Health Plans’ Early Response to Federal Parity Legislation for Mental Health and Addiction Services

Insurance coverage for mental and substance use disorders (behavioral health disorders) has historically been more limited than coverage for general medical conditions (1,2), contributing to high rates of untreated disease and significant costs to individuals and society. Congressional enactment of the Mental Health Parity and Addiction Equity Act (MHPAEA) (PL 110–343) in 2008 was a landmark. The law required important changes to private health insurance, with the goal of equalizing coverage of behavioral health and general medical conditions (3). The law is expected to transform behavioral health care delivery by expanding access and improving financial protection (2). Since release of initial regulations in 2010, the law has been evolving. Recent lawsuits indicate lack of clarity in the legislation and challenges in making parity determinations between behavioral health care and general medical care.

MHPAEA prohibits health plans that cover behavioral health conditions from imposing more restrictive financial requirements (for example, copayments) or treatment limitations on behavioral health care than on general medical care. The legislation deferred many details of implementation to the federal rulemaking process. Interim final rules, released in January 2010, clarified how comparisons should be made when employers have multiple general medical or behavioral health plans. The rules also clarified that parity is required not only for benefit design features (or “quantitative treatment limitations”) but also for a wider set of health plan practices that affect access, now known as “nonquantitative treatment limitations” (NQTLs). This means that plans cannot manage behavioral health care more stringently than general medical care in terms of such practices as prior authorization, network design, and provider fee determination. The final regulations, issued in November 2013, provide further clarification and reinforce the application of parity for both quantitative and nonquantitative limitations.

There is concern that MHPAEA could have unintended effects—for example, health plans could drop coverage of behavioral health care, drop coverage of certain diagnoses (for example, substance use disorders), tighten utilization management approaches, cut fees, or change organizational structures. Some of these responses could generate new barriers to accessing services. Further, the challenge of comparing behavioral health care with general medical care to ensure parity may continue to result in barriers that go unrecognized. One initial analysis of MHPAEA identified many changes in health plan benefits between 2009 and 2010; however, the analysis also found that some products still had unequal benefits (4). MHPAEA continues to be clarified through the legal system. For example, the attorney general of New York recently settled with three health plans (5), and state insurance departments across the country are focusing on parity enforcement (6).

The aim of this study was to examine private health plans’ early responses to MHPAEA, in terms of both the intended effects on benefits and the unintended effects, in a nationally representative survey of private health plans. This first systematic, national look at health plans’ early responses to MHPAEA will be valuable to policy makers, health plans, and providers, particularly as MHPAEA provisions continue to be clarified via investigation and litigation.

Methods

Data were from the third round of a nationally representative survey of commercial health plans regarding behavioral health services in the 2010 benefit year and the preparity 2009 benefit year. The telephone survey was conducted with senior health plan executives from September 2010 to June 2011. Typically the medical director or the behavioral health medical director addressed clinical questions (for example, utilization management), and another official responded regarding administrative issues (for example, plan characteristics and benefit design). Occasionally, plans referred us to their managed behavioral health organization (MBHO) contractor. For some national or regional plans, respondents were interviewed regarding multiple sites. Items were asked at the product level (for example, health maintenance organization or preferred provider organization) within each market area–specific plan. For all products, we asked whether they covered behavioral health services and the proportion of members with behavioral health coverage. All other questions were asked in regard to the plan’s top three commercial products. The Institutional Review Board of Brandeis University approved the study.

We employed a panel survey design with replacement. The national sample from 2003 (round 2) was augmented with plans not previously operating in the market areas. The primary sampling units were the 60 market areas selected by the Community Tracking Study to be nationally representative (7). The second stage sampled plans within markets. Plans serving multiple markets were defined separately, and data were collected by market area. We screened for eligibility by verifying health plan operation in the market area and coverage of behavioral health services for a commercial population with more than 300 subscribers or 600 covered lives. This approach identified 438 eligible plans, of which 389 responded (89%) and reported on 939 insurance products. For the clinical portion of the survey, 385 plans (88%) responded, reporting on 925 products. Findings reported are national estimates. Data were weighted to be representative of health plans’ commercial managed care products in the continental United States (weighted sample N=8,431 products).

The parity law was passed in 2008 and went into effect for plan years beginning on or after October 3, 2009. We examined health plan policies for the 2010 benefit year. All plans in our study had a January 1 start date so that this was the first benefit year under MHPAEA. We also examined the 2009 benefit year, prior to MHPAEA. The interim final rules went into effect in July 2010 and were not in effect for plans during the study period. In our study, 2% of products were sold exclusively in the individual market or to small groups and therefore were not subject to parity requirements.

Health plans may deliver and manage behavioral health services internally or may contract with an MBHO for this service. Some have hypothesized that the implementation of parity would be difficult for plans that use an MBHO because of challenges comparing benefits across companies. Therefore, we examined whether plans changed their contracting approach for behavioral health services between 2009 and 2010. Contracting approaches fall into one of three categories: external (contracted with an MBHO for delivery and management of behavioral health), hybrid-internal (behavioral health services are managed by a specialty behavioral health organization that is part of the same parent organization as the health plan and that also contracts with other health plans), and internal (all behavioral health services are provided by plan employees or a network of providers administered by the plan). We examined whether plans switched categories between 2009 and 2010.

