Introduction
The magnitude of the OUD public health crisis
The United States remains in the midst of an epidemic of opioid-related overdose and mortality driven by use of heroin and synthetic opioids. In 2020, an estimated 2.7 million Americans had an opioid use disorder (OUD).Footnote 1 Opioid-related mortality has increased by more than four-hundred percent since 1999, surpassing that of traffic fatalities.Footnote 2 The COVID-19 pandemic has exacerbated these challenges, giving rise to a dramatic increase in overdose deaths. In 2020 alone, the nation witnessed over 93,000 estimated drug-related overdose deaths--the highest number of overdose deaths ever recorded in a twelve-month period.Footnote 3 In addition to the devastating personal consequences for individuals and families, the epidemic has also exacted an astounding economic toll: sequalae resulting from SUD, including unemployment, incarceration, and poverty have resulted in an economic burden of over $1 billion annually.Footnote 4
Decades of research show that evidence-based treatment for opioid misuse, including medication and psychosocial intervention, reduces the risk of relapse, overdose, and death.Footnote 5 Effective OUD treatment encompasses a broad range of services designed to address OUD, including four levels of care (outpatient, intensive outpatient, residential, and inpatient), as well as several medications approved by the Food and Drug Administration (FDA) for treatment of OU: methadone, buprenorphine, and extended-release injectable naltrexone.Footnote 6 Numerous studies have also revealed a high prevalence of mental health conditions coinciding with opioid use,Footnote 7 and the National Institute for Health has created the HEAL Initiative specifically to address these co-occurring illnesses,Footnote 8 underscoring the importance of structuring Medicaid policy to address the high rates of OUD co-occurring with mental health conditions.
Remarkably, most Americans with OUD do not receive any treatment for their condition. In 2021, only one-quarter of Americans with an OUD secured treatment.Footnote 9 Because the condition disproportionately affects poor working-age adults,Footnote 10 treatment affordability is an enormous issue. While there are other accessibility barriers, such as the lack of available sources of care, the inability to pay for care remains a serious obstacle to access, even among insured Americans.Footnote 11
Medicaid and OUD Treatment
The prevalence of OUD is higher among Medicaid enrollees than among other insured populations. For this reason Medicaid assumes an outsized role where access to effective treatment is concerned.Footnote 12 In 2020, Medicaid covered roughly forty percent of all Americans with OUD, and over half of people with OUD who are below two-hundred percent of the federal poverty level.Footnote 13 Medicaid also financed OUD medication-assisted treatment for over half of all Americans who received it nationwide.Footnote 14 Medicaid funds an estimated $12 billion in treatment services annually—more than six times the combined annual funding provided by the federal Substance Abuse Prevention and Treatment Block GrantFootnote 15 and the grants appropriated under the 2018 SUPPORT Act, which contains significant reforms aimed at expanding access to care.Footnote 16
Given states’ central role in Medicaid administration, how states design OUD treatment programs is highly consequential. Research has documented extensive gaps in coverage and heavy use of utilization management protocols that together can impede access.Footnote 17 While curbing unnecessary use of care is important, overuse can restrict access to the point at which essential care is being denied and people with OUD cannot initiate and remain in treatment. Footnote 18 Indeed, removing utilization controls on FDA-approved medications for opioid use disorder (MOUD) has been shown to improve health outcomes by increasing relevant prescribing rates and decreasing hospitalization rates.Footnote 19
In recent years, Congress has taken steps to strengthen state Medicaid response to OUD. The Affordable Care Act’s (ACA’s) Medicaid eligibility expansion opened coverage to millions of poor working-age adults, including those with an elevated risk for OUD, providing a foundation for OUD-specific reforms within Medicaid. In addition to the ACA’s Medicaid eligibility expansion, reforms contained in the 2018 SUPPORT Act, which among its many provisions, enables states to use Medicaid funds for treatment of individuals with SUDs at institutions for mental disease (IMDs) as well as to use waivers to increase the SUD treatment workforce. It also requires Medicaid programs to cover medication assisted treatment, fulfill new drug utilization review conditions, and report on behavioral health quality measures.Footnote 20
Moreover, because the majority of state contract with managed care organizations (MCOs), the road to OUD treatment runs through Medicaid managed care, which in 2020 represented seventy-two percent of all enrollment and fifty-two percent of total Medicaid spending. Footnote 21 State MCO contracts lie at the heart of the managed care model and create a performance blueprint for plan operations, so we undertook what appears to be the first-ever effort to systematically examine state Medicaid MCO contracts to determine how their terms align with what are considered the standard in the field for the treatment of OUD. We begin with an overview of Medicaid managed care and state MCO contracts. Following a summary of our research approach, we present principal findings and discuss their implications.
