Mental Health Ghost Networks in Commercial Insurance Plans

Simon Haeder; Yuqin Lou; Wendy Xu; Benjamin F Miller

doi:10.1001/jamahealthforum.2026.3746

. 2026 Oct 16;7(10):e263746. doi: 10.1001/jamahealthforum.2026.3746

Mental Health Ghost Networks in Commercial Insurance Plans

Simon Haeder ^1,^✉, Yuqin Lou ¹, Wendy Xu ¹, Benjamin F Miller ²

¹Division of Health Services Management & Policy, College of Public Health, The Ohio State University, Columbus, Ohio

²Department of Psychiatry and Behavioral Sciences, Stanford University School of Medicine, Stanford, California

Accepted for Publication: August 7, 2026.

Published: October 16, 2026. doi:10.1001/jamahealthforum.2026.3746

Open Access: This is an open access article distributed under the terms of the CC-BY-NC-ND License, which does not permit alteration or commercial use, including those for text and data mining, AI training, and similar technologies. © 2026 Haeder S et al. JAMA Health Forum.

^✉

Corresponding Author: Simon Haeder, PhD, MPA, Division of Health Services Management & Policy, College of Public Health, The Ohio State University, 1841 Neil Ave, Cunz Hall, 250, Columbus, OH 43210 (haeder.1@osu.edu).

Author Contributions: Dr Haeder had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Haeder, Xu, Miller.

Acquisition, analysis, or interpretation of data: All authors.

Drafting of the manuscript: Haeder, Miller.

Critical review of the manuscript for important intellectual content: All authors.

Statistical analysis: Haeder, Lou.

Obtained funding: Haeder, Xu.

Administrative, technical, or material support: All authors.

Supervision: Haeder, Xu.

Conflict of Interest Disclosures: Dr Haeder reported receiving grants from Inseparable, the Robert Wood Johnson Foundation, and the Pennsylvania Insurance Department during the conduct of the study; receiving personal fees from Grant & Eisenhofer, P.A., Bradley Bernstein Sands LLP, McGuireWoods LLP, Cera LLP, and Watts Guerra LLP outside the submitted work. Dr Xu reported receiving grants from Inseparable during the conduct of the study. No other disclosures were reported.

Funding/Support: This work was funded by Inseparable.

Role of the Funder/Sponsor: The funder participated in discussions regarding the selection of states for study. The funder had no role in the conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; or the decision to submit the manuscript for publication.

Data Sharing Statement: See Supplement 2.

Additional Contributions: Cecilia Glancy, Camille Richardson, Katherine Anderson, Rute Admasu, Keerti Gudla, Jiyi Liu, Tyler Zevallos, Sydney Rizzo, CJ Marcella-Sill, Kayli Rego, and Sean Custy (College of Public Health, The Ohio State University) conducted the telephone surveys. They were compensated for this work.

AI Acknowledgement: The authors used Claude Opus 4.8 (Anthropic) during the revision stage of this manuscript to assist with cross-checking the manuscript for internal consistency, such as verifying that numerical values reported in the text matched those in the corresponding tables, that terminology was used consistently, proofreading, and final editing from July 28 to August 2, 2026. The authors reviewed, verified, and take full responsibility for all content. The paper represents the authors’ own ideas, analysis, and interpretation of the data.

^✉

Corresponding author.

PMCID: PMC13675814 PMID: 42886497

Key Points

Question

How effectively do network directories in commercial insurance plans connect patients with mental health care?

Findings

In this cross-sectional study with a secret shopper design of 5822 calls to psychiatrists and psychologists listed in the directories of the largest commercial carriers across 5 states, 61.3% of calls yielded no definitive outcome. Among evaluable calls, only 23.6% resulted in a scheduled appointment while 55.6% were subject to directory errors, and 21.6% were subject to capacity constraints; a single call attempt yielded an appointment in 12.4% of cases.

Meaning

Results of this study suggest that mental health barriers previously documented in public insurance programs were found to extend to the privately insured.

Abstract

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.

This cross-sectional study assesses mental health care access via an audit of secret shopper calls to psychiatrists and psychologists listed in online network directories.

Introduction

Access to outpatient mental health care in the US remains severely constrained despite decades of policy efforts.^{1,2,3,4,5,6,7,8,9,10} Evidence suggests that patients often encounter numerous barriers when seeking mental health care as access to mental health networks is generally inadequate and clinicians are less likely to accept insurance.^{7,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25} These barriers have drawn increasing scrutiny through secret shopper studies, which capture access as experienced by consumers. Using this approach, researchers have documented low rates of appointment scheduling, frequent inability to reach listed clinicians, extended wait times, and substantial variation across various markets and payers.^{7,24,26,27,28,29,30,31,32,33,34}

A related body of evidence has identified pervasive deficiencies in network directories, with large numbers of phantom or ghost clinicians who appear in directories but are not meaningfully available to patients.^{7,11,15,24,28,35,36,37,38,39,40,41,42,43,44,45,46} These directories are crucial because they are the only valid public-facing source of information on network participation, which carries direct implication for various patient protections.^47,48 Importantly, inaccuracies not only create barriers for patients but also undermine the regulatory frameworks to assess network adequacy.^49,50,51

In combination, these access barriers can impose substantial financial and nonfinancial costs on patients.^{20,47,50,52,53} Federal and state regulators have responded with expanded directory accuracy requirements,^{24,49,50,54,55,56,57,58,59,60,61,62,63,64} including the No Surprises Act mandate that insurers verify and update directory information at least every 90 days. However, evidence suggests that these provisions have not yet translated into meaningful improvements.^36,64,65

Despite a growing body of secret shopper research documenting barriers to mental health care,^{26,27,31,32,33,66,67,68,69} the commercial insurance market, has only seen limited analysis through this method.^27,28 Existing research has generally assessed narrow outcome sets, focused primarily on psychiatrists, and covered a small number of predominantly urban markets.^{7,11,15,24,27,28,35,36,37,38,39,40,41,42,43,44,45,46} We are not aware of prior work that simultaneously assesses directory accuracy by error type, capacity constraints, wait times, and travel burden for both psychiatrists and psychologists. Moreover, 4 of the 5 states examined have never been included in any secret shopper study of mental health access, and Virginia is the sole state that appeared in a single prior study limited to psychiatry.²⁶ This study addresses these gaps through a secret shopper audit of calls, the largest such study to date to our knowledge, to psychiatrists and psychologists listed in the network directories of the largest commercial carriers across 5 geographically diverse states.

Methods

This cross-sectional study with a secret shopper design was approved by the Ohio State University institutional review board. The informed consent requirement was waived given the minimal risk to clinician offices, the design of the study, and the public interest in monitoring network activity. This study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline. We conducted secret shopper surveys of psychiatrists and psychologists listed in the online network directories of the largest commercial carrier in each of 5 states between October 2025 and March 2026. States were purposively selected to capture geographic, demographic, and market diversity spanning the Mid-Atlantic (Delaware, New Jersey, Virginia), New England (Rhode Island), and the Mountain West (Montana) and to include states that had received limited attention in prior secret shopper research. We identified the carrier with the largest commercial market share in each state (eTable 1 in Supplement 1).⁷⁰ Because Anthem Blue Cross Blue Shield does not offer coverage across the entire commonwealth of Virginia, we selected CareFirst BlueCross BlueShield as a secondary carrier. A standard preferred provider organization (PPO) plan was selected for each carrier. Calls were distributed proportional to population at the county level within each state, and each state received approximately the same number of calls.

