States Should Measure Opioid Use Disorder Treatment to Improve Outcomes

Overview

The most effective treatments for opioid use disorder (OUD) are medications: methadone, buprenorphine, and naltrexone.¹ However, there are substantial gaps between the number of people who need these medications and those who receive them,² and the United States continues to experience a devastating number of lives lost to the opioid epidemic.

To close these gaps, improve treatment overall, and save lives, states need a set of core metrics to track relevant data and provide a comprehensive picture of care for OUD—from diagnosis through recovery. Such metrics would allow policymakers to measure successes, identify areas for improvement, and take appropriate actions.

At present, many states have dashboards that track overdose deaths. But far too few track a full range of metrics, including how many people are diagnosed with OUD, the number of providers available to treat them, use of FDA-approved medications, or treatment retention and outcomes. In fact, just a third of states publicly report any type of metric related to OUD treatment.³

Faced with a lack of consensus among state health officials and policymakers on which metrics should be used to assess care for OUD, The Pew Charitable Trusts convened an expert panel in September 2021 comprising state and federal officials, people with lived experience, advocates, treatment providers, and experts in health measurement and analytics. (See Appendix for a full list of panelists.) The panel worked to identify core OUD treatment measures using a method known as the “cascade of care,” which examines key measurements in the life cycle of a chronic disease, cascading from the initial diagnosis to the final outcome.

This method has been applied successfully worldwide to manage hepatitis C, HIV, diabetes, and other chronic diseases. In the United Kingdom, for example, health authorities adopted the United Nations Joint Programme on HIV/AIDS’ goals of increasing the rates of each stage of the HIV care cascade (diagnosis, initiation of treatment, and viral suppression) to 90%.⁴ They reached these goals in 2017 and, based on this success, set a new goal of zero HIV transmission by 2030.⁵

Pew’s expert panel agreed on the following core measures for tracking opioid use disorder. The measures cascade from the initial diagnosis of OUD through treatment, retention in treatment, and recovery, and are explained in detail in this brief.

• OUD diagnosis: Percentage of people who had a documented OUD diagnosis (e.g., on an insurance claim).
• Percentage of people assessed using a standardized screening tool: Rate of assessment using a tool such as the Drug Abuse Screening Test (DAST) or the Tobacco, Alcohol, Prescription Medication, and Other Substance Use (TAPS) tool.
• Use of pharmacotherapy for OUD: Percentage of people diagnosed with an OUD who received medication to treat it.
• OUD provider availability: Number of providers and treatment programs that can provide medication for opioid use disorder (MOUD).
• Continuity of pharmacotherapy for OUD: Percentage of people receiving MOUD who use it for at least six months.
• Initiation of OUD treatment and engagement in OUD treatment: Percentage of people who initiate treatment within two weeks of diagnosis; percentage of people with two or more services within the first month after initiating treatment.
• Follow-up after an emergency department visit for substance use: Percentage of people who receive follow-up care for substance use disorder or an overdose within seven days, and 30 days after visiting an emergency department for a substance use disorder (SUD)-related issue.
• One or more patient-reported outcome measures to be determined by each state: Percentage of individuals who achieve an improved level of functioning or quality of life.

The panel recommended that all states implement these core measures to give state health departments, Medicaid directors, policymakers, and other stakeholders better information in order to improve practices and target scarce resources where they will do the most good—whether that means increased screening, increased access to proven medications, improved transitions of care from the hospital to the community, or other data-driven interventions.

Further, Pew recommends that states develop action plans to review the data, make it public, and adjust policies over time. Specifically, states should:

• Create a data-use plan that includes:
  • Reporting data publicly to create accountability on the effectiveness of the state’s efforts to address the opioid crisis.
  • Reviewing and acting on the data regularly.
  • Working with people with OUD and treatment providers to understand the numbers and develop quality improvement initiatives.
• Disaggregate the data by race/ethnicity, age, gender, and other demographic categories to uncover and address health inequities.

