, State Medicaid Agencies' Multi-Faceted Response to the Opioid Epidemic State Medicaid Agencies' Multi-Faceted Response to the Opioid Epidemic Abstract Context: Medicaid is the primary payer for substance use disorder Policy Points (SUD) treatment in the United States. While some policy changes • Policies and operational strategies adopted by have been well documented, the operational decisions that guide state Medicaid programs to expand access and the implementation of these policies have received insufficient improve substance use disorder treatment quality attention. The objective of this analysis is to describe the roles that have received insufficient attention. Medicaid programs have taken to address the opioid epidemic and their policy and operational decisions. • In interviews with 9 state Medicaid agencies, we observed substantial policy convergence between Methods: We conducted 27 key informant interviews with state states, including the removal of prior authoriza- agency representatives in 9 states, all of which have been substan- tion for buprenorphine/naloxone, an expansion tially impacted by the opioid epidemic. We focused our interviews of access through coverage and delivery system on 3 distinct state roles: Regulator, Monitor, and Enforcer; Payer and reforms, and the establishment of intrastate cross- Contractor; and Collaborator, Evaluator, and Educator. Within those agency collaboratives. roles, we aimed to synthesize the degree of variation of the policies implemented across these states from 2014-2019, given the breadth of • Our findings are relevant to state policymakers policy levers available to them. Interviews were recorded and tran- and health services researchers interested in iden- scribed, responses were summarized categorically where possible, and tifying effective approaches to address the opioid the transcripts were reviewed to identify areas of variation. crisis, and establishing metrics to advance policies in development, such as value-based purchasing Findings: We observed substantial convergence in the policies and for substance use disorder treatment. actions taken by states. All 9 states relaxed or eliminated utilization management policies, such as prior authorization of medications for opioid use disorder, that may be a barrier to access. Most states expanded SUD treatment coverage to align with the American Conclusions: State Medicaid agencies are engaging in roles that go Society of Addiction Medicine continuum of care. As collabora- beyond that of just a payer of health services, especially in expand- tors, Medicaid programs participated in interagency efforts such as ing access to SUD treatment; however, further support is needed to opioid task forces, including various levels of data-sharing between advance future policy goals, such as value-based payment. agencies. Interviewees discussed ongoing evaluative activities; how- ever, OUD treatment quality measurement remains an area in need of development to support state policymakers. 1 State Medicaid Agencies' Multi-Faceted Response to the Opioid Epidemic Introduction Background Medicaid is the single most important payer for health care services Evidence-based treatment and clinical guidelines related to the opioid epidemic. State Medicaid programs collectively SUDs in the United States have been treated historically as social covered an estimated 38% of nonelderly adults with opioid use dis- disorders stemming from moral or spiritual failings.4 Only in 1987 order (OUD) in 2017, and 54% of those who received treatment for did the American Medical Association (AMA) classify all drug OUD.1 Thus, the OUD treatment policies set by Medicaid programs addictions as medical diseases.5 Since then, 40-50 different treat- can shape how health care systems treat all individuals with OUD. ment approaches emerged for clinical application, some with poorly developed theories of change and questionable efficacy.6,7 Payers Federal policies afford state Medicaid programs substantial flex- like state Medicaid programs were left to decide which set of criteria ibility as payers of substance use disorder treatment, both in terms were most effective or appropriate, without sufficient research or of the services they cover and the utilization management poli- clinical consensus upon which to base their decisions. cies they use to govern which patients receive care and how it is delivered. In response to the opioid epidemic, Medicaid programs In 1988, the American Society of Addiction Medicine (ASAM) have expanded coverage of SUD treatments, and reformed deliv- was accepted into membership by the AMA as a national medical ery and payment systems. Some of these policy changes, such as specialty society, with addiction medicine becoming a self-desig- Section 1115 SUD Demonstration Waivers, have been well-doc- nated practice specialty in 1990.8 Since 1992, ASAM has developed umented although most have not yet been rigorously evaluated.2 patient placement criteria for treating SUD, publishing its first set The operational decisions that guide how policies are implemented in 1994 and the most recent version, The ASAM Criteria: Treatment and enforced are nuanced and have received less attention from re- Criteria for Addictive, Substance-Related, and Co-Occurring Condi- searchers. For example, while coverage of services is generally well tions, in 2013. The "ASAM Criteria" (as that document is known) documented, much less is known about the utilization management created a standardized comprehensive assessment model from policies that may affect access and quality, the flexibility provided to which placement recommendations for appropriate treatments Medicaid Managed Care Organizations (MCOs) in applying those are derived.9 The clinically-derived system assesses patients over 6 policies, or the reasoning and evidence that informs the enactment biopsychosocial domains, focusing on outcomes, team-based ap- of those policies. Understanding these operational decisions is proaches and recommendations.9 In 2015 and 2017, the Centers for critical to understanding how states have acted to counter increases Medicare and Medicaid Services (CMS) encouraged state Medicaid in OUD-related morbidity and mortality, and may explain vary- programs to improve SUD treatment coverage and delivery systems ing trends in access, quality and outcomes of OUD and other SUD by using the ASAM Criteria.10-12 As of 2017, 33 state Medicaid treatment across states. agencies required the use of the ASAM Criteria at least in part or in principle by their Medicaid MCOs and providers.9,13,14 To advance knowledge about the multiple facets of state Medic- aid policy adoption and implementation, we conducted in-depth Federal Legislation and Regulation interviews with policy officials drawn from a multi-state Medicaid Federal legislation enacted in recent years further informed and collaborative project (Medicaid Outcome Distributed Research shaped SUD treatment and recovery efforts by state Medicaid