|Home | About | Journals | Submit | Contact Us | Français|
Congress is considering enactment of comprehensive parity legislation. The intent of parity is to equalize private coverage of behavioral and general medical care, thereby improving efficiency and fairness in insurance markets. One issue is whether to extend parity to substance abuse (SA) benefits. In the past, inclusion of substance abuse has been a hurdle to passage of parity. We examine the politics of SA parity, compare coverage trends for substance abuse and mental health, and assess the rationale for equalizing benefits. We conclude that the justification for SA parity is as compelling as it is for mental health parity.
The aim of parity regulation is to equalize private insurance coverage for behavioral health and general medical care. Under private insurance, substance abuse (SA) and mental health (MH) benefits are offered on a more limited basis than other services are.1 Parity has long been viewed as a means of improving efficiency by reducing insurers’ ability to select for risk and by increasing fairness in the employer-based health insurance market.
Consideration of comprehensive federal parity in 2007 constitutes the most recent episode in a decade-long effort to expand a narrow parity law enacted by Congress in 1996. The 1996 Mental Health Parity Act (P.L. 104-204) prohibited the use of special annual and lifetime dollar limits on coverage for MH services but did not apply to other benefit limits (for example, annual inpatient day and outpatient visit limits, higher cost sharing, and special deductibles) and did not extend to SA services. The Government Accountability Office (GAO) found that health plans circumvented this law.2 Attempts have since been under way to pass a stronger bill. In 2002, President George W. Bush expressed support for equalizing coverage.3 With the return of the Democrats to power in Congress in 2006, proponents of comprehensive parity face their best chance of enactment in a decade.
One area of contention has been whether to include treatment for SA problems under a more comprehensive parity law. The term SA problems refers to both abuse of and dependence on alcohol and illicit drugs based on Diagnostic and Statistical Manual, Fourth Edition (DSM-IV), criteria.4 In this paper we examine the politics of parity and compare trends in private MH/SA insurance coverage. We then examine whether the same rationale for equalizing benefits for mental health applies to substance abuse.
The 2007 comprehensive parity debate in Congress constitutes the most recent effort to regulate the scope of MH and sometimes SA benefits. Beginning in the 1970s, states enacted mandated benefit laws establishing minimum coverage levels for MH/SA services under private insurance. Data from the Blue Cross Blue Shield Association indicate that forty-four states passed alcoholism treatment mandates, thirty-one states passed drug abuse mandates, and thirty-four passed MH mandates.5 By the 1990s, legislative efforts shifted to advocating for parity. By requiring equity in coverage, MH advocates framed parity explicitly in terms of anti-discrimination and fairness. More than forty states have enacted parity laws, but these policies vary in scope, and most are not comprehensive.
Inclusion of SA benefits has been a hurdle to the passage of comprehensive parity legislation in the past. Over the years, most comprehensive parity bills introduced in the Senate have applied only to MH benefits. House sponsors have tended to support legislation that is more expansive. Yet during a debate over comprehensive parity in 2002, House sponsors dropped the SA provision from their bill to increase the likelihood of success by putting forward an identical MH-only bill in both chambers of Congress.
Similar wrangling over inclusion of SA provisions has occurred in the states. Only a subset of state parity laws include provisions regulating the scope of SA services (Exhibit 1). Even when substance abuse is included in a state’s parity law, the language tends to be more restrictive than is the case for mental health. In the 2000 Massachusetts law, for example, SA parity applied only to the subset of enrollees with a co-occurring mental illness. SA provisions were sometimes added to existing parity laws by means of expansion legislation (Exhibit 1). However, the reverse has also occurred. For example, in 2004 the West Virginia legislature repealed the alcohol provision included in the state’s original 2002 parity law.