Applying the Cascade of Care to OUD

The cascade of care model was first developed to assess the HIV epidemic and has since been successfully applied worldwide to improve the treatment of chronic diseases such as hepatitis C, hepatitis B, diabetes, and tuberculosis.⁶ 

In 2017, researchers funded by the National Institute on Drug Abuse (NIDA) made a general recommendation that the cascade of care framework be applied to the opioid epidemic.⁷ Several states, localities, and Tribal nations have since used the method to establish metrics for their individual programs, but no nationwide consensus on the appropriate set of measurements emerged.⁸

To formulate its recommendations, Pew’s panel applied the framework and defined the stages of the cascade  as follows:

• OUD identification or diagnosis: Having a formal diagnosis of OUD in a medical record or claim.
• Initiation of OUD treatment: Receiving OUD treatment services.
• Retention in OUD treatment: Staying engaged in OUD treatment for at least six months.
• Recovery: Reaching a state defined by the Substance Abuse and Mental Health Services Administration (SAMHSA) as “a process of change through which individuals improve their health and wellness, live self-directed lives, and strive to reach their full potential.”⁹

Using these stages, the panel laid out the core treatment measures in Table 1.

Table 1

The Core OUD Treatment Measures

Additional detail on the core OUD treatment measures selected for each stage of care follows, including the expert panel’s rationale for selecting each measure and how the cascade and supporting measures work together.

1. OUD Identification and Diagnosis

1a. Diagnosed OUD

The starting point for evaluating care is measuring the population diagnosed with an OUD. Comparing diagnosis rates, such as those captured on claims data, to prevalence estimates from the state’s own data or SAMHSA’s National Survey on Drug Use and Health will provide policymakers a sense of how well their treatment systems are identifying people with OUD—the first step toward engaging them in care. For example, efforts to apply the cascade of care in Denver indicated that fewer than half of the estimated residents with OUD, opioid misuse, or opioid poisoning had been diagnosed with OUD in the Denver Health system, a safety net provider.¹³

1b. Percentage Assessed for SUD Using a Standardized Screening Tool

States should measure how widely they are screening for OUD in order to identify more people with the condition. The U.S. Preventive Services Task Force, authorized by Congress to make recommendations for preventive services in primary care settings, recommends that providers ask all adults if they partake in unhealthy drug use.¹⁴

However, research suggests that such widespread screening does not happen. A study of rural Pennsylvania counties found that fewer than 20% of people with OUD in the Medicaid program were diagnosed in primary care, although they regularly sought care in these settings.¹⁵ Screening rates are lacking in prisons, as well. Interviews with representatives from 21 state prison systems, representing 583 individual facilities, found that fewer than half of the prisons conducted screening.¹⁶ Mental health specialists also frequently underdiagnose substance use disorders among their patients.¹⁷

Using these two measures—diagnosed OUD and patients assessed for SUD using a standardized screening tool— will provide states with an understanding of how effectively they are identifying people in need of care. Policymakers can then target efforts to increase screening rates in settings where they will have the greatest impact.

2. Initiation of OUD Treatment

2a. Use of pharmacotherapy for OUD

Medication is the most effective treatment for opioid use disorder. Methadone and buprenorphine, in particular, reduce the risk of overdose, improve treatment retention, and, by reducing risky behaviors such as the injection of illicit drugs, also decrease the transmission of infectious diseases such as HIV and hepatitis C.¹⁸ Injectable naltrexone is effective in helping people reduce opioid cravings, reduce illicit opioid use, and stay in treatment.¹⁹

However, in 2020, fewer than 12% of people ages 12 or older in the U.S. with OUD in the past year received medications.²⁰ States that measure the use of pharmacotherapy for OUD have also found treatment gaps. In Rhode Island, just 27% of people at risk for OUD (those who “report heroin and/or extra-medical opioid use”) had initiated medication, while in Florida 40% of those with an OUD diagnosis received MOUD. Although these numbers are not directly comparable because Rhode Island includes people without a diagnosis in the rate, and Florida only includes those with a diagnosis, these findings indicate a need for strategies to increase medication uptake.²¹

2b. OUD provider availability

A lack of available providers is one factor that prevents people with OUD from receiving medication. As of July 2020, more than one-third of all rural counties lacked a single buprenorphine prescriber.²² People with OUD also face challenges accessing opioid treatment programs (OTP), the only place where they can obtain methadone. A study of five states found the average drive time to reach an OTP was nearly 40 minutes.²³

However, these access challenges are not evenly distributed within states. The same study found that, when comparing urban and rural communities, urban patients had a much shorter average drive to an OTP—less than 10 minutes—compared with nearly 50 minutes for people in rural areas.²⁴ Buprenorphine access also varies within states; while some counties don’t have any treatment providers, neighboring jurisdictions could have several.²⁵ 

Further, the medication someone uses may be determined by where they live, rather than what works best for them. Although OTPs that distribute methadone are more common in predominantly Black and Hispanic communities, White communities are more likely to have buprenorphine providers.²⁶

By measuring provider availability and stratifying this data by geography, states can identify the areas with greatest need and develop plans to increase the number of providers in those places.