Medicaid managed care and contracts on which managed care rests
How states organize and structure their Medicaid programs becomes central to any discussion of Medicaid’s effectiveness in building access to OUD treatment. In today’s world, Medicaid-financed investment in health care, including OUD treatment, largely rests on the large managed care systems that have emerged as the program’s dominant strategy for organizing, covering, delivering, and paying for care.
Managed care (originally known as prepaid health care) has been a state option for five decades, but few states in the 1970s and 1980s pursued this model owing to its novelty, the dearth of sellers, early evidence of corruption, and provider resistance to any model—in any insurance market—that would give payers direct control over care. As the ground shifted in the commercial insurance market, federal Medicaid policy also shifted toward a formal embrace of managed care as a program organizing principle.Footnote 22 Earlier reforms culminated with enactment of 1997 legislation that dramatically expanded states’ ability to contract with comprehensive managed care plans built exclusively for Medicaid beneficiaries (Medicaid-only plans); Congress authorized use of such plans to cover both outpatient and inpatient care and to operate like other insurers, on a financial risk basis. The 1997 amendments thus marked Medicaid’s full emergence as a distinct insurance market.
At the heart of the model, as with all insurance, lies the contract between the sponsor (in this case, the Medicaid agency) and health plans known as managed care organizations (MCOs). Federal law gives states broad leeway in structuring their contracts, which must address enrollment, coverage, access, networks, program integrity, and quality performance. But the scope and breadth of the agreement lies with each state, and as a result, state managed care markets vary dramatically. Most importantly perhaps, federal law permits states to select which services covered under the state plan will be included in their contracts and which to maintain as a directly paid state benefit (e.g., long term nursing home care; highly specialized treatments for certain populations). Although states remain obligated to directly cover state plan services not included in the contract, as extracontractual benefits,Footnote 23 no study has ever tested whether services not explicitly addressed in the contract remain available to beneficiaries.
There are several other plan requirements that states can include in MCO contracts, but the most common include utilization management requirements, inclusion of provider types, and carve-outs for certain benefits. Most importantly, states can contract for a care model in which the contractor is obligated to conduct policies that vary from the state’s own practices. For example, states can require plans to employ utilization management strategies that vary from the state’s own practices. Conversely, states can allow MCOs discretion over utilization management that allows variation from the state’s own practices. Under federal law, the only bottom line is that utilization management be reasonable and efficient—a standard that is virtually never enforced by federal officials, if they satisfy the federal “reasonableness standard” governing Medicaid utilization management generally.Footnote 24
As mentioned, states can assign other state plan administration duties to their contractors, such as making the supplemental payments to which certain providers (e.g., federally qualified health centers or disproportionate share hospitals) are entitled, assuming that plans do not incorporate special payment rules into their own schedules. Footnote 25
Finally, states can operate their managed care systems through multiple related contracts, such as coupling a contract covering physical health services and basic mental health care with special “carve-out” arrangements with specialty plans. Bifurcation can be complex, since multiple contractors may find that they must coordinate coverage and treatment. For this reason, carve-outs are specified less often in state contracts, though prime contractors may have their own internal subcontracts with major subsidiaries that carry their own cross-plan navigation complexities.
Drafting managed care contracts is difficult. A crucial decision on any aspect of the contract is how specifically to address its coverage and care terms—that is, how closely to manage the work of the contractor. Explicit provisions such as detailed utilization management guidelines give states control over plan decision-making; it also means that states do not get the benefit of plans’ own experiences and innovations. By contrast, silence on the meaning of terms and fewer detailed requirements offer plans considerable leeway to substitute their own approach to care and their own utilization management techniques. In the absence of overarching law limiting plan discretion, silence and deference effectively signals that states desire its contractors to use their judgement about how to manage care. A purchaser presumably might tie a deferential approach to specific and measurable performance outcomes—that is a recognized standard of care to which plans are expected to adhere. This approach is what controls Medicaid and Medicaid plans in the case of childhood immunizations, which under federal Medicaid law must be furnished in accordance with recommendations of the Advisory Committee on Immunization Practice (ACIP). Such universally recognized gold standards of care that have the force and effect of law tend to be extremely few in number, and embedding such standards into contract as a binding performance expectation is quite uncommon.
