The Effects of Federal Parity on Mental Health Services Use and Spending: Evidence From the Medical Expenditure Panel Survey
Abstract
Objective:
This study evaluated the effects of the federal Mental Health Parity and Addiction Equity Act (MHPAEA) of 2008 on the use of outpatient and clinic-based mental health services and spending on those services.
Methods:
Data came from the 2005–2013 Medical Expenditure Panel Survey. The analytic sample included adults ages 26–64 who were continuously enrolled in employer-sponsored insurance for a calendar year (N=66,602 person-year observations). A difference-in-differences study design was used to compare changes in outcomes before and after implementation of the MHPAEA between people whose insurance plan was or was not affected by the law.
Results:
The federal parity law was not significantly associated with changes in the likelihood of using mental health services, the amount of mental health services used, or total or out of-pocket spending for mental health services. The law was marginally significantly associated with a shift toward more use of mental health specialty services rather than primary care services among individuals who used ambulatory mental health care.
Conclusions:
Consistent with other research using different study designs and data, this study found that the MHPAEA had at most small effects on patterns of mental health services use and spending through 2013. Understanding whether these effects were small because most employer-sponsored plans were already parity compliant or because plans were noncompliant with the law has major implications for mental health policy and parity enforcement.
Enactment of the 2008 Mental Health Parity and Addiction Equity Act (MHPAEA) was intended to end the disparity between private insurance benefits for mental health and general medical services. The MHPAEA builds on previous federal and state parity laws. The 1996 federal parity act was limited in scope, and research indicates that state parity laws had little or no effect on mental health service use (1). State parity laws, however, were associated with meaningful reductions in out-of-pocket spending for mental health services (1). In response to earlier parity efforts, employers and insurers may have increased the extent of utilization management techniques, which may explain why the laws did not lead to increased service use (2).
The MHPAEA could have a stronger effect on mental health services than prior laws, for two reasons. First, the MHPAEA applies to private-sector enrollees in self-insured plans that are exempted by ERISA from state-level insurance regulation, an estimated 60% of all private-sector enrollees (3). Second, the MHPAEA requires parity between coverage for mental health and general medical services, not just in quantitative insurance benefits (e.g., deductibles, copayments, and annual visit limits) but also in nonquantitative treatment limits (e.g., rationale for prior authorization for service use and criteria for provider admission to network). Given that only 43% percent of the 44.7 million adults with mental illness used any services in 2016, understanding the effects of the MHPAEA on service use is critical (4).
HIGHLIGHTS
•
The authors evaluated the effects of the Mental Health Parity and Addiction Equity Act (MHPAEA) of 2008 on the use of outpatient and clinic-based mental health services and spending.
•
The MHPAEA was not significantly associated with changes in the likelihood of using mental health services, the amount of mental health services used, or total or out-of-pocket spending for mental health services.
•
The MHPAEA was marginally associated with a shift toward more use of mental health specialty services rather than primary care services among mental health services users.
•
Understanding whether these small effects were due to employer-sponsored plans already being parity compliant or to employer noncompliance with the law has major implications for mental health policy and parity enforcement.
Several important trends in mental health services use occurred in the period before and after the MHPAEA’s passage. Between 2008–2009 and 2014–2015, the proportion of U.S. adults using outpatient specialty mental health services rose modestly, with no change in mean number of visits among mental health service users (5). Emphasis on integrating mental health services into primary care, in some cases in conjunction with specialty consultation, has increased. For example, between 1995–1998 and 2007–2010, per capita mental health visits to primary care physicians increased from 8.75 to 13.23 per 100 visits. Mental health visits to psychiatrists increased by a smaller amount, from 7.60 to 8.95 per 100 visits (6). The extent to which these trends were affected by MHPAEA is not known.
The literature evaluating the effects of the MHPAEA is small but growing. These studies show that after the implementation of the MHPAEA, most commercial plans quickly complied with provisions related to quantitative treatment limits (7–9), but research has been less clear about the effects of the MHPAEA on nonquantitative treatment limits, partly because nonquantitative limits are relatively difficult to measure (10, 11). With the exception of Busch et al. (12), who focused exclusively on substance use disorders, and Ettner et al. (13), whose analyses were limited to enrollees in carve-out plans, most evaluations of the effects of the MHPAEA on service use and spending used interrupted time-series research designs with no comparison group. The studies found that the MHPAEA was associated with little to no change in utilization of and out-of-pocket spending for carved-out mental health benefits and with modest increases in utilization of mental health benefits that are integrated with the plan (13, 14). In addition, studies of individuals with eating disorders and children with autism found that the MHPAEA was associated with increases in service use but not with changes in out-of-pocket spending (15, 16).
