Abasic premise in mental health policy

Transcription

1 MarketWatch Design Of Mental Health Benefits: Still Unequal After All These Years Benefit differentials persist between mental health coverage and coverage for other health conditions, employer survey data show. by Colleen L. Barry, Jon R. Gabel, Richard G. Frank, Samantha Hawkins, Heidi H. Whitmore, and Jeremy D. Pickreign ABSTRACT: This paper examines recent trends in the design and organization of coverage for mental health care using data from a Henry J. Kaiser Family Foundation and Health Research and Educational Trust (KFF/HRET) national employer survey. Legislation and changes in the delivery of mental health services have altered how mental health insurance is bought and sold. However, our findings reveal that mental health coverage is still typically not offered at a level equivalent to coverage for other medical conditions. We attempt to synthesize these data with prior research as a foundation for informed debates. Abasic premise in mental health policy is that markets, if left alone, may not supply efficient and fair levels of insurance coverage against the financial risks associated with treatment of mental disorders. Historically, mental health coverage has been more limited than coverage for other medical services. Health plans have tended to require higher copayments and more stringent limits on inpatient hospital days and outpatient visits for behavioral health treatment. Plans have also limited mental health coverage through the use of annual and lifetime dollar limits. In recent years legislation regulating insurance and fundamental changes in the delivery of mental health services have altered the environment in which mental health coverage is bought and sold. Parity regulations aim to eliminate the difference in insurance coverage for mental health care and other medical conditions. Likewise, managed care has transformed the delivery of mental health care services over the past decade. Employers and health plans now frequently contract out administration of their mental health benefits to specialty managed behavioral health care organizations (MBHOs). These carve-outs use specialized expertise to manage utilization and to direct care to efficient treatment in provider networks. A national survey conducted by the Henry J. Kaiser Family Foundation and Health Research and Educational Trust (KFF/HRET) of 2,014 randomly selected public and private firms provides an opportunity to examine the impact of these regulatory and market changes on the composition of private insurance cover- Colleen Barry is a doctoral candidate in the Department of Health Policy, Harvard Medical School, in Boston. Jon Gabel is vice-president, Health System Studies, at the Health Research and Educational Trust (HRET) in Washington, D.C. Richard Frank is the Margaret T. Morris Professor of Health Economics at Harvard Medical School. Samantha Hawkins is a research assistant, Heidi Whitmore a senior research associate, and Jeremy Pickreign a statistician at HRET. HEALTH AFFAIRS ~ Volume 22, Number /hlthaff Project HOPE The People-to-People Health Foundation, Inc.

2 Health Tracking age for mental health care in 2002 and over the past decade. We focus specifically on whether people have coverage for mental health care, the nature of this coverage, and the organization of managed mental health care. The evolution of the market has resulted in smaller differentials in cost sharing between mental health and other medical services. However, the use of annual day and visit limits on mental health benefits may have increased somewhat over the past few years. In theory, MBHOs and health maintenance organizations (HMOs) might be less likely to impose benefit design limitations, opting instead to achieve their cost objectives via management. In fact, we find that this is not the case. We present data from the KFF/HRET survey in the context of other studies on the design and organization of mental heath benefits. Prior research has contributed greatly to public understanding of the nature of mental health coverage. 1 Yet the field continues to rely on a wide range of definitions and data collection strategies. Nonetheless, by evaluating our results in the context of prior research, we aim to synthesize the current state of knowledge about mental health insurance coverage. Given the substantial impact of benefit design and organization on access to health services, we believe that a clear picture of the nature of the mental health benefits being offered to workers provides a foundation for informed debates in this area. Study Data And Methods Data source and sampling. The principal database for this paper is the 2002 KFF/ HRET Employer Health Benefits Survey. To track trends in mental health benefits, we also used data from earlier years, when other organizations sponsored the survey: data from the 1991 Health Insurance Association of America (HIAA) survey, the 1995 KPMG Peat Marwick survey, and the 1999 KFF/HRET survey. For all years of the survey, employers