Political progress on SA parity has been hampered by both low public support and the absence of well-organized advocacy groups. Public opinion polls indicate that addiction does not generate as much compassion as mental illness does. Less sympathetic views result in part from societal ambivalence over whether to regard SA problems as medical conditions to be treated or personal failings to be overcome. More so than with mental health, addiction is often viewed as a moral failing, and the illegality of drug use reinforces this perspective. Social stigma associated with addiction remains strong, and poor behavior accompanying substance use (for example, crime and impaired driving) heightens society’s condemnation. In her public opinion review, Kristina Hanson cited a Mellman-Lazarus-Lake survey indicating that only 16 percent of respondents strongly favored inclusion of SA coverage in a basic insurance package under national health reform, compared with 29 percent favoring MH coverage.6 Two-fifths characterized SA coverage as less important than covering services for physical problems, compared with only about one-fifth describing MH coverage as less important.
Likewise, there is less interest-group support for SA parity than for MH parity. Only two of the fifty-two groups affiliated under the MH Liaison Group (the Association for Addiction Professionals and the National Association of State Alcohol and Drug Abuse Directors) focus primarily on substance abuse. This coalition, founded in 1969, has been a key force in advocating for comprehensive federal parity.
The political strategy on SA parity shifted in 2007 with the return of the Democrats to control of Congress. A decision was made to include SA benefits in comprehensive parity legislation introduced in the 110th Congress. A headline-catching debate over two competing father-son versions of the federal bill has brought attention to this issue.7 The House bill (H.R. 1424), sponsored by Rep. Patrick Kennedy (D-RI) and Rep. Jim Ramstad (R-MN) would apply to all components of the MH and SA benefit design (such as cost sharing, deductibles, and special inpatient day and outpatient visit limits). This bill, modeled on a parity directive implemented in the Federal Employees Health Benefits (FEHB) program in 2001, would cover all medically necessary conditions listed in the DSM-IV. The Senate version (S. 558), sponsored by Sen. Ted Kennedy (D-MA), Patrick Kennedy’s father, is similar to the House bill except that it would not specify the diagnoses that must be covered, leaving this decision to insurers. Senator Kennedy described the Senate approach as more pragmatic because of its inclusion of compromise language developed in discussions with insurance and business groups concerned about the cost of parity. It is unclear how equity in coverage for SA services might fare under the Senate bill. In the context of lower public support and weaker organized interests, advocates express concern that granting insurers discretion would lead to unequal coverage for SA services.
On 18 September 2007 the Senate passed its version of parity by unanimous consent. While Rep. Kennedy expressed a preference to wait until after the 2008 election to push for the House version of the bill, he acknowledged that mental health parity advocates and members like Sen. Pete Domenici (R-NM), who have pushed for comprehensive parity for years, would resist waiting until 2009.8 Depending on the outcome in the current session of Congress, individual states may wish to take additional action.
According to the National Survey on Drug Use and Health, about twenty-two million Americans (9 percent of the population age twelve and older) had a substance problem in 2005.9 Of those, four million were dependent on or abused illicit drugs, fifteen million were dependent on or abused alcohol, and three million were dependent on or abused both.
Financing of substance dependence and abuse treatment services occurs through a patchwork of public and private sources.10 Private health insurance constitutes an important (albeit declining) source of financing for treatment of SA problems. Only a quarter of total health care spending to treat substance abuse is paid by private insurance, compared with about 32 percent in 1987.11 Private benefits include pharmacological, inpatient, outpatient, and residential care to combat drug and alcohol abuse and dependence.12 From 1992 to 2001, use of any SA services among those with employer-based health insurance declined by 23 percent; reductions occurred across the board in the inpatient, outpatient, and pharmaceutical sectors.13 A third of payments for drug abuse treatment and almost half of payments for alcohol treatment were paid out of pocket by consumers in 2004.14
Reduction in the scope of SA benefits under private insurance may, in part, explain declines in service use. Although the overwhelming majority of the privately insured are offered some SA coverage, most insurers impose special limits on these services, and use of these limits appears to be increasing over time.15 Data from the U.S. Bureau of Labor Statistics (BLS) indicate that in 2003, 94 percent of workers in private industry had some level of health insurance coverage for inpatient alcohol and drug detoxification, 85 percent had some inpatient alcohol and drug abuse rehabilitation coverage, and 88 percent had some outpatient alcohol and drug abuse coverage. Availability of private coverage has greatly increased since the early 1980s, when only 37 percent of full-time workers had any coverage for drug abuse and 50 percent had coverage for alcohol abuse treatment.16
BLS trend data also indicate that the use of limits on SA benefits has increased over time. In 1988, among those with private SA coverage, only 49 percent were subject to special limits on outpatient alcohol services, and 46 percent, on outpatient illicit drug abuse services. Similarly, in 1988, less than 60 percent of insured workers were subject to limits on inpatient alcohol and drug rehabilitation services.17 By 2002, 89 percent of workers were subject to special limits on outpatient alcohol abuse services, and 91 percent, on outpatient illicit drug abuse services.18 Ninety percent were subject to limits on inpatient alcohol and drug rehabilitation services. Comparable increases in restrictions on inpatient detoxification were observed over this period.
Consistent with this, a survey by the Henry J. Kaiser Family Foundation and Health Research and Educational Trust (HRET) reported that average inpatient deductibles and median outpatient cost sharing for substance abuse were much higher than for general medical care in 2006.19 Higher deductibles and cost sharing reduce the likelihood that a substance abuser will initiate treatment.
In 1988, less than 25 percent of workers were subject to special annual outpatient visit limits for alcohol and drug abuse services, and 36 percent of workers had outpatient MH limits (Exhibit 2). By 2002, almost 70 percent of workers had insurance coverage that included such limits on alcohol and drug abuse services, and 75 percent had these limits on mental health treatment.
Dollar limits on both MH and SA services are the only categories of restrictions that appear to be used less frequently by insurers than in the past. In 1997, BLS data show that more than 50 percent of workers were subject to special annual and lifetime dollar limits for alcohol and drug abuse coverage (Exhibit 2). By 2000, only a small proportion of workers had dollar limits on coverage for these services. This trend is also observed for MH coverage. The elimination of dollar limits on MH coverage was required by law under the federal partial parity law implemented in 1998. Although the law did not apply to SA coverage, BLS data indicate a substantial decline in the use of SA dollar limits after implementation.
In summary, data on benefit design trends collected by the BLS over the past twenty-five years suggest that although a much higher proportion of privately insured families have access to some coverage to treat SA problems, the scope of these benefits has grown more restrictive. Furthermore, coverage levels for MH and SA benefits are similar. One explanation for this similarity may be expediency. Less than 0.5 percent of those with employment-based health insurance use any SA services, and these services constitute only about 0.20 percent of total health care spending in the private market.20 Therefore, it may be administratively simpler for payers to treat SA benefits similarly to MH benefits.
Here we examine whether the rationale for equalizing benefits for MH treatment applies to SA treatment.
Similarities along key dimensions may justify a consistent approach to regulating private coverage for both MH and SA services.
First, these disorders often accompany each other. The recently released National Epidemiological Survey on Alcohol and Related Conditions showed that among those with a twelve-month mood disorder, 20 percent have an SA disorder, and among those with a twelve-month anxiety disorder, 15 percent have an SA disorder.21 In the absence of treatment, co-occurrence may be attributable in part to self-medicating by use of substances.