In this study, we complemented the existing MHPAEA evaluation literature in several ways. We used nationally representative data on mental health services use and, in contrast with most prior work, our analysis controlled for secular trends by including a comparison group of individuals with insurance plans from employers with fewer than 50 employees. These types of plans were initially exempted from the MHPAEA’s regulations. We used survey data, which, unlike claims data, include services not covered by insurance, and we observed more data in the pre-MHPAEA period compared with most prior work.
Historically, only mental health specialty care has been associated with increased cost sharing, treatment limits, and other management techniques to curb use. Thus, one would expect that the direct effects of parity would be centered on specialty treatment, possibly with shifts away from primary care. Therefore, our analyses examined trends in office- or clinic-based specialty and nonspecialty mental health services, whereas prior MHPAEA evaluations focused on specialty services. To provide evidence about whether MHPAEA led to a reallocation of services to those with the most need, we also considered the effects of the law on subgroups of people with and without evidence of a more serious mental illness.
Methods
Study Design
We compared changes in mental health services use and spending before (2005–2009) and after (2011–2013) MHPAEA implementation among people with employer-sponsored health insurance who were affected by the MHPAEA (employees and their dependents at companies with ≥50 employees) and people with employer-sponsored health insurance who were exempt from MHPAEA requirements (employees and their dependents at companies with <50 employees)—a difference-in-differences research design (17). Because the interim final rules for MHPAEA were not published until February 2, 2010, we considered 2010 as a transition year.
Study Data and Population
We used data from the 2005–2013 Medical Expenditure Panel Survey (MEPS), a nationally representative, household-level survey. MEPS respondents are surveyed regarding their health care utilization and expenditures on five occasions over a 2-year period, and expenditures are checked against supplementary information from surveys of respondents’ medical providers. Our analytic sample included respondents ages 26 to 64 who were continuously enrolled in employer-sponsored insurance in a calendar year. Respondents between the ages of 18 and 25 and data from 2014 onward were excluded to avoid confounding with the Affordable Care Act’s dependent-coverage mandate (18) or major coverage provisions, including the essential health benefit rules, which requires small employers to comply with the MHPAEA.
We restricted the sample to the random subsample of MEPS respondents who responded to the MEPS self-administered questionnaire (SAQ), yielding an analytic sample of 66,602 person-year observations. The SAQ includes the Kessler Psychological Distress Scale (K6) survey instrument, which is one way that we identified respondents with evidence of a more serious mental illness (score of ≥13) (19). We also identified respondents as having a serious mental illness if they had an office-based or outpatient visit associated with a clinical classification code (CCC) of mood disorder (657), schizophrenia and other psychotic disorder (659), or suicide and intentional self-inflicted injury (662). In addition, respondents with a mental health–related inpatient or emergency department event were classified as having serious mental illness. An event was defined as related to mental health if it had at least one of the clinical classification codes listed above (657, 659, or 662) or a CCC for adjustment disorder (650); anxiety disorder (651); attention-deficit, conduct, and disruptive behavior disorder (652); personality disorders (658); impulse control disorders (656); or screening and history of mental health and substance abuse codes (663). These CCCs include all nonsubstance use disorder, psychiatry-related codes. The analysis focused on visits with psychiatric diagnoses but did not exclude visits that also had a substance use disorder diagnosis, and as such, the mental health visits may have included some substance use services, as well. On the basis of these criteria, 3,653 person-year observations (5.5%) were classified as having evidence of a serious mental illness. Of these observations, 1,834 had a qualifying CCC, 1,401 had a K6 score of 13 or greater, and 418 had both.
Study Variables
Outcomes.
Dependent variables included use of any outpatient or office-based services (excluding emergency departments), total expenditures, out-of-pocket expenditures, and number of visits for mental health treatment, defined at the calendar-year level. For each category, we estimated total, mental health specialty care, and primary care services. Visits were classified as related to mental health if any of the conditions listed above were noted by the respondent, and as such the mental health concern may not be the primary reason for the visit. Mental health specialist care was defined as outpatient or office-based treatment provided by psychiatrists, psychologists, and social workers. Family practitioners, general practitioners, general internists, and OB-GYNs were all categorized as primary care providers (PCPs).