were asked a core set of questions about their largest conventional, HMO, preferred provider organization (PPO), and point-of-service (POS) plan. The 2002 survey asked a special series of questions about mental health benefits for each plan. National Research LLC conducted telephone interviews with employee benefit managers during January May The 2002 sample consisted of 2,014 randomly selected public and private employers with three or more workers, 1,092 of whom participated in the 2001 survey. KFF/HRET selects its sample firms from a listing of U.S. firms constructed by Dun and Bradstreet. The sample was stratified by number of workers in the firm and industry. Throughout the paper we use the term large firms to refer to those with 200 or more workers; small firms are those with workers. In 2002 the survey drew a 50 percent response rate. The sampling methods for prior surveys were similar to the 2002 survey. Sample sizes varied from 3,231 firms in 1991 to 1,939 in Response rates have ranged from 76 percent in 1991 to 60 percent in Weights. Since firms are selected randomly, it is possible to use statistical weights to calculate national, regional, firm-size, and industry figures each year. We calculated weights by determining the basic weight, and then applied a nonresponse adjustment followed by a post-stratification adjustment. In calculating standard errors for statistical tests, we used the statistical program SUDAAN to adjust for survey design effects. Study Findings Basic elements of mental health insurance coverage. The Kaiser/HRET survey results suggest that percent of workers with employer-sponsored health insurance had coverage for mental health care in 2002 (Exhibit 1). Ninety-six percent of covered workers had inpatient mental health coverage, while percent had outpatient mental health coverage. These findings are consistent with prior research. For example, the 2000 Bureau of Labor Statistics (BLS) National Compensation Survey of employee benefits in private industry found that 93 percent of employees were covered for both inpatient and outpatient mental health benefits. 2 Large firms were significantly more likely than small ones were to offer mental health 128 September/October 2003

3 EXHIBIT 1 Percentage Of Covered Workers Offered Mental Health Benefits And Having Benefit Limits, By Firm Size And Plan Type, 2002 Firm size Type of plan Total Small Large Conventional HMO PPO POS Offered any mental health benefits Offered inpatient mental health benefits Offered outpatient mental health benefits Annual inpatient day limit Annual outpatient visit limit Higher cost sharing (copayment or coinsurance) % 95% % a 99 a 99 a % 94 % % 64 99% a 78 a 24 a 36 b 52 b b b SOURCE: Henry J. Kaiser Family Foundation/Health Research and Educational Trust Employer Health Benefits Survey, NOTES: Small firms have workers, while large firms have 200 or more workers. HMO is health maintenance organization. PPO is preferred provider organization. POS is point-of-service plan. a Significantly different from small firms at the.05 level. b Significantly different from total at the.05 level. benefits (at the.05 level). Ninety-nine percent of covered employees in large firms were offered mental health benefits, compared with 95 percent in smaller firms. This finding also appears consistent with prior research. Using the 19 Mercer/Foster Higgins survey, Jeff Buck and colleagues found that 99 percent of employees in large firms were offered coverage for both mental health and substance abuse, compared with 91 percent of employees in smaller firms. 3 Furthermore, we find little variation in the proportion of covered workers offered mental health insurance by plan type, industry, region of the country, or earnings level. From 1991 to 2002 the proportion of covered workers offered mental health benefits increased (Exhibit 2). This finding is consistent with earlier research by Gail Jensen and colleagues. 4 Across the three study years (1991, 1995, and 2002), coverage for mental health benefits increased, regardless of plan type or region. Large firms appear to have most consistently offered mental health benefits. The BLS has tracked the proportion of midsize and large firms offering mental health benefits dating back to 10; these data show a similarly high level of coverage among large firms over the past few decades. 5 While most workers with insurance were offered some coverage for mental health services in 2002, firms continued to place special limits on mental health benefits. Federal and state parity laws were enacted over the past decade in an effort to curb these insurance limitations. The Mental Health Parity Act (P.L ), passed by the U.S. Congress in 19, barred health plans from using annual or lifetime dollar limits on coverage for mental illnesses unless equal dollar limits were placed on other medical services. However, the law does not apply to other types of benefit restrictions, such as day and visit limits or higher copayments, and does not regulate firms with fewer than fifty employees. Thirty-four states have also enacted some form of parity regulations. These state laws vary greatly according to the population covered, types of limitations removed, and diagnoses included. For example, twenty state parity regulations apply exclusively to a subset of mental disorders identified as serious mental illnesses (SMI). Gaps in federal and state parity laws explain the persistent differences in coverage between mental health and general medical benefits. The KFF/HRET survey found that a large HEALTH AFFAIRS ~ Volume 22, Number 5 129

