept pace with the time psychiatric visits take, especially initial evaluations and complex follow-ups. Second, administrative burden, including pre-authorization for medications and network paperwork, adds unpaid overhead. Third, demand exceeds supply in most metropolitan areas, so a psychiatrist can fill a schedule without accepting any insurance at all.

The result is a market in which cash-pay and out-of-network practice is more sustainable in psychiatry than in most of medicine. Bishop and colleagues' follow-up work on network adequacy documented that many psychiatrists listed as in-network are unreachable, not accepting new patients, or unwilling to accept the plan they are listed under, patterns often described as ghost networks.

For patients, the practical consequence is higher out-of-pocket spending. AHRQ Medical Expenditure Panel Survey data show that mental health services have a higher out-of-pocket share than most outpatient medical categories. Some patients pay cash and submit for out-of-network reimbursement; some pay cash and don't submit; and some don't start care because the cost gap is prohibitive. This is a piece of the treatment cascade described in the U.S. Anxiety Access Report, and it interacts with wait times covered in the report on psychiatric wait times.

The Mental Health Parity and Addiction Equity Act of 2008 requires group health plans that offer mental health and substance use benefits to cover them at parity with medical and surgical benefits. The Consolidated Appropriations Act of 2021 added a requirement that plans conduct and document a comparative analysis of non-quantitative treatment limitations, and the 2024 amendments strengthened enforcement. See the US Department of Labor's MHPAEA overview for the current framework.

Parity law changes what benefits look like on paper. It does not directly change the number of in-network psychiatrists, and enforcement is uneven. The Department of Labor's annual reports to Congress have repeatedly identified compliance gaps, and state insurance regulators have brought actions against insurers whose behavioral-health networks fell short of adequacy tests.

Even when a plan lists a robust behavioral-health network, patients often can't reach the people in it. Secret-shopper studies calling insurer-published directories have found high rates of wrong numbers, providers not accepting the plan, providers not accepting new patients, or providers no longer in practice. These findings have been documented enough times to have a name: ghost networks. The result is that on-paper access, which is what parity law regulates, can meaningfully overstate real access.

Acceptance-rate surveys measure the practice, not the patient. They don't say why a particular patient couldn't find in-network care, and they don't identify which policy lever moves the number fastest. The Bishop numbers are national averages; the reality varies by state, insurer, and whether behavioral health is carved out to a managed behavioral health organization. Any local number should be read against the local plan and network.

Psychiatry accepts insurance at lower rates than most other medical specialties. The most-cited evidence is a 2014 study by Bishop and colleagues in JAMA Psychiatry, which found that psychiatrists accepted new patients with private non-capitated insurance at 55.3 percent, compared with 88.7 percent for physicians in other specialties. Medicare acceptance among psychiatrists was 54.8 percent versus 86.1 percent for other specialties. Medicaid acceptance among psychiatrists was 43.1 percent versus 73.0 percent. The gap was large enough to be a structural feature of the field rather than a rounding difference.

The reasons are debated but include reimbursement that has not kept pace with the time an evaluation and management visit takes, administrative burden, and demand high enough that psychiatrists can fill their schedules without contracting with insurers. The result is a larger cash-pay and out-of-network segment in psychiatry than in most of medicine. AHRQ Medical Expenditure Panel Survey data on out-of-pocket spending consistently show mental health as a higher-cost category per user than most outpatient care.

Mental health parity law, principally the Mental Health Parity and Addiction Equity Act of 2008 and the Consolidated Appropriations Act of 2021's amendments to it, requires insurers to cover mental health benefits comparably to medical and surgical benefits. Parity has improved coverage on paper, but enforcement and network adequacy remain ongoing concerns. A covered benefit is of little practical use if few local psychiatrists are in network. The US Department of Labor's annual MHPAEA reports to Congress have repeatedly found compliance gaps, and states have taken enforcement action against insurers whose behavioral-health networks failed adequacy tests.

The Bishop study is from 2014, so the exact percentages will have shifted. The direction of the gap, psychiatry below other specialties, has been stable across the available evidence, including newer secret-shopper studies that reach the same conclusion.

"In network" and "accepting new patients" are different questions, and both matter separately.