Another goal was to determine how many insurance products had higher cost sharing for outpatient behavioral health care than for general medical care. Some plans required copayments for medical care and coinsurance for behavioral health care (or vice versa). For these plans, comparing enrollee cost sharing across types of care required assuming a typical provider fee per outpatient visit. We assumed a fee of $130, based on the average observed in the Medical Expenditure Panel Survey for 2008 (8). At this fee, a coinsurance rate of 20% equates to a copayment of $26.

Special limits on behavioral health care were defined as benefit limits that apply only to behavioral health diagnoses and not to general medical conditions. This definition excludes general limits that some plans apply to all medical care (including behavioral health), which are not affected by MHPAEA. In addition, plans were asked about their use of the various practices defined as NQTLs. These included use of prior and concurrent authorization requirements, network size, and provider fees.

Statistical analyses were implemented with SUDAAN 11.0.1 software for accurate estimation of the sampling variance given the complex sampling design. Significant differences reported are based on pairwise t tests with a .05 significance level.

Results

Coverage for Behavioral Health Services

For all products, behavioral health care was covered in the most commonly purchased package in 2009 and 2010 (Table 1). Plans reported that across their membership, an average of 94.7% of members were covered for behavioral health care in 2010; the percentage for 2009 was not significantly different. Out-of-network behavioral health coverage was available in 81.6% of products in 2010, a slight but significant decline since 2009. Under MHPAEA, products that cover behavioral health care may exclude specific diagnoses. In 2010, some respondents reported exclusion of specific behavioral health diagnoses, namely eating disorders (22.4% of products), autism (7.6%), and attention-deficit hyperactivity disorder (1.5%). No products excluded alcohol or drug use disorders.

Among the 89% of products that had a deductible in 2010, 74.7% reported having a common deductible for behavioral health and general medical care, rather than separate ones. All products studied retained the same type of contracting arrangement for behavioral health services in 2010 as in 2009.

Quantitative Treatment Limitations

The proportion of products with special annual limits on mental health care dropped from 27.8% in 2009 to 4.0% in 2010, and a similar decrease was observed for substance use disorders (Table 2). For in-network outpatient behavioral health care in 2010, adding across rows, 75.5% of products required copayments, with the remainder requiring coinsurance. The pattern differed for in-network outpatient medical care, where 46.9% of products required copayments and the remainder required coinsurance. In 2010, the mean in-network copayment was higher for behavioral health care than for general medical care ($25.40 versus $21.50). However, the mean in-network coinsurance rate was lower in 2010 for behavioral health care than for general medical care (13.3% versus 20.8%). Similar patterns were observed in 2009.

In 2010, 10.1% of products had higher in-network cost sharing for behavioral health care than for general medical care, assuming a typical visit fee of $130. However, this result is sensitive to our visit-fee assumption. If the fee were set only $4 higher ($134), 6.7% of products (not 10.1%) had higher cost sharing for behavioral health care than for general medical care in 2010 (data not shown). This sensitivity results from the difference in coinsurance rates between behavioral health and general medical visits.

NQTLs

Trends in the use of administrative practices that might serve as NQTLs are shown in Table 3. Prior authorization requirements were less common for behavioral health care than for general medical care in both 2009 and 2010. There were substantial decreases in the proportion of products requiring prior authorization for outpatient behavioral health care. Between 2009 and 2010, this proportion decreased from 14.2% to 4.7% for mental health care and from 13.2% to 4.8% for substance use disorder treatment. The trend was not unique to behavioral health care, because over the same period the proportion of products requiring prior authorization for specialty outpatient medical care decreased from 27.7% to 16.3%. In both years, about 79% of products required continuing review, either at strict intervals or with a frequency depending on or varying by patient (coded “as needed”). Between 2009 and 2010, a slight shift was noted from requiring continuing review “as needed” to requiring reviews with specified frequencies.

Trends in plans’ use of other administrative controls showed a mixed pattern, in some cases tending toward increased stringency and in others toward less. Most products (79.8%) reported a larger behavioral health provider network in 2010. For each of type of specialty visit considered, about two-thirds of products reported that their fee schedule was unchanged, and more than one-fifth reported that they had increased fees in 2010.

Discussion

The main goals of MHPAEA are to increase access to behavioral health services and ensure that they are treated equitably with general medical services. We found in general that the law had the intended effect of eliminating specific quantitative limitations that applied only to behavioral health care. These early results indicate that the law did not have unintended consequences, including elimination of behavioral health coverage, shrinkage of provider networks, or decreasing fees. Some evidence was found of more strict continuing review. There was a dramatic decrease in special limits on behavioral health care since 2003, when nearly all plans used them (9), and a continued decline between 2009 and 2010. Similarly, between 2009 and 2010, fewer plans reported higher cost sharing for behavioral health than for general health services. These findings suggest that in general plans complied with the law.