4 Health Tracking EXHIBIT 2 Percentage Of Workers In Employer-Sponsored Health Plans With Mental Health Benefits, 1991, 1995, And 2002 Inpatient care Outpatient care All insured workers 87% a 93% a % 86% a 92% a % Firm size Fewer than 50 workers workers 200 or more workers By type of plan Conventional HMO PPO POS By region Northeast South Midwest West 65 a 73 a a a 75 a 87 a 82 a a a 86 a 73 a a a 86 a 85 a 76 a a 87 a 82 a a a 83 a 94 a 95 a SOURCE: 1991 Health Insurance Association of America Employer Health Benefits Survey; 1995 KPMG Peat Marwick Survey of Employer-Sponsored Health Benefits; and 2002 Henry J. Kaiser Family Foundation/Health Research and Educational Trust Employer Health Benefits Survey. NOTES: HMO is health maintenance organization. PPO is preferred provider organization. POS is point-of-service plan. a Significantly different from 2002 at the.05 level (one sample t-test on 2002). majority of covered workers were still subject to special mental health day and visit limits in Exhibit 1 shows that 74 percent of covered workers were subject to an annual outpatient visit limit, and 64 percent were subject to an annual inpatient day limit. These data are relatively similar to the most recent 2000 BLS benefit survey, which found that among employees with mental health coverage, 72 percent had an annual outpatient visit limit, and 76 percent had an annual inpatient day limit. Inpatient and outpatient service limits were somewhat less common among those employed in smaller firms and those with conventional insurance coverage (Exhibit 1). Annual mental health day and visit limits may have increased somewhat over the past few years. The proportion of covered workers subject to outpatient visit limits increased from 65 percent in 1999 to 74 percent in 2002 (Exhibit 3). Similarly, 57 percent of covered workers had inpatient day limits in 1999, compared with 65 percent in This finding is equivocal, however, because of the comparatively higher level of firms reporting don t know in response to questions about limits in Cost sharing for mental health services constitutes a third dimension of benefit design. Health plans have historically attempted to control costs by requiring that enrollees pay more at the point of service for mental health care compared with other medical services. During the 10s the BLS found that a majority of enrollees in medium and large firms paid a 50 percent coinsurance rate for outpatient mental health care rather than the usual 20 percent paid for other illnesses. 7 The use of higher mental health coinsurance and copayments continued during the 1990s. 8 In contrast, using the KFF/HRET survey we found that in 2002 only 22 percent of covered workers had higher cost sharing for mental health benefits (Exhibit 1). 9 HMO and POS plans were somewhat more likely to require higher cost sharing. Overall, this finding suggests that health plans are relying less on higher cost sharing as a means of curbing use of mental 130 September/October 2003

5 EXHIBIT 3 Percentage Of Covered Workers In Firms With Service Limits On Mental Health Care, 2002 And 1999 Days or visits or more Don t know Unlimited Any limit Annual outpatient visit limits Annual inpatient day limits % a 25 24% 24 18% 16 14% 17 12% a 18 74% a a 9 36 a a a a 57 SOURCE: Henry J. Kaiser Family Foundation/Health Research and Educational Trust Employer Health Benefits Survey, a Significantly different from 1999 at the.05 level (one sample t-test); two sample test could eliminate several significant results. health services. State parity laws could be more influential in the benefit design for employees of smaller firms, since many large firms self-insure and are thereby exempt from state-mandated insurance benefit statutes. The data indicate that covered workers in smaller firms were significantly less likely to be subject to service limits and higher cost sharing than their largefirm counterparts (Exhibit 1). However, it is difficult to attribute this firm-size difference solely to parity laws. 10 Finally, it is useful to gauge whether these three categories of benefit design features inpatient day limits, outpatient visit limits, and higher cost sharing tend to be set in tandem. We find that 48 percent of covered workers were enrolled in plans that subjected them to all three types of limits on mental health care simultaneously. In contrast, only 8 percent of covered workers were enrolled in plans with none of these restrictions. Organization of mental