Network [MODRN]), that includes many of the states hardest hit programs. The Mental Health Parity Act of 1996, the Paul Wellstone by the opioid crisis, including 5 of the 10 states with the highest and Pete Domenici Mental Health Parity and Addiction Equity Act overdose death rates.3 We sought to answer the following questions: of 2008 (MHPAEA) and the Affordable Care Act (ACA) each took How have state Medicaid programs changed their coverage and incremental steps towards broadening the coverage of behavioral reimbursement policies in response to the opioid epidemic? In what health services by health insurers.15-19 This required states and ways do these policies vary across states? Beyond the typical role Medicaid MCOs to implement and analyze mental health parity in of payer, what do state Medicaid agency officials see as their role their own programs.20 More recently, the 21st Century Cures Act of in responding to the epidemic? What do Medicaid agency officials 2016 and The Substance Use Disorder Prevention that Promotes describe as the next frontier of policy changes? Opioid Recovery and Treatment for Patients and Communities Act (SUPPORT Act) of 2018 have more directly targeted the opioid epi- In this report, we first present relevant background on recent policy demic. The former included $1 billion to support states in a variety history that has affected current SUD policy decisions, followed by of programs to improve access to OUD treatment and to prevent a description of our methods and findings from the interviews we future OUD from occurring.21 The latter mandated coverage of conducted with state Medicaid agency officials, and finally a discus- buprenorphine, methadone and naltrexone for OUD (otherwise sion of implications of our findings and ways to assist state policy- known as Medications for Opioid Use Disorder (MOUD)) by state makers in serving the needs of their enrollees with OUD. Medicaid programs, among other OUD-related policies.22 2 State Medicaid Agencies' Multi-Faceted Response to the Opioid Epidemic Methods Key Informant Interviews Information on state Medicaid SUD policies was collected through Sample a series of interviews with state officials in each of the 9 MODRN The state Medicaid agencies represented in this report participate states. We developed and refined an interview guide with input in the MODRN project.23 MODRN is a multi-state collaboration from MODRN investigators in the 9 participating universities, founded by participants in AcademyHealth's State-University Partner- Medicaid clinical leaders (i.e., Chief Medical Officers), and state ship Learning Network (SUPLN)24 and Medicaid Medical Directors SUD policy experts from 2 National Institute on Drug Abuse Network (MMDN).25 The objectives of the MODRN-OUD project are (NIDA)-funded centers of excellence. to provide a comprehensive assessment of OUD treatment quality and outcomes in Medicaid, and to inform policy decisions on coverage The interview guide focused on 10 key policy domains where and payment for evidence-based OUD treatments. states have substantial discretion in how they address the needs of Medicaid enrollees with OUD or other SUDs. We present each do- We first present the characteristics of the 9 participating state Med- main in Table 2. We focused primarily on state Medicaid program icaid programs (Kentucky, Maryland, Michigan, North Carolina, policy and the administrative decisions Medicaid agencies made Ohio, Pennsylvania, Virginia, West Virginia, and Wisconsin) in to address the opioid epidemic. Specifically, we asked states about Table 1. Next, we describe specific policies related to each state's re- MOUD and OUD treatment utilization management and payment sponse to the opioid epidemic specifically and, more broadly, SUD policies as they have been shown to affect access to treatment.28-31 treatment. As death and addiction rates have continued to rise, 3 Due to its relevance towards both quality of care and access,32-35 of the 9 states have used emergency declarations to address the cri- we asked states about delivery system reforms, care coordina- sis.26,27 Our sample of states also reflected the diversity of Medicaid tion, transition, and integration with other health care services. In programs in the US. Seven of the states expanded Medicaid under addition, we asked if Medicaid agencies had partnered with other the ACA, and 4 states carved-out behavioral health care manage- organizations of state government to address the opioid epidemic, ment from their physical health MCO contracts (2 other states not and in what capacity. Finally, we inquired about activities that often included in the analysis do this as well), decisions that predate the fall outside of the responsibility of a state Medicaid agency, such opioid crisis. All 9 states received approval for 1115 SUD demon- as licensing of SUD providers, but closely intersect with Medicaid stration waivers between 2016 and 2019. agency responsibilities and other interagency efforts. Table 1: Key Characteristics of State Medicaid Programs Included in the Policy Inventory Related to Substance Use Disorder Policies Age-Adjusted Drug Number of Percent of Percent of Medicaid Medicaid Behavioral Overdose Mortality States Enrollees as of population in population in Managed Expansion, Health Rate per 100,000 Dec 201947 Medicaid48 Care49 Date of Effect50 Carve-out (Rank), 20193 Kentucky 1,187,843 26.8% 89% Yes, Jan 2014 32.5 (7) Maryland 1,328,704 21.9% 85.7% Yes, Jan 2014 X 38.2 (4) Michigan 2,320,304 22.8% 76% Yes, Apr 2014 X 24.4 (20) North Carolina 1,772,156 16.8% NA No X 22.3 (23) Ohio 2,609,614 14.9% 89% Yes, Jan 2014 38.3 (3) Pennsylvania 2,938,411 22.8% 77% Yes, Jan 2015 X 35.6 (5) Virginia 1,414,239 15.9% 82%51 Yes. Jan 2019 18.3 (28) West Virginia 507,398 29.3% 77% Yes, Jan 2014 52.8 (1) Wisconsin 1,046,309 17.8% 72% No 21.1 (25) 3 State Medicaid Agencies' Multi-Faceted Response to the Opioid Epidemic Table 2: Ten Domains from Key Informant Interview Guide General overview of the state's SUD treatment landscape PAYER & CONTRACTOR Medications for opioid use disorder (MOUD) coverage and utilization management Other SUD treatment coverage and utilization management SUD provider payment Delivery system reforms SUD-related care coordination, managing care transitions, and integration of SUD COLLABORATOR, EVALUATOR, & EDUCATOR Coordinated interagency and multiagency state efforts Naloxone coverage and availability REGULATOR, MONITOR, & ENFORCER Network development and licensing of SUD providers Quality and outcome measurement improvement initiatives for SUD providers Participants included state agency representatives, representing Medicaid, behavioral health, and public health agency staff. The Results roles of participants included Medicaid clinical leaders, senior Regulator, Monitor and Enforcer Medicaid pharmacy directors, Medicaid data analytics manag- In their roles as regulators, monitors and enforcers, states made de- ers, senior Medicaid strategists, behavioral