Calls were conducted by trained research assistants following a standardized protocol. Callers contacted clinicians on behalf of a simulated patient (a close relative).^12,28,71,72 Each caller was assigned a random simulated patient profile including a home address, name, date of birth, insurance plan, and a nonemergency mental health condition. Callers used the assigned carrier’s online network directory to identify and contact the geographically closest in-network clinician of the assigned specialty to request the first available appointment. Calls in which a specific date was not provided because staff requested additional information before scheduling were included as successful appointments but excluded from wait time analyses. When a listed clinician was unavailable, callers recorded any appointment with an alternate clinician at the listed number that staff volunteered. No actual appointments were scheduled.

Data Analysis

Callers verified network directory information in a fixed order: contact information, then specialty, then network status. Calls generally ended once an inaccuracy was identified, consistent with the experience of real consumers, although staff occasionally volunteered more than 1 inaccuracy in a single response. Calls were classified as evaluable when the caller obtained a definitive outcome regarding appointment availability, network directory accuracy, or clinician capacity, either by reaching office staff or by receiving a disconnection or out-of-service message. Primary analyses were restricted to evaluable calls. Formal outcome definitions are provided in eTable 2, call script in eTable 3, distribution of calls in eFigures 1-5 in Supplement 1. We examined variation by state, carrier, specialty, caller rurality, and clinician rurality. Adjusted differences are reported from linear regression models with robust SEs, which permit interpretation on the percentage-point scale. Predicted probabilities and predicted means with 95% CIs were derived from logit models for binary outcomes and ordinary least squares regression with robust SEs for continuous outcomes. Statistical analyses, including pairwise comparisons across subgroups, were conducted using Stata version 19.5 (StataCorp). A 2-sided P < .05 was considered statistically significant. Percentages and means reported throughout this study are adjusted (predicted) with 95% CIs, based on the regression models.

Results

The survey encompassed 5822 calls across 5 states (Table 1), of which 532 (9.1%) yielded an appointment with the listed clinician and 724 (12.4%) with any clinician. However, 3570 calls (61.3%) were nonevaluable, most commonly because the caller reached only an answering machine or automated system with no option to speak to staff (2576 calls [44.2%]). The evaluable rate was higher for psychiatry (44.0%) than psychology (33.3%) and varied across states, from 30.7% in Rhode Island to 48.1% in Virginia.

Table 1. Secret Shopper Data Overview^a.

Variable	No. (%)
Variable	All calls (N=5822)	Evaluable calls (n=2252)
Psychiatry	2913 (50.0)	1283 (57.0)
Psychology	2909 (50.0)	969 (43.0)
Anthem (Virginia)	1022 (17.6)	487 (21.6)
Blue Cross Blue Shield of Rhode Island	1249 (21.5)	384 (17.1)
Blue Cross and Blue Shield of Montana	1226 (21.1)	444 (19.7)
CareFirst BlueCross BlueShield (Virginia)	78 (1.3)	42 (1.9)
Highmark Blue Cross Blue Shield Delaware	1164 (20.0)	485 (21.5)
Horizon Blue Cross Blue Shield of New Jersey	1083 (18.6)	410 (18.2)
Nonrural resident	4428 (76.1)	1731 (76.9)
Rural resident	1394 (23.9)	521 (23.1)
Nonrural clinician	4699 (80.7)	1868 (83.0)
Rural clinician	1123 (19.3)	384 (17.1)

Open in a new tab

^{^a}

Results based on data from a secret shopper survey (5822 calls) of psychiatrists and psychologists listed in commercial network directories for the largest commercial carrier in each of 5 states. Carriers surveyed include Anthem Blue Cross Blue Shield (Virginia), Blue Cross Blue Shield of Rhode Island, Blue Cross and Blue Shield of Montana, CareFirst BlueCross BlueShield (Virginia), Highmark Blue Cross Blue Shield (Delaware), and Horizon Blue Cross Blue Shield of New Jersey. In Virginia, two carriers were surveyed because they serve different regions of the state. Evaluable calls are those in which the caller obtained a definitive outcome regarding appointment availability, network directory accuracy, or clinician capacity (eTable 2 in Supplement 1 includes definitions). Calls distributed proportional to population at the county level within each state.

Adjusted differences in access outcomes among evaluable calls are given in Table 2. Among evaluable calls, 23.6% (95% CI, 21.9%-25.4%) resulted in a successfully scheduled appointment with the listed clinician (Table 3). Rates were similar for psychiatry (24.5%; 95% CI, 22.2%-26.9%) and psychology (22.4%; 95% CI, 19.8%-25.1%) and ranged from 18.0% (95% CI, 6.1%-29.9%) for CareFirst in Virginia to 32.9% (95% CI, 28.2%-37.6%) for Horizon Blue Cross Blue Shield in New Jersey (P = .02). Rural clinicians were more likely to offer appointments than nonrural clinicians (33.2% [95% CI, 25.9%-40.5%] vs 21.9% [95% CI, 19.9%-23.9%]; P = .006). When appointments with alternate clinicians at the same number were included, the overall success rate increased to 32.1% (95% CI, 30.2%-34.1%) (N = 724).

Table 2. Adjusted Differences in Access Outcomes Among Evaluable Calls^a.

Variable	Adjusted difference (95% CI)
Variable	Appointment^b	Wait time, d	Travel time, min	Directory error^b	Contact information^b	Specialty^b	Network status^b	Capacity limitation^b
Psychology	−2.1 (−5.6 to 1.5)	−5.2 (−10.1 to −0.3)	5.3 (−0.7 to 11.3)	−5.0 (−9.1 to −0.8)	0.4 (−3.5 to 4.3)	−5.0 (−8.4 to −1.5)	0.6 (−0.4 to 1.7)	6.4 (2.9 to 9.8)
Blue Cross Blue Shield of Rhode Island	−3.3 (−9.1 to 2.4)	−3.9 (−12.4 to 4.6)	−0.6 (−3.9 to 2.6)	−7.8 (−14.5 to −1.0)	−6.0 (−12.3 to 0.3)	3.9 (−2.0 to 9.8)	−3.1 (−4.7 to −1.6)	6.8 (1.5 to 12.1)
Blue Cross and Blue Shield of Montana	−6.6 (−12.5 to −0.6)	−13.0 (−26.4 to 0.4)	29.2 (14.8 to 43.6)	−0.6 (−7.7 to 6.5)	−5.7 (−12.4 to 1.0)	4.4 (−1.9 to 10.6)	−2.1 (−4.1 to −0.2)	5.0 (−0.8 to 10.9)
CareFirst BlueCross BlueShield (Virginia)	−7.2 (−19.1 to 4.7)	−9.2 (−20.5 to 2.1)	1.9 (−3.2 to 7.0)	10.4 (−4.3 to 25.0)	−0.8 (−15.7 to 14.1)	15.0 (−0.2 to 30.2)	−3.2 (−4.8 to −1.7)	1.1 (−10.7 to 12.8)
Highmark Blue Cross Blue Shield Delaware	−5.0 (−10.3 to 0.3)	−0.8 (−10.6 to 9.0)	−1.0 (−5.8 to 3.7)	−3.5 (−9.7 to 2.7)	−0.3 (−6.2 to 5.7)	−2.0 (−7.2 to 3.3)	−1.1 (−3.0 to 0.8)	9.0 (3.9 to 14.0)
Horizon Blue Cross Blue Shield of New Jersey	7.2 (1.2 to 13.2)	−2.6 (−11.1 to 5.8)	−1.9 (−5.0 to 1.2)	−15.2 (−21.8 to −8.6)	−0.7 (−7.0 to 5.6)	−12.4 (−17.3 to −7.5)	−1.5 (−3.4 to 0.5)	9.0 (3.7 to 14.3)
Rural caller	−1.6 (−8.3 to 5.1)	6.6 (−4.3 to 17.6)	18.1 (3.7 to 32.5)	5.7 (−2.4 to 13.7)	1.9 (−5.9 to 9.7)	4.2 (−2.7 to 11.1)	−0.6 (−3.6 to 2.4)	6.7 (−0.6 to 13.9)
Rural clinician	10.3 (3.5 to 17.0)	2.5 (−8.7 to 13.7)	−7.5 (−22.4 to 7.5)	−13.5 (−21.6 to −5.5)	−7.0 (−14.6 to 0.6)	−7.3 (−14.3 to −0.4)	0.2 (−2.8 to 3.2)	−5.0 (−12.2 to 2.1)
Constant	24.9 (20.6 to 29.3)	24.2 (16.6 to 31.7)	7.5 (4.1 to 11.0)	63.6 (58.7 to 68.4)	34.0 (29.3 to 38.7)	25.5 (21.3 to 29.7)	2.9 (1.3 to 4.5)	12.4 (8.8 to 16.0)
Observations	2252	437	531	2252	2252	2252	2252	2252
R ²	0.02	0.03	0.23	0.02	0.01	0.03	0.01	0.01
Outcome, mean (SD)	23.6 (42.5)	20.5 (27.2)	17.1 (32.9)	55.6 (49.7)	31.0 (46.3)	22.2 (41.6)	1.6 (12.4)	21.6 (41.1)