Other factors further complicate the drafting process. First, a state may have its own policy or political preferences that bind plans. These detailed priorities may encourage the state to lessen demands for other populations and services, giving them lower priority. A second consideration is that ultimately, managed care systems can be only as strong as the underlying health care system on which they rest. If a plan’s service area is also an addiction treatment desert, the state might afford the contractor more discretion to impose strict treatment limits and other controls in order to not unduly stress whatever treatment is available. While certain innovations such as telehealth can mitigate shortages, for very acute health problems that demand intensive in-person care in specialized settings, health care shortages are an ever-present problem for Medicaid agencies, beneficiaries, and plans.
A third consideration relates to the very nature of managed care and why states use it. States buy managed care because they want more budgeting certainty through an all-inclusive per-member-per-month premium and because they want the twin benefits of access (the desire for which overlooks the second consideration) and cost containment. This desire to make use of what managed care presumably has to sell means giving plans a fair amount of leeway to control utilization—more, perhaps, than what the state historically has done.
Fourth, Medicaid managed care enrollees are disproportionately health and socially burdened and without the resources to supplement current coverage through additional out-of-pocket payments. This dependence creates especially high needs among a population without the means to meet those needs and exerts serious upward pressure on plans. Rather than trying to deal with health needs for which there are insufficient resources, plans may respond through aggressive efforts to shield heightened need by imposing tougher utilization controls as a means of classifying the demand as simply social needs that do not satisfy medical necessity criteria.
Finally, as with any product, managed care plans function according to their own internal operating rules. States can write their contracts with specificity and yet find that plans are more restrictive than they intended. This tendency on the part of plans to operate according to their own rules has been shown in both government studiesFootnote 26 and litigationFootnote 27 that reveal instances in which plans are denying care that should have been furnished under law or under the terms of state laws, policies, and plan documents. The evidence shows that, effectively, managed care is a product like any other and that efforts to customize the product amount to an uphill climb.
Methods
This project focuses on Medicaid managed care purchasing agreements that specify SUD treatment as a covered benefit. This analysis of state Medicaid contracts is part of a larger multi-state study of the effects of Medicaid MCO coverage and utilization management design on OUD treatment receipt and outcomes. Contracts were collected from fifteen states for in-depth analysis: the fifteen states selected for this study are intended to be illustrative of the level of discretion given plans as well other factors including (1) the prevalence of OUD in the state; (2) the quality of MMC claims data submitted to the federal government (CMS) for each state; and (3) the availability of plan-specific design data in each state. The states selected also show a mix of contracting approaches including both all-inclusive plans offering both comprehensive physical and behavioral health care and multi-plan arrangements where OUD treatment is managed by specialty managed care arrangements.
While the contracts are point-in-time, they do not change materially from year to year. Managed care purchasing is done through multi-year bids that set the basic framework in effect over the multi-year time period. Thus, while the contracts evaluated here come from 2021, they reflect the basic performance blueprint in effect over a longer time period. States may supplement initial documents with clarifying policies, but they do not deviate from the requirements in effect, at least not without a formal re-bid.
