Assessment of Perceptions of Mental Health vs Medical Health Plan Networks Among US Adults With Private Insurance

This survey study assesses how patients with private insurance perceive the adequacy of their health plan’s mental health and medical provider networks, whether practitioners frequently leave plans, and whether patients’ choice of plan depends on participation by a specific practitioner.


Introduction
Specialty mental health practitioners are more likely to opt out of participation in private insurance provider networks (provider networks include physicians, clinicians, other health care professionals, and their institutions that constitute the network), compared with practitioners in other specialties because of a combination of workforce shortages, low reimbursement compared with other specialties, and high demand for services. [1][2][3] One study estimated that 35% of psychiatrists do not participate in managed care networks compared with 8% to 12% of other specialists. 4 This has resulted in substantially higher rates for out-of-network mental health care compared with general medical or surgical care not related to mental health (hereinafter referred to as medical care), as well as concomitant higher out-of-pocket health care spending. 1,5,6 Implications for access to care for individuals with mental health disorders may be substantial.
Network adequacy has been recognized as a necessary component of a high-quality plan. 7,8 If there are too few practitioners in the network, enrollees may have difficulty locating an in-network practitioner who is accepting new patients within an acceptable time frame. Most states (29) have adopted at least 1 required measure of network adequacy for private health insurance plans, such as geographic distance to practitioners, time to appointment for new patients, or practitioner to enrollee ratios. 9 However, there is a lack of consensus on how network adequacy should be measured and regulated. Even with mental health provider network requirements in place, state oversight of these regulations can be inconsistent, and compliance is often difficult to measure. [10][11][12] Limited practitioner participation also has implications for continuity of care. Patients often have ongoing relationships with practitioners. When these practitioners no longer participate in the network either because of voluntary exit or termination by the plan, patient treatment plans may be interrupted. This may be particularly problematic for patients receiving mental health care if treatment plans are of longer duration or if practitioners need time to gain patient trust for patients to be more likely to reveal relevant sensitive information, increasing the value of continuity. 13 Patients' responses to a practitioner's exit or termination from a plan may include continuing to see the same practitioner (often at higher out-of-pocket cost and less often); switching practitioners, which requires a new relationship be developed; or stopping treatment, perhaps because of frustration.
Plan choice may also be affected by limited or inadequate plan networks. Before choosing a plan, individuals may attempt to identify whether a specific practitioner is in the network. This may lead to risk selection, or a concentration of patients with the most serious conditions in a single plan.
For example, plans that cover the most expensive, or star, hospitals are often chosen by the patients with the most serious illnesses or those who more frequently use these high-priced hospitals. 14 Thus, including these hospitals can lead the cost of premiums to spiral upward, with only individuals who have the highest health care costs ultimately choosing the plan. In the case of mental health, insurance companies may attempt to avoid coverage of certain patients who require high-cost treatment by skimping on certain services or distorting network choices to avoid inclusion of certain patients (ie, excluding practitioners who provide high-cost care). [15][16][17] Both of these concerns are particularly worrisome if patients actively consider the inclusion of specific practitioners when they choose plans. Delivery systems have responded to mental health workforce shortages through adoption of team-based models that effectively treat conditions such as anxiety and depression through primary care practices, 18 although in real-world settings, implementation of and fidelity to these models may be challenging because of lack of financial resources, technical guidance, and staff. [19][20][21] In practice, primary care practitioners may address deficiencies in provider networks in several ways: they may prescribe medication for patients with mild to moderate symptoms while working collaboratively with a therapist who provides counseling, they serve as a trusted referral source and provide initial treatment while helping patients locate a specialist, or they may prescribe medication out of necessity if a patient is unable to find a specialist. Few studies have assessed whether treatment of mental health disorders by primary care practitioners has helped address mental health workforce supply issues, particularly from the perspective of patient experiences with plan provider networks.

JAMA Network Open | Health Policy
Our objective was to compare patient perceptions and experiences with mental health and medical provider networks. We conducted a survey study of patients in the US who were receiving both types of care in the past year. We assessed patient ratings of the inadequacy of the provider network, whether in the past 3 years a treating practitioner left their network and the resulting responses, and whether practitioner participation in the network affected plan choice.

Survey Development
Data were obtained from a de novo 2018 national internet survey of adults with private insurance on their experiences with access to outpatient mental health services, conducted from August to September 2018. Data analysis was performed from November 12, 2020, to May 12, 2021. Survey methods have been described previously. 22 We developed the survey, which was then tested using A series of screener questions was used to identify this study sample of US adults aged 18 to 64 years who were enrolled in private insurance with a provider network and who were treated by both an outpatient mental health practitioner and an outpatient medical practitioner in the preceding 12 months. We selected participants receiving treatment from both practitioner types for these analyses to allow for within-person comparisons in patient experiences with their mental health and medical provider networks. Mental health practitioners were defined as "professionals specifically trained to diagnose and treat emotional or mental health problems, including psychiatrists, therapists, psychologists, mental health nurse practitioners, and social workers." Medical practitioners were defined as "doctors, nurse practitioners, and physician assistants." We explored patient experiences with network composition through a series of questions on network adequacy, network continuity, and plan choice. When assessing network adequacy, we also considered whether results differed for individuals receiving mental health care from a primary care practitioner in addition to a specialty practitioner, those with and without serious psychological distress, those with and without self-reported fair or poor health status, and those receiving and not receiving any out-of-network mental health treatment in the past year. Serious psychological distress was assessed as a score of 13 or higher on the Kessler 6-Item Psychological Distress Scale. 24 First, we assessed perception of mental health network adequacy by asking participants to rate, on a 5-point Likert scale-from strongly agree to strongly disagree-the statement, "My insurer has done a good job of making enough in-network mental health providers available." We specifically noted to only include experiences with mental health practitioners or networks. We asked the same question, with medical substituted for mental health. We coded ratings of inadequate or insufficient networks by creating a binary variable equal to 1 if the participant chose disagree or strongly disagree.

JAMA Network Open | Health Policy
Next, to assess network continuity, we asked whether the participant had a practitioner leave their network in the last 3 years and, if so, how it affected treatment (eg, the participant stopped treatment, switched practitioners, or continued to see the same practitioner). We did not distinguish whether practitioners were terminated from a network or exited voluntarily. We asked this question for 3 practitioner types (ie, mental health, specialist, and primary care), and participants could answer yes for more than 1 practitioner type. Specialists were defined as "a provider who specializes in a particular medical field. For example, dermatologists specialize in skin disorders; cardiologists specialize in problems of the heart." Finally, we assessed whether networks affected plan choice by asking participants if they looked up a practitioner or provider before choosing a health insurance plan, and, if the answer was yes, the practitioner or provider type (ie, mental health, specialist, or primary care practitioner or hospital).
Participants could choose more than 1 practitioner or provider type. Those who looked up a practitioner or provider were then asked whether it affected their choice of plan.
Demographic information had previously been collected by KnowledgePanel. This included Health Insurance Marketplace plan participation and self-reported health status. Participants selected their race and ethnicity from a list of 5 options defined by KnowledgePanel: Black Non-Hispanic; Hispanic; White Non-Hispanic; Other, Non-Hispanic; or 2 or more races Non-Hispanic.
Because of the small sample sizes we combined the Black Non-Hispanic, Other Non-Hispanic, and 2 or more races Non-Hispanic categories. Questions used for these analyses are in the eMethods in the Supplement.

Survey Administration
We recruited participants from KnowledgePanel, a validated panel of approximately 50 000 households constructed through high-quality probability-based sampling (eFigure in the Supplement). 25 Among the 29 854 sampled panelists, 19 602 (excluding breakoffs) completed the screening questionnaire, resulting in a survey completion rate of 66% based on the AAPOR standard definition for probability-based internet panels. 26 All reported analyses were weighted to match participants to the US population based on current population survey data in terms of sex, age, race and ethnicity, educational level, census region, household income, home ownership, and metropolitan area. Weights provided by KnowledgePanel were also adjusted for panel recruitment, attrition, oversampling, and survey nonresponse.

Statistical Analysis
Conditional (fixed-effects) logistic models with statistical significance set to P < .05 (2-sided) were used to examine relevant within-person associations. All analyses used Stata, version 16.1

Results
Of

Network Adequacy
The final sample for the network adequacy analysis omitted individuals who responded "don't know" or refused to answer, leaving 615 participants for the analysis. Overall, participants were more likely to rate their plan's mental health provider network as not adequate compared with their plan's medical provider network (Figure 1)

Network Discontinuity and Consequences
Of the 523 participants with a choice of plan, 168 (21%) reported that at least 1 practitioner had left their plan's insurance network in the past 3 years (

Network Composition and Plan Choice
Among the 523 participants who had a choice of plan in our sample, 302 (57%) considered whether at least 1 practitioner participated in the network before choosing a plan (Figure 2 considered whether a specific hospital was in a network before choosing a plan. For 200 participants (37%), the availability of a practitioner in the network influenced the choice of plan. However, among our sample of individuals who received mental health services in the past year, only 46 (8%) responded that the inclusion of a specific mental health practitioner ultimately influenced plan choice.

Discussion
We found that patients with private insurance were more likely to rate their mental health provider Although the high rates of treatment by out-of-network mental health practitioners and patient perception of disparities in network adequacy between mental health and medical plan networks  that we found in this study are not de facto evidence of inadequate mental health networks, they may be a signal of inadequate networks. States often rely on information from consumers to identify network adequacy issues, although few enrollees submit formal complaints. 22 Stronger quantifiable network adequacy standards that include wait times, whether practitioners are accepting new patients, and whether practitioners have submitted a behavioral health claim in the past 6 months may improve access. 28 We did not find differences in the ratings of mental health and medical network adequacy among those receiving treatment from primary care practitioners, and, to our knowledge, this finding has not been reported earlier. Increased resources devoted to primary care practitioners may help insurers address deficiencies in mental health networks for plan enrollees, including payment mechanisms that adequately reimburse primary care physicians for providing mental health care, increased technical guidance for practitioners, and additional resources for staff to provide care management services. This finding suggests that when mental health network adequacy is being assessed, the availability of primary care practitioners to fill this role, such as those participating in collaborative care models, might be considered. 28 We found that patients were no more likely to experience the exit of a mental health practitioner than the exit of a primary care or specialty practitioner. Regulations related to consumer protections when practitioners are terminated vary considerably by state and do not apply to selffunded plans. 29 Although the final regulations have not been released to date, the No Surprises Act (2020) addresses this issue in group health plans in part, allowing some patients in active treatment to continue under in-network cost sharing for 90 days when a practitioner is terminated without cause. 30 Whether all mental health care will be covered under this provision has not yet been determined. Other policy solutions are needed to retain (and recruit) practitioners in networks, possibly through reimbursement and a reduction in administrative burden. 31 The high rate of denials of claims for mental health treatment under private insurance suggests this may be a relevant factor underlying low participation among mental health practitioners. 32,33 Overall, our results were consistent with those of a survey among individuals enrolled in a single health plan, 34 which showed that approximately 60% of participants considered inclusion of a specific doctor or hospital when choosing a plan, although the researchers did not examine whether participants considered network status of mental health practitioners. We found that individuals were no more likely to check the network status of mental health practitioners than other specialists.
Checking the status of either type of practitioner was relatively uncommon; only about 20% of individuals did so, and even fewer noted that this information affected their plan choice. This suggests that concerns about plan selection through network construction may be less common than originally believed. Researchers have previously found that enrollees are willing to pay more than $1000 annually for a plan that includes a current practitioner, although those researchers also found evidence of high rates of "inattention bias," with many enrollees failing to choose a lower-cost plan that includes their current practitioner. 35

Strengths and Limitations
Our study has several strengths. Network adequacy has been defined differently by states, the federal government, and plan type. We sidestepped this issue by asking patients directly whether they believe their plan's provider network is adequate. This consumer-centric focus allowed us to directly compare at least 1 measure of network adequacy across mental health and medical networks.
Comparing responses for the same participants in our sample allowed us to address concerns that individuals receiving different types of care may differ in person-level characteristics that also affect responses to care experience questions. Another strength is that we examined a national sample, and most survey participants were enrollees in non-Marketplace commercial plans, the focus of current policy. 30 This study has limitations related to issues common to all surveys, including potential biases related to self-report and response bias. To address these issues, we used cognitive interviewing to pretest the survey instrument, and responses were weighted by demographic characteristics, yet some biases may remain. In some cases, our sample size was limited, making it more likely that we failed to detect meaningful differences between groups. If a participant received treatment more often from one practitioner type than another (eg, more treatment from specialists than from primary care practitioners), the patient might have been more likely to have a negative experience related to a provider network. When we considered whether the provider network affected plan choice, we were unable to identify the patients whose treatment had the highest cost.

JAMA Network Open | Health Policy
It is notable that our study sample does not consider the experiences of patients who attempted but did not ultimately receive treatment, arguably the group that had the most difficulty locating a practitioner. The survey included only individuals with private insurance; thus, perceptions of network adequacy for Medicaid managed care plans or Medicare Advantage plans were not assessed.

Conclusions
In this survey study, more respondents perceived their mental health provider networks to be inadequate than those who perceived their medical provider networks to be inadequate.