health coverage. Managed care has greatly altered the organization and delivery of mental health services. As noted earlier, firms increasingly opt to carve out the administration of mental health benefits to specialty carve-out companies. These organizations use specialized expertise to establish networks of providers, negotiate volume-related discount contracts, identify evidence-based treatment protocols, and develop other incentive programs to manage utilization and costs. Many health plans have also begun contracting with MBHOs to gain access to specialty mental health networks and to manage utilization. Thus, a covered worker or dependent could receive mental health care through a carve-out at either the firm or the plan level. A variety of studies indicate that the introduction of a behavioral health carve-out can reduce costs. 11 Less research has examined the affects of carve-outs on quality of care, and existing studies have produced mixed results. 12 Employers and health plans can structure carve-out contracts in various ways. For example, a carve-out contract with a purchaser might be risk-based or nonrisk (administrative services only); it might provide utilization review or case management services only or employee assistance program (EAP) services only; or it might contract to provide both managed behavioral health services and EAP services. Larger MBHOs might offer all of these separate services, depending on the needs of a given purchaser. Variation in the types of services provided through carve-outs creates certain challenges in terms of quantifying industry growth. A number of methods exist for collecting information on carve-out market penetration. Approaches include eliciting information from carve-out companies, from firms, or directly from people with mental health insurance coverage. The latter strategy is particularly prob- HEALTH AFFAIRS ~ Volume 22, Number 5 131

6 Health Tracking lematic, since most employees are unaware of their employer s contracting arrangements with various insurers. The approach used by KFF/HRET entails surveying employers to determine the proportion of workers covered under behavioral health carve-out contracts at either the plan or the firm level. A limitation of this approach is that many employee benefit managers possess little knowledge about health plan level carve-outs. Data from the KFF/HRET survey suggest that about a third of people with employersponsored health insurance were offered mental health care through carve-outs at either the firm (16 percent) or the health plan level (20 percent) in 2002 (Exhibit 4). Those employed by small firms were significantly less likely (at the.05 level) than those at larger firms were to have their mental health benefits carved out at the firm level. Also, workers enrolled in POS plans were more likely to be covered under a carve-out than workers overall. The KFF/HRET survey provides an opportunity to compare benefit design characteristics by carve-out status that is not afforded through other data sources (Exhibit 5). We found relatively little difference in the benefit design features of people with carved-out versus integrated mental health coverage. No clear pattern emerged in terms of the comparative likelihood of carve-outs to restrict the number of inpatient days and outpatient visits compared with integrated plans. Likewise, the percentage of covered workers with higher cost sharing for outpatient mental health services did not differ notably in carve-outs compared with integrated plans. Only those in conventional plans, who constitute just 6 percent of enrollees, appeared significantly less likely (at the.05 level) to have a higher costsharing requirement under a carve-out compared with an integrated plan. Finally, people in carve-outs and integrated plans appeared equally likely to be subject to both higher outpatient cost sharing and service limits. Likewise, receiving mental health coverage through a carve-out was not associated with the likelihood that a person would be free of such restrictions. The KFF/HRET estimates of the market penetration of carve-outs are somewhat lower than those based upon other data sources. For example, of the 336 Fortune 500 companies surveyed by Elizabeth Merrick and colleagues, 39 percent reported firm-level contracts with MBHOs. 13 This study did not examine employees with carved-out benefits at the plan level. Our estimates of carve-out penetration are also lower than those reported by Open Minds, an organization that surveys MBHOs directly. 14 In 2002 Open Minds surveyed more than 800 behavioral health programs and EAPs using mixed-mode techniques to improve its response rate after initial mail contact. 15 According to this resource, about 164 million people with health coverage (66 percent of the insured population) were enrolled in some type of managed behavioral health EXHIBIT 4 Percentage Of Covered Workers In Firms Carving Out Mental Health Benefits, 2002 Firm size Type of plan Total Small a Large a Conventional HMO PPO POS Either firm or health plan level Firm level Health plan level 36% % % 20 c 19 52% % % % b b SOURCE: Henry J. Kaiser Family Foundation/Health Research and Educational Trust Employer Health Benefits Survey, NOTES: HMO is health maintenance organization. PPO is preferred provider organization. POS is point-of-service plan. a Because of the methodology used to aggregate categorical variables across plan types, component estimates might not sum to total estimates. b Significantly different from total at the.05 level. c Significantly different from small firms at the.05 level. 132 September/October 2003