Consistent with some earlier findings (4,10), the survey data indicated that a few plans still reported using special annual limits and higher behavioral health cost sharing in 2010, although this is prohibited under MHPAEA. There are a number of possible explanations. Some products are not subject to the law (for example, individual and small group), and others may not have been able to implement the requirements quickly. Some of the plans that we classified as having higher cost sharing for behavioral health services were using a different type of cost sharing for behavioral health care than for general medical care—usually coinsurance for general medical and copayments for behavioral health. The parity law and regulations do not directly address this type of difference and specify that comparisons should be made only between copayment and copayment and between coinsurance and coinsurance.

We found only limited evidence of the unintended consequences of the parity law that some had feared. After implementation of MHPAEA, plans reported a small decline in the proportion of employers offering coverage of behavioral health services. This agrees with the finding of the Government Accountability Office that 2% of employers stopped covering behavioral health services in 2010 (10). Another concern was that fees for behavioral health providers would decline. A small proportion of products reported a decrease in provider fee schedules in 2010 compared with 2009, but this was not the dominant response.

MHPAEA regulations address NQTLs, but during the study period (2009–2010) plans were not yet required to comply. Even so, we identified a decrease in the use of prior authorization. This could be a continuation of a trend observed since 2003 away from tight initial management, because plans realized it was not cost-effective to tightly manage initial access to outpatient behavioral health services. Alternatively, plan administrators may have realized that the forthcoming parity regulations would apply to managed care techniques.

More limited state and federal parity laws existed before MHPAEA. Research findings on their impact and the parity requirement in the Federal Employees Health Benefit Plan (FEHBP) are mixed, with some reports of improved access and others of increased use of utilization management controls (11–18). Studies of the Oregon parity law, the only state law that includes restrictions on utilization management similar to those in MHPAEA, did not identify significant cost increases or changes in behavioral health care utilization (19,20). A study of the early impact of MHPAEA in one health plan found no effect on the proportion of enrollees using substance use disorder treatment but a modest increase in spending per enrollee (21).

Of note, we did not identify a change in use of specialty behavioral health organizations or MBHOs. One might expect that with more extensive coverage of behavioral health services, plans would turn to external organizations that specialize in managing behavioral health care, as happened after parity implementation in the FEHBP and when Vermont passed a parity law, where plans responded by carving out behavioral health services (12,22). However, our findings showed no change in approaches to contracting for behavioral health services between 2009 and 2010. This may be because the law was still in flux and regulations were not yet finalized in 2010.

A controversial issue has been the existence of separate deductibles for behavioral health, which were not initially addressed in MHPAEA but were subsequently prohibited by regulations. Some plans and MBHOs argued that this provision would impose prohibitive administrative costs and put MBHOs out of business (23). By 2010, 75% of plans reported having a common deductible for behavioral health and general medical care, implying that separate deductibles were by no means predominant immediately before the regulation. Separate deductibles may have been more common in direct contracts between MBHOs and employers, which were not addressed in this analysis.

Our study had several limitations. We surveyed health plans regarding typical plan designs in their top three commercial products. Our findings may not fully capture variation within products and do not represent less popular products. However, 81% of all products were included in the top three products. Data were self-reported by health plan executives and were not otherwise verified. Although we report the frequency with which authorizations are conducted and how medical necessity criteria are made available, we were not able to determine the stringency with which criteria were applied. It was also not possible to definitively attribute observed changes to the effect of parity. Changes may have resulted from other factors in plans’ environments between 2009 and 2010. However, in requiring the removal of special benefit limits and the changes in cost sharing, the law presumably played a large part. Our survey addressed health plan policies in general, and individuals may have different experiences with medical necessity approvals. We did not collect information on denial rates, appeals, or complaints about individual experiences. Finally, improved access to coverage is necessary to reduce the treatment gap, but other factors, including enrollee awareness of coverage and desire to use services, are also important (24) and were not measured in this study.

Conclusions

It is important to evaluate these findings in terms of the goal of parity legislation—to ensure equitable coverage—and ask whether it is enough to regulate benefits. Prior state laws relied on regulation of benefits to try to ensure equitable access. However, health plans have other techniques to manage care and control costs, including utilization management and selective contracting with providers. The interim final regulations recognized this and tried to make it more difficult to restrict care through nonquantitative means. Our results suggest that between passage of the law and publication of the interim final regulations, plans did not rush to greater use of nonregulated cost control techniques. However, recent lawsuits and settlements indicate that there are challenges with compliance and that continued vigilance is needed.

Health plans’ responses to parity legislation directly influence enrollees’ access to the full range of behavioral health services. When MHPAEA was initially passed, it applied to private health plans. The final regulations clarify application of MHPAEA to Medicaid managed care plans. The 2010 Patient Protection and Affordable Care Act (ACA) (2,25) extends the reach of MHPAEA to additional parts of the commercial insurance market—to ACA health insurance marketplaces, Children’s Health Insurance Program plans, Medicaid managed care plans, and plans in the FEHBP (26,27). Thus it is important to continue to monitor plans’ evolving response to the parity law and regulations.