Mental Health Ghost Networks in Commercial Insurance Plans

Importance: Whether network directories in commercial insurance plans, which cover the majority of insured US residents, reliably connect patients to mental health care has only seen limited testing. Objective: To assess network directory accuracy, appointment availability, wait times, and travel burden for mental health clinicians listed in commercial insurance online plan directories. Design, Setting, and Participants: A cross-sectional study with a secret shopper design of 5822 calls to psychiatrists and psychologists listed in the online network directories of the largest commercial carrier in 5 diverse states (Delaware, Montana, New Jersey, Rhode Island, and Virginia), conducted between October 2025 and March 2026. Callers contacted the geographically closest listed clinician on behalf of a simulated patient assigned a standard preferred provider organization plan and a nonemergency mental health condition. Calls in which the caller could not obtain a definitive outcome were classified as nonevaluable. Main Outcomes and Measures: Appointment availability with the listed clinician and with alternate clinicians, network directory inaccuracy by error type (contact information, specialty, network status), capacity limitations, wait time to appointment, and travel time to appointment. Reported percentages and means are adjusted (predicted) values with 95% CIs. Results: Of 5822 calls, 3570 (61.3%) were nonevaluable, most commonly because callers reached only an answering machine or automated system. Among the 2252 evaluable calls, 23.6% (95% CI, 21.9%-25.4%) resulted in a scheduled appointment with the listed clinician, increasing to 32.1% (95% CI, 30.2%-34.1%) when alternate clinicians were included. Network directory inaccuracies affected 55.6% (95% CI, 53.6%-57.7%) of evaluable calls, with contact information errors (31.0%; 95% CI, 29.1%-32.9%) and specialty errors (22.2%; 95% CI, 20.6%-23.9%) being most prevalent. Capacity limitations prevented scheduling in 21.6% (95% CI, 19.9%-23.3%) of evaluable calls. Mean wait time was 20.5 days (95% CI, 18.0-23.0 days), with longer waits for psychiatry than psychology (22.5 vs 17.3 days). Mean travel time was 17.1 minutes (95% CI, 14.6-19.6 minutes) overall but 41.8 minutes (95% CI, 27.5-56.0 minutes) in Montana and 30.8 minutes (95% CI, 19.8-41.7 minutes) for rural residents. Across all initial call attempts, only 12.4% resulted in an appointment. Conclusions and Relevance: In this study, access barriers previously documented in public insurance programs extend to the privately insured population, undermining the assumption that commercial coverage automatically translates into meaningful access to mental health care.