Open in a new tab

^{^a}

Estimates are from linear regression models with robust SEs; 95% CIs are shown in parentheses. Columns 2 and 3 report adjusted differences in days and minutes, respectively. Linear models were used for binary outcomes to permit interpretation on the percentage-point scale. Logistic models are reported in eTable 4 in Supplement 1 and yield substantively similar results. Reference categories are Psychiatry, Anthem (Virginia), Nonrural Caller, and Nonrural Clinician. The constant represents the estimated outcome for a nonrural caller contacting a nonrural psychiatrist listed by Anthem (Virginia). Outcome means and SDs are for each model’s estimation sample. Wait time models are restricted to evaluable calls in which an appointment was secured, and a specific appointment date was provided (n = 437). Travel time models are restricted to evaluable calls in which an appointment was secured (n = 532); in 1 case the geospatial software did not return a travel time estimate. Estimates for these 2 outcomes are conditional on successfully obtaining an appointment. Network status errors were observed in fewer than 2% of evaluable calls; estimates in column 7 are correspondingly imprecise and the linear specification can yield fitted values outside the 0-100 range. These estimates should be interpreted with caution.

^{^b}

Binary outcomes in columns 1 and 4-8 were scaled by 100, so estimates represent adjusted differences in percentage points.

Table 3. Appointment Success, Wait Times, and Travel Times Among Evaluable Calls^a.

Characteristic	Measurement (95% CI)
	Original clinician			Any available clinician^b
	Appointment success, %^c	Wait time to appointment, d^d	Travel time, min^d	Appointment success, %^c	Wait time to appointment, d^d	Travel time, min^d
Total	23.6 (21.9- 25.4)	20.5 (18.0-23.0)	17.1 (14.6-19.6)	32.1 (30.2-34.1)	19.4 (17.2-21.7)	16.4 (14.5-18.4)
Psychiatry	24.5 (22.2-26.9)	22.5 (18.9-26.1)	14.9 (12.9-17.0)	32.6 (30.0-35.1)	21.1 (18.0-24.3)	14.3 (12.6-16.0)
Psychology	22.4 (19.8-25.1)	17.3 (14.0-20.7)	20.3 (14.8-25.8)	31.6 (28.6-34.5)	16.7 (13.8-19.6)	19.4 (15.2-23.5)
Anthem (Virginia)	25.4 (21.5-29.3)	24.1 (16.3-31.8)	12.6 (10.5-14.7)	34.4 (30.2-38.7)	23.5 (16.8-30.1	12.5 (10.8-14.3)
Blue Cross Blue Shield of Rhode Island	22.1 (17.7-26.5)	20.2 (14.6-25.7)	12.0 (9.0-15.0)	30.8 (26.0-35.6)	19.8 (14.8-24.8)	11.8 (9.5-14.2)
Blue Cross and Blue Shield of Montana	19.0 (14.9-23.1)	11.1 (2.7-19.5)	41.8 (27.5-56.0)	28.5 (23.6-33.3)	10.4 (3.2-17.7)	37.6 (26.9-48.3)
CareFirst BlueCross BlueShield (Virginia)	18.0 (6.1-29.9)	14.9 (5.8-24.0)	14.5 (9.3-19.7)	22.3 (9.6-35.0)	15.0 (6.1-23.9)	15.6 (11.4-19.8)
Highmark Blue Cross Blue Shield Delaware	20.4 (16.7-24.0)	23.3 (17.5-29.2)	11.6 (7.5-15.6)	28.3 (24.2-32.4)	21.5 (16.3-26.6	11.4 (8.5-14.4)
Horizon Blue Cross Blue Shield of New Jersey	32.9 (28.2-37.6)	21.5 (16.1-26.8)	10.7 (7.9-13.5)	40.4 (35.5-45.3)	19.8 (15.0-24.5)	10.9 (8.7-13.2)
Nonrural resident	24.1 (21.6-26.6)	19.0 (1.8-22.3)	12.7 (8.2-17.1)	33.9 (31.2-36.6)	17.9 (15.1-20.8	11.8 (8.4-15.3)
Rural resident	22.2 (16.8-27.5)	25.7 (16.4-34.9)	30.8 (19.8-41.7)	26.9 (21.4-32.5)	25.0 (16.3-33.8)	31.7 (22.3-41.2)
Nonrural clinician	21.9 (19.9-23.9)	20.1 (17.2-23.0)	18.6 (14.9-22.3)	29.8 (37.7-52.6)	19.3 (16.7-21.9)	17.8 (14.8-20.8)
Rural clinician	33.2 (25.9-40.5)	22.6 (12.6-32.6)	11.2 (−1.3 to 23.6)	45.1 (37.7-52.6)	20.2 (11.2-29.1)	10.9 (0.3-21.5)

Open in a new tab

^{^a}

Results based on data from a secret shopper survey (5822 calls) of psychiatrists and psychologists listed in commercial network directories for the largest commercial carrier in each of 5 states: Highmark Blue Cross Blue Shield Delaware, Blue Cross and Blue Shield of Montana, Horizon Blue Cross Blue Shield of New Jersey, Blue Cross Blue Shield of Rhode Island, and CareFirst BlueCross BlueShield and Anthem Blue Cross Blue Shield in Virginia. Analyses restricted to 2252 evaluable calls. Full models and CIs are available in eTables 5-7, 13-15 in Supplement 1. Calls distributed proportional to population at the county level within each state.

^{^b}

Includes cases where callers were offered an appointment with either the original clinician they sought to contact or, if unreachable or unavailable, an alternative clinician.

^{^c}

Predicted probabilities based on logit coefficients for binary outcomes.

^{^d}

Predicted means based on ordinary least squares coefficients for continuous outcomes.

Network directory inaccuracies were the dominant barrier to access, affecting 55.6% (95% CI, 53.6%-57.7%) of evaluable calls ( Table 4). Directory error rates were highest for CareFirst in Virginia (70.8%; 95% CI, 56.8%-84.8%) and lowest for Horizon in New Jersey (45.2%; 95% CI, 40.3%-50.2%; P = .001). Psychiatry listings had higher error rates than psychology (57.8% [95% CI, 55.1%-60.5%] vs 52.8% [95% CI, 49.7%-55.9%]; P = .02) and rural clinicians fared better than their nonrural counterparts (44.3% [95% CI, 37.2%-51.3%] vs 57.9% [95% CI, 55.5%-60.3%]; P = .001).