7 EXHIBIT 5 Percentage Of Workers With Mental Health Benefit Limits, By Carve-Out (CO) Or Integrated (Int) Plan Status, 2002 Total Conventional HMO PPO POS CO Int CO Int CO Int CO Int CO Int Any annual outpatient visit limit Any annual inpatient day limit Higher outpatient mental health cost sharing All of the above mental health benefit restrictions None of the above mental health benefit restrictions 78% % % 26 8 a 54% % % % % % % SOURCE: Henry J. Kaiser Family Foundation/Health Research and Educational Trust, Employer Health Benefits Survey, NOTES: HMO is health maintenance organization. PPO is preferred provider organization. POS is point-of-service plan. a Carve-out (CO) is significantly different from integrated (Int) plan at the.05 level. program in Open Minds documents a substantial increase in enrollment in managed behavioral health programs during the 1990s, from 70 million in 1993 to 164 million in In 2000 the industry reached peak enrollment of 169 million and has declined slightly since then. 17 The absence of comprehensive trend data on carve-outs in the KFF/HRET survey prevents direct comparison with the Open Minds data. Like the other methods of estimating carveout penetration, Open Minds strategy has its own limitations. One key issue relates to the likelihood that a firm and an MBHO might characterize carve-outs differently. For example, while an HMO contracting with an MBHO solely for utilization review services would likely represent itself to a purchaser as carving-in mental health coverage (that is, benefits provided by the same organization that provides the rest of the health benefits), it would count as a carve-out in the Open Minds survey. Prevalence estimates are relatively sensitive to such definitional issues. For example, if stand-alone utilization review and case management services are excluded from Open Minds estimates, the prevalence of carve-outs drops from 66 percent to 49 percent of the insured population. 18 Prior authorization constitutes an additional organizational feature of managed behavioral health care used to control utilization. Prior authorization rules require physicians to obtain approval from a health plan or a carveout prior to prescribing treatment. In 2002 KFF/HRET results indicate that half of all covered workers were required to obtain authorization prior to using outpatient specialty mental health services. Prior authorization for these services was significantly more common (at the.05 level) among people employed by larger firms (53 percent) than among those employed by smaller firms (42 percent). As expected, covered workers enrolled in HMOs (69 percent) and POS plans (65 percent) were significantly more likely (at the.05 level) to be subject to prior authorization for outpatient mental health services than those in conventional (33 percent) and PPO (35 percent) plans. Prior authorization was also significantly more widespread (at the.05 level) among workers enrolled in behavioral health carve-outs. Sixty-one percent of covered workers in carve-outs were required to obtain prior authorization overall, compared with 43 percent of workers with integrated mental health coverage. Discussion Virtually all insured workers were covered for mental health in 2002, and differential cost sharing appears to be less prevalent than in the past. However, a large majority of workers HEALTH AFFAIRS ~ Volume 22, Number 5 133

8 Health Tracking continue to be subject to inpatient day and outpatient visit limits. Overall, the KFF/HRET findings on mental health benefit design suggest that parity laws have not broadened access to mental health services as advocates intended. Limits under managed care. Furthermore, despite the rise of managed care and behavioral health carve-outs, employees still carry a greater out-of-pocket cost burden if they use mental health services in comparison with other illnesses. HMOs and carve-outs are thought to transform how firms manage mental health costs by using supply-side utilization controls rather than relying solely on cost sharing and benefit limits to lower demand. As a result, mental health benefits could be expanded, even as firms maintained control over costs. Thus, we might expect to find that people covered under HMOs or carve-outs would face fewer nominal benefit limits than those in traditional plans face. However, our results suggest that nominal benefits under HMOs and carve-outs tend to be at least as restrictive as conventional plans. Similarly, in their 19 employer survey, David Salkever and colleagues found that HMOs and carve-outs were more likely to use some types of benefit limits for mental health care compared with traditional insurance arrangements. 