3. Retention in OUD Treatment

3a. Continuity of pharmacotherapy for OUD

To achieve good outcomes, people with OUD must do more than just start medication; they need to stay on it.²⁷ Studies have shown that long-term treatment can lead to better outcomes in employment, health, and criminal justice involvement.²⁸ In addition, it is safer for patients to continue prescribed medications for OUD than to suddenly stop treatment, which can increase the risk of overdose. For these reasons, federal guidelines discourage the removal of patients from treatment programs.²⁹

To evaluate whether people with OUD are continuing to use medications, states should use the metric “Continuity of pharmacotherapy for OUD” to measure the percentage of people who receive MOUD for 180 days or six months.

An important aspect of this measure: It reports treatment retention rates for each of the three FDA-approved medications for OUD. Because the medications are available in different settings and under different conditions (e.g., methadone is available only in OTPs; prescribers must get permission from federal agencies to prescribe buprenorphine), tracking differences in retention across them could indicate a need to modify medicationspecific policies and practices. ³⁰ For example, if data showed that patients taking methadone were less likely to stay in treatment, policymakers might determine that they need to remove barriers to the medication.

3b. Initiation of OUD treatment and engagement in OUD treatment

In Medicaid, many people who begin treatment for OUD leave care in the early stages. The measure “Initiation of OUD treatment and engagement in OUD treatment” indicates whether people continue to receive services during the first six to seven weeks after diagnosis. Among the 39 states reporting this measure to the Centers for Medicare and Medicaid Services in 2020, an average of 55% of people between the ages of 18 and 64 began treatment within 14 days of a diagnosis, but only 39% received two or more additional treatment services over the following 34 days.³¹ These rates varied greatly by state, suggesting that some states may have more promising approaches to retaining people in care early on.³²

Using this measure and stratifying it by treatment setting, geography, race, gender, and other characteristics will help states identify who is at the greatest risk for leaving treatment in the early stages and develop strategies to help them stay.

3c. Follow-up after an emergency department (ED) visit for substance use

People visiting an ED for a substance use disorder are at a high risk for a subsequent fatal overdose.³⁵ As a result, the American College of Emergency Physicians recommends that EDs provide medication to people with SUD and link them to community-based treatment. Unfortunately, however, data shows that few people who present at the ED with an overdose obtain follow-up treatment.³⁶

By measuring the rates of follow-up care, states can design interventions to keep people connected to the treatment system after visiting the ED, improving their retention in care, and ultimately saving lives. However, it is important to note that this measure is not designed to specifically evaluate OUD-related care linkages—instead, it looks at substance use disorders more broadly.

4. Recovery from OUD

Patients report that one of the things they most want out of substance use treatment is to improve the quality of their lives, a central element of recovery. ³⁷ Yet people with OUD face many barriers to achieving this goal: stigma for both addiction and use of medication to manage it; difficulty finding a job or stable housing that supports recovery; interactions with the criminal legal system; co-occuring mental health conditions; and other challenges that may require support from the treatment system.³⁸

It is critical for state policymakers to know if their treatment systems are improving lives and helping people recover, but because they are not systematically collecting and using data, their ability to do so is currently limited.

As a baseline, states can use data collected by SAMHSA in its National Survey on Drug Use and Health. This survey has since 2020 asked respondents who indicated that they have a problem with drug use whether they perceived themselves to be in recovery.³⁹ However, although this survey provides an overall estimate of the number of residents in recovery, it does not provide the targeted information that policymakers need to improve the treatment system, such as where they received care.

One potential source for more detailed data is the SAMHSA National Outcome Measures, which states are required to collect and report for individuals receiving treatment in addiction programs that receive federal block grant funding.⁴⁰ However, this data has multiple limitations. First, it does not include people receiving treatment in doctor’s offices. Second, it is collected using a tool called the GPRA (which stands for Government Performance and Results Act), which state officials have criticized for its length and potential for retraumatizing patients with questions about sensitive subjects—although a recent update to the tool attempts to address the latter concern.⁴¹ Finally, one of the outcome measures is successful discharge, which may be misinterpreted to imply that MOUD should be time-limited, despite research supporting long-term treatment.⁴²

Other data that may be readily available to states, such as employment status and abstinence from drugs, may not accurately reflect the goals of each individual client.