Table 4. Network Directory Errors and Capacity Limitations Among Evaluable Calls^a.

Characteristic	% (95% CI)
Characteristic	Any directory error	Contact information	Specialty information	Network status	Capacity limitation
Total	55.6 (53.6-57.7)	31.0 (29.1-32.9)	22.2 (20.6-23.9)	1.6 (1.0-2.1)	21.6 (19.9-23.3)
Psychiatry	57.8 (55.1-60.5)	30.9 (28.3-33.4)	24.4 (22.1-26.7)	1.3 (0.7-1.9)	18.8 (16.7-21.0)
Psychology	52.8 (49.7-55.9)	31.3 (28.4-34.2)	19.4 (16.9-21.9)	1.9 (1.1-2.8)	25.2 (22.5-27.9)
Anthem (Virginia)	60.4 (56.1-64.8)	33.4 (29.2-37.6)	23.1 (19.4-26.9)	3.1 (1.7-5.0)	15.8 (12.6-19.1)
Blue Cross Blue Shield of Rhode Island	52.6 (47.4-57.8)	27.5 (22.9-32.0)	27.0 (22.4-31.5)	0	22.7 (18.3-27.2)
Blue Cross Blue Shield of Montana	59.9 (54.4-65.3)	27.6 (22.5-32.7)	27.5 (22.4-32.5)	0.7 (0.1-2.0)	20.9 (16.3-25.4)
CareFirst BlueCross BlueShield (Virginia)	70.8 (56.8-84.8)	32.5 (18.4-46.6)	38.2 (23.5-52.9)	0	16.9 (5.5-28.3)
Highmark Blue Cross Blue Shield Delaware	56.9 (52.5-61.3)	33.1 (28.9-37.4)	21.1 (17.4-24.8)	2.1 (1.0-3.8)	24.8 (20.9-28.7)
Horizon Blue Cross Blue Shield of New Jersey	45.2 (40.3-50.2)	32.6 (28.0-37.3)	10.9 (7.8-13.9)	1.7 (0.7-3.5)	24.9 (20.6-29.2)
Nonrural resident	54.3 (51.5-57.1)	30.6 (28.1-33.1)	21.3 (19.1-23.5)	1.7 (1.1-2.4)	20.2 (18.1-22.3)
Rural resident	60.0 (53.6-66.5)	32.6 (26.1-39.1)	25.4 (19.7-31.0)	1.2 (0.4-2.5)	27.0 (20.6-33.4)
Nonrural clinician	57.9 (55.5-60.3)	32.2 (29.9-34.6)	23.6 (21.4-25.7)	1.6 (1.1-2.3)	22.4 (20.3-24.6)
Rural clinician	44.3 (37.2-51.3)	25.3 (19.2-31.3)	17.1 (12.6-21.7)	1.3 (0.4-3.0)	17.9 (12.9-22.9)

Open in a new tab

^{^a}

Analyses restricted to 2252 evaluable calls. Columns show predicted probabilities based on logit coefficients, except network status errors, which are reported as observed percentages because 2 carriers exhibited no such errors and could not be included in a logit model. Directory error subcategories are not mutually exclusive, as staff occasionally reported more than 1 inaccuracy in a single response, and a small number of calls involved other definitively identified errors not shown separately. Errors were assessed in a fixed order (contact information, then specialty, then network status) and calls generally ended once an inaccuracy was identified; rates for specialty and network status are therefore conservative. Full models and CIs are in eTables 4, 8-12 in Supplement 1. Calls distributed proportional to population at the county level within each state.

Contact information errors were the most prevalent type (31.0%; 95% CI, 29.1%-32.9%), including 257 calls (11.4% of evaluable calls) that reached a disconnected or out-of-service number, followed by specialty errors (22.2%; 95% CI, 20.6%-23.9%) and network status errors (1.6%; 95% CI, 1.0%-2.1%). Because errors were assessed in a fixed order and calls generally ended once an inaccuracy was identified, rates for the latter two categories are conservative. Among the 1553 calls at which specialty was assessed, 29.7% (95% CI, 27.5%-31.9%) involved a specialty error. Among the 1092 calls at which network status could be assessed, 3.0% (95% CI, 2.0%-4.0%) involved a network status error. We identified differences between psychology and psychiatry with regard to specialty listing errors (19.4% [95% CI, 16.9%-21.9%] vs 24.4% [95% CI, 22.1%-26.7%]; P = .004) and between rural and nonrural clinicians (17.1% [95% CI, 12.6%-21.7%] vs 23.6% [95% CI, 21.4%-25.7%]; P = .02). Moreover, differences for this type of error ranged from 10.9% (95% CI, 7.8%-13.9%) to 38.2% (95% CI, 23.5%-52.9%) across carriers (P < .001).

Capacity limitations (ie, clinicians not taking in new patients or offering appointments) prevented appointment scheduling in 21.6% (95% CI, 19.9%-23.3%) of evaluable calls, with higher rates for psychology (25.2%; 95% CI, 22.5%-27.9%) than psychiatry (18.8%; 95% CI, 16.7%-21.0%) (P < .001).

Among the 437 callers who secured appointments with a specific date, mean wait time was 20.5 days (95% CI, 18.0-23.0 days), with shorter wait times for psychologists than psychiatrists (17.3 [95% CI, 14.0-20.7] days vs 22.5 [95% CI, 18.9-26.1] days; P = .04). Mean travel time was 17.1 minutes (95% CI, 14.6-19.6 minutes) overall but reached 41.8 minutes (95% CI, 27.5-56.0 minutes) in Montana compared with 10.7 minutes (95% CI, 7.9-13.5 minutes) in New Jersey (P < .001) and 30.8 minutes (95% CI, 19.8-41.7 minutes) for rural residents compared with 12.7 minutes (95% CI, 8.2-17.1 minutes) for nonrural residents (P = .01).

Discussion

Results of this cross-sectional study with secret shopper design of mental health access in commercial insurance plans suggest that the pathway from network directory to actual care may often be broken. For one, more than 60% of calls to directory-listed psychiatrists and psychologists were nonevaluable. Among the calls that were not, fewer than 1 in 4 resulted in an appointment with the listed clinician. Directory inaccuracies affected more than half of evaluable calls, and capacity constraints blocked access in an additional 1 in 5. Taken together, a single call to a directory-listed clinician yielded an appointment with the specific clinician listed 9.1% of the time, expanding to 12.4% if appointments with any available clinician were accepted. While patients who call repeatedly and those who leave messages for callbacks would likely fare better, commercial coverage did not translate into ready access to mental health care.

Our work confirms and expands the literature on mental health access, particularly as related to commercial coverage.^27,28,73 Appointment success rates across this literature have been consistently low.^{26,27,31,33,68} The barriers documented in prior work also parallel our findings. For example, Rhodes et al⁶⁹ reported in 2009 that 45% of calls to mental health clinics reached only an answering machine (compared with 8% for medical clinics) virtually matching our own rate of 44%, suggesting a persistent and mental health-specific pattern rather than a transient consequence of post-pandemic demand. However, our contribution also expands on existing research due to the breadth of outcomes we assess. In combination, our analyses reveal that these barriers compound rather than substitute for one another. Our deliberate focus on the most prominent carriers adds an important benchmark because if the largest, and arguably most reputable,⁷⁴ commercial carrier in each state cannot maintain accurate directories or provide timely access for mental health clinicians, the scope of the problem across smaller and less resourced carriers and other plan types is likely worse.