19 These researchers speculated about why employers might offer managed care products while continuing to impose special limits and substantial cost sharing on mental health. They hypothesized that lack of employer education about the cost advantages of behavioral health care management, minimal risk sharing under many carve-out contracts, or a single-minded focus on cost containment could explain resistance to eliminating these restrictions. Likewise, it seems plausible that adverse-selection incentives could play a role in explaining the endurance of benefit limits. While the advent of managed care has attenuated fears that coverage expansions would exacerbate cost control problems, benefit restrictions could be motivated by a health plan s desire to avoid enrollees with a propensity to avail themselves of mental health care. Mix of benefit limits. These findings also indicate that the mix of benefit limits has changed over time. Nearly three-quarters of covered workers were subject to mental health service limits in 2002, while fewer than onefourth had higher cost sharing. Decreased use of differential cost sharing could be attributable in part to managed care. However, the persistent use of day and visit limits is particularly noteworthy, given evidence from simulations suggesting that this design feature is not as useful in controlling use and spending. 20 Given this, whether day and visit limits continue to serve any function beyond facilitating adverse selection merits careful consideration. Growth in drug coverage and spending. The one mental health benefit consistently on a par with general medical insurance is coverage for prescription drugs. The percentage of insured workers with prescription drug coverage has expanded over time, from 93 percent in 18 to 99 percent in During our study years, cost sharing for prescription drugs declined. About 70 percent of workers in 1991 had to meet a deductible. By 2002 only 60 percent faced a deductible, and the average deductible was statistically equivalent to the 1991 figure. 22 Outpatient psychotropic drugs now constitute the fastest-growing cost in mental health care. According to a recent study by Tami Mark and Rosanna Coffey, prescription drug use grew from 22 percent of total behavioral health care spending in 1992 to 48 percent in 1999 among people with employer-based health insurance. 23 Growth in the use of psychotropic drugs is attributable not only to the comparatively generous coverage for this portion of the mental health benefit, but also to the increased availability of effective medications to treat mental disorders. Drug treatment options for most major mental illnesses have grown substantially over the past ten years, and the delivery system for dispensing psychotropic drugs also has expanded. Psychiatrists now prescribe only a third of psychotropic medications; primary care physicians and other medical specialists prescribe the remaining two-thirds. 24 However, it is also not surprising that mental health 134 September/October 2003

9 spending has grown most in the area where special benefit restrictions are absent. Importance of carve-outs. Finally, data reviewed in this paper illustrate the important institutional role of managed behavioral health carve-outs. Given this industry s rapid evolution over the past decade, it is perhaps not surprising that a dominant mechanism for systematically tracking and analyzing change within this sector has yet to emerge. We have tried to triangulate the various data sources on carve-out market penetration with an eye to developing a more lucid picture of the role of this sector in organizing mental health coverage. Clearly, definitional inconsistencies and measurement limitations explain much of the variation in estimates across surveys. It will be crucial for the mental health field to work toward developing consistent methods of measurement, given increasing reliance on carve-outs. Study limitations. A number of limitations in using KFF/HRET survey data to assess mental health coverage are worth noting. The survey did not examine certain aspects of benefit design such as special deductibles and dollar limits for mental health care. However, Buck and colleagues found that plans rarely use mental health deductibles, and special dollar limits on mental health are now prohibited under the 19 federal parity law. 25 The survey also does not provide information on certain components of the mental health benefit solely used in the context of treating mental conditions such as partial hospitalization, short-term mental health crisis intervention services, and residential treatment. These specialty services are seldom discussed in the context of parity-related benefit expansion, yet they constitute an important component of evidence-based treatment. The fact that parity laws are narrowly tied to medical benefits means that the laws do not recognize certain services that are uniquely geared to the treatment of mental conditions. Thus, because of the limited scope of mental health benefit information included in the KFF/HRET survey, there may be important differences across insurance products that remain undetected in our study. For example, prior research indicates that carve-outs could be more likely to include coverage for partial hospitalization. 