We also examined percentage of total visits with a mental health specialist and percentage of total expenditures, specialist expenditures, and PCP expenditures that were out of pocket. These outcomes were calculated for each respondent only when the denominator (either total visits or expenditures) was greater than zero.
Independent variables.
We categorized respondents as covered under a large employer if their insurance was obtained through an establishment (i.e., a physical location with 50 or more employees). Person-year characteristics that were controlled for in the regressions included age (26–34, 35–44, 45–54, and 55–64), an indicator of whether the person’s household income was over 200% of the federal poverty level, five education categories (less than high school, high school or GED, bachelor’s degree, graduate or professional degree, and missing), race (non-Hispanic white, non-Hispanic black, Asian, Hispanic, and other), marital status, whether the person has children (0, 1, 2, or ≥3), sex, and Census region.
Statistical Analysis
We estimated the impact of the MHPAEA with multivariate regression models that used the following difference-in-differences model:
where 
denotes a given outcome for person 
in year 
. 
is a categorical variable that indicates whether MHPAEA was implemented in time period 
(2005–2009, 2010, or 2011–2013), 
indicates whether the person was covered by a large establishment in year 
, and 
is the interaction between those two variables. To account for 2010 being a transition year in the MHPAEA’s implementation, the MHPAEA variable also included a separate category for observations in 2010. We did not report those results. 
is a vector of person-level covariates as listed in Table 1, and 
denotes year fixed effects. Our coefficient of interest, the difference-in-differences term 
represents the change among individuals covered by large employers versus those covered by small employers
denotes a given outcome for person in year . is a categorical variable that indicates whether MHPAEA was implemented in time period (2005–2009, 2010, or 2011–2013), indicates whether the person was covered by a large establishment in year , and is the interaction between those two variables. To account for 2010 being a transition year in the MHPAEA’s implementation, the MHPAEA variable also included a separate category for observations in 2010. We did not report those results. is a vector of person-level covariates as listed in Table 1, and denotes year fixed effects. Our coefficient of interest, the difference-in-differences term represents the change among individuals covered by large employers versus those covered by small employers| All respondents | Serious mental illness | No serious mental illness | ||||
|---|---|---|---|---|---|---|
| Characteristic | Small employer (N=21,243) | Large employer (N=45,359) | Small employer (N=1,073) | Large employer (N=2,580) | Small employer (N=20,170) | Large employer (N=42,779) |
| Annual income <200% of the federal poverty level | 10.2 | 9.1 | 17.0 | 12.9 | 9.8 | 8.9 |
| Married | 71.0 | 73.0 | 64.0 | 63.9 | 71.4 | 73.6 |
| Female | 50.2 | 51.3 | 63.0 | 64.2 | 49.5 | 50.5 |
| Age | ||||||
| 26–34 | 21.7 | 19.3 | 18.0 | 15.6 | 21.9 | 19.5 |
| 35–44 | 25.9 | 26.9 | 29.4 | 26.2 | 25.8 | 26.9 |
| 45–54 | 29.2 | 29.3 | 27.8 | 30.9 | 29.2 | 29.2 |
| 55–64 | 23.2 | 24.6 | 24.8 | 27.2 | 23.1 | 24.4 |
| Education | ||||||
| No high school | 4.4 | 4.4 | 4.4 | 5.5 | 4.4 | 4.4 |
| High school or GED | 43.9 | 41.4 | 49.0 | 45.4 | 43.6 | 41.1 |
| College graduate | 24.8 | 25.0 | 19.7 | 19.2 | 25.1 | 25.4 |
| Graduate degree | 21.3 | 23.3 | 20.1 | 24.8 | 21.4 | 23.2 |
| Missing data | 5.6 | 5.8 | 6.9 | 5.2 | 5.5 | 5.9 |
| Race-ethnicity | ||||||
| Non-Hispanic white | 76.4 | 72.0 | 84.1 | 79.5 | 76.0 | 71.5 |
| Non-Hispanic black | 7.6 | 10.9 | 4.9 | 6.7 | 7.8 | 11.2 |
| Asian or Pacific Islander | 4.9 | 6.2 | 1.6 | 2.9 | 5.0 | 6.4 |
| Other | 1.6 | 1.4 | 1.3 | 2.0 | 1.6 | 1.4 |
| Hispanic | 9.5 | 9.5 | 8.1 | 8.9 | 9.6 | 9.5 |
| Children | ||||||
| 0 | 58.7 | 58.5 | 60.3 | 66.4 | 58.6 | 58.0 |
| 1 | 18.1 | 17.5 | 20.5 | 15.2 | 18.0 | 17.6 |
| ≥2 | 23.2 | 24.0 | 19.2 | 18.4 | 23.4 | 24.3 |
a
The sample included individuals who responded to the self-administered questionnaire in the Medical Expenditure Panel Survey (MEPS) and who were covered by employer-sponsored insurance for the entire year. Small employers, <50 employees; large employers, ≥50 employees. Percentages represent the sample mean for 2005–2013. Means were calculated with self-administered questionnaire survey weights.