health/drug & alcohol terminations on how services for SUD were defined, who may pro- policy advisors and analysts, and other state representatives with vide them, and what requirements must be met by those providers. programmatic knowledge. The interviews (n = 27) were conduct- These functions included both Medicaid and other state agencies. ed with each state via a series of phone calls (median 3-4 hours per state Medicaid program). The interviews were conducted Clinical Criteria from June to October 2019 and were recorded with permission Seven of nine states reported using ASAM criteria to guide SUD from the state agency representatives and then transcribed. After coverage decisions. Similarly, 6 states used ASAM criteria to the interviews, we followed up with interview participants and guide SUD treatment placement decisions, whereas another state their designated subject matter experts to clarify details (e.g., reported the use of criteria similar to that of ASAM. Among the 2 specific dates of policy implementation) and compile supplemen- states that did not use ASAM criteria, Wisconsin commented that tal information cited on the call (e.g., opioid treatment guidelines, the MCOs may use alternative criteria that cannot be more restric- Medicaid preferred drug lists). tive than ASAM, and West Virginia clarified that while providers should use ASAM, some choose not to and opt for the Clinical Following the completion of the interviews, we organized the infor- Opiate Withdrawal Scale36 instead. mation in 3 ex ante identified state roles: 1) Regulator, Monitor and Enforcer, 2) Payer and Contractor; and 3) Collaborator, Evaluator, SUD Provider Licensing and Educator. Within those roles, we aimed to synthesize the degree All states reported that SUD inpatient, residential, and most outpa- of variation of the policies implemented across this group of states, tient providers had to be licensed by the state, often using national given the breadth of policy levers available to them. We summarized credentialing standards. In addition, some states reported that answers categorically where possible, such as the coverage of certain each MCO was permitted to have their own credentialing process. SUD treatment services. Transcripts from the interviews for other Specific to residential treatment facilities, states that had adopted questions were analyzed to identify areas of variation. This included ASAM used the criteria to license and/or credential facilities (i.e., identifying programs, policies, and approaches discussed by each state ensure that the facility offers the services as defined in the criteria). for specific questions in the interview guide and comparing against re- All states required (or were moving towards requiring) residential sponses to the same question from other states. Our analysis was then treatment facilities to either provide MOUD onsite or facilitate re- reviewed by interviewees and university partners to ensure accuracy. ferrals so that their patients had access to MOUD while in residen- tial treatment, which is meaningful as many residential programs abide by abstinence-based treatment.37 4 State Medicaid Agencies' Multi-Faceted Response to the Opioid Epidemic Though licensing administration and enforcement typically falls to quirements to patients either beginning or maintaining treatment. a separate state agency, many Medicaid agencies described a role in Only 1 had a requirement related to step therapy which was ap- informing licensing standards in their state. Some states reported plied to all MOUD medications. This required enrollees to receive that Medicaid representatives play an advisory role to the licensing 4 hours of counseling per month and 2 urine drug screens during agency, such as ensuring that those with licensing responsibilities an initial phase of treatment. If the enrollee had complied with are fully aware of ASAM criteria. One state commented that their therapy and did not have a positive urine drug test, they could Medicaid program works with the licensing agency by informing drop to 1 counseling session and 1 urine drug screen per month. them of providers who have been reported by patients or other The majority of states did not have extensive monitoring require- providers for delivering low-quality or inappropriate care. ments for continuation of MOUD either, and during the interview would often point to the other utilization management policies as Payer and Contractor sufficiently restricting MOUD prescribing to appropriate cases, As a health insurer, Medicaid programs can affect access and qual- as well as relying on the provider to adhere to what is medically ity through the use of delivery and payment reforms, as well as appropriate, as satisfactory. coverage decisions. The most commonly used utilization management policies were MOUD Coverage and Access 2 clinical requirements: quantity limits and dosage requirements. We observed many areas where the states included in this analysis These policies tended to be used in conjunction with one another, appeared to be converging in their OUD treatment policies. First, as states referenced restrictions such as a maximum of 24 mg of bu- MOUD coverage had been expanded, and accompanying utili- prenorphine for induction, and 16 mg for maintenance.39 Dosages zation management policies had been relaxed, typically in both above those levels were available but subject to prior authorization. fee-for-service (FFS) and MCO programs. All 9 state Medicaid In contrast, Virginia's Board of Medicine required that enrollees be programs covered buprenorphine, naltrexone, and methadone, initiated on buprenorphine starting at 8 mg per day and increase to although West Virginia and Kentucky began covering methadone higher dosages as necessary. Duration limits or caps on the amount only recently, in 2018 and 2019, respectively. A common change we of time an enrollee may receive MOUD were not used by any state. observed was that all states removed prior authorization policies for buprenorphine/naloxone,38 and all but one state had done so SUD Treatment Services within the last 5 years. Interviewees provided varying reasons for States had made changes to expand the services for SUD treat- the use and operationalization of their previous prior authorization ment covered by the Medicaid program. In Figure 1 we present policies. Interviewees in one state described the flexibility that was indicators of coverage for each service and each state, representing initially given to MCOs to set their own prior authorization policies 4 broad categories: coverage without limits, limits on coverage for on MOUD; however, this created a burden on providers to under- certain subgroups, quantity limits, or if the service is not covered. stand and comply with an array of policies. Another state's prior Services were broadly covered across states, with the exception of authorization policy was implemented with the intent to control partial hospitalization and residential