Unmet need for mental health care is common across insurance market segments in the United States

Abstract

A substantial proportion of individuals with depression in the United States do not receive treatment. While access challenges for mental health care have been documented, few recent estimates of unmet mental health needs across insurance market segments exist. Using nationally representative survey data with participant-reported depression symptom severity and mental health care use collected in Spring 2023, we assessed access to mental health care among individuals with similar levels of depression symptom severity with commercial, Medicare, Medicaid, and no insurance. Among individuals who reported symptoms consistent with moderately severe to severe depression, 37.8% did not have a diagnosis for depression (41.0%, 28.1%, 33.6%, and 56.3% with commercial, Medicare, Medicaid, and no insurance), 51.9% did not see a mental health specialist (49.7%, 51.7%, 44.9%, and 91.8%), and 32.4% avoided mental health care due to affordability in the past 12 months (30.2%, 34.0%, 21.1%, and 54.8%). There was substantial unmet need for mental health treatment in all insurance market segments, but especially among individuals without insurance.

Introduction

Depression has become increasingly prevalent in the United States over the last decade¹ and this increase was likely accelerated by the COVID-19 pandemic.^2,3 Although there are many effective treatments for depression,^4,5 approximately 40% of people with a major depressive episode do not receive treatment.⁶ While there are many reasons that individuals with depression may not get treatment (eg, stigma, not perceiving a need for treatment), affordability is a contributor to this treatment gap.⁶ Among those who might benefit from mental health treatment but did not receive any, the most commonly cited reasons are affordability and other insurance-related factors.^6,7

Health insurance can influence access to care in a number of ways, including through cost-sharing for mental health care services and by facilitating access to mental health specialists. While national policy efforts in the last 2 decades improved coverage of mental health care for some,^8,9 different health insurance market segments are subject to variable regulations that may result in differing levels of access. Commercial insurance and Medicaid are subject to the Mental Health Parity and Addiction Equity Act of 2008, which required mental health coverage to be no more restrictive than other medical benefits. However, “non-quantitative treatment limits” have been difficult to enforce (eg, ensuring equivalent reimbursement rates for mental health providers), which may limit access.¹⁰ Further, the Parity Act does not apply to Medicare and Medicare Advantage plans, although these plans are subject to other requirements that may directly or indirectly affect mental health coverage (eg, network requirements for psychiatrists). Contributing to the uneven regulatory landscape, many states enforce their own mental and behavioral health standards to Medicaid and portions of the commercial insurance market segment.^10,11

Despite the potential for these factors to translate into differences in access to care across insurance market segments, there is relatively limited peer-reviewed evidence directly comparing levels of unmet mental health need using full-scale mental health screeners across insurance market segments and among those without any health insurance after 2020.^12-15 Commonly used data sources to study mental health care use (eg, health insurance claims data) are either siloed across insurance market segments, do not contain encounters not paid for with insurance, or do not capture measures of mental health beyond diagnosis codes. Using these data sources, for example, studies have found higher dissatisfaction with out-of-pocket expenses among Medicare Advantage beneficiaries with mental illness compared with those without a mental illness¹⁶ and narrower provider networks for mental health specialists in commercial and Medicare Advantage networks compared with non-specialist networks.^17-19 Reports using 2019 survey data demonstrate that a high proportion of individuals with moderate to severe symptoms of anxiety and/or depression do not receive care, with the highest rates among individuals who are uninsured.¹⁴ However, data from these reports precede the COVID-19 pandemic and substantial developments in the financing of mental health care (eg, expansions of coverage for telemedicine). Understanding the burden of unmet mental health needs across different market segments can inform policy priorities and efforts to improve access to mental health care.

In this paper, we quantify levels of unmet need for mental health care across insurance types using the most recent available nationally representative sample of US adults.