We estimated linear probability models for the outcomes relating to any mental health service use and ordinary least squares for all other outcomes. We used an inverse hyperbolic sine transformation on total expenditures, out-of-pocket expenditures, and the number of visits. This transformation behaves like a simple natural log transformation, but it is able to accommodate outcomes with a value of zero (20, 21). We used the MEPS’s SAQ survey weights, strata, and primary sampling units to account for the complex sample design.
The key assumption of the difference-in-differences model was that both the large and small employer groups would have had parallel trends in outcomes in the absence of the policy change. We tested the tenability of this assumption by estimating a model with an interaction between year dummies and the large employer group in the pre-MHPAEA years (2005–2009) and then testing whether those interaction terms were jointly significantly different from zero.
Results
The characteristics of the analytic sample are described in Table 1. Table 2 shows the mean outcomes by employer size and presence of serious mental illness. Use of specialty mental health services and total spending were somewhat higher for people in large-employer plans than for those in small-employer plans, whereas the two groups were more similar in their use of mental health services in primary care. [Sample sizes for all outcomes reported in percentages, which are conditional upon the given outcome being greater than zero, are shown in Appendix Table 1 of the online supplement to this article.]
| All respondents | Serious mental illness | No serious mental illness | ||||
|---|---|---|---|---|---|---|
| Outcome | Small employer | Large employer | Small employer | Large employer | Small employer | Large employer |
| Use of mental health services (%) | ||||||
| Any | 6.9 | 7.3 | 67.0 | 68.6 | 3.6 | 3.4 |
| Mental health specialist | 3.8 | 4.1 | 38.7 | 41.1 | 1.8 | 1.7 |
| Primary care provider | 2.8 | 3.1 | 30.4 | 32.9 | 1.3 | 1.2 |
| Visits and expenditures | ||||||
| Visits (M±SD) | .44±3.24 | .45±3.27 | 4.76±10.70 | 5.29±11.26 | .20±1.91 | .15±1.33 |
| Total expenditures (M±SD $) | 60.46±509.69 | 65.33±587.79 | 681.12±1741.06 | 783.12±2142.9 | 26.00±288.95 | 19.78±203.37 |
| OOP expenditures (M±SD $) | 19.06±212.14 | 16.74±227.62 | 192.60±710.21 | 202.48±875.15 | 9.43±133.26 | 4.96±64.70 |
| Visits with a specialist as a percentage of all visits (M±SD %)b | 49.1±47.6 | 52.4±46.1 | 50.5±47.2 | 46.5±48.0 | 51.7±47.1 | 48.0±49.0 |
| OOP expenditures as a percentage of total expenditures (M±SD %)c | 33.8±32.9 | 29.9±29.0 | 30.7±30.8 | 37.1±35.0 | 30.7±31.0 | 31.7±33.0 |
| OOP expenditures as a percentage of total specialist expenditures (M±SD %)c | 36.2±33.6 | 32.0±31.5 | 32.3±33.0 | 38.2±34.4 | 34.6±32.9 | 32.8±34.0 |
| OOP expenditures as a percentage of total PCP expenditures (M±SD %)c | 27.9±27.6 | 25.7±25.4 | 26.2±26.2 | 29.8±28.8 | 26.4±26.6 | 27.2±27.5 |
a
The sample included individuals who responded to the self-administered questionnaire in the Medical Expenditure Panel Survey and who were covered by employer-sponsored insurance for the entire year. Small employers, <50 employees; large employers, ≥50 employees. Means were calculated with self-administered questionnaire survey weights. Expenditures were inflated to 2013 dollars by using the medical consumer price index. OOP, out of pocket. PCP, primary care provider.
b
Proportions reflect only respondents with a visit to a specialist.
c
Proportions reflect only respondents with OOP mental health expenditures.