26 Continued benefit limits combined with powerful managed care techniques to control utilization help explain reduced growth in mental health spending compared with health spending overall; they also could signal access problems. Mental health advocates have worried that strict benefit limitations expose people with serious mental illnesses to catastrophic expenses. To date, the scope of parity laws might be too limited to greatly reduce the disparity between mental health benefits and general medical benefits. For example, a U.S. General Accounting Office (GAO) study found that when the 19 federal parity law eliminated the use of mental health specific dollar limits, 87 percent of employer plans complying with the law had at least one other benefit design feature differentially limiting coverage for mental health in their benefit package. 27 Also, about two-thirds of compliant employers changed at least one other mental health benefit design feature to be more restrictive. Likewise, the majority of state parity laws apply only to a subset of mental health benefit design features or only to certain diagnoses. Legislation pending in Congress would broaden the existing federal parity law to include a prohibition on day and visit limits and higher cost sharing for mental health. This law is patterned on the parity regulation mandated in 1999 for federal employees. The findings of our study suggest that in the absence of broader parity laws, progress toward benefit parity could prove difficult to achieve. The authors gratefully acknowledge grant support from the John D. and Catherine T. MacArthur Foundation. Colleen Barry received training grant support from the National Institute of Mental Health (T32 MH133-08). They are also grateful to Howard Goldman, Don Metz, and two anonymous reviewers for constructive comments on an earlier draft of this paper. HEALTH AFFAIRS ~ Volume 22, Number 5 135

10 Health Tracking NOTES 1. See, for example, U.S. Bureau of Labor Statistics, National Compensation Survey: Employee Benefits in Private Industry in the United States (Washington: BLS, 2002); M. Oss et al., Open Minds Yearbook of Managed Behavioral Health and Employee Assistance Program Market Share in the U.S., (Gettysburg, Pa.: Open Minds, 2003); E.L. Merrick et al., Benefits in Behavioral Health Carve-Out Plans of Fortune 500 Firms, Psychiatric Services 52, no. 7 (2001): ; J.A. Buck et al., Behavioral Health Benefits in Employer-Sponsored Health Plans, 19, Health Affairs (Mar/Apr 1999): 67 78; D.S. Salkever, J. Shinogle, and H. Goldman, Mental Health Benefit Limits and Cost Sharing under Managed Care: A National Survey of Employers, Psychiatric Services 50, no. 12 (1999): ; and G.A. Jensen et al., Mental Health Insurance in the 1990s: Are Employers Offering Less to More? Health Affairs (May/June 19): BLS, National Compensation Survey: Employee Benefits in Private Industry in the United States, Bulletin 2555 (Washington: BLS, 2000). 3. See Buck et al., Behavioral Health Benefits. The KFF/HRET survey defines large firms as those with more than 200 covered lives, while Buck and colleagues define large firms as those with 500 or more employees. 4. Jensen et al., Mental Health Insurance. 5. Based on the authors calculations of data from BLS, National Compensation Survey, on the percentage of employers with medical care benefits who had coverage for mental health care from 10 to Assuming that don t know responses would follow a comparable distribution, we conducted a second analysis on limits excluding these responses. We found a statistically significant increase in the proportion of covered workers subject to outpatient visit limits from 78 percent in 1999 to 86 percent in On the inpatient side, 76 percent of covered workers had limits in 1999 compared with 78 percent in 2002; however, this increase was not statistically significant. 7. See the following BLS publications: National Compensation Survey: Employee Benefits in Medium and Large Firms, 11, Bulletin 2140 (Washington: BLS, 12); 12, Bulletin 2176 (13); 13, Bulletin 2143 (14); 14, Bulletin 2262 (15); 15, Bulletin 2262 (16); 16, Bulletin 2281 (17); and 18, Bulletin 2236 (19). 8. For example, in 1993, 53 percent of enrollees in medium and large firms paid a separate coinsurance rate for mental health services, and 18 percent paid a separate copayment. By 19, 36 percent paid a separate coinsurance rate, and 30 percent, a separate copayment. See BLS, National Compensation Survey: Employee Benefits in Medium and Large Firms, 1993, Bulletin 2456 (Washington: BLS, 1994); and National Compensation Survey: Employee Benefits in Medium and Large Firms, 19, Bulletin 2517 (Washington: BLS, 1999). 