treatment in Wisconsin.40 quality and required prescribers to attest that the patient had been Officials in 6 states commented that they had adopted policies diagnosed with OUD, that an informed consent was signed, that to formally cover or enhance access to residential SUD services the prescription drug monitoring program had been checked, and within the last 5 years. Three states began covering inpatient that the patient had been referred to counseling. Similarly, another SUD care, 3 states adopted policies to reimburse for peer support state's policy included a requirement that the enrollee demonstrate services, and 3 states added or enhanced access to partial hospi- they were in or were seeking an active treatment program. In these talization services. Interviewees from Kentucky described major cases, any potential quality control benefit or flexibility granted changes to coverage of SUD services in their state that occurred to MCOs were counterbalanced by increased provider or enrollee in 2014, which included expanding coverage to all Medicaid burden that limited access. enrollees for residential services, partial hospitalization, intensive outpatient services, psychotherapy, peer support services, and tar- Beyond prior authorization, states did not widely use enrollee- geted case management. Relative to before the period in question based utilization management policies (e.g., patient compliance (2014 and later), these states broadened the coverage of inpatient requirements), rather clinically-based utilization management and residential SUD services to a degree previously not available (e.g., prescribing dosage and quantity limits) were commonly to enrollees. applied. We asked interviewees specifically if their state applied patient compliance, completion of treatment, or step therapy re- States frequently reported utilization management policies for 5 State Medicaid Agencies' Multi-Faceted Response to the Opioid Epidemic Figure 2: State Coverage of Other SUD Services as of 2019 Figure 1: State Coverage of Other SUD Treatment Services as of 2019 Behavoral Intensive Peer Care SUD Urine Drug Health Outpatient Partial Inpatient Residential State SBIRT Support Management Telehealth Screens Counseling Services Hospitalization SUD Treatment OH KY WI MI VA WV NC Not covered Subgroup and quantity limits PA Subgroup limits Quantity limits MD Covered, No limits Uncategorized care management and residential treatment that either limited the program42 for Medicaid participants who have serious and per- quantity of services an enrollee could receive or required a docu- sistent mental illness; OUD and are at risk of additional chronic mented diagnosis to receive the service (i.e., subgroup coverage conditions due to tobacco, alcohol, or other non-opioid substance limits). Examples of subgroup coverage limits for care management use; or children with serious emotional disturbances. In addition, included a diagnosis of SUD and chronic physical pain, moderate Michigan launched an opioid health home pilot program with to severe SUD, or a mental health diagnosis and an intellectual dis- federally qualified health centers located in rural areas, funded ability. Quantity limits for residential treatment were often specified through a hybrid of federal, grant, and waiver resources. While as a 30-day length of stay, but that could be extended with prior enhancing access to treatment, these programs also sought to authorization. improve care coordination and transitions for specific vulnerable populations (e.g., pregnant women, co-occurring serious mental Delivery Reforms illness), and to reduce acute care utilization for OUD. States reported the use of specific innovative delivery reforms with the goal of improving OUD treatment as well as the coordi- Delivery reforms intended to improve transitions across care set- nation of care for Medicaid enrollees in specific programs rather tings were a common theme across states. Two care settings where than system-wide changes. We define the key terms related to states reported major efforts included the hospital (both inpatient these reforms in the call out box. Pennsylvania and Maryland and the ED) and residential treatment centers. For example, Michi- have both implemented health homes in the last 5 years. Penn- gan used various grants, including the SAMHSA-funded State sylvania's Centers of Excellence program includes a range of SUD Opioid Response (SOR) grant, to support ED-initiated buprenor- providers, primary care providers, hospitals, and county agencies phine and warm handoffs to providers in addition to identifying that were funded through lump sum payments to provide SUD community providers for referral. The State also utilized their treatment directly while coordinating care for other physical contracts with prepaid inpatient health plans to require providers to and mental health conditions.41 Maryland utilized a state plan incorporate a warm handoff as patients are transitioned from one amendment (SPA) to implement a comprehensive health home 6 State Medicaid Agencies' Multi-Faceted Response to the Opioid Epidemic level of care to the next. In residential settings, states required MCOs Value-Based Purchasing to implement warm handoffs from discharge to outpatient treat- We observed that value-based purchasing was of great interest to states, ment programs, so that MOUD treatment is coordinated while the but this payment model had not been widely implemented. Interview- enrollee is being served in a residential treatment facility. Through ees broadly reported that FFS payments either from the state or the a SPA, Kentucky added coverage of methadone for the treatment of MCOs was the primary method of reimbursement to SUD providers. OUD, included care coordination in residential treatment locations, Most states reported that MCOs must, at a minimum, reimburse what and allowed peer support services to bridge the transition from ED the FFS program would pay, but were then granted flexibility to set rates to treatment. These ED-Bridge clinics were funded using a SAMHSA with contracted providers. While MCOs could create reimbursement State Targeted Response (STR) grant, which are aimed towards sup- arrangements other than FFS, interviewees believed that MCOs were port services that address the continuum of care. typically reimbursing providers via FFS payments, and at a similar level to what the state would reimburse in its own FFS program. While val- Similar to certain delivery reforms, interviewees often reported that ue-based purchasing was of interest to states to incentivize providers to care coordination activities targeted priority populations, such as improve quality of care and outcomes for Medicaid enrollees, no state pregnant women. Other reported priority populations included par- had fully implemented such a policy at the time of the interviews. Ex- ents at risk of losing custody of their children, people with injection amples of reimbursement policies that states had recently