The unadjusted trends for several study outcomes, broken out by small- and large-employer insurers, are shown in Figure 1. During the period before the MHPAEA, the trends for use of mental health services and number of mental health–related visits for the two groups appeared to be relatively comparable. Our statistical tests largely supported the interpretation that the pre-MHPAEA trends in the outcomes were similar in the treatment (employees of large employers) and comparison (employees of small employers) groups. We failed to reject the null hypothesis of no differential changes at the level of p<0.05 in all but five of 30 comparisons: mental health use and mental health use with a specialist for the overall population, out-of-pocket expenditures for the serious mental illness sample subsample, and mental health use with a specialist and number of mental health visits for the sample without a serious mental illness. In these cases we interpret the results with caution.
FIGURE 1. Trends in utilization of and expenditures for office-based and outpatient mental health services among MEPS respondents, 2005–2013, by employer sizea
aMEPS, Medical Expenditure Panel Survey. A: proportion of respondents who used mental health services. B: mean N of mental health–related visits. C: proportion of all mental health–related visits that were classified as a visit with a specialist. D: proportion of all mental health–related expenses that were paid out of pocket. The vertical dotted line marks a transition year for the Mental Health Parity and Addiction Equity Act of 2008, in which interim final rules were issued and the transition toward full implementation of the law began.
The results of the difference-in-differences models are shown in Table 3. We did not find any significant associations between the MHPAEA and the likelihood of using any mental health services, either overall or from a specific source. For the overall population, the associations between the MHPAEA and the various outcomes were neither significant nor substantial in magnitude, although results for two outcomes came from models in which the parallel trends test was rejected. For the serious mental illness subsample, the probability of using a PCP for mental health specialist services decreased by 4.65 percentage points and the probability of using a mental health specialist increased by 7.83 percentage points. However, these estimates were not statistically significant.
| All respondents | Serious mental illness | No serious mental illness | |||||||
|---|---|---|---|---|---|---|---|---|---|
| Outcome | Coeff. | SE | 95% CI | Coeff. | SE | 95% CI | Coeff. | SE | 95% CI |
| Use of mental health services | |||||||||
| Any | –.041 | .691 | –1.398 to 1.316 | 3.989 | 4.826 | –5.497 to 13.476 | –.365 | .552 | –1.45 to.721 |
| Mental health specialist | .457 | .573 | –.668 to 1.583 | 7.829 | 5.018 | –2.033 to 17.691 | –.021 | .375 | –.758 to .717 |
| Primary care provider | –.124 | .404 | –.918 to .670 | –4.648 | 4.695 | –13.875 to 4.579 | .050 | .262 | –.466 to .566 |
| Visits and expenditures | |||||||||
| N of visits | .001 | .016 | – .030 to .032 | .081 | .142 | – .198 to .360 | – .007 | .011 | – .029 to .016 |
| Total expenditures ($) | .002 | .047 | – .091 to .095 | .253 | .357 | – .449 to .955 | – .019 | .036 | – .09 to .052 |
| Out-of-pocket expenditures ($) | – .018 | .036 | – .089 to .054 | .085 | .298 | – .500 to .670 | – .030 | .027 | – .083 to .024 |
| Visits with a specialist as a percentage of all visits | 8.771† | 4.847 | −.757 to 18.300 | 9.836 | 6.062 | –2.081 to 21.753 | 9.149 | 6.134 | –2.912 to 21.209 |
| OOP expenditures as a percentage of total expenditures | –3.205 | 3.261 | –9.615 to 3.205 | −.502 | 4.211 | –8.781 to 7.777 | –6.776 | 4.435 | –15.495 to 1.943 |
| OOP expenditures as a percentage of total specialist expenditures | –1.150 | 4.509 | –10.019 to 7.718 | −.118 | 6.116 | –12.147 to 11.912 | –3.846 | 6.018 | –15.691 to 7.998 |
| OOP expenditures as a percentage of total PCP expenditures | –4.860 | 3.709 | –12.155 to 2.434 | –5.529 | 4.595 | –14.566 to 3.508 | –3.793 | 5.786 | –15.178 to 7.592 |
a
Mental health use models are estimated with linear probability models. All other models are estimated with ordinary least squares. Total expenditures, out-of-pocket (OOP) expenditures, and visits are transformed with an inverse hyperbolic sine transformation, which behaves similarly to a log transformation. All models were estimated with MEPS self-administered questionnaire survey weights and were adjusted for survey design with probability sampling units and strata. MEPS, Medical Expenditure Panel Survey. MHPAEA, Mental Health Parity and Addiction Equity Act. PCP, primary care provider.