Collection of contracts was conducted online (documents are publicly available) supplemented by outreach to individual states where documents were missing. A list of topics to be researched was developed that reflects a review of the literature regarding the range of treatments crucial to effective care. These treatments are by and large captured by what are widely considered the definitive standards in the field of OUD treatment developed by the American Society of Addiction Medicine (ASAM). Numerous validation studies have established that matching the severity of a patient’s substance use disorder to the levels of care specified in the ASAM criteria optimizes treatment processes and outcomes.Footnote 28 This research has supported the ASAM guidelines, which are now the most widely used and evaluated set of guidelines for treating patients with substance use disordersFootnote 29 and are formally recognized by the Centers for Medicare and Medicaid Services (CMS), which requires states to adopt ASAM guidelines as a condition of approving Medicaid 1115 demonstrations aimed at improving SUD treatment.Footnote 30 The ASAM criteria specify four levels of treatment encompassing intensive inpatient care, residential treatment, and varying levels of outpatient care as patients progress.Footnote 31 Additionally, the Food and Drug Administration (FDA) has approved four medications that can be used in combination with psychosocial treatment for effective treatment of OUD.Footnote 32 These medications are methadone, buprenorphine, and both oral and extended-release injectable naltrexone. The ASAM guideline also recommends that all OUD medications be offered in conjunction with the appropriate level of psycho-social treatment, typically furnished through the appropriate level of outpatient care. In sum, we examine state MCO contracts against the ASAM standards to determine how contractual terms align with what are considered the standard of care in the field for the treatment of OUD.
Results
As shown in Table I, approximately half (eight) of the states examined include any mention of ASAM. Most commonly, ASAM was identified as the standard applicable to medical necessity determinations. (5). Fewer states specified the actual ASAM continuum of care (from intensive inpatient care to residential treatment and then through stepdown intensive and ongoing outpatient care) in regard to coverage or actually ensuring that care is furnished in a manner consistent with ASAM recommendations. West Virginia is the only state of the fifteen studied that discusses ASAM in relation to specific services covered, staffing, and medical necessity, a finding consistent with the fact that their OUD system operates under 1115 demonstration authority.Footnote 33 Plans may well be covering and furnishing ASAM-level care to members, but the requirement is not an express condition of performance. Similarly, states may aspire to plan coverage at the ASAM level through additional transmittals,Footnote 34 but our study classifies ASAM as a formal expectation only if references are found in the formal plan document, as an express performance condition.
In Table II, we examine seven coverage dimensions, three of which directly address ASAM continuum of care services: outpatient services, intensive outpatient/partial hospitalization services, residential services, and inpatient services.Footnote 35 The least common service to be explicitly included in the contracts is residential and inpatient services, followed closely by intensive outpatient and partial hospitalization services, withdrawal management services, and methadone maintenance. The most common services discussed were outpatient services, followed by recovery and peer support services. Six states—Michigan, New Hampshire, New Mexico, New York, Virginia, and West Virginia—discuss coverage of at least six ASAM treatment dimensions.
The states that employ separate “carve-out” arrangements to cover and deliver OUD care show a range of approaches in describing plan obligations. This is the case for Michigan, Pennsylvania and Maryland. In Maryland’s case, the basic MCO contract thus explicitly clarifies that plans are “not responsible for reimbursing … regardless of diagnosis (1) Services delivered by a [certified] community-based provider [furnishing specified] Alcohol and/or drug services; ambulatory detoxification.”Footnote 36
The contracts show variable degrees of deference in how covered treatments are described. For example, New Hampshire specifies that partial hospitalization must be offered within a set number of days from the date of ASAM level of care assessment, while Pennsylvania makes partial hospitalization an optional MCO service furnished “in lieu of” other covered services as permitted under federal managed care rules.Footnote 37 (The New Hampshire and Pennsylvania contract language is shown in Appendix A)
Even when the contract specifies inclusion of a particular service, the actual descriptions vary significantly. This is especially the case for peer support services, considered an important element of care under ASAM standards. West Virginia details the duties of peer support specialists and required trainings that MCOs must follow, while New Mexico provides only a short statement of coverage and reference to its 1115 waiver, leaving more ambiguity within the bounds of the purchasing agreement. New Hampshire’s language suggests peer support services is a state priority. Whereas West Virginia focused on training and New Mexico is in an earlier demonstration phase, New Hampshire is the only state studied that requires MCOs to “actively promote” its peer support program.Footnote 38
The types of covered withdrawal management services described can also vary substantially by state. Kentucky’s language is limited to “management of symptoms during the acute withdrawal phrase,”Footnote 39 while New York includes outpatient withdrawal services with details about what those services must entail.Footnote 40 Michigan pays particular attention to coverage of withdrawal services for individuals with co-occurring mental health disorders.Footnote 41
Discussion
The differences in language and emphasis identified in this study may point to a state’s priorities in addressing OUD. These study findings show that among the states selected here—states that make extensive use of managed care and have a high incidence of OUD—only some expressly reference the ASAM guidelines. Among these states, references are scattered, and no state unequivocally specifies that in all coverage and treatment decisions involving people diagnosed with OUD, the plan is bound by the ASAM guidelines as the standard of both coverage and care. The language differences also suggest the absence of consensus regarding exactly how to translate ASAM guidelines into contract language other than specifying the guidelines as the standard of coverage and care that binds plans, whether full-service or behavioral carve-out.