implemented drug use, and incarcerated or criminal justice-involved individuals. or were developing were cited during the interviews. In Pennsylvania, the Medicaid agency implemented a per-member-per-month (PMPM) rate for their Centers of Excellence, which both FFS and MCOs must pay. Another state utilized a similar payment mechanism to pay for care • Health Home – Health homes provide coordinated coordination for enrollees with OUD. Officials in a third state reported they were in the process of drafting a policy that would reimburse care for physical and behavioral health conditions outpatient OUD treatment using a weekly bundled rate. for vulnerable populations with multiple comor- bidities,52 and their use accelerated after the ACA. Integration Some health homes focus on individuals with SUD, States described efforts to make systemic changes in the financing and and coordinate community supports, care man- delivery of behavioral health services to better integrate care with the agement services, referrals, along with treatment management and delivery of physical health care. For example, in 2018, for both physical and behavioral health needs. Ohio undertook a behavioral health redesign that included a number of policy changes to improve integrated care. The state carved-in behav- • Care Coordination – Form of care delivery where ioral health services, which also allowed behavioral health providers to multiple providers coordinate health services to render existing physical health services if they had qualified clinicians address the physical and behavioral health needs on staff. The redesign expanded coverage for a myriad of treatment ser- of the patient.53 vices and adopted ASAM criteria as guidance for levels of care. Another state used a SAMHSA-funded Young Adult Substance Abuse Treat- • Integrated Care – A systematic approach to ment grant to develop a comprehensive strategic plan to improve SUD blending physical and behavioral health care, treatment services for adolescents and/or transition-age youth with extending to mental health, substance abuse, and SUD or co-occurring substance use and mental health disorders. This primary care, in a single setting.54,55 state was planning to use the SUPPORT Act to further increase access to evidence-based treatment for OUD and SUD, especially for preg- • Warm Handoff – An intervention where a provider nant, parenting, and justice-involved enrollees. conducts an in-person introduction between the patient and the behavioral health provider they Collaborator, Evaluator, and Educator are being referred to, with the goal of improving States can engage as a collaborator through working and sharing the initiation and coordination of behavioral health data with agencies at the state and county levels, as an evaluator by treatment.56 leveraging their large amount of claims data to track key metrics and outcomes for enrollees with OUD, and as an educator by work- ing with providers to ensure the needed skills to treat OUD are available in the community. 7 State Medicaid Agencies' Multi-Faceted Response to the Opioid Epidemic Education To improve transitions within the community, some interviewees Interviewees described their participation in numerous interagency cited collaborations between Medicaid and a broad spectrum of training initiatives including those intended to increase the number health and social services, including the state's public health agency, of providers prescribing MOUD to Medicaid enrollees. For example, child welfare, or county-level agencies. Partnerships with counties Virginia's Medicaid agency in collaboration with their state's Depart- included initiatives with county jails, local public health depart- ment of Health offered a free Project ECHO initiative to provide ments, and social service front-line workers. Partnerships between training on buprenorphine prescribing where providers could earn the states and multiple stakeholders at the county-level were built Continuing Medical Education credits. States also highlighted that to ensure that their approach to address the opioid crisis was as these efforts were often in collaboration with their mental health agen- comprehensive as possible. cy using federal STR or SOR funds. Three states specifically noted the use of SOR funds to educate the provider workforce to offer MOUD, including support in obtaining their DATA 2000 waivers. Discussion The objective of our study was to understand how state Medic- Quality Measurement aid programs responded to the opioid epidemic through policy Interviewees pointed to limited development and use of standard- changes, as well as how those changes were operationalized. Since ized SUD quality measures across state Medicaid programs. Most Medicaid programs play a disproportionate role in covering in- states continued to monitor levels of prescribed opioids, and some dividuals with SUD, they have a significant impact on their state's specifically highlighted tracking acute events such as inpatient and SUD delivery system, and thus Medicaid policies have the poten- ED use for OUD. MOUD-specific measures included duration of tial to mold the way care is delivered for all patients with SUD. pharmacotherapy, retention in treatment, counseling rates, along While our aim was to describe the degree of variation in policies, with other national measures calculated by their state university we broadly found a high degree of convergence in the approaches partners. In some cases, states focused their outcome measures on a taken by these 9 states over recent years. particular initiative, such as programs specific to pregnant enrollees with OUD, and non-claims-based measures, such as average wait States have the potential to serve as the laboratories of the US time to assessment and client experience. Other states reported that federal system, and in the case of Medicaid programs, have the flex- they were still identifying outcome or quality measures to target for ibility to use federal and state funds to test policies and programs to ongoing monitoring. better serve their populations. Given the latitude that state policy- makers have to shape Medicaid, the degree of policy convergence Collaboration we observed in our sample of states is noteworthy. The experience Multi-agency collaboration within states to address the opioid of dramatic increases in OUD rates and overdose in this sample of epidemic was novel in its extensiveness. All 9 states had estab- states may have led them to relax utilization management policies lished interagency task forces or command centers, which were and expand covered SUD treatment services to prioritize access typically created and overseen by the state's governor and were above other considerations. For example, all 9 states had removed being used to bring leaders from multiple state agencies together prior authorization for buprenorphine/naloxone by 2019; however, on a regular basis to identify problems, share information and ad- according to a report to