Data and methods

Data and sample

We used the fourth wave of the COVID-19 and Life Stressors Impact on Mental Health and Well-being (CLIMB) Study, which was collected between March 24, 2023, and April 17, 2023 (survey completion rate, 31.8%). The CLIMB study is sourced from the AmeriSpeak standing panel, run by NORC at the University of Chicago. Participants in the AmeriSpeak panel were recruited using a non-zero probability, multi-stage, address-based approach with a sampling frame that covers 97% of US households. The survey was administered online and via telephone. The sample was limited to those with a non-missing insurance value that included commercial, Medicare, Medicaid, or no insurance.

In total, our sample included 2365 individuals (1338 with commercial, 636 with Medicare, 250 with Medicaid, and 141 with no insurance). For each of the 3 access measures described in the next section, we also required a non-missing response for the outcome (sample of 2287, 2276, and 2131 for each of the 3 outcomes, respectively). Survey weights were used to account for nonresponse and to align the sample with the US population using Current Population Survey estimates.^2,20,21

Key measures

In all analyses, access was assessed conditional to an individual's level of reported depression symptom severity. Depression symptom severity was captured via the 9-item Patient Health Questionnaire (PHQ-9), a commonly used and validated instrument for screening symptoms of depression.²² The PHQ-9 results in a score between 0 and 27 that maps to a level of depression symptom severity. We used the following categories of depression symptom severity, which we will refer to as “depression severity” in the remainder of the paper: none to mild (0–9), moderate (10–14), and moderately severe to severe (15–27) based on cutoffs used in validation studies.^22,23 Additionally, we included a more detailed version that further broke down these categories into none (0), minimal (1–4), mild (5–9), moderate (10–14), moderately severe (15–19), and severe (20–27) in regression analyses. A treatment plan is generally recommended for individuals with a PHQ-9 of 10 or higher.²²

We assessed 3 measures of access to mental health care within these depression severity categories. The first measure was whether the individual received a diagnosis in the last 12 months for depression from a health professional. Given that we compared individuals with similar levels of survey-measured depression symptom severity, having had a diagnosis within the past 12 months was viewed as a measure of access to providers that identify and diagnose the individual's depression. The second measure was whether the individual saw or talked to a mental health professional, such as a psychiatrist, psychologist, psychiatric nurse, or clinical social worker, in the last 12 months. For this measure, we also examined whether the individual had at least 1 of their visits by telemedicine. Telemedicine was defined as seeing a health care professional using phone or video. The third measure was whether the individual ever wanted to see a health professional about their mental health but did not due to the cost of getting care in the past 12 months. The survey questionnaire language is included in Appendix S1. (To access the Appendix, click on the Details tab of the article online.)

The key independent variable of interest was the type of insurance that the individual had, categorized into the following: commercial insurance (employer-sponsored or individual), Medicare (including fee-for-service or Medicare Advantage), Medicaid (including fee-for-service or Medicaid managed care), and no insurance. In regression analyses, described in the next section, we also included demographic controls for gender and an individual's personal savings, age (continuous), highest level of educational attainment, and race and ethnicity (categorical, with categories shown in Appendix S2). A summary of all measures across the 4 insurance groups is shown in Appendix S2.

Statistical analysis

First, we compared the unadjusted percentages for each of the 3 access outcome measures by type of insurance within the primary categories of depression severity. Percentages are weighted with CLIMB survey weights. See Appendix S3 for unadjusted percentages for mental health visits via telemedicine.

Then, we estimated regression-adjusted differences in each outcome measure by insurance type among individuals with at least moderate depression symptom severity (PHQ-9 of 10 or greater), controlling for the category of depression severity. In the simplest model, we controlled only for the aggregated primary depression severity categories, then added the more detailed categories described above. Finally, we added controls for an individual's level of personal savings, age, education, race and ethnicity, and their reported gender. The preferred specification, shown in the main text, includes the more detailed depression severity category and demographic controls. Model estimates from the other specifications are shown in Appendix S4.