A central question raised by these findings is to what degree they reflect failures of directory maintenance or characteristics of mental health practices themselves. Calls reaching voicemail, going unanswered, or encountering automated systems were classified as nonevaluable and excluded from all error and appointment rates because their cause is ambiguous. However, among the errors we do document, many are not. For example, 257 calls (11.4% of evaluable calls) reached a disconnected or out-of-service number. Such listings cannot reflect practice staffing or call-handling and detecting them is precisely what network regulations such as the No Surprises Act 90-day verification requirement were meant to accomplish. At the same time, the high nonevaluable rate likely does reflect a structural feature of behavioral health care delivery. Many mental health clinicians operate solo or small-group practices without dedicated front desk staff, making them systematically harder to reach.⁷⁵ Both patterns point the same direction: directory accuracy standards developed for medical specialties may be ill-suited for mental health.

The implications of our findings extend beyond directory accuracy to the broader question of what constitutes a functioning network. Federal and state regulators have relied heavily on clinician-to-enrollee ratios and time-and-distance standards to assess network adequacy.^50,61 These metrics count clinicians who appear in directories regardless of whether they can be reached, are accepting patients, or are practicing the listed specialty. The findings here demonstrate that such nominal measures may dramatically overstate the actual availability of mental health care. A directory listing a psychiatrist who cannot be reached by phone, no longer practices psychiatry, or has no available appointments provides no meaningful access. Yet that listing may be used to satisfy regulatory requirements.^16,61 Travel burden warrants attention because time-and-distance standards are expressed in the units we measure. Since callers contacted the geographically closest listed clinician and travel was recorded only where an appointment was obtained, these estimates approximate realized access from a patient perspective. That mean travel nonetheless reached 41.8 minutes in Montana and 30.8 minutes for rural residents suggests standards satisfied on paper may not be satisfied in practice. Telehealth may mitigate some of this burden, but directories list physical addresses, and initial evaluations and controlled substance management often require in-person care. Various burdens may limit the utility of such care.⁷⁶

The variation across states and carriers, while present, was less striking than the consistency of the overall pattern. Appointment success ranged from 18.0% to 32.9% across carriers, and directory error rates ranged from 45.2% to 70.8%. However, even the best-performing carrier in the study, Horizon Blue Cross Blue Shield of New Jersey, failed to provide appointments for two-thirds of evaluable calls. This consistency suggests that the barriers documented are systemic and pervasive. Moreover, although rural clinicians were more likely to offer appointments than nonrural clinicians (potentially reflecting lower panel demand), rural residents faced travel times more than double those of nonrural residents, and the nonevaluable rate suggests that reaching these clinicians remains a substantial barrier.

The distinction between psychiatry and psychology also merits attention. Psychologists were harder to reach than psychiatrists (33.3% evaluable rate vs 44.0% for psychiatrists) and faced higher capacity constraints (25.2% vs 18.8%), while psychiatrist listings had higher directory error rates (57.8% vs 52.8%). These patterns likely reflect different practice structures, as psychologists more often operate in solo practice settings with limited administrative infrastructure. More broadly, this finding suggests that access interventions may need to be tailored to clinician type.

These findings also carry implications for clinicians on both sides of the referral pathway. For one, directory inaccuracies mean that patients who need and are actively seeking care may not find it. For primary care and other referring clinicians, who serve as the principal gateway to mental health services for most patients, these findings are particularly consequential. The breakdown in the referral pathway described by our findings may contribute to untreated mental illness, crisis care utilization, and the well-documented burden on primary care clinicians who must manage psychiatric conditions beyond their training when specialist referrals fail.^26,33 This may also mean that mental health clinicians may experience more severe cases when patients eventually connect. At the same time, our findings offer little reason to think this leaves mental health practices short of patients as capacity limitations in 21.6% of evaluable calls suggest the opposite. Ultimately, much of the consequences may fall on individual patients, who are turned away from clinicians who would have seen them, and on listed clinicians, who may bear the reputational consequences of directories they do not control.^49,50

Lastly, these findings also carry several policy implications. First, the persistence of directory inaccuracies more than 3 years after the No Surprises Act directory accuracy provisions took effect suggests that existing regulatory requirements are insufficient or lack meaningful enforcement mechanisms.^36,49,50 Regulators should consider shifting from self-reported compliance to audit-based verification, incorporating claims-based measures of network engagement, and tying directory accuracy to plan certification and quality ratings. Second, network adequacy standards should be grounded in consumer experience rather than clinician counts alone. Appointment availability, wait times, and the ability to reach a clinician by telephone should be incorporated into adequacy assessments.^50,61 Third, the mental health clinician workforce and practice structure present unique challenges that current regulatory frameworks may not adequately address. Standards designed around medical practice groups with front desk staff and appointment scheduling systems may be structurally inappropriate for behavioral health clinicians. Regulators and carriers should explore alternative verification and access mechanisms, including integration of scheduling platforms and real-time availability indicators in network directories.

Limitations

Several limitations merit acknowledgment. First, the study surveyed the largest commercial carrier in each state. Because Blue Cross Blue Shield (BCBS) affiliates hold the largest commercial market share in each state studied, all surveyed carriers are BCBS licensees operated by six distinct parent organizations (3 multistate insurers and 3 single-state or regional plans). We are not aware of evidence that BCBS affiliates maintain systematically more or less accurate directories than other large commercial carriers. Because the largest carrier typically has the broadest network and greatest administrative capacity, these results likely represent a conservative estimate of access barriers in the commercial market. Our findings may nonetheless not generalize to smaller carriers or other plan types. Second, the study was conducted across 5 states, which may limit generalizability to markets with different regulatory environments or clinician supply conditions. Third, the high nonevaluable rate, while itself an important finding, means that primary analyses are restricted to a limited number of calls, and outcomes among nonevaluable calls are unknown. Fourth, the cross-sectional design captures directory accuracy and appointment availability at a single point in time. Fifth, because calls generally ended once an inaccuracy was identified, the unconditional error rates we report are conservative; we therefore also report rates conditional on each element having been assessed. Moreover, callers made a single attempt per listing and did not leave messages. The resulting yield is therefore a lower bound on what a persistent patient might achieve, and we could not assess whether practices returned calls. Finally, callers contacted clinicians on behalf of a simulated patient, and some offices declined to provide information to third-party callers, which may have contributed to the nonevaluable rate.

Conclusions

The findings of this study suggest that even for commercially insured US residents seeking mental health care, network directories frequently do not lead to care. The gap between nominal network inclusion and realized access is not confined to public insurance programs but extends to the commercial market that has long been assumed to offer superior access. Closing this gap across markets may require enforceable directory accuracy standards, network adequacy metrics grounded in consumer experience, and a recognition that the structural features of mental health care delivery demand regulatory approaches distinct from those applied to medical specialties. Given the evolving nature of US insurance markets, future research should examine whether these patterns extend to other commercial carriers, states, and plan types and should assess the impact of emerging regulatory interventions in access to mental health care.

