Federal Parity and Access to Behavioral Health Care in Private Health Plans

In 2014, nearly 14.8% of Americans received treatment for psychiatric disorders, and 1.6% did so for substance use disorders (1). For many of those individuals, their care was paid for through commercial health plans in which they were enrolled (1). Historically, commercial coverage in the United States often included higher copayments and coinsurance, more stringent managed care practices, or benefit limits that applied only to behavioral health (mental and substance use disorders) care (2,3). Over time, many observers became concerned that discriminatory coverage of behavioral health was impeding access to care (4,5). In response, 45 states passed laws requiring parity in benefits for behavioral health (6), but only 20% of U.S. employees with employer-sponsored health insurance were subject to strong state parity laws; this result occurred because many large employers self-insured, making them exempt from state insurance laws (7).

In 2008, Congress enacted the Mental Health Parity and Addiction Equity Act (MHPAEA), which sought to improve access to behavioral health care by regulating health plans’ coverage and management of services for mental and substance use disorders. The law did not require commercial health plans to cover behavioral health (although more than 90% do so) (8). However, it specified that if coverage is provided, the limits and financial requirements (for example, copayments) cannot be more stringent for behavioral health care than for general medical care. The federal regulations implementing the law also specified that various health plan policies (referred to as “nonquantitative treatment limitations”) cannot be applied more stringently to behavioral health than to general medical care. These policies include prior authorization requirements and the criteria for design of provider networks and medication formularies (9). Unlike the Affordable Care Act (ACA) of 2010, MHPAEA does not appear threatened with repeal, but any repeal of ACA might reduce parity protections established by ACA in small-group and individual insurance markets and some Medicaid plans (10).

Some employers and insurers expressed concern about possible unintended effects of MHPAEA (2). First, plans could still drop behavioral health coverage or coverage for certain diagnoses. Second, although legislators had seen the law as a tool to improve behavioral health coverage (by “leveling up” to match coverage of general medical care), plans could alternatively achieve compliance by reducing their coverage of general medical care (“leveling down”). The actual impact of the law is thus by no means a foregone conclusion. Early evidence suggested that in 2010, 4% of plans still had special limits on mental health care, but few dropped behavioral health coverage, shrank networks, or cut provider fees (8). Two surveys of employers in 2011 had similar findings (11,12), whereas another study found decreases in in-network outpatient cost-sharing between 2008 and 2013 (along with increases for intermediate care) (6). However, the early surveys found persistence of higher patient cost-sharing for outpatient behavioral health than for general medical care (11) and exclusion of some diagnoses (12). More recently, many plans have substantially increased deductible levels and replaced copayments with coinsurance (13). Both these trends encourage patients to select lower cost treatments and providers to avoid paying more out of pocket.

In this study, we examine how plans’ coverage and management of behavioral health care changed after full implementation of the federal parity law in 2011. Study questions include whether behavioral health benefits had improved since 2010, whether they were now more similar to benefits provided for general medical care, and whether plans had relaxed use of prior and concurrent authorization for behavioral health care since 2010.

Methods

Sample

Data are from a nationally representative survey of commercial health plans regarding behavioral health services in the 2010 and 2014 benefit years. The telephone surveys were conducted during September 2010–June 2011 and August 2014–April 2015. The surveys were administered to an executive-level director or a person most knowledgeable about behavioral health service delivery (14).

Items were asked at the product level (preferred provider organizations or health maintenance organizations) 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 about the three commercial products that the plan reported as having the highest enrollment. The Brandeis University Institutional Review Board approved the study.

For 2010, 389 plans (89% response rate) reported on 939 insurance products for the administrative module, and 385 plans (88%) reported on 925 products for the clinical module. In 2014, 274 plans responded (80%) and reported on 705 products (14). [Details of the sampling approach are available as an online supplement to this article.]

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, 2010; N=6,974, 2014). The weights accounted for probability of selection (of site and plan) and for nonresponse. In 2010, the response rate was lower in the West (80%), whereas in 2014, it was lower in the Midwest (69%).

Measures

For both 2010 and 2014, we examined the percentage of members with behavioral health coverage, the percentage of products excluding coverage of specific diagnoses, the percentage of products covering certain specific services, types of cost-sharing, and the percentage of products requiring prior authorization for in-network outpatient general medical and behavioral health care. We also compared 2014 responses with responses from our 2010 survey to examine the impact of parity on quantitative and nonquantitative treatment limitations.

In addition, the 2014 survey included a series of questions asking respondents whether their health plan had changed certain approaches since 2010, and if so, whether changes were “in part because of parity.” For both behavioral health and general medical care, these questions asked about changes in covered services in cost-sharing and in prior-authorization requirements.

Statistical Analysis

Statistical analyses were implemented with SUDAAN, version 11.0.1, to allow accurate estimation of the sampling variance. Student’s t tests for independent samples were used to examine changes in services covered, cost-sharing, and prior authorization requirements for both behavioral health and general medical care. Tests for partially dependent samples (15) were feasible only for those comparisons that had outcome variation in both years and they yielded results that were qualitatively similar to the independent-samples approach presented here.