9. The 2002 BLS employee benefit survey results are not yet available to directly compare with our KFF/HRET findings. However, the 2000 BLS survey (not limited to medium and large firms) also reported lower levels of special cost sharing, with only 20 percent of enrollees paying a separate coinsurance rate and 30 percent paying a separate copayment for mental health services. 10. This issue is further complicated by the fact that firms with fewer than fifty workers are exempt from the federal parity law. Thus, some small firms still could retain annual and lifetime dollar limits and therefore could be less likely to rely on service limits and high cost sharing to control mental health service use. 11. See, for example, C.A. Ma and T.G. McGuire, Costs and Incentives in a Behavioral Health Carve-Out, Health Affairs (Mar/Apr 19): 53 69; W. Goldman, J. McCulloch, and R. Sturm, Costs and Use of Mental Health Services before and after Managed Care, Health Affairs (Mar/Apr 19): 40 52; J.B. Christianson et al., Utah s Prepaid Mental Health Plan: The First Year, Health Affairs (Fall 1995): ; S.M. Essock and H.H. Goldman, States Embrace of Managed Mental Health Care, Health Affairs (Fall 1995): 34 44; and R.G. Frank, T.G. McGuire, and J.P. Newhouse, Risk Contracts in Managed Mental Health Care, Health Affairs (Fall 1995): See, for example, S.H. Busch, Specialty Health Care, Treatment Patterns and Quality: A Case Study of Treatment for Depression, Health Services Research 37, no 6 (2002): ; E.L. Merrick, Effects of a Behavioral Health Carve- Out on Inpatient-Related Quality Indicators for Major Depression, Medical Care 37, no. 10 (1999): ; W.G. Manning et al., Outcomes for Medicaid Beneficiaries with Schizophrenia under a Prepaid Mental Health Carve-Out, Journal of Behavioral Health Services Research 26, no 4 (1999): ; B. Dickey et al., Managing the Care of Schizophrenia: Lessons from a Four-Year Massachusetts Medicaid Study, Archives of General Psychiatry 53, no. 10 (19): ; and J.J. Callahan et al., Mental Health/Substance Abuse Treatment in Managed Care: The Massachusetts Medicaid Experience, Health Affairs (Fall 1995): Merrick et al., Benefits in Behavioral Health Carve-Out Plans. 14. Oss et al., Open Minds Yearbook. 136 September/October 2003

11 15. While not directly reported in the Open Minds 2003 published market share survey, a response rate of 80 percent was provided through personal communication with Open Minds staff. For the remaining 20 percent of MBHOs that did not respond to the survey, enrollment figures were estimated on the basis of prior-year data, publicly available information, and interviews. 16. Open Minds uses a denominator of 250 million Americans with health insurance, citing the U.S. Census Bureau Current Population Survey March Supplement in 2000, to develop its estimate. 17. Open Minds reported that the mix of services offered by these programs has shifted somewhat. In 2002 risk-based contracts accounted for 36 percent of carve-out enrollment, with non-riskbased contracts representing 28 percent. Likewise, stand-alone utilization review/case management and integrated managed behavioral health care/eap constituted 26 percent and 10 percent of the market, respectively. Contracts solely for services provided by EAPs were not included in this count. Over the past year enrollment in risk-based, nonrisk, and integrated managed behavioral health care/eap carve-out contracts has declined slightly, while utilization review/case management program enrollment has risen. 18. Again, this estimate uses a denominator of 250 million Americans, as in Note Salkever et al., Mental Health Benefit Limits. 20. R.G. Frank, H.H. Goldman, and T.G. McGuire, A Model Mental Health Benefit in Private Health Insurance, Health Affairs (Fall 1992): J.R. Gabel, Job-Based Health Insurance, 17 19: The Accidental System under Scrutiny, Health Affairs (Nov/Dec 1999): 68; and Henry J. Kaiser Family Foundation/Health Research and Educational Trust, Employer Health Benefits: 2002 Annual Survey (Menlo Park, Calif.: Kaiser Family Foundation, 2002), Authors calculations from C.B. Sullivan et al., Employer-Sponsored Health Insurance in 1991, Health Affairs (Winter 1992): 181; and KFF/HRET, Employer Health Benefits: 2002 Annual Survey, T.L. Mark and R.M. Coffey, What Drove Private Health Insurance Spending on Mental Health and Substance Abuse Care, ? Health Affairs (Jan/Feb 2003): H.A. Pincus et al., Prescribing Trends in Psychotropic Medications: Primary Care, Psychiatry, and Other Medical Specialties, Journal of the American Medical Association 279, no. 7 (19): See Buck et al., Behavioral Health Benefits. 26. Salkever et al., Mental Health Benefit Limits. 27. U.S. General Accounting Office, Mental Health Parity Act: Despite New Federal Standards, Mental Health Benefits Remain Limited, Pub. no. GAO/HEHS (Washington: GAO, May 2000). HEALTH AFFAIRS ~ Volume 22, Number 5 137