Supplement 1.

eTable 1. Carriers Included in Secret Shopper Survey, by State

eTable 2. Outcome Definitions and Call Disposition Categories

eTable 3. Excerpt from Standardized Call Script: Directory Verification Sequence

eTable 4. Adjusted Regression Results, Evaluable Calls Analysis: Logit Coefficients (Binary Outcomes) and OLS Coefficients (Continuous Outcomes), Estimated Using Robust Standard Errors

eTable 5. Call Success Rates, Predicted Probability

eTable 6. Wait Time to Appointment in Days, Predicted Means

eTable 7. Travel Time to Appointment in Minutes, Predicted Means

eTable 8. Percentage of Calls with Any Provider Directory Errors, Predicted Probabilities

eTable 9. Percentage of Calls with Errors Related to Clinician Contact Information, Predicted Probabilities

eTable 10. Percentage of Calls with Errors Related to Clinician Specialty Information, Predicted Probabilities

eTable 11. Percentage of Calls with Errors Related to Clinician Network Status, Predicted Probabilities

eTable 12. Percentage of Calls with Capacity Limitations, Predicted Probabilities

eTable 13. Call Success Rates, Predicted Probability, Including Alternative Clinicians

eTable 14. Wait Time to Appointment in Days, Predicted Means, Including Alternative Clinicians

eTable 15. Travel Time to Appointment in Minutes, Predicted Means, Including Alternative Clinicians

eFigure 1. Home Addresses and Addresses for Contacted Clinicians, Virginia

eFigure 2. Home Addresses and Addresses for Contacted Clinicians, Rhode Island

eFigure 3. Home Addresses and Addresses for Contacted Clinicians, New Jersey

eFigure 4. Home Addresses and Addresses for Contacted Clinicians, Montana

eFigure 5. Home Addresses and Addresses for Contacted Clinicians, Delaware

jamahealthforum-e263746-s001.pdf^{(1.2MB, pdf)}

Supplement 2.