Results

Sample Characteristics

The insurance products described are from plans that in 2010 were mostly for profit (88%), with 47% from the South census region (Table 1). In both years, health maintenance organizations and preferred provider organizations accounted for a large share of products (>64%), and most products (>84%) managed behavioral health care internally rather than through carve-outs. However, by 2014, 18% of products were consumer-directed plans with high deductibles (exceeding $1,250 per individual or $2,500 per family), a substantial increase from 2010.

Covered Services

Recalling changes from 2010 to 2014, respondents said that two-thirds of insurance products had expanded behavioral health coverage (68%) (Table 2). For another 32%, covered services stayed the same, and for less than 1% of products, covered services either underwent multiple changes or respondents said they could not recall. Among plans that were reported as having expanded their covered services, for 96% this result was in part because of parity (data not shown).

When asked about changes in covered services for general medical care between 2010 and 2014, respondents reported no change for 63% of products. In one-quarter of products (26%) respondents reported “other” changes, such as both increases and decreases depending on which services. In another 8% of products, covered services were described as having expanded.

Comparing contemporaneous (nonrecall) responses across the two years, in both 2010 and 2014, 100% of products surveyed included behavioral health benefits (Table 3). When respondents were asked about their entire enrollment (not just the top three products), they estimated on average (mean±SE) that 94.7%±9.6% of members had behavioral health coverage in 2010, which increased to 98.8%±3.5% in 2014. In both survey years, few products excluded benefits for the specific psychiatric diagnoses we asked about. Eating disorders were the most frequently excluded diagnostic group in 2010 (23% of products); however, by 2014, no health plan reported excluding them anymore. However, the proportion reporting exclusion of autism treatment increased from 8% in 2010 to 24% in 2014. Very few products excluded attention-deficit hyperactivity disorder in 2010 (2%) or 2014 (.7%). No products excluded treatment for alcohol or drug use disorders in either year.

For the seven specific substance use services we asked about, coverage rates were high in both years of the survey, exceeding 98% for detoxification, inpatient hospital care, and intensive outpatient care. In both years, outpatient counseling services were covered by 100% of products. The two least frequently covered services in 2010, opioid treatment (69%) and residential rehabilitation (84%), became covered by virtually all products by 2014 (97% and 97%, respectively). Similar patterns were observed for mental health services.

Cost-Sharing

In 2014, more than half of products (54%) responded that behavioral health cost-sharing had decreased since 2010 (Table 2). In contrast, general medical cost-sharing was described as having decreased in only 2% of products. The proportion of products reporting increased cost-sharing was the same for behavioral health (34%) and for general medical care (34%). Although only 6% of products reported no changes for behavioral health cost-sharing, levels for general medical care remained unchanged in 33% of products. Similarly, just 2% of products had “other” reported changes for behavioral health, but many more (25%) did for general medical care. All the products that reduced cost-sharing for behavioral health said that this result was in part due to parity. Almost none (<1%) of the products that increased cost-sharing for behavioral health said that this was in part due to parity (data not shown).

In 2014, 64% of products reported using copayments for both behavioral health and general medical care (Table 4). Most of the remainder (33%) used copayments for behavioral health and coinsurance for general medical care. Comparison of these results with those from the 2010 survey suggest that for behavioral health, many plans shifted from using coinsurance to copayments, as is more typical for general medical care. The mean±SD copayment for behavioral health increased slightly from $25±$7 to $28±$9, but the mean coinsurance rate (patient share) increased considerably more, from 13%±8% to 31%±22%. However, by 2014, only 15% of plans were using coinsurance for behavioral health.

Prior Authorization

In 2014, a majority of respondents reported that their prior-authorization policies remained unchanged, for both behavioral health (57%) and general medical care (65%; Table 2). Few respondents said that prior authorization for behavioral health and general medical products had been tightened since 2010 (2% and 8%, respectively). Reports of relaxed prior authorization were much more likely for behavioral health (34%) than general medical products (.9%). Among products that relaxed prior authorization for behavioral health, 98% said this result was in part due to parity (data not shown).

Comparing 2014 with 2010 responses, in both years, prior authorization was rarely required for outpatient care, whether behavioral health or medical specialty (Table 4). Moreover, in both years, prior authorization was less often required for outpatient behavioral health than for outpatient specialty medical care. The proportions requiring prior authorization for outpatient substance use disorder and mental health treatment were nearly identical in both years (about 4.7% in 2010 and 4.6% in 2014), whereas for medical care the proportion dropped from 16.3% in 2010 to 10.2% in 2014. However, we found that in 2014, plans still required prior authorization for behavioral health care in inpatient settings (94% of plans) and for intensive outpatient substance use disorder care (88% of plans; data not shown).

Discussion

The results of this study suggest that the implementation of federal parity legislation was accompanied by continuing improvement in behavioral health coverage. Some of these gains had already been achieved early in implementation, as indicated by the results for 2010 (8) as well as by other studies in 2011 (11,12) and in one large managed behavioral health care organization in 2013 (6,16).