Congress, MACPAC found that 30 states vance programs. At least one state included community represen- (including those less impacted by the opioid epidemic) at that time tatives in their opioid task force or worked with community-based still required prior authorization.43 The expansion of coverage of organizations on local initiatives. the continuum of care in these states appears to be facilitated by the recent adoption of the ASAM criteria, which defines what is One of the most common agencies that Medicaid officials report- included in each type of service with greater clarity, and thus makes ed ongoing collaborations with were departments of corrections. it easier for state agencies to operationalize these coverage policies. In these collaborations, corrections and state Medicaid programs While the states in this study diverged somewhat in the delivery typically were working together to ensure MOUD access for models they have tested, developing and piloting models for SUD incarcerated individuals within jails and prisons or for recently care delivery was the norm. released individuals. For example, one state discussed an initia- tive to offer MOUD in county jails so that induction could occur Our experience in collecting data on operational decisions and there, and then to have an outpatient appointment arranged for implementation can be instructive to other researchers and policy- the individual upon release. makers at the state and federal level who seek to understand the im- pact of various policies. Generally, it is simultaneously infeasible to conduct secondary data analyses to study Medicaid programs with 8 State Medicaid Agencies' Multi-Faceted Response to the Opioid Epidemic precision without understanding the way they are implemented, Thus, the need for more research on SUD treatment effectiveness to and impractical to collect the information required to completely support policymakers is essential for multiple reasons. First, given understand the implementation process, as the necessary informa- the expansion of SUD treatment services now available to many tion is often not formally documented and/or publicly available. Medicaid enrollees (especially residential treatment), studies on In particular, where MCOs do and do not have flexibility in the what treatment yields the best outcomes and for whom would in- timing or method of adopting a policy is unclear, but potentially form how Medicaid programs apply and refine utilization manage- consequential for evaluating the impact of certain policies. Various ment policies for different levels of care. The priority for the states organizations admirably document variation in Medicaid policies in our sample has been improving access; the next logical steps by state and should continue to do so; however, researchers should would include evaluating trajectories in recovery, coordination and partner with state Medicaid agencies whenever possible to ensure outcomes after residential or inpatient treatment, and evidence- that they completely understand the inevitable caveats that exist for based prescribing of MOUD, among others, to refine policies that these policies. ultimately improve the lives of individuals with SUD. Second, deliv- ery and reimbursement approaches that best incentivize coordina- Our finding of where states have converged will be valuable to state tion across settings were of great interest to interviewees, who were and federal policymakers, as Medicaid programs often look to using grant funds to support such services. Building a sustainable each other to learn how to improve care for their own enrollees. By system of coordination, either through MCOs or provider groups, identifying and disseminating areas of policy convergence, along that integrates physical and behavioral health and limits the possi- with the reasons provided by interviewees that the potential gain bility that enrollees "fall through the cracks" was a high priority for in quality was not worth the reduced access to evidence-based all stakeholders involved. And third, studying the degree to which treatment, policymakers in Medicaid agencies can better weigh recent changes to utilization management policies and treatment their options regarding MOUD treatment moving forward. Exist- coverage have affected access and quality would inform other states ing learning collaboratives including state policymakers, such as who have yet to take such action. Similarly, disseminating the les- the Medicaid Medical Director's Network,25 the State University sons learned by states as they develop their SUD treatment policies Partnership Learning Network,24 and the Medicaid Demonstration will likely be of great benefit to all states. Waiver Evaluation Learning Collaborative44 can facilitate shared policy practices and dissemination so that states can learn from Our findings can be viewed in light of recent changes to and each other. disruptions in OUD treatment due to the COVID-19 pandemic for 2 primary reasons. The first is that data from the early months We found that states are interested in but have not yet adopted of the pandemic suggest an acceleration in opioid-related over- value-based purchasing for SUD treatment providers. Our findings dose deaths, potentially reversing previous gains made against indicate that the lack of standardized quality metrics on SUD treat- the opioid epidemic.46 The second is the subsequent emphasis ment may be holding back state Medicaid agencies from pursuing on telemedicine for all treatment types, including for SUD. In value-based purchasing and limiting states' understanding of the addition to the research gaps described above, policymakers must outcomes associated with the care that SUD providers are deliver- now understand how telemedicine can be efficiently used in the ing. Namely, in the 2019 Core Set of Adult Health Care Quality continuum of SUD treatment. Further, the relaxation of other Measures for Medicaid collected by CMS,45 just two measures were OUD treatment requirements during the pandemic, such as the related to the quality of care for SUD. Many states focused pre- frequency of urine drug tests and limits to take-home methadone dominantly on measuring access or utilization rather than quality quantities, will need to be evaluated to determine if these changes or outcomes, and states reported that they were still identifying the were associated with poor treatment outcomes, or is worth the measures beyond rates of opioid prescribing they want to focus on tradeoff for enhanced access to MOUD. as indicators of improved quality. However, state Medicaid agen- cies are often not equipped to create and validate their own quality Our analysis is primarily limited in that our sample of states may measures, and each state doing so may be problematic towards not be generalizable to other states in the US. The states are not comparing quality and outcomes across states. Our findings point distributed evenly across the US geographically, had higher rates to an opportunity for states to lead the demand for such measures of overdose death rates, and were more likely to expand Medicaid and work with researchers to produce policy-relevant metrics. eligibility under the ACA. The views on and policies that affect State policymakers and researchers must collaborate as to which SUD treatment in the southeast or southwest (for example) may measures are both feasible and meaningful to support policy efforts. not reflect those of this group of states. In addition, our analysis is This is one objective of the MODRN project. best viewed as a series of case studies rather than other qualitative 9 State Medicaid Agencies' Multi-Faceted Response to the Opioid Epidemic studies that seek to achieve thematic saturation. Our interviewees Marian Jarlenski, PhD, MPH were from a convenience sample of Medicaid agency staff who were University of Pittsburgh Graduate School of Public Health connected to our broader project, and those who were referred to us by those staff members. Paul Lanier, PhD University of North Carolina School of Social Work Conclusion Alice Middleton, JD The Hilltop Institute, University of Maryland Baltimore County The trend of convergence towards less restrictive utilization man- agement policies for MOUD, broader coverage of SUD treatment Nathan Pauly, PhD services, and push to innovate in the delivery and coordination of West Virginia University Office of Health Affairs care is a meaningful shift for Medicaid programs. States will need to continue to modify and test policies to counter the negative Logan Sheets, BA economic and health effects of the COVID-19 pandemic on SUD. AcademyHealth We believe that the cross-state policy and quantitative analyses that Jeff Talbert, PhD are the aim of the MODRN project will help inform policymakers University of Kentucky College of Pharmacy as to which policies have been most successful in serving Medicaid enrollees with OUD. Julie M. Donohue, PhD University of Pittsburgh Graduate School of Public Health Acknowledgements: References This work was funded by a grant from the National Institute on 1. Kaiser Family Foundation. The Opioid Epidemic and Medicaid's Role in Facilitat- Drug Abuse (1R01DA048029-01). The authors would like to thank ing Access to Treatment. https://www.kff.org/medicaid/issue-brief/the-opioid- David Kelley, Bradley Stein, and Maureen Stewart for their con- epidemic-and-medicaids-role-in-facilitating-access-to-treatment/. Published tribution to the development of the interview guide used in this 2019. Accessed February 28, 2020. analysis. 2. Centers for Medicare & Medicaid Services. 1115 Substance Use Disorder Demonstrations. https://www.medicaid.gov/state-resource-center/innovation- accelerator-program/program-areas/reducing-substance-use-disorders/1115- Authors sud-demonstrations/index.html. Published 2020. Accessed March 4, 2020. Evan S. Cole, PhD (Corresponding Author) 3. National Center for Health Statistics. Drug Overdose Mortality by State. https:// University of Pittsburgh Graduate School of Public Health www.cdc.gov/nchs/pressroom/sosmap/drug_poisoning_mortality/drug_poison- ing.htm. Published 2021. Accessed February 21, 2021. Crabtree Hall A616, 130 DeSoto Street Pittsburgh, PA 15261 4. Leshner AI. Addiction Is a Brain Disease, and It Matters. Science. 1997;278(5335):45-47. 412-383-2087 evancole@pitt.edu 5. Bettinardi-Angres K, Angres DH. Understanding the Disease of Addiction. Journal of Nursing Regulation. 2010;1(2):31-37. Susan Kennedy, MPP, MSW 6. The American Psychiatric Publishing textbook of substance abuse treatment, 4th AcademyHealth ed. Arlington, VA, US: American Psychiatric Publishing, Inc.; 2008. Amy Raslevich, MBA, MPP 7. Morey LC. Patient Placement Criteria: Linking Typologies to Managed Care. Alcohol Health Res World. 1996;20(1):36-44. University of Pittsburgh Graduate School of Public Health 8. Smith DE. The Evolution of Addiction Medicine as a Medical Specialty. Virtual Marguerite Burns, PhD Mentor. 2011;13(12):6. University of Wisconsin – Madison School of Medicine 9. American Society of Addiction Medicine. About the ASAM Criteria. American and Public Health Society of Addiction Medicine. https://www.asam.org/asam-criteria/about. Ac- cessed February 25, 2021. Sarah Clark, MPH 10.Overview of Substance Use Disorder (SUD) Care Clinical Guidelines: A Resource University of Michigan School of Medicine for States Developing SUD Delivery System Reforms. In: Program CMIA, ed2017:26. Dushka Crane, PhD, LSSBB 11.Neale, B. (2017). Strategies to Address the Opioid Epidemic. CMS Center for The Ohio State University College of Medicine Medicaid and CHIP Services. Baltimore, MD: 14. Peter Cunningham, PhD 12.Wachino, V. (2015). New Service Delivery Opportunities for Individuals with a Sub- stance Use Disorder. Centers for Medicare and Medicaid Services. Baltimore, MD. Virginia Commonwealth University School of Medicine 10 State Medicaid Agencies' Multi-Faceted Response to the Opioid Epidemic 13.Isaacson R, Shea-Delaney E, Waldman B. Implementing the ASAM Criteria 33.Miller BF, Ross KM, Davis MM, Melek SP, Kathol R, Gordon P. Payment reform for SUD Treatment through Medicaid Managed Care. Princeton, NJ: Princeton in the patient-centered medical home: Enabling and sustaining integrated behav- University; 11/2017 2019. ioral health care. American Psychologist. 2017;72(1):55-68. 14.Tager S, Watkins D, Chatlos J, Nelson L. Enhanced substance use disorders 34.Saunders EC, Moore SK, Walsh O, et al. "It's way more than just writing a pre- standards for behavioral health organizations. The Joint Commission R3 Report. scription": A qualitative study of preferences for integrated versus non-integrated 2019(25). treatment models among individuals with opioid use disorder. Addiction science & clinical practice. 2021;16(1):1-16. 15.Text - H.R.6983 - 110th Congress (2007-2008): Paul Wellstone and Pete Do- menici Mental Health Parity and Addiction Equity Act of 2008 | Congress.gov | 35.Woltmann E, Grogan-Kaylor A, Perron B, Georges H, Kilbourne AM, Bauer Library of Congress. MS. Comparative effectiveness of collaborative chronic care models for mental health conditions across primary, specialty, and behavioral health care set- 16.Text - H.R.4058 - 104th Congress (1995-1996): Mental Health Parity Act of 1996 | tings: systematic review and meta-analysis. American Journal of Psychiatry. Congress.gov | Library of Congress. 2012;169(8):790-804. 17.Text - H.R.3590 - 111th Congress (2009-2010): Patient Protection and Affordable 36.Wesson DR, Ling W. The clinical opiate withdrawal scale (COWS). Journal of Care Act | Congress.gov | Library of Congress. psychoactive drugs. 2003;35(2):253-259. 18.Frank RG, Beronio K, Glied SA. Behavioral health parity and the Affordable Care 37.Goodnough A. In Rehab, 'Two Warring Factions': Abstinence vs. Medication. Act. J Soc Work Disabil Rehabil. 2014;13(1-2):31-43. The New York Times2018. 