Data Sharing Statement

jamahealthforum-e263746-s002.pdf^{(14.9KB, pdf)}

References

1.Barry CL, Venkatesh M, Busch SH. Assessing new patient access to mental health providers in HMO networks. Psychiatr Serv. 2008;59(12):1413-1418. doi: 10.1176/ps.2008.59.12.1413 [DOI] [PubMed] [Google Scholar]
2.Stefl ME, Prosperi DC. Barriers to mental health service utilization. Community Ment Health J. 1985;21(3):167-178. doi: 10.1007/BF00754732 [DOI] [PubMed] [Google Scholar]
3.Knaak S, Mantler E, Szeto A. Mental illness-related stigma in healthcare: barriers to access and care and evidence-based solutions. Healthc Manage Forum. 2017;30(2):111-116. doi: 10.1177/0840470416679413 [DOI] [PMC free article] [PubMed] [Google Scholar]
4.Ahrnsbrak R, Bose J, Hedden SL, Lipari RN, Park-Lee E. Key Substance Use and Mental Health Indicators in the United States: Results from the 2016 National Survey on Drug Use and Health. Substance Abuse and Mental Health Services Administration; 2017. [PubMed] [Google Scholar]
5.Cook BL, Trinh NH, Li Z, Hou SSY, Progovac AM. Trends in racial-ethnic disparities in access to mental health care, 2004-2012. Psychiatr Serv. 2017;68(1):9-16. doi: 10.1176/appi.ps.201500453 [DOI] [PMC free article] [PubMed] [Google Scholar]
6.Zhu JM, Huntington A, Haeder S, Wolk C, McConnell KJ. Insurance acceptance and cash pay rates for psychotherapy in the US. Health Aff Sch. 2024;2(9):qxae110. doi: 10.1093/haschl/qxae110 [DOI] [PMC free article] [PubMed] [Google Scholar]
7.Burman A, Haeder SF, Xu WY. Provider directory inaccuracy and timely access for mental health care. Am J Manag Care. 2023;29(2):96-102. doi: 10.37765/ajmc.2023.89318 [DOI] [PubMed] [Google Scholar]
8.Berkowitz SA, Basu S. Unmet social needs and worse mental health after expiration of COVID-19 federal pandemic unemployment compensation. Health Aff (Millwood). 2021;40(3):426-434. doi: 10.1377/hlthaff.2020.01990 [DOI] [PMC free article] [PubMed] [Google Scholar]
9.Patel SY, Mehrotra A, Huskamp HA, Uscher-Pines L, Ganguli I, Barnett ML. Variation in telemedicine use and outpatient care during the COVID-19 pandemic in the United States. Health Aff (Millwood). 2021;40(2):349-358. doi: 10.1377/hlthaff.2020.01786 [DOI] [PMC free article] [PubMed] [Google Scholar]
10.Anderson S, Haeder S, Caseman K, Mathur A, Ulmen K. When adolescents are in school during COVID-19, coordination between school-based health centers and education is key. J Adolesc Health. 2020;67(6):745-746. doi: 10.1016/j.jadohealth.2020.09.005 [DOI] [PMC free article] [PubMed] [Google Scholar]
11.Burman A, Haeder SF. Potemkin protections: assessing provider directory accuracy and timely access for four specialties in California. J Health Polit Policy Law. 2022;47(3):319-349. doi: 10.1215/03616878-9626866 [DOI] [PubMed] [Google Scholar]
12.Polsky D, Richards M, Basseyn S, et al. Appointment availability after increases in Medicaid payments for primary care. N Engl J Med. 2015;372(6):537-545. doi: 10.1056/NEJMsa1413299 [DOI] [PubMed] [Google Scholar]
13.Rhodes KV, Kenney GM, Friedman AB, et al. Primary care access for new patients on the eve of health care reform. JAMA Intern Med. 2014;174(6):861-869. doi: 10.1001/jamainternmed.2014.20 [DOI] [PubMed] [Google Scholar]
14.Tipirneni R, Rhodes KV, Hayward RA, Lichtenstein RL, Reamer EN, Davis MM. Primary care appointment availability for new Medicaid patients increased after Medicaid expansion in Michigan. Health Aff (Millwood). 2015;34(8):1399-1406. doi: 10.1377/hlthaff.2014.1425 [DOI] [PubMed] [Google Scholar]
15.Elton TI IV, Xu WY, Haeder SF. Provider directory inaccuracy and timely access to physical therapy. World Med Health Policy. 2024;16(3):447-459. doi: 10.1002/wmh3.607 [DOI] [Google Scholar]
16.Zhu JM, Rowland R, Polsky D, et al. Medicaid managed care organizations’ experiences with network adequacy. Health Aff Sch. 2025;3(4):qxaf049. doi: 10.1093/haschl/qxaf049 [DOI] [PMC free article] [PubMed] [Google Scholar]
17.Zhu JM, Rumalla KC, Polsky D. New opportunities to strengthen Medicaid managed care network adequacy standards. JAMA Health Forum. 2023;4(10):e233194-e233194. doi: 10.1001/jamahealthforum.2023.3194 [DOI] [PMC free article] [PubMed] [Google Scholar]
18.Haeder SF, Weimer D, Mukamel DB. A consumer-centric approach to network adequacy: access to four specialties in California’s marketplace. Health Aff (Millwood). 2019;38(11):1918-1926. doi: 10.1377/hlthaff.2019.00116 [DOI] [PubMed] [Google Scholar]
19.Haeder SF, Weimer DL, Mukamel DB. Integrating travel distance into assessments of provider networks using a dyadic approach: the case of California’s Affordable Care Marketplace. SAGE Research Methods Cases: Medicine and Health. Published online January 15, 2020. doi: 10.4135/9781529723021 [DOI] [Google Scholar]
20.Haeder SF. A tale of two programs: access to high quality providers for Medicare Advantage and Affordable Care Act beneficiaries in New York State. World Med Health Policy. 2019;11(3):212-230. doi: 10.1002/wmh3.309 [DOI] [Google Scholar]
21.Bishop TF, Press MJ, Keyhani S, Pincus HA. Acceptance of insurance by psychiatrists and the implications for access to mental health care. JAMA Psychiatry. 2014;71(2):176-181. doi: 10.1001/jamapsychiatry.2013.2862 [DOI] [PMC free article] [PubMed] [Google Scholar]
22.Melek S, Davenport S, Gray TJ. Addiction and Mental Health vs. Physical Health: Widening Disparities in Network Use and Provider Reimbursement. Milliman; 2019. [Google Scholar]
23.Xu WY, Song C, Li Y, Retchin SM. Cost-sharing disparities for out-of-network care for adults with behavioral health conditions. JAMA Netw Open. 2019;2(11):e1914554-e1914554. doi: 10.1001/jamanetworkopen.2019.14554 [DOI] [PMC free article] [PubMed] [Google Scholar]
24.Busch SH, Kyanko KA. Incorrect provider directories associated with out-of-network mental health care and outpatient surprise bills. Health Aff (Millwood). 2020;39(6):975-983. doi: 10.1377/hlthaff.2019.01501 [DOI] [PMC free article] [PubMed] [Google Scholar]
25.Benson NM, Myong C, Newhouse JP, Fung V, Hsu J. Psychiatrist participation in private health insurance markets: paucity in the land of plenty. Psychiatr Serv. 2020;71(12):1232-1238. doi: 10.1176/appi.ps.202000022 [DOI] [PMC free article] [PubMed] [Google Scholar]
26.Sun CF, Correll CU, Trestman RL, et al. Low availability, long wait times, and high geographic disparity of psychiatric outpatient care in the US. Gen Hosp Psychiatry. 2023;84:12-17. doi: 10.1016/j.genhosppsych.2023.05.012 [DOI] [PubMed] [Google Scholar]
27.Tenner NL, Reddy M, Block AE. Secret shopper analysis shows getting psychiatry appointment in New York City is well kept secret. Community Ment Health J. 2023;59(2):290-293. doi: 10.1007/s10597-022-01006-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
28.Haeder SF, Weimer DL, Mukamel DB. Secret shoppers find access to providers and network accuracy lacking for those in marketplace and commercial plans. Health Aff (Millwood). 2016;35(7):1160-1166. doi: 10.1377/hlthaff.2015.1554 [DOI] [PubMed] [Google Scholar]
29.Holstein R, Paul DPI III. ‘Phantom networks’ of managed behavioral health providers: an empirical study of their existence and effect on patients in two New Jersey counties. Hosp Top. 2012;90(3):65-73. doi: 10.1080/00185868.2012.714689 [DOI] [PubMed] [Google Scholar]
30.Holstein RM, Paul DPI III. Access to behavioral health care services in New Jersey. Hosp Top. 2017;95(3):51-56. doi: 10.1080/00185868.2017.1300481 [DOI] [PubMed] [Google Scholar]
31.Blech B, West JC, Yang Z, Barber KD, Wang P, Coyle C. Availability of network psychiatrists among the largest health insurance carriers in Washington, D.C. Psychiatr Serv. 2017;68(9):962-965. doi: 10.1176/appi.ps.201600454 [DOI] [PubMed] [Google Scholar]
32.Cama S, Malowney M, Smith AJB, et al. Availability of outpatient mental health care by pediatricians and child psychiatrists in five U.S. cities. Int J Health Serv. 2017;47(4):621-635. doi: 10.1177/0020731417707492 [DOI] [PubMed] [Google Scholar]
33.Malowney M, Keltz S, Fischer D, Boyd JW. Availability of outpatient care from psychiatrists: a simulated-patient study in three U.S. cities. Psychiatr Serv. 2015;66(1):94-96. doi: 10.1176/appi.ps.201400051 [DOI] [PubMed] [Google Scholar]
34.Haeder S, Wapner A. Provider directory accuracy and access to pediatric primary care in Pennsylvania’s Affordable Care Act Marketplace: a simulated patient study. Acad Pediatr. 2026;26(6):103356. doi: 10.1016/j.acap.2026.103356 [DOI] [PubMed] [Google Scholar]
35.Zhu JM, Charlesworth CJ, Polsky D, McConnell KJ. Phantom networks: discrepancies between reported and realized mental health care access in Oregon Medicaid. Health Aff (Millwood). 2022;41(7):1013-1022. doi: 10.1377/hlthaff.2022.00052 [DOI] [PMC free article] [PubMed] [Google Scholar]
36.Haeder SF, Zhu JM. ACA network regulatory filings are inaccurate, poorly match provider directories. Am J Manag Care. 2025;31(9):e265-e269. doi: 10.37765/ajmc.2025.89791 [DOI] [PMC free article] [PubMed] [Google Scholar]
37.Xu WY, Raver EW, Elton T IV, Davis M, Haeder SF. Disconnects between provider network directories and patient preferences. Am J Manag Care. 2024;30(12):660-666. doi: 10.37765/ajmc.2024.89638 [DOI] [PubMed] [Google Scholar]