Several findings support the idea that parity is being achieved at least in part through improving behavioral health coverage, not through reductions to general medical coverage (by leveling up rather than down). Specifically, a majority of plans reported having reduced cost-sharing for behavioral health, whereas very few reported doing the same for general medical care. Similarly, many more respondents reported having expanded their covered services for behavioral health than for general medical care. However, one-third of plans reported increasing their cost-sharing for behavioral health and one third reported increasing cost-sharing for general medical care. This result corresponds to trends reported elsewhere over the period studied, toward higher deductibles and cost-sharing in health insurance generally (17). Those trends make the parity law less effective in protecting behavioral health coverage because when cost-sharing is increasing for general medical care, there is no parity basis to challenge similar increases for behavioral health care. Thus, some leveling down of coverage could eventually occur.

It is also important that no plans reported reducing behavioral health covered services and that more than two-thirds reported expanding them. The expansion of behavioral health covered services could refer to the addition of diagnoses, settings, or services. Our study was not able to determine which of these approaches was used, although we note that coverage of residential and outpatient opioid treatment has improved since 2010. The reduced use of coinsurance for behavioral health contrasts with another recent study that found the opposite, using a sample of plans from a single large health insurer (6).

The finding that in 2014 almost all plans covered behavioral health is reassuring, in that plans do not appear to have dropped behavioral health coverage as some had feared that they would. Studies from the early implementation period (8,11,12) had already reported similar findings, which are now confirmed for 2014, after full implementation.

The rapid growth in exclusion of coverage for autism is concerning. The high and persistent cost of applied behavior analysis for autism may have led some plans to drop coverage. Several states are now mandating coverage of autism (18), although (as in the case of parity) state mandates have less reach than federal ones because (unlike federal laws) they do not apply to the many large employer plans that are self-insured. However, the disappearance of exclusion for eating disorders is noteworthy because this is another condition in which plans might have been concerned about rapid spending growth.

In terms of nonquantitative treatment limits, it is striking that fewer plans now require prior authorization for outpatient care of behavioral health conditions than for general medical conditions. However, this situation predates the federal parity law, as many plans had already abandoned prior authorization for outpatient behavioral health care before 2009, earlier than they did for general medical care (8). The federal parity law does not appear to have affected the prevalence of prior authorization, although it could have affected the stringency of management, which we did not measure. Most plans still require prior and continuing authorization for inpatient and intermediate settings of care.

Before MHPAEA, no state parity laws addressed utilization management and provider network design, except in Oregon. McConnell et al. (19) found limited impact of Oregon’s law on prior authorization requirements for outpatient behavioral health care. However, nationally, one potential indicator is that consumers continue to complain about parity violations (20). For example, in one survey of patients based on a convenience sample, 29% of respondents reported that they or their family member had been denied mental health care on the basis of medical necessity, twice the percentage who reported being denied general medical care (21). Managed care techniques appear to result in more complaints than do benefits issues, perhaps reflecting the greater difficulty of monitoring plan policies in this area (21,22). Gaps in plans’ provider networks are also cited as a problem (23).

Our results may help explain the somewhat mixed findings regarding the effects of MHPAEA on spending and utilization for behavioral health care. Recent studies have reported either small reductions or no effect on patients’ out-of-pocket burden for behavioral health and small increases or no change in utilization (24–28). Our finding that most plan policies did not change dramatically after MHPAEA could help explain the lack of large effects on these outcomes.

Our study had several limitations. First, the study evaluated a law that was implemented nationwide, and we therefore lacked a comparison group to control for other concurrent changes separate from the law. These changes could have included improving economic conditions as the Great Recession ended; the initial implementation of the ACA; and new developments in care management and integration, such as the spread of accountable care organizations. As a result, we are not able to confidently attribute the changes observed to the federal parity law, although it is a plausible influence on plans’ decisions. We asked respondents whether changes they made were in part due to parity, and this item was strongly endorsed for several types of change in plans’ management of behavioral health (including expansion of covered services, reductions in cost-sharing, and relaxation of prior authorization). However, we did not ask respondents about other potential influences on their decisions.

Second, data were self-reported by health plan officials and were not otherwise verified. In addition, some questions asked respondents to recall how their plans’ benefits and other policies had changed since four years earlier (2010), which could have introduced some recall bias, or inability to recall, if the respondent was new to the organization.

Finally, we asked only about each plan’s top three commercial products in each market area studied, and some plans’ approaches may differ in their smaller insurance products. This limitation is more likely to be true for benefits than for utilization management, which is less easy to customize across purchasers.

Conclusions

Our research suggests that on the whole, plans appear to have responded as intended to MHPAEA. We did not find evidence either of widespread noncompliance or of the unintended effects that some had feared, such as dropping coverage of behavioral health care altogether. It remains to be seen how plans’ coverage of behavioral health care evolves if some aspects of parity regulation are weakened by a potential repeal of the ACA.