†
p<.1.
We did not find any significant association between the MHPAEA and the overall number of visits (Table 3). In the full sample, the associations between the MHPAEA and both total spending and out-of-pocket spending were relatively small and nonsignificant. The magnitude of those associations were larger for the serious mental illness subsample but were not statistically significant.
We found evidence suggesting that the MHPAEA had an effect on the composition of and out-of-pocket spending for mental health services (Table 3). In the overall population, the MHPAEA was associated with a marginally significant increase in the percentage of ambulatory mental health visits provided by a specialist (8.77 percentage points, p<0.1). Comparable increases were also found among the subgroups of people with (9.84) or without (9.15) severe mental illness, although the estimates were not significant. In the overall population and for both subgroups, the MHPAEA was negatively associated with out-of-pocket spending as a proportion of all mental health expenditures, suggesting that the MHPAEA was associated with a decrease in out-of-pocket spending. However, none of those estimates were statistically significant.
Our results were robust to several sensitivity analyses. We used K6 scores in regressions to examine whether levels of distress affected outcomes. In separate analyses, we estimated nonuse outcomes as nonlinear models that may improve model fit. Neither of these changes affected our baseline results [see Appendix Tables 2 and 3 in the online supplement ]. We also conducted a separate analysis that defined severe mental illness strictly on the basis of K6 score. This also did not change the results, with two exceptions: MHPAEA was associated with a marginally significant decrease in out-of-pocket spending as a proportion of all expenditures in the sample of persons with serious mental illness and with a significant increase in the proportion of visits with a specialist in the sample of persons without a serious mental illness [see Appendix Table 4 in the online supplement .]
Discussion and Conclusions
The MHPAEA has the potential to improve access to and affordability of mental health services by restricting the ability of insurers to underprovide benefits for mental health services. Evaluating the effects of the MHPAEA continues to be important to help understand whether the law is having its intended effects and to better understand whether stricter enforcement mechanisms are needed.
We found that the MHPAEA had no significant effects on use of outpatient or office-based mental health services, number of mental health–related visits, total spending for mental health services, or out-of-pocket spending for mental health services. We found marginally significant evidence that the MHPAEA shifted the share of mental health services used by the overall population from primary care to specialty care. However, we interpret this result cautiously, given that it is marginally significant and is just one of many statistical tests that were conducted. Consistent with other recent research on the law’s effects, our findings showed that the MHPAEA had—at most—modest effects (13).
Our study had several limitations. First, our measure of whether insurance was obtained from a large or a small employer likely creates some measurement error, given that employer size in the MEPS is based on the size of the establishment, regardless of whether the establishment is part of a multiestablishment firm. Second, we did not examine use of prescription drugs for treatment of mental illness. Third, we did not examine inpatient and residential services, which are used less commonly relative to outpatient services, although they are much costlier for both patients and insurers.
Fourth, our research design relied on the untestable assumption that the trends for large and small employers would have remained similar without the MHPAEA. To the extent that other trends affected employees of small and large employers differently, our results may be incorrect. That said, our statistical tests and visual inspection of pre-MHPAEA trends support, but cannot prove, the assumption of parallel trends. Fifth, MEPS respondents are surveyed for at most two years, meaning that a repeated, cross-sectional sample was used, rather than a longitudinal sample. Sixth, compared with other MHPAEA evaluations, we had a smaller sample size (13). Therefore, for some of our results, and especially for our analysis of the subgroup with severe mental illness, we lack the ability to rule out meaningful effects.