38.Tipirneni R, Rhodes KV, Hayward RA, et al. Primary care appointment availability and nonphysician providers one year after Medicaid expansion. Am J Manag Care. 2016;22(6):427-431. [PubMed] [Google Scholar]
39.Melnikow J, Evans E, Xing G, et al. Primary care access to new patient appointments for California Medicaid enrollees: a simulated patient study. Ann Fam Med. 2020;18(3):210-217. doi: 10.1370/afm.2502 [DOI] [PMC free article] [PubMed] [Google Scholar]
40.Burman A, Haeder SF. Directory accuracy and timely access in Maryland’s Medicaid managed care program. J Health Care Poor Underserved. 2022;33(2):597-611. doi: 10.1353/hpu.2022.0050 [DOI] [PubMed] [Google Scholar]
41.Blumenberg E, Agrawal AW. Getting around when you’re just getting by: transportation Survival Strategies of the Poor. J Poverty. 2014;18:355-378. doi: 10.1080/10875549.2014.951905 [DOI] [Google Scholar]
42.Burman A, Haeder SF. Provider directory accuracy and timely access to mammograms in California. Women Health. 2022;62(5):421-429. doi: 10.1080/03630242.2022.2083284 [DOI] [PubMed] [Google Scholar]
43.Hamlyn GS, Hutchins KE, Johnston AL, Thomas RT, Tian J, Kamal AH. Accessibility and barriers to oncology appointments at 40 National Cancer Institute-designated comprehensive cancer centers: results of a mystery shopper project. J Oncol Pract. 2016;12(10):e884-e900. doi: 10.1200/JOP.2016.014159 [DOI] [PubMed] [Google Scholar]
44.Tsang MW, Resneck JS Jr. Even patients with changing moles face long dermatology appointment wait-times: a study of simulated patient calls to dermatologists. J Am Acad Dermatol. 2006;55(1):54-58. doi: 10.1016/j.jaad.2006.04.001 [DOI] [PubMed] [Google Scholar]
45.Resneck JS Jr, Quiggle A, Liu M, Brewster DW. The accuracy of dermatology network physician directories posted by Medicare Advantage health plans in an era of narrow networks. JAMA Dermatol. 2014;150(12):1290-1297. doi: 10.1001/jamadermatol.2014.3902 [DOI] [PubMed] [Google Scholar]
46.Lim SE, Krajewski CM. A mystery shopper study identifying practice-level barriers to adolescent IUD access in western Pennsylvania. Contraception. 2020;101(2):130-131. doi: 10.1016/j.contraception.2019.10.011 [DOI] [PubMed] [Google Scholar]
47.Haeder SF, Xu WY. Consumer experiences navigating health care provider directories and support of federal policy action. World Med Health Policy. 2024;16(4):577-599. doi: 10.1002/wmh3.625 [DOI] [Google Scholar]
48.Haeder SF, Xu WY. What methods do patients use to schedule medical appointments? Health Aff Sch. 2025;3(4):qxaf077. doi: 10.1093/haschl/qxaf077 [DOI] [PMC free article] [PubMed] [Google Scholar]
49.Burman A, Haeder SF. Without A Dedicated Enforcement Mechanism, New Federal Protections Are Unlikely To Improve Provider Directory Accuracy. Health Affairs Forefront. November 5, 2021. [Google Scholar]
50.Haeder SF, Weimer DL, Mukamel DB. A knotty problem: consumer access and the regulation of provider networks. J Health Polit Policy Law. 2019;44(6):937-954. doi: 10.1215/03616878-7785835 [DOI] [PubMed] [Google Scholar]
51.Haeder SF, Xu WY. When is a network adequate? consumer perspectives on network adequacy definitions. Am J Manag Care. 2024;30(9):424-432. doi: 10.37765/ajmc.2024.89601 [DOI] [PubMed] [Google Scholar]
52.Haeder SF, Weimer DL, Mukamel DB. Going the extra mile? how provider network design increases consumer travel distance, particularly for rural consumers. J Health Polit Policy Law. 2020;45(6):1107-1136. doi: 10.1215/03616878-8641591 [DOI] [PubMed] [Google Scholar]
53.Kim J, Norton EC, Stearns SC. Transportation brokerage services and Medicaid beneficiaries’ access to care. Health Serv Res. 2009;44(1):145-161. doi: 10.1111/j.1475-6773.2008.00907.x [DOI] [PMC free article] [PubMed] [Google Scholar]
54.Giovannelli J, Lucia KW, Corlette S. Implementing the Affordable Care Act: State Regulation of Marketplace Plan Provider Networks. The Commonwealth Fund. May 5, 2015.Accessed May 1, 2026. https://www.commonwealthfund.org/publications/issue-briefs/2015/may/implementing-affordable-care-act-state-regulation-marketplace [PubMed]
55.Giovannelli J, Lucia K, Corlette S. Regulation of health plan provider networks. Health Affairs. July 28, 2016.Accessed May 1, 2026. https://www.healthaffairs.org/cms/10.1377/hpb20160728.898461/asset/media/healthpolicybrief_160.pdf [Google Scholar]
56.Hall MA, Ginsburg PB. A better approach regulating provider adequacy. Brookings. September 14, 2017. Accessed May 1, 2026. https://www.brookings.edu/articles/a-better-approach-to-regulating-provider-network-adequacy/
57.Holahan J, Zuckerman S, Evans A, Rangarajan S. Medicaid managed care in thirteen states. Health Aff (Millwood). 1998;17(3):43-63. doi: 10.1377/hlthaff.17.3.43 [DOI] [PubMed] [Google Scholar]
58.Brown VY, Hartung BR. Managed care at the crossroads: can managed care organizations survive government regulation? Ann Health Law. 1998;7:25-72. [PubMed] [Google Scholar]
59.Miller TE. Managed care regulation: in the laboratory of the states. JAMA. 1997;278(13):1102-1109. doi: 10.1001/jama.1997.03550130076041 [DOI] [PubMed] [Google Scholar]
60.Hall MA, Brandt C. Network Adequacy Under The Trump Administration. Brookings. September 14, 2017. Accessed May 1, 2026. https://www.brookings.edu/articles/network-adequacy-under-the-trump-administration/
61.Haeder SF, Weimer DL, Mukamel DB. Mixed signals: the inadequacy of provider-per-enrollee ratios for assessing network adequacy in California (and elsewhere). World Med Health Policy. 2023;15(3):258-272. doi: 10.1002/wmh3.466 [DOI] [Google Scholar]
62.Haeder SF, Xu WY, Elton T IV, Pitcher A. State efforts to regulate provider networks and directories: lessons for the future. J Health Polit Policy Law. 2023;48(6):951-968. doi: 10.1215/03616878-10852610 [DOI] [PubMed] [Google Scholar]
63.Wishner JB, Marks J. Ensuring compliance with network adequacy standards: lessons from four states. Urban Institute. March 16, 2017. Accessed May 1, 2026. https://www.urban.org/research/publication/ensuring-compliance-network-adequacy-standards-lessons-four-states
64.Haeder SF, Zhu JM. Persistence of provider directory inaccuracies after the No Surprises Act. Am J Manag Care. 2024;30(11):584-588. doi: 10.37765/ajmc.2024.89627 [DOI] [PMC free article] [PubMed] [Google Scholar]
65.Haeder SF, Zhu JM. Inaccuracies in provider directories persist for long periods of time. Health Aff Sch. 2024;2(6):qxae079. doi: 10.1093/haschl/qxae079 [DOI] [PMC free article] [PubMed] [Google Scholar]
66.Olin SC, O’Connor BC, Storfer-Isser A, et al. Access to care for youth in a state mental health system: a simulated patient approach. J Am Acad Child Adolesc Psychiatry. 2016;55(5):392-399. doi: 10.1016/j.jaac.2016.02.014 [DOI] [PMC free article] [PubMed] [Google Scholar]
67.Chen V, Barrera CA, Golden K, Perez-Stable C, Carreon RL, Boyd JW. Availability of outpatient child psychiatric care during COVID-19: a simulated-patient study in three U.S. cities. Psychiatr Serv. 2023;74(12):1277-1280. doi: 10.1176/appi.ps.20220623 [DOI] [PubMed] [Google Scholar]
68.Brahmbhatt D, Schpero WL. Access to psychiatric appointments for Medicaid enrollees in 4 large US cities. JAMA. 2024;332(8):668-669. doi: 10.1001/jama.2024.13074 [DOI] [PMC free article] [PubMed] [Google Scholar]
69.Rhodes KV, Vieth TL, Kushner H, Levy H, Asplin BR. Referral without access: for psychiatric services, wait for the beep. Ann Emerg Med. 2009;54(2):272-278. doi: 10.1016/j.annemergmed.2008.08.023 [DOI] [PubMed] [Google Scholar]
70.KFF . Market share and enrollment of largest three insurers- large group market. 2024. Accessed May 1, 2026. https://www.kff.org/state-health-policy-data/state-indicator/market-share-and-enrollment-of-largest-three-insurers-large-group-market/?currentTimeframe=0&sortModel=%7B%22colId%22:%22Location%22,%22sort%22:%22asc%22%7D
71.Haeder SF, Xu WY. Inaccuracies related to accepting new patient flags in Affordable Care Act plans’ provider directories. Health Aff Sch. 2025;3(5):qxaf082. doi: 10.1093/haschl/qxaf082 [DOI] [PMC free article] [PubMed] [Google Scholar]
72.Haeder SF, Lou Y, Xu W. Assessing access to cardiologists and endocrinologists in the Texas ACA market. Health Aff Sch. 2026;4(4):qxag068. doi: 10.1093/haschl/qxag068 [DOI] [PMC free article] [PubMed] [Google Scholar]
73.Williams MO, Gilroy JR, Chang TY, Seymour DJ. Challenges for insured patients in accessing behavioral health care. Ann Fam Med. 2017;15(4):363-365. doi: 10.1370/afm.2092 [DOI] [PMC free article] [PubMed] [Google Scholar]
74.Cunningham RI, Cunningham RMJ. The Blues: A History of the Blue Cross and Blue Shield System. Northern Illinois University Press; 1997. [Google Scholar]
75.Kuntz L. New report highlights independent mental health practice, access to care, and evolving role of AI. Psychiatr Times. May 13, 2026. Accessed May 1, 2026. https://www.psychiatrictimes.com/view/new-report-highlights-independent-mental-health-practice-access-to-care-and-evolving-role-of-ai
76.Olfson M, McClellan C, Zuvekas SH, Blanco C. Telemental Health, Hybrid, and In-Person Outpatient Mental Health Care in the US. JAMA Psychiatry. 2026;83(1):59-67. doi: 10.1001/jamapsychiatry.2025.3575 [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials