Achieving Mental Health Care Parity Might Require Changes In Payments And Competition

Abstract

One of the most prominent features of the Affordable Care Act has been the promotion of individual health plans chosen by consumers in the Marketplaces. These plans are subject to regulation and paid by risk-adjusted capitation, a set of policies known as managed competition. Individual health insurance markets, however, are vulnerable to what economists describe as efficiency problems stemming from adverse selection, and Marketplaces are no exception. Health plans have incentives to discriminate against services used by people with certain chronic illnesses, including mental health conditions. Parity regulations, which dictate coverage for mental health benefits on par with medical and surgical benefits, can eliminate discrimination in coverage but redirect discrimination toward hard-to-regulate tactics from managed care such as restrictive network design and provider payment. This article reviews policy options to contend with ongoing selection issues. “Better enforcement” of parity has less chance of success than more fundamental but feasible changes in the way plans are paid or in the way competition among plans is structured.

In Better but Not Well, Richard Frank and Sherry Glied concluded that in the fifty years between 1950 and 2000, “substantial progress” had been made in access to care, financial protection, and meeting basic needs of people with mental illnesses in the United States.¹ Changes in the financing and organization of mental health care, not new treatment technologies, made the difference: “Improvements…evolved through…more money, greater consumer choice, and the increased competition among technologies and providers that these forces unleashed.”¹(p4) Furthermore, “people with mental illnesses of all types have their medical care paid for by the same organizations that pay for physical health care. Mental health care has been mainstreamed.”¹(p5) According to Frank and Glied, then, two broad changes in the financing of care—the introduction of markets and choices and integration of mental with physical health care—were responsible for improving the well-being of people with mental illnesses in the second half of the twentieth century. While full integration remains aspirational, the trends identified by Frank and Glied continue to move mental health into the mainstream of health care financing and organization.

In the shorthand of health policy, health insurance for mental health care in some sectors has been folded into the “managed competition” model espoused by Alain Enthoven.² Managed competition features individual choice among private health plans competing on the quality of coverage and the premium charged to enrollees. Coverage and premium categories are regulated. A public authority collects funds in the form of premiums or taxes and makes risk-adjusted capitated payments to plans. Ideally, covering mental health care at parity with other coverage for medical and surgical benefits in comprehensive insurance plans offered by competing health plans creates the markets, choices, and integration of mental and physical health care lauded in Better but Not Well.

The most prominent example of the principles of managed competition is the state Marketplace concept created as part of the Affordable Care Act (ACA). Other important sectors including Medicare Advantage, the private-plan alternatives to traditional Medicare, and state Medicaid managed care programs also feature elements of managed competition. The major sectors in which managed competition is not followed are traditional Medicare, which is a form of social insurance, and employer-sponsored health plans, in which the employer preselects one or a small set of plans and pays plans at negotiated prices without formal risk adjustment.

This article is concerned with the part of the story from Better but Not Well that deals with health insurance within the managed competition paradigm. Mental health and substance use disorders are prevalent and costly, and health insurance is needed to provide financial protection and secure access to care. In terms of Better but Not Well, this article asks, is managed competition the right policy prescription for the lingering weaknesses in the organization and financing of services for mental health and substance use disorders? If so, what directions should policy management take? If not, what is the alternative? Focus is on the Marketplaces—the epitome of managed competition, and the newest and most fluid health insurance sector.

The article first reviews the problem of equal coverage and access for mental health and substance use disorders with competing managed care plans within the managed competition environment. The second part considers steps beyond benefit regulation to move toward full equality in access for people with these conditions.

Parity and ACA policies have largely succeeded in equalizing coverage, in terms of deductibles, copayments, and limits, for mental health and substance use disorders compared to coverage for medical and surgical conditions in managed competition. While nominal coverage has been equalized, because of underlying forces of adverse selection working against enrollees with certain chronic illnesses, including mental health and substance use disorders, equal access has not been secured.

Managed Care And Coverage For Mental Health And Substance Use Disorders

Through several cycles of national health insurance reform, dating back decades, mental health policy has been concerned with parity in coverage for mental health and substance use disorders.³ One barrier to parity was a consensus that use of mental health care was unduly responsive to provisions of insurance and payment—the so-called moral hazard problem.⁴ Prior to the widespread adoption of managed care techniques in private health insurance in the 1990s, research in health economics—including that from the RAND Health Insurance Experiment—found that the demand response by consumers was greater for outpatient mental health care than for outpatient medical services.⁴ Later, however, when parity was implemented in the context of managed care, research found that demand response had largely disappeared,⁴ a finding that continues to hold in more recent studies.⁵ Because managed care techniques were sufficient to control spending, parity in coverage for mental health and substance use disorders could be mandated without fear of suboptimal allocation of resources or what could be termed inefficient cost increases.

Although managed care defanged moral hazard as an argument against parity, managed care did the same to parity regulation itself, equipping managed care plans with a plethora of new ways to control the use of services by enrollees—methods that are also potential mechanisms for discriminating against certain sets of enrollees or types of services. A health plan loses money on some enrollees and makes money on others. When these financial patterns are systematic and predictable, as they are in the cases of people with certain chronic illnesses, plans have strong incentives to discriminate against services used by the high-cost patients on which they lose money.⁶ Parity regulations attempt to contend with this problem, but the health plan strategies are difficult to monitor and regulate.

Parity, Essential Health Benefits, And Marketplaces

The Mental Health Parity and Addiction Equity Act of 2008 extended parity to self-insured health plans. In recognition of the managed care tools that health plans have to control usage, the law also required parity in the so-called nonquantitative treatment limitations such as medical management methods, criteria for inclusion in provider networks, and provider payment levels. Specifically, any nonquantitative treatment limitation applied to behavioral health care must be “comparable to and no more stringently applied than” the comparable limitation applied to medical and surgical care. The Parity Act did not, however, actually require mental health and substance use disorder coverage; rather, it required only the provision of parity in existing coverage. ACA regulations went further than the Parity Act and required plans to cover “essential health benefits” including services for mental health and substance use disorders. The Department of Health and Human Services has estimated that the ACA and the Parity Act will extend parity protection to sixty-two million Americans.⁷

Mental health spending as a share of total health spending has fallen steadily in the past twenty years, down from 9.4 percent of total spending in 1986 to 7.0 percent in 2014, even as parity has improved nominal benefits.⁸ It should not be surprising that parity regulations would do little to increase total spending on mental health care. Parity regulations target services supplied by specialty mental health providers, but as of 2009, half of all patients treated for a mental health disorder received only drug therapy and did not access care from a specialty provider.⁹ Even those patients who received nondrug therapy from a provider often received professional care in a primary care setting. Thus, the majority of patients treated for a mental health or substance abuse disorder in outpatient settings would be little affected by parity in coverage for mental health specialty providers.

Health plans can tailor networks, provider payment levels, management protocols, drug formularies, and other policies to attract profitable enrollees and deter those for whom revenues fall short of costs. These are policies that are difficult to monitor and regulate. A case study of two states’ implementation of parity regulations exposes difficulty in enforcement to guard against discrimination and deterrence of enrollees with mental health and substance use disorders, particularly around the nonquantitative treatment limitations.¹⁰

Paying more for enrollees expected to need various types of care is intended to encourage plans to want to enroll and serve the sick as well as the healthy. The federal risk-adjustment system planned for Marketplaces in 2017, however, would underpay for patients with cancer, diabetes, heart disease, and behavioral health conditions—and among these, the largest gap between average cost and average payment, almost 20 percent, is for behavioral health.⁶ Variables selected for the risk-adjustment formula pick up some diagnoses but miss about 80 percent of the enrollees with a mental health or substance use disorder diagnosis, as shown by Ellen Montz and colleagues.¹¹ Health plans in Marketplaces are protected by reinsurance (which pays a share of costs above a threshold such as $80,000) and risk corridors set up by the federal government to offset high losses but also share large profits. Both features dampen incentives to avoid high-cost enrollees, but both are set to be phased out after 2016.¹²

Medicaid Managed Care And Medicare Advantage

Since Frank and Glied closed the books in 2000 on their fifty-year review of mental health treatment, Medicaid’s role as the primary financier of mental health care has continued to grow. In 2009, 41 percent of beneficiaries younger than age sixty-five receiving Supplemental Security Income from Social Security qualified because of mental illness.⁹ Medicaid has accounted for nearly two-thirds of all new state mental health agency spending since 2001.¹³ Medicaid rules increased the scope-of-coverage benefit for people with mental health and substance use disorders. Targeted case management and the rehabilitation option in Medicaid gave states a way to include medication management, assertive case management, and other services for people with mental health disorders.¹⁴ States prefer Medicaid instead of direct state funding to take advantage of federal financial participation.

The picture for substance abuse is somewhat different than that for mental health. Because Medicaid has had a less generous definition of disability attributable to addiction and less generous coverage of addiction services, it has played a smaller role in funding treatment for addiction and substance use disorders compared to its role in mental health.¹⁴ The ACA might be changing this through Medicaid expansion for US citizens with income up to 138 percent of the federal poverty level and extension of federal parity rules to Medicaid managed care plans.¹⁵ Medicaid spending for substance use disorders is projected to grow to $12 billion by 2020, making Medicaid the largest US payer for addiction services.¹⁶

About 80 percent of the nation’s roughly seventy million Medicaid enrollees¹⁷ are served through managed care plans.¹⁸ New York’s Medicaid program—at 6.6 million enrollees, the largest in the country—has moved aggressively to managed care, including auto-assigning eligible enrollees to plans if they do not make an active choice.¹⁹ To partially address concerns with adverse selection related to people with mental illnesses, a number of states have “carved out” the mental health benefit to take decisions about spending on specialty mental health care out of the hands of competing health plans.²⁰ The carve-outs represent a deviation from the managed competition model. Medicaid managed care deviates from the managed competition model in another respect in that enrollees pay no premiums for the plans they choose. Thus, any competition among plans is channeled into provider networks and other elements of plan management.

Moving from how managed competition affects Medicaid, this article looks at Medicare Advantage plans—the private alternative to traditional Medicare that enrolls about 30 percent of the more than fifty million Medicare beneficiaries.²¹ Medicare Advantage must provide at least the same benefits as traditional Medicare. In terms of behavioral health, traditional Medicare has moved by regulation to close to full parity. Medicare beneficiaries are still subject to a 190-day lifetime cap on days in a psychiatric hospital.²² Implementation of the highly subsidized prescription drug coverage through Medicare Part D in 2006 broadened the scope of Medicare coverage significantly for people with mental health and substance use disorders.

Research identifies long-standing concerns for selection incentives to limit exposure to high-cost enrollees in Medicare Advantage, including results specific to people with mental health or substance use disorders.²³ The magnitude of these issues in Medicare Advantage appears to be falling as a result of improved risk adjustment and longer lock-in periods.²⁴ Better risk adjustment tightens the match between revenues and likely health care costs; longer lock-in periods prevent beneficiaries from moving into and out of plans easily in response to unanticipated health care needs. One could view Medicare Advantage as an amended form of managed competition because of the presence of the default option of traditional Medicare. Traditional Medicare serves as a refuge for any beneficiary anticipating poor service in a Medicare Advantage plan and also serves as a form of market discipline for the private plans in Medicare since they know beneficiaries can leave if their services are not satisfactory.

Regulation Is Partial And Imperfect

Regulation within the context of managed competition and markets for individual health insurance is partial and imperfect. Innovations in health insurance associated with managed care—principally, how payers structure and contract with networks of providers, and how they impose protocols on use of services (usage management protocols)—have had profound effects on the form of health insurance available to consumers, the prices consumers pay for plans, and the access they receive under the plans. Paradoxically, requiring benefit parity aggravates instead of ameliorates selection incentives for insurers to avoid these high-cost enrollees with mental health or substance use disorders because insurers are responsible for a larger share of costs. Parity in benefit design also redirects discrimination to how the provider networks are structured and how utilization is managed. In spite of risk adjustment, which should help health plans calibrate their offerings, they have incentives to discriminate against services used by people with mental health and substance use disorders (and some other chronic illnesses).

Moving Forward

Numerous alternatives can contend with the problem of selection incentives for chronic illnesses and mental health and substance use disorders in managed competition. I am not optimistic about direct regulation of contracting and network adequacy (such as the ratio of providers per enrollee). Direct regulation for parity works well with benefit design because plan benefits can be readily observed (does the deductible apply to both medical/surgical and mental health and substance use disorder costs?) and monitored. Networks and other managed care methods are difficult to measure, let alone monitor and regulate. I believe that more promise rests with changing the underlying incentives to discriminate against certain types of care, by redesigning plan payment schemes or by restructuring the form of competition.

Regulate and Monitor Networks

Network regulation must first take care not to do more harm than good. Plans have the authority to negotiate selective provider contracts. These can lead to big savings,²⁵ and plans’ authority to negotiate these contracts is essential to effective functioning of managed competition. Plans face large, powerful integrated provider organizations looking for high prices for reimbursement of services. However, if regulation of provider networks led to de facto very broad provider networks, that would undercut one effective tool the plans have for partially taming provider pricing.²⁶

While they are an essential component of efficient health plan management, plan network strategies are problematic. One basic threat is the trade-off between price (premium) and coverage.²⁷ When choosing a plan, consumers see premiums but might not understand or pay enough attention to network quality,²⁸ leading plans to offer lower-price options at the expense of provider network quality. States are attempting network regulation in the form of enrollee-to-provider ratios and distance-to-provider regulations, which attempt to make sure the patient has reasonable access, but these metrics are obviously highly imperfect.²⁹ There is no consensus on how to measure network quality—a precondition for effective regulation.

Plans also have incentives to discriminate among different services. Service use by enrollees for whom plan revenues fall short of plan costs are subject to this “service-level” selection, which can lead to inefficient restrictions on access for particular services. Regulation here is also problematic. What does it mean for there to be “parity” across services in access to providers? This can be answered in the abstract,³⁰ but theoretical concepts have not so far been translated to measurement and operational policy.⁵

Redesign Plan Payment

An entirely different strategy to address network design and other managed care techniques gives plans the incentives to build good networks for mental health and substance use disorders (and other chronic conditions). Instead of trying to specify what a good network looks like and then checking to see if it is there, the idea is to structure competition and pay plans so that they are given incentives to set up a good network in the first place. In the managed competition model, the job of creating the right incentives falls to risk adjustment of plan payments, but, as noted above,⁶ current risk adjustment schemes cannot be relied upon to solve the problem.

The first policy to consider is whether modifications of risk adjustment can do better for people with mental health and substance use disorders. In fact, the federal government has developed several risk-adjustment systems all on the platform of Hierarchical Condition Categories: one for Marketplaces, one for Medicare Advantage, and one for the Medicare Part D (prescription drug) program. All of these systems recognize only a subset of the diagnoses and claims types for purposes of risk adjustment. Montz and colleagues show that the risk-adjustment scheme used in Medicare Part D recognizes 50 percent of people with a mental health or substance use disorder, compared to the 20 percent in the current system noted above, so some clear improvements are possible even within the current framework of risk adjustment of payments to Marketplace plans.¹¹

More broadly, improving the empirical methodologies used to derive risk-adjustment weights could improve incentives for care for people with chronic illnesses. Empirical methods for estimating risk-adjustments weights—basically, ordinary least squares regressions with indicator variables for the Hierarchical Conditions Categories and some demographic characteristics—are unchanged from the 1980s. Innovations in empirical research, such as constraining risk-adjustment weights to ensure payment for groups not recognized in the risk-adjustment system³¹ or application of modern estimation methods developed for “big data” contexts, can improve plan payment incentives for people with chronic illnesses. For example, using data from the same source set to be used for updating risk adjustment in Marketplaces for 2017, Sherri Rose found that the predictive weight of major depression and bipolar illness (one of the mental health and substance use disorder Hierarchical Conditions Categories) is about 50 percent greater when predictive models are estimated with modern machine-learning methods instead of an ordinary least squares regression.³² Risk-adjustment schemes can be supplemented with risk-sharing features, such as reinsurance. By transferring funds to cover costs of expensive cases, reinsurance is highly effective at improving the fit of plan payments to predicted costs.¹²

A highly effective policy for contending with selection-related incentives for plans to avoid enrollees with mental health or substance use disorders is to write a separate contract to a specialized managed care company to manage the risk. A carve-out contract can protect funding for a clinical area such as mental health and substance use disorders.²⁰ Although carve-outs at the program level are in place for some Medicaid programs, the carve-out approach would mesh poorly with the current organization and financing of Marketplaces and Medicare Advantage.

Finally, an idea proposed in a working paper using the risk-adjustment system and data from the Netherlands is to “risk adjust to the system you want, not the system you have.”³³ The idea is this: Suppose policy makers believe that not enough funds are devoted to office-based care for certain chronic conditions. Instead of estimating a risk-adjustment model on the system with underspending for certain conditions, before estimating, move the money to where you want it—for example, by increasing spending for the target diagnoses. Then proceed to estimating the risk adjustment relative weights. Estimating weights on data from an unsatisfactory system generates incentives to recreate the existing patterns of care. Estimating weights on the preferred spending patterns conveys better incentives for plans to return those outcomes by their independent decisions. Potential enrollees with the target chronic illnesses will be more financially attractive to plans (the purpose of risk adjustment), encouraging plans to spend on the conditions to attract them.

Restructure Competition

The managed competition model calls for vigorous competition among plans as individuals choose their insurers. This is not the only way, or even the most common way, to structure competition among plans. “Payer-driven,” not “patient-driven,” competition is used by large employers to provide subsidized health insurance to their workers. In payer-driven competition, health plans compete to be among the small set of plans offered as a choice by the employer.

The arguments for limiting plan choices are analogous to the arguments for limited provider networks in managed care. Limited choice of plans dictated by payers allows them to weed out bad choices to avoid “mistakes” by consumers, especially those with cognitive impairments,³⁴ in confusing choice settings and to exercise negotiating power over prices.³⁵ A third argument for payer-driven competition is important in context of concern for network adequacy. A bad network may be like pornography: It might be hard to define, but you know it when you see it. In the polite language of contract theory, network adequacy may be “observable” but not “contractible.” A buyer’s ability to observe and choose on the basis of quality and price is the basic mechanism to enforce efficiency in markets—not regulation and monitoring. Large employers and other large buyers of health insurance put considerable effort into qualifying and contracting with health plans, a form of insurance procurement that has proved to be highly resilient.³⁶

The state Marketplace authority would play the role of the large employer. In 2014 ten state-based Marketplaces were “active” purchasers, working, like employers, to qualify health plans before allowing them to offer products on the Marketplace.³⁷ California, for example, directly negotiates prices and other plan characteristics, leveraging its power to ensure offered plans deliver on observable, not just contractible, dimensions.³⁷ Other states operate on a “clearinghouse” model closer to the managed competition philosophy that allows any health plan meeting regulatory standards to offer products. We currently lack broad-based evidence on what actions the “active” Marketplaces are taking to ensure and monitor quality insurance offerings and on their effects.

Conclusion

Health insurance markets in which individuals choose among competing private plans paid by capitation pose risks for people with mental health and substance use disorders and other chronic illnesses. Managing competition through benefit regulation and risk adjustment can move market outcomes in the right direction to be more efficient and allocate resources fairly to people with mental health disorders, substance use disorders, and some other chronic conditions. However, unless discriminatory managed care techniques are effectively prohibited by regulation or dissuaded by the right incentives, access and quality for some of these conditions will be restricted. “Consumer choice” and competition will fail to protect consumers if all competing health plans are subject to the same set of incentives that adversely affect access to care.

Current approaches to parity regulation and risk-adjustment policies do not appear to be up to the job of ensuring nondiscriminatory access in Marketplaces. “Better enforcement of parity” or “incorporating more risk adjusters” might help but not do enough. Other policy options, however, can change incentives in more than incremental ways. Within the paradigm of managed competition, redesign of payment methodologies can improve incentives to enroll and treat people with chronic illnesses. The alternative paradigm used in employer-based health insurance would move Marketplaces more toward a sponsor-managed market. Choosing among a small number of good plans is better for consumers than choosing among a large set of poor ones. Some combination of new plan payment mechanisms and more active Marketplace authorities might be necessary to keep the mental health and substance use disorder treatment system on the road to recovery.