We estimated similar regressions among all individuals across the full spectrum of PHQ-9 scores and among individuals with scores of 9 and below (none to mild depression severity). All regressions were estimated using ordinary least-squares regression and weighted with CLIMB survey weights. Linear probability models were used for ease of interpretation of the mean marginal effect without transformation and because predictions did not lie below 0 or above 1. Data collection for the CLIMB study was deemed exempt by the Institutional Review Board (IRB) at NORC at the University of Chicago; data analysis for the CLIMB study was deemed not human subjects research by the IRBs at Boston University Medical Center and Johns Hopkins Bloomberg School of Public Health.

Limitations

While the CLIMB survey offers several important advantages over other potential data sources, including its recency and inclusion of key study measures (eg, PHQ-9, personal savings, multiple insurance categories, and utilization of telemedicine for mental health), our study is subject to limitations. First, our sample size was limited to respondents to the CLIMB survey with responses to the key measures included in this study, which results in small samples in some of the evaluated subgroups. Relatedly, our primary categorization of depression severity used aggregated categories of PHQ-9 depression symptom severity to enable adequate samples for comparison. For the same reason, we did not investigate subcategories of insurance types (eg, employer-sponsored insurance and individual). Second, the PHQ-9 is a screening instrument and not a formal diagnosis of depression made by a provider. However, in validation studies, the PHQ-9 has a sensitivity of 88% and specificity of 88% compared to formal diagnosis from a provider when using a cutoff score of 10 or greater.²² Additionally, given that not all persons have access to a provider, the PHQ-9 is an effective tool to estimate depressive symptoms to document unmet need for mental health care. Third, while we control for depression symptom severity using the PHQ-9, there may still be unmeasured differences in depressive symptoms that could contribute to differences in the use of mental health care across groups. Relatedly, the cross-sectional nature of our data makes it challenging to disentangle if individuals with lower levels of depression symptom severity are experiencing fewer depressive symptoms as a result of mental health care. Therefore, we assessed each outcome measure among individuals with low levels of depression symptom severity, in addition to individuals with at least moderate depression symptom severity. Fourth, the CLIMB survey does not currently collect information on whether respondents receive medication for their depression, which may represent another important source of access to mental health care.

Results

Among US adults with moderately severe to severe depressive symptoms, 62.2% had a formal diagnosis of depression from a health care professional in the last 12 months (Figure 1). In this depression severity group, individuals with Medicare had the greatest percentage with a diagnosis at 71.9%, compared to 59.0%, 66.4%, and 43.7% of those with commercial, Medicaid, or no insurance, respectively. Among individuals with moderate depression severity, 34.1% had a diagnosis for depression in the last 12 months, with a similar frequency in commercial (32.7%), Medicare (30.4%), Medicaid (39.5%), and no insurance (35.9%).

Figure 1.

Percentage of US adults who were diagnosed with depression by a health professional in the last 12 months, by depression severity and insurance type: 2023. Source: Authors’ analysis of the fourth wave of the COVID-19 and Life Stressors Impact on Mental Health and Well-being (CLIMB) study questionnaire. Error bars reflect 1 standard error in either direction of the mean. Depression symptom severity was captured via the PHQ-9. The score was categorized into 3 categories of depression symptom severity: none to mild (0–9), moderate (10–14), and moderately severe to severe (15–27). Percentages are weighted with CLIMB survey weights. Abbreviation: PHQ-9, 9-item Patient Health Questionnaire.

Fewer than half (48.1%) of individuals with moderately severe to severe depression saw a mental health specialist in the last 12 months (Figure 2). More than half of individuals in the group that did see a mental health specialist (59.1%) had at least 1 mental health visit in the past 12 months via telemedicine. Within the group of individuals with moderately severe to severe depression, among individuals with no insurance, 8.2% saw a specialist compared to 50.3%, 48.3%, and 55.1% with commercial, Medicare, and Medicaid insurance, respectively. Over half of individuals with commercial (56.5%), Medicare (73.5%), and Medicaid (51.1%) insurance who saw a mental health specialist in the past year had at least 1 mental health visit via telemedicine, whereas 43.6% of individuals with no insurance who saw a mental health specialist had at least 1 visit via telemedicine.

Figure 2.

Percentage of US adults who saw or spoke to a mental health specialist in the last 12 months, by depression severity and insurance type: 2023. Source: Authors’ analysis of the fourth wave of the COVID-19 and Life Stressors Impact on Mental Health and Well-being (CLIMB) study questionnaire. Error bars reflect 1 standard error in either direction of the mean. Depression symptom severity was captured via the PHQ-9. The score was categorized into 3 categories of depression symptom severity: none to mild (0–9), moderate (10–14), and moderately severe to severe (15–27). Percentages are weighted with CLIMB survey weights. Abbreviation: PHQ-9, 9-item Patient Health Questionnaire.

Among individuals with moderate depression severity, 30.8% saw a mental health specialist in the last 12 months. The frequency was similar for commercial (32.2%), Medicare (32.1%), and Medicaid (35.6%), but was less than 20% among individuals with no insurance (18.3%).

Approximately one-third (32.4%) of individuals with moderately severe to severe depression said that there was a time in the past 12 months where they wanted to see a health provider about their mental health but did not due to affordability (Figure 3). Within this depression severity category, 54.8% with no insurance said that they skipped care due to affordability, compared to 30.2%, 34.0%, and 21.1% among those with commercial, Medicare, and Medicaid insurance, respectively. Among those with moderate depression severity, 23.7% reported avoiding mental health care due to affordability; however, 59.2% of individuals with no insurance and moderate depression severity reported doing so.

Figure 3.

Percentage of US adults who wanted to see a health professional about their mental health but did not due to costs in last 12 months, by depression severity and insurance type: 2023. Source: Authors’ analysis of the fourth wave of the COVID-19 and Life Stressors Impact on Mental Health and Well-being (CLIMB) study questionnaire. Error bars reflect 1 standard error in either direction of the mean. Depression symptom severity was captured via the PHQ-9. The score was categorized into 3 categories of depression symptom severity: none to mild (0–9), moderate (10–14), and moderately severe to severe (15–27). Percentages are weighted with CLIMB survey weights. Abbreviation: PHQ-9, 9-item Patient Health Questionnaire.

Among individuals with at least moderate depression symptom severity with similar demographic characteristics and depression severity level, access was similar between individuals with Medicare and Medicaid compared with those with commercial insurance across all 3 measures (Figure 4). Individuals without health insurance were 24.4 percentage points (95% CI: −39.3 to −9.4) less likely to have had a mental health specialist visit and 32.3 percentage points (95% CI: 16.3–48.3) more likely to have avoided care due to costs compared with individuals with commercial insurance. They were 17.9 percentage points less likely to have seen a mental health specialist via telemedicine (95% CI: 4.8–30.9). They were also 14.9 percentage points less likely to have had a depression diagnosis than individuals with commercial insurance, although the difference was not statistically significant (95% CI: −30.1 to 0.003).

Figure 4.

Regression-adjusted relationship between mental health access and insurance type relative to commercial insurance among US adults with at least moderate depression severity: 2023. Source: Authors’ analysis of the fourth wave of the COVID-19 and Life Stressors Impact on Mental Health and Well-being (CLIMB) study questionnaire. The figure shows coefficients for Medicare, Medicaid, and no insurance (relative to having commercial insurance) and 95% CIs from 3 separate regressions that estimate (1) whether the individuals received any depression diagnosis in the last 12 months (n = 513), (2) whether the individual saw or talked to a mental health (MH) specialist in the last 12 months (n = 505), and (3) whether the individual wanted to see a health professional about their MH but did not due to costs in last 12 months (n = 441). All regressions were estimated with ordinary least-squares regression and controlled for the depression severity, reported age, education, race and ethnicity, gender, and personal savings levels. Depression severity was captured via the PHQ-9. The sample included only individuals with a PHQ-9 of 10 or greater with the following categories: moderate (10–14), moderately severe (15–19), and severe (20–27). All regressions were estimated using ordinary least-squares regression and weighted with CLIMB survey weights. Abbreviation: PHQ-9, 9-item Patient Health Questionnaire.