19.Rosenbaum S, Westmoreland TM. The Supreme Court's surprising decision on 38.We note that in many states prior authorization is still required for high doses of the Medicaid expansion: how will the federal government and states proceed? buprenorphine (often > 24 or 16 mg) as well as sub-populations such as pregnant Health affairs (Project Hope). 2012;31(8):1663-1672. women. 20.McMullen E. Implementation of the Mental Health Parity and Addiction Equity 39.Clark RE, Baxter JD. Responses of state Medicaid programs to buprenor- Act in Medicaid and CHIP. Washington DC2021. phine diversion: doing more harm than good? JAMA internal medicine. 2013;173(17):1571-1572. 21.H.R.34 - 114th Congress (2015-2016): 21st Century Cures Act | Congress.gov | Library of Congress. Congressgov. 2016. 40.Interviewees from Wisconsin noted that residential treatment would be covered once their 1115 SUD waiver was implemented in 2020, and that although partial 22.Text - H.R.6 - 115th Congress (2017-2018): SUPPORT for Patients and Com- hospitalization by that name is not covered, coverage for intensive outpatient/day munities Act | Congress.gov | Library of Congress. treatment is similar. . 23.AcademyHealth. The Medicaid Outcomes Distributed Research Network 41.Pennsylvania Department of Human Services. Centers of Excellence. https:// (MODRN). https://academyhealth.org/about/programs/medicaid-outcomes- www.dhs.pa.gov/Services/Assistance/Pages/Centers-of-Excellence.aspx. Pub- distributed-research-network-modrn. Published 2020. Accessed January 5, 2020. lished 2020. Accessed March 17, 2020. 24.AcademyHealth. State-University Partnership Learning Network. https://acad- 42.Mohamoud S, Idala D, Perez R, Malomo-Paris K. Health Home program evalua- emyhealth.org/about/programs/state-university-partnership-learning-network. tion: CY 2013 to CY 2018. Baltimore MD2021. Published 2020. Accessed January 5, 2020. 43.MACPAC. Report to Congress: Utilization Management of Medication-Assisted 25.AcademyHealth. Medicaid Medical Directors Network (MMDN). https://acad- Treatment in Medicaid. Washington DC2019. emyhealth.org/about/programs/medicaid-medical-directors-network-mmdn. Published 2020. Accessed January 5, 2020. 44.AcademyHealth. Medicaid Demonstration Evaluation Learning Collaborative. https://academyhealth.org/about/programs/Medicaid-Demonstration-Evalua- 26.Dedon L. Using Emergency Declarations to Address the Opioid Epidemic: Les- tion-Learning-Collaborative. Published 2021. Accessed March 18, 2021. sons Learned from the States. 2018. 45.Centers for Medicare & Medicaid Services. Adult Core Set Reporting Resources. 27.House USW. Ending America's Opioid Crisis. https://www.whitehouse.gov/ https://www.medicaid.gov/medicaid/quality-of-care/quality-of-care-perfor- opioids/. Published 2017. Accessed. mance-measurement/adult-and-child-health-care-quality-measures/adult-core- set-reporting-resources/index.html. Published 2020. Accessed March 9, 2020. 28.A.G. Schneiderman Announces National Settlement With Anthem To Discon- tinue Pre-Authorization For Opioid Addiction Treatment Drugs [press release]. 46.Centers for Disease Control and Prevention. Overdose Deaths Accelerating 1/29/2017 2017. During COVID-19: Expanded Prevention Efforts Needed. https://www.cdc.gov/ media/releases/2020/p1218-overdose-deaths-covid-19.html. Published 2020. Ac- 29.Barenie RE, Kesselheim AS, Tsacogianis T, Fischer MA. Associations Between cessed January 5, 2020. Copays, Coverage Limits for Opioid Use Disorder Medications, and Prescribing in Medicaid, 2018. Med Care. 2021;59(3). 47.Kaiser Family Foundation. Total Monthly Medicaid and CHIP Enrollment. https://www.kff.org/health-reform/state-indicator/total-monthly-medicaid- 30.Mark TL, Parish W, Zarkin GA. Association Between Medicare and FDA Policies and-chip-enrollment/?currentTimeframe=3&selectedRows=%7B%22state and Prior Authorization Requirements for Buprenorphine Products in Medicare s%22:%7B%22kentucky%22:%7B%7D,%22michigan%22:%7B%7D,%22m Part D Plans. Jama. 2019;322(2):166-167. aryland%22:%7B%7D,%22ohio%22:%7B%7D,%22north-carolina%22:%7B %7D,%22pennsylvania%22:%7B%7D,%22virginia%22:%7B%7D,%22west- 31.Park Y, Raza S, George A, Agrawal R, Ko J. The effect of formulary restrictions on virginia%22:%7B%7D,%22wisconsin%22:%7B%7D%7D%7D&sortModel=%7B patient and payer outcomes: a systematic literature review. Journal of managed %22colId%22:%22Location%22,%22sort%22:%22asc%22%7D. Published 2019. care & specialty pharmacy. 2017;23(8):893-901. Accessed July 6, 2020. 32.Bauer MS, Weaver K, Kim B, et al. The collaborative chronic care model for 48.United States Census Bureau. 2018 National and State Population Estimates. mental health conditions: from evidence synthesis to policy impact to scale-up https://www.census.gov/newsroom/press-kits/2018/pop-estimates-national-state. and spread. Med Care. 2019;57(10 Suppl 3):S221. html. Published 2018. Accessed February 7, 2020. 11 State Medicaid Agencies' Multi-Faceted Response to the Opioid Epidemic 49.Kaiser Family Foundation. Medicaid MCO Enrollment by Plan and Parent Firm, 52.Substance Abuse & Mental Health Services Administration. What is a Health March 2019. https://www.kff.org/medicaid/state-indicator/medicaid-mco-enroll- Home? 2013. ment-by-plan-and-parent-firm-march-2019/?currentTimeframe=0&selectedRow s=%7B%22medicaid-mcos-by-state%22:%7B%22kentucky%22:%7B%22all%22:% 53.Substance Abuse & Mental Health Services Administration. Care Coordination. 7B%7D%7D,%22maryland%22:%7B%22all%22:%7B%7D%7D,%22michigan%2 https://www.integration.samhsa.gov/workforce/care-coordination. Published 2:%7B%22all%22:%7B%7D%7D,%22ohio%22:%7B%22all%22:%7B%7D%7D,%2 2020. Accessed April 24, 2020. 2pennsylvania%22:%7B%22all%22:%7B%7D%7D,%22virginia%22:%7B%22all% 22:%7B%7D%7D,%22west-virginia%22:%7B%22all%22:%7B%7D%7D,%22wisc 54.Agency for Healthcare Research and Quality. What is Integrated Behavioral onsin%22:%7B%22all%22:%7B%7D%7D%7D%7D&sortModel=%7B%22colId% Health Care? https://integrationacademy.ahrq.gov/about/what-integrated-behav- 22:%22State%22,%22sort%22:%22asc%22%7D. Published 2019. Accessed July 7, ioral-health. Published 2020. Accessed April 29, 2020. 2020. 55.Substance Abuse & Mental Health Services Administration. What is Integrated 50.Kaiser Family Foundation. Status of State Action on the Medicaid Expansion De- Care? https://www.integration.samhsa.gov/resource/what-is-integrated-care. cision. https://www.kff.org/health-reform/state-indicator/state-activity-around- Published 2020. Accessed April 24, 2020. expanding-medicaid-under-the-affordable-care-act/?currentTimeframe=0&so rtModel=%7B%22colId%22:%22Location%22,%22sort%22:%22asc%22%7D#. 56.Substance Abuse & Mental Health Services Administration. SAMHSA-HRSA Published 2020. Accessed February 7, 2020. Center for Integrated Health Solutions - Glossary. https://integration.samhsa.gov/ glossary. Published 2020. Accessed March 2, 2020. 51.Centers for Medicare & Medicaid Services. 2018 Share of Medicaid Enrollees in Managed Care. https://data.medicaid.gov/Enrollment/2018-Share-of-Medicaid- Enrollees-in-Managed-Care/cfcx-qyg7/data. Published 2018. Accessed July 6, 2020. 12