As expected, all states studied here cover OUD treatment, and there is no question regarding states’ commitment to providing at least some forms of treatment for OUD. Indeed, in some states—regardless of what plan language specifies—care may be happening at the ASAM level. Certain services furnished in connection with OUD are mandatory for the population such as physician and hospital care, and states cannot discriminate in the provision of coverage based on diagnosis. All states cover prescription drugs in their plans, making FDA-approved OUD drugs a basic element of their formularies, since federal Medicaid law requires that state plans include all FDA-approved prescribed drugs. States effectuate their coverage either through their general service agreements or via a multi-plan approach that utilizes the services of specialty care and coverage arrangements that supplement what the basic plan offers. But because ASAM is not specified as the unconditional coverage and treatment standard, plans’ approaches may diverge in terms of both what treatments they cover and the standard of care they follow when making coverage decisions.
CMS has tried to move states in this direction by tying approval of 1115 Medicaid demonstrations focused on OUD treatment to use of ASAM guidelines. But even here, state 1115 demonstrations are generally not unequivocal on this point by binding contractors to the standard as a matter of compliance.
Many factors may drive states’ decisions not to expressly and unequivocally make ASAM the standard of care when managing coverage and treatment. ASAM-level care is undoubtedly costly, even if cost effective. Cost-effectiveness studies that assume multi-year savings have limited utility for states that are under pressure in any single year to keep costs down. As a result, even a gold standard such as ASAM might be cost-prohibitive.
An additional consideration is that states and plans simply lack the on-the-ground resources to offer gold-standard care. The shortage of residential treatment and outpatient care—both generally and at the prices Medicaid can afford to pay—can defeat any effort by states and plans to increase the quality of care. The states whose contracts were examined in this study constituted twenty-one percent of pain reliever use disorders (328 thousand individuals) and twenty-six percent of heroin use (201 thousand individuals) in the United States in 2019.Footnote 42 The high prevalence of these conditions within this sample of states results in a state average of 757 individuals with pain reliever disorders and 431 heroin users per specialty facility with opioid treatment programs, though, on average, these facilities saw a median of only 325 outpatient clients in total in 2020.Footnote 43 In 2021, almost 5.6 million individuals had an opioid use disorder, but only 1.2 million received treatment for one.Footnote 44 Critical shortages of appropriate treatment could result in coverage decisions meant to maintain care at accessible levels in order to avoid (or at least mitigate) the problem of long waiting lists for services.
Ensuring that Medicaid managed care plans perform at the ASAM standard of care means saying so unequivocally in plan documents. It also means a massive investment in on-the-ground care and sufficient financing to enable states to buy, and plans to provide, treatment at the definitive standard of care. In particular, lack of attention to coverage of residential treatment appears to be an ongoing issue.Footnote 45 This level of investment is not likely to materialize. Multiple states continue to reject basic Medicaid eligibility for the low income working-age adult population, and those that have expanded coverage must confront the high cost of care, even at the enhanced federal funding rates available for the Medicaid expansion population.Footnote 46 What might help address this problem beyond the obvious of lessening the incidence of OUD is the possibility of guidelines that offer an effective but less costly alternative strategy for managing OUD.
Of course, it is not only in the case of OUD that there exists a gulf between known, effective clinical standards of care and what binding plan documents specify. This is true across the care spectrum, from preventive pediatric care to maternity care, to guidelines governing treatment for cancer, heart conditions or diabetes. And Medicaid is hardly alone in this gap; insurers generally are not held to express standards of care. The gap between what is known and what is permitted means that plans are free to impose their own judgment and preferences, which may or may not reflect such standards. Changing Medicaid in this regard means changing the way insurance works more broadly: an advance in American health care that likely is a long way off.
Supplementary material
The supplementary material for this article can be found at http://doi.org/10.1017/amj.2023.35.