Nevertheless, our results have implications for policy. The MHPAEA may have only modest effects on use of mental health services. Most employer-sponsored plans were already in compliance with MHPAEA’s requirements (22), and factors outside insurance design were important for use of mental health services, such as limited availability of specialty mental health providers, stigma, and an inelastic demand for mental health services (23). Our results are consistent with the view that the MHPAEA may have had small effects on the use of specialty mental health services, with spending being modestly shifted from patients to insurers. Our results should also be interpreted as a reflection of the effects of MHPAEA, net of any existing state parity law, although the majority of individuals in the large-employer insurance group were in self-insured plans that were exempted from state parity laws because of ERISA (24).
Two reports using administrative data from large employer plans found that after passage of the MHPAEA, there were differences in the share of in-network versus out-of-network services and in reimbursement rates between general medical and specialty behavioral health providers (10, 25). This might suggest that insurers are not fully complying with the MHPAEA’s rules for nonquantitative treatment limitations, although recent research found that the MHPAEA was associated with reductions in nonquantitative treatment limitations for in-network mental health benefits (11). In addition, the U.S. Department of Labor’s Employee Benefits Security Administration, which enforces the MHPAEA, cited over 300 MHPAEA violations between October 2010 and December 2017, many of which involved inappropriate use of nonquantitative treatment limitations (26). States have become involved in enforcing the MHPAEA. For example, the New York Attorney General reached settlement agreements with a number of insurers that were found to be noncompliant with the MHPAEA.
To the extent that insurers were noncompliant during the study period, our results do not necessarily represent the full potential of the law’s effects on mental health services use. Were parity regulations enforced more effectively, we would expect to see larger effects. If, on the other hand, most plans were, in fact, compliant with the law, we would conclude that the MHPAEA’s effects on ambulatory mental health services use have been modest, and efforts to improve the appropriate use of mental health services might need to consider approaches other than insurance benefit design. As such, tracking the effectiveness and enforcement of the MHPAEA will remain a priority for mental health services research.
Supplementary Material
File (appi.ps.201800313.ds001.pdf)
- View/Download
- 67.60 KB
References
1.
Peterson E, Busch S: Achieving mental health and substance use disorder treatment parity: a quarter century of policy making and research. Annu Rev Public Health 2018; 39:421–435
2.
Barry CL, Ridgely MS: Mental health and substance abuse insurance parity for federal employees: how did health plans respond? J Policy Anal Manage 2008; 27:155–170
3.
Fronstin P: Self-insured health plans: recent trends by firm size, 1996–2015. Notes, July2016, pp 2–6. Available at file:///C:/Users/kstearman/AppData/Local/Microsoft/Windows/Temporary%20Internet%20Files/Content.IE5/C7PMXIP9/SSRN-id2815653.pdf
4.
Key Substance Use and Mental Health Indicators in the United States: Results From the 2016 National Survey on Drug Use and Health. HHS pub no SMA 17-5044. NSDUH series H-52. Rockville, MD, Substance Abuse and Mental Health Services Administration, Center for Behavioral Health Statistics and Quality, 2017
5.
Han B, Olfson M, Huang L, et al: National trends in specialty outpatient mental health care among adults. Health Aff 2017; 36:2062–2068
6.
Olfson M, Kroenke K, Wang S, et al: Trends in office-based mental health care provided by psychiatrists and primary care physicians. J Clin Psychiatry 2014; 75:247–253
7.
Hodgkin D, Horgan CM, Stewart MT, et al. Federal parity and access to behavioral health care in private health plans. Psychiatr Serv 2018; 69:396–402
8.
Thalmayer AG, Friedman SA, Azocar F, et al: The Mental Health Parity and Addiction Equity Act (MHPAEA) Evaluation Study: impact on quantitative treatment limits. Psychiatr Serv 2017; 68:435–442
9.
Friedman SA, Thalmayer AG, Azocar F, et al: The Mental Health Parity and Addiction Equity Act Evaluation Study: impact on mental health financial requirements among commercial “carve-in” plans. Health Serv Res 2018; 53:366–388
10.
Mark TL, Olesiuk W, Ali MM, et al: Differential reimbursement of psychiatric services by psychiatrists and other medical providers. Psychiatr Serv 2018; 69:281–285
11.
Thalmayer AG, Harwood JM, Friedman S, et al: The Mental Health Parity and Addiction Equity Act Evaluation Study: impact on nonquantitative treatment limits for specialty behavioral health care. Health Serv Res 2018; 53:4584–4608
12.