Second, both disorders impose costs on society.22 Beyond direct treatment costs, alcohol use and psychiatric disorders ranked among the ten leading causes of disability worldwide in 1990.23 As does having a mental illness, being a heavy drinker or dependent on illicit drugs has been shown to lower earnings and reduce the likelihood of being employed.24 Substance abuse in particular confers major negative externalities including those associated with driving impaired, transmitting communicable diseases through unprotected sex, and committing crimes (although the latter may be less of an issue in a privately insured population).25
Third, treatment of SA and MH disorders still occurs largely in specialty sectors. In 2001, 84 percent of total SA spending and 53 percent of total MH spending occurred in specialty sectors.26 New evidence-based SA treatments administered in offices of generalist physicians (for example, buprenorphine) are being developed, but specialized clinics are likely to remain the primary site of care.27
Fourth, large gains have been made in advancing the evidence base for treating SA and MH conditions, although both sectors face challenges in translating these advances to routine care.28 Clinical trial and observational studies have demonstrated the efficaciousness of a range of pharmacological (for example, methadone, disulfiram, and buprenorphine) and outpatient (such as cognitive behavioral therapy, family education, and brief interventions) treatments for SA problems.29 Studies are beginning to demonstrate the cost-effectiveness of treatment as well.30 Likewise, the 1999 surgeon general’s report on mental health concluded that evidence-based treatments were available to treat most MH-related diagnoses.31
Fifth, a large share of those with SA and MH conditions do not receive treatment at all or receive inadequate care.32 It is estimated that only 10–17 percent of those who need SA treatment receive specialty care.33 Among adolescents, only about 9 percent of those classified as needing specialty treatment for illicit drug use and 7 percent needing alcohol treatment receive it.34 Rates of service use among those with an MH diagnosis are also low.35 The problem of unmet need is attributed in part to stigma and the marginalized role of these groups in society. Likewise, many in treatment do not receive appropriate care. Elizabeth McGlynn and colleagues found that those with medical records indicating alcohol dependence received recommended care 10 percent of the time, and patients treated for clinical depression received recommended care 58 percent of the time.36
A key role of health insurance is to protect consumers against high medical expenses. As noted above, SA and MH coverage takes the opposite form and exposes enrollees to potentially large costs of illness by limiting benefits. Although consumer groups view benefit limits as evidence of discrimination, economic principles also explain them.37 Insurers use benefit design to combat moral hazard and protect against adverse selection. Comprehensive parity, in turn, has been offered as a regulatory response to protect consumers against the adverse consequences of benefit limits. These issues have been explained with regard to mental health; we compared their applicability to substance abuse and found the economic arguments for parity to be as strong for SA as for MH benefits.
Moral hazard is one reason why insurance coverage is more limited for MH conditions than for general medical conditions. Moral hazard refers to people’s tendency to use more health services when they face a reduced out-of-pocket price because of health insurance than when they face the full price. The RAND Health Insurance Experiment (HIE) provided evidence that consumers have a greater demand response to coverage of outpatient MH services than general medical services.38 The elasticity for ambulatory mental health care was estimated to be more than twice that for general medical care. Increased use of services in response to coverage expansion can represent a misuse of society’s resources. That is, consumers use treatment until the value to them is less than the value of the resources to society at large. The limits on MH services were developed in part to reduce this loss.
It is important to point out that the demand response noted in the RAND HIE may no longer be a valid justification for discrepancies in coverage for care in the era of managed care. Managed care uses supply-side rationing as an alternative to benefit rationing to control health care costs. Instead of influencing demand for health care services through consumer cost sharing, managed care limits access by the use of provider payment incentives, preferred provider networks, utilization management, and other mechanisms that affect the supply of medical care.
The RAND HIE did not study the question of demand response for SA services. However, some characteristics of alcohol and drug addiction may naturally lead to less demand response to insurance coverage, and therefore less concern about moral hazard even in an indemnity insurance context. For example, some modes of SA treatment (such as rapid opiate detoxification) are physically unpleasant and therefore unlikely to be overused. Also, those with drug and alcohol problems are often in denial about their problems and do not seek treatment. Social stigma may also be a barrier to seeking care.
In addition, workers may be concerned about the repercussions of an employer’s becoming aware of an addiction and might prefer to pay for treatment costs out of pocket to avoid using employer-based coverage. Workers with illicit drug abuse problems may be particularly wary of detection by employers, since this form of substance use is illegal and is not protected by workplace laws. Fear of workplace consequences would not be at issue in the context of dependent or spousal treatment.
Given that only a small proportion of substance abusers seek treatment for their addictions, insurance expansion alone is unlikely to prompt large increases in demand for services. Privately insured people spent less than $6 annually, on average, for SA coverage in 2001.39 Therefore, moving to parity-level coverage would likely not appreciably change insurance premiums. Roland Sturm and colleagues estimate, for example, that removing an annual limit of $10,000 per year on SA treatment would increase private insurance payments by about six cents a year under managed care.40 Finally, as with mental health, supply-side rationing presumably plays a role in controlling the use of SA services under managed care.
Evidence on demand response to expansion of both MH and SA benefits in a managed care environment comes from an evaluation of the effects of comprehensive parity in the FEHB program.41 When a comparison group was used to control for secular trends, after parity, the probability of using SA services increased in four of nine plans studied and did not change in the other plans.42 In contrast, the FEHB parity directive increased MH services use slightly in only one plan. FEHB parity had no effect on SA spending per user in eight plans, nor did it lead to an increase in per user spending for MH services. Another evaluation of a state parity law found a reduction in use of SA services after parity.43 Thus, findings on the sensitivity of service use to SA benefit changes are mixed. In sum, we do not expect moral hazard to be a pressing concern for SA parity, and, given the sizable externalities imposed on society by SA, some increased use may be beneficial.
A second reason why insurers limit coverage is adverse selection. Adverse selection occurs in health insurance markets when potential enrollees know more about their health risk profiles than competing health plans do and when plan premiums do not fully capture these risk differences. Because private insurance is often obtained as a fringe benefit at work, premiums do not reflect people’s systematically greater use of services. Those with high expected spending for themselves or their dependents can select more-generous coverage without paying its full expected cost. Selection incentives will be strongest with conditions that are expensive and predictable (chronic), such as mental illness.44 Those with MH conditions also tend to use other general medical services at higher rates. Such enrollees will have a propensity to opt for plans with more-generous coverage. Therefore, health plans can achieve favorable selection by limiting MH benefits to discourage those with MH problems from enrolling. A rationale for comprehensive parity regulation is to counteract wasteful competition among plans to underprovide services because of selection fears.45
We consider whether the same concern about adverse selection that motivates MH benefit limits applies to SA benefit limits. A worker with an SA disorder might not perceive a need for treatment; therefore, the generosity of SA benefits might not be a key determinant when choosing among jobs or health plan options. Consistent with this, Katherine Harris and Sturm did not find evidence of selection into plans with more generous alcohol treatment benefits.46
However, concern about adverse selection seems warranted because of certain characteristics of substance abuse. SA conditions are chronic, and costs are often high for persons in treatment. Sturm and colleagues found a much larger portion of high-cost users in private SA treatment than in all behavioral health care.47 Although two-thirds of MH treatment costs were for outpatient care, about 70 percent of SA costs were for more costly intermediate forms of treatment. The authors concluded that the typical treatment mode for substance abuse is more costly per unit of service than the typical treatment mode for mental health. Therefore, although the total costs to the employer of SA treatment are low, costs for individual users can be high.
In addition, employers concerned about poor work performance or illegal activity at the workplace have a strong motivation to discourage hiring people with alcohol and drug abuse problems. If SA insurance benefits do have signal value, offering limited coverage may be one strategy to avoid problem workers. Limiting SA coverage confers little risk from the employer’s perspective, since substance-abusing workers are unlikely to advocate for broader benefits. Furthermore, opportunities to select for risk might be greater, given that insecure labor-market attachments lead to more frequent insurance changes among substance abusers.
In 2007, whether to require equal benefits to treat all medically necessary DSM disorders, including substance abuse, distinguishes House and Senate comprehensive parity bills. Given lower public support for SA coverage, advocates have expressed concern that granting insurers discretion will lead to unequal coverage for SA services. In our assessment, the adverse selection and fairness arguments for comprehensive parity appear to be at least as strong for substance abuse as for mental health. The commonalities between substance abuse and mental health, including similar benefit structures, stigma, and development of increasingly effective treatments, provide further justification for a consistent regulatory approach. Thus, the argument for inclusion of SA services under federal parity appears to be as compelling for substance abuse as it is for mental health.
The authors gratefully acknowledge research support from the National Institute on Drug Abuse (Grant no. R01-DA14471). Allison Brenner provided expert research assistance for this paper.