Among all depression severity levels, individuals with Medicare were 6.3 percentage points more likely to have had a mental health specialist visit (95% CI: 1.9–10.6) than those with commercial insurance. These differences were driven by individuals with mild or lower depression severity. See Appendixes S5 and S6 for regressions across all depression symptom severity levels and in individuals with mild to no depression symptom severity, respectively, and Appendix S7 for telemedicine mental health visit regressions.

Discussion

Using data that combined participant-reported depression symptom severity and health care use in a nationally representative sample of US adults, we found that, among individuals who reported symptoms consistent with moderately severe to severe depression, 37.8% did not have a diagnosis for depression, 51.9% did not see a mental health specialist, and 32.4% avoided mental health care due to affordability in the past 12 months. Unmet mental health care needs were found to be similar across the 3 investigated insurance market segments (commercial, Medicare, and Medicaid), while individuals with no health insurance reported lower levels of access across all 3 measures.

We found, first, that among adults with moderate to severe depression, for whom follow-up for diagnosis and monitoring would be recommended, more than one-third had not been diagnosed with depression in the last year, with the greatest percentage of undiagnosed persons in the uninsured group. Our findings are consistent with overall estimates of undiagnosed depression in the United States.²⁴ A formal diagnosis of depression may be the first step to treating depression and its core causes. Receiving a depression diagnosis can help individuals name what they have been feeling and make them eligible for additional interventions or services that can prevent symptoms from worsening.

Second, we found that fewer than half (48.1%) of all adults with moderate to severe depression had spoken with a health care provider about their mental health within the last year, with uninsured people having the lowest percentage and adults with Medicaid reporting the greatest percentage of mental health care use. These findings document slightly higher unmet need than other estimates; for example, the KFF estimated that 39% of persons with moderate to severe depression or anxiety¹⁴ had unmet mental health care needs. Our findings are also consistent with this KFF report and prior work,^13,25 which found the greatest levels of unmet need among people who were uninsured. These latter estimates, however, included individuals with anxiety symptoms and were from 2019. Our 2023 estimates of unmet mental health care needs in all groups may be more aligned with the current landscape, as we find, for example, that over half of individuals with insurance who saw a mental health specialist had at least 1 of these visits via telemedicine. In particular, we found that 73.5% of Medicare beneficiaries with moderate to severe depression symptoms saw a mental health specialist via telemedicine, suggesting that telemedicine may represent an important form of access for older adults who may face greater mobility and transportation constraints.

Third, we found that cost was identified as preventing respondents from receiving mental health care across all insurance groups. Our findings are consistent with evidence from across market segments that access to mental health care can be limited even for those with health insurance^16-19,26 and that affordability is an important barrier to access to mental health care.^6,7,27 Our 2023 estimates align with recent data from the Commonwealth Fund 2023 Health Care Affordability Survey²⁸ that reported high levels of unaffordability across all types of care, even among those with health insurance.

However, we found that being uninsured is associated with the much higher levels of unmet need across all 3 measures. Medicaid expansion and other policies to extend coverage to persons still uninsured could substantially improve access to mental health care. We show that any form of health insurance is associated with greater levels of access, and Medicaid is associated with similar (or greater) levels of access compared with Medicare and commercial insurance.

Our findings point to the importance of improving access to mental health care among those with health insurance across each insurance market segment as well. While mental health parity has been shown to increase the utilization of mental health care and decrease out-of-pocket costs,^8,9 its implementation has still left gaps in coverage and access.¹⁰ Expanding mental health parity to Medicare; improving reimbursement rates to mental health providers, which are often lower than payments to other providers²⁹; developing mental health–specific network adequacy criteria across market segments³⁰; as well as bolstering the mental health workforce could all potentially improve access.^31,32 Improving access to care, through improving accurate diagnosis of depression, mental health care use, and affordability of mental health care may improve population mental health.

Supplementary material

Supplementary material is available at Health Affairs Scholar online.

Funding

This study was funded in part through support from the De Beaumont Foundation.