Busch SH, Epstein AJ, Harhay MO, et al: The effects of federal parity on substance use disorder treatment. Am J Manag Care 2014; 20:76–82
13.
Ettner SL, M Harwood J, Thalmayer A, et al: The Mental Health Parity and Addiction Equity Act Evaluation Study: impact on specialty behavioral health utilization and expenditures among “carve-out” enrollees. J Health Econ 2016; 50:131–143
14.
Harwood JM, Azocar F, Thalmayer A, et al: The Mental Health Parity and Addiction Equity Act Evaluation Study: impact on specialty behavioral health care utilization and spending among carve-in enrollees. Med Care 2017; 55:164–172
15.
Stuart EA, McGinty EE, Kalb L, et al: Increased service use among children with autism spectrum disorder associated with mental health parity law. Health Aff 2017; 36:337–345
16.
Huskamp HA, Samples H, Hadland SE, et al: Mental health spending and intensity of service use among individuals with diagnoses of eating disorders following federal parity. Psychiatr Serv 2018; 69:217–223
17.
Dimick JB, Ryan AM: Methods for evaluating changes in health care policy: the difference-in-differences approach. JAMA 2014; 312:2401–2402
18.
Saloner B, Lê Cook B: An ACA provision increased treatment for young adults with possible mental illnesses relative to comparison group. Health Aff 2014; 33:1425–1434
19.
Kessler RC, Barker PR, Colpe LJ, et al: Screening for serious mental illness in the general population. Arch Gen Psychiatry 2003; 60:184–189
20.
MacKinnon JG, Magee L: Transforming the dependent variable in regression models. Int Econ Rev 1990; 31:315–339
21.
Burbidge JB, Magee L, Robb AL: Alternative transformations to handle extreme values of the dependent variable. J Am Stat Assoc 1988; 83:123–127
22.
Goplerud E: Consistency of Large Employer and Group Health Plan Benefits With Requirements of the Paul Wellstone and Peter Domenici Mental Health Parity and Addiction Equity Act of 2008. Washington, DC, US Department of Health and Human Services, Assistant Secretary for Planning and Evaluation, 2013
23.
Meyerhoefer CD, Zuvekas SH: New estimates of the demand for physical and mental health treatment. Health Econ 2010; 19:297–315
24.
Percent of Private-Sector Enrollees That Are Enrolled in Self-Insured Plans at Establishments That Offer Health Insurance by Firm Size and Selected Characteristics: United States, 2011. Rockville, MD, Agency for Healthcare Research and Quality, 2011. https://meps.ahrq.gov/data_stats/summ_tables/insr/national/series_1/2011/tib2b1.htm
25.
Melek SP, Perlman D, Davenport S: Addiction and Mental Health vs Physical Health: Analyzing Disparities in Network Use and Provider Reimbursement Rates. Seattle, Milliman, 2017
26.
Report to Congress: Pathway to Full Parity. Washington, DC, US Department of Labor, 2018
Information & Authors
Information
Published In
Cover: XXXX
History
Received: 4 July 2018
Revision received: 3 October 2018
Revision received: 14 November 2018
Accepted: 20 November 2018
Published online: 29 January 2019
Published in print: April 01, 2019
Keywords
Authors
Competing Interests
The authors report no financial relationships with commercial interests.
Funding Information
This work was supported by a grant-in-aid from the Office of the Vice President for Research, University of Minnesota.
Metrics & Citations
Metrics
Citations
Export Citations
If you have the appropriate software installed, you can download article citation data to the citation manager of your choice. Simply select your manager software from the list below and click Download.
For more information or tips please see 'Downloading to a citation manager' in the Help menu.
View Options
View options
PDF/EPUB
View PDF/EPUBLogin options
Already a subscriber? Access your subscription through your login credentials or your institution for full access to this article.
Personal login Institutional Login Open Athens loginNot a subscriber?
PsychiatryOnline subscription options offer access to the DSM-5-TR® library, books, journals, CME, and patient resources. This all-in-one virtual library provides psychiatrists and mental health professionals with key resources for diagnosis, treatment, research, and professional development.
Need more help? PsychiatryOnline Customer Service may be reached by emailing [email protected] or by calling 800-368-5777 (in the U.S.) or 703-907-7322 (outside the U.S.).