Measures that strive to reflect whether clients are meeting their own goals—known as patient-reported outcome performance measures (PRO-PMs)—are currently limited in the field of substance use disorder. A 2020 report from the National Quality Forum identified the development of “quality of life/level of functioning measures” as a priority area of measurement for CMS.⁴³

Progress is being made in establishing these measures. With funding from CMS, the American Psychiatric Association is developing measures for the “improvement or maintenance of symptoms, functioning, and recovery.” These measures may be available to states in the future.⁴⁴

However, because states don’t have access to high-quality, actionable data on whether treatment is improving lives, the expert panel decided not to endorse a recovery measure at this time. Instead, the panel recommended that states collaborate with people with lived experience and providers to select one or more patient-reported outcomes and test strategies to collect and report the data.

As a starting point, state officials can select from existing recovery assessment tools developed to ask people with SUD about their quality of life. For example, the Brief Assessment of Recovery Capital (BARC-10) consists of 10 questions about substance use, mental and physical health, community involvement, social support, and other aspects of recovery capital—defined as “a variety of psychological, physical, social, and environmental resources” that mitigate the stress of achieving and maintaining recovery.⁴⁵

For more options, see the Core OUD Treatment Measures Toolkit.    

Recommendations

Implementing the core measures will give policymakers the data they need to understand the existing treatment landscape, including the number of people in a state with OUD who are receiving medication, as well as potential intervention points to increase the use of MOUD. However, to effectively close the gap between those who need medication and those who receive it, state officials need to think carefully about how the data is used.

Pew developed the following recommendations for state officials implementing the Core OUD Treatment Measures informed by the expert panel and a review of the literature:

Measures Must Be Accompanied by Planning and Action

Publicly report OUD treatment data

States commonly report opioid overdose deaths on public dashboards, but they should also report the core measures described here, including differences in MOUD availability and utilization by demographic groups and geographic areas.⁴⁶ A systematic review and meta-analysis of public reporting on clinical measures found that public reporting is associated with improvements in outcomes.⁴⁷

Design a plan to use OUD treatment data

In addition to publicly reporting OUD treatment metrics, state officials charged with responding to the opioid crisis should develop a plan that involves regularly reviewing the data and using it to inform program and policy decisions. This plan should include identifying a point person to take responsibility for ensuring that data is updated and reviewed; convening stakeholders across state agencies, provider groups, and communities who can help interpret and act on the data; and reviewing the resources available to support quality improvement activities.

The case study below illustrates the steps that leaders in one state took when establishing a data-driven approach to the opioid crisis. For additional implementation considerations, see the toolkit.

Integrate community members and providers into data-driven decision-making

To understand the data, states need to collaborate with those in the best position to understand what it means and how to act on it—people with OUD and providers. People with lived experience can provide powerful insights into aspects of the OUD treatment system that are working well and areas that need improvement.

For example, researchers in Minnesota collaboratively developed a cascade of care for an American Indian Tribal Nation by establishing a community advisory board of members of the nation. This board collaborated with the researchers to interpret findings, noting that members of the Tribe with OUD had a lot of distrust and fear of the health care system while navigating treatment. They also found that treatment did not always incorporate or even acknowledge traditional healing practices.⁵¹ These insights provided a starting point for changing the treatment system to better meet the needs of Tribal members.

When establishing these collaborations, state officials should be careful to treat participants respectfully and not cause harm. Ethical collaborations benefit everyone involved. For state officials, the benefits are clear: a deeper understanding of how to serve their residents. For community members, the desired benefits vary but include helping others, having their voices heard, developing leadership opportunities, gaining new skills, and receiving monetary compensation.⁵² Community members should also have access to the data and findings; after all, it is their lives that are represented in the numbers.⁵³ 

States should also involve treatment providers when analyzing these metrics. As people progress across the cascade of care, they will interact with many different providers and treatment settings. For example, they may be screened in an emergency department, receive support from a peer recovery coach, and go to an OTP for methadone. All these providers have a role to play in improving the OUD treatment system, and they must work together to support transitions from one setting to another.⁵⁴ 

To Improve Health Equity, States Must Disaggregate the Data

In just one year, from 2019 to 2020, overdose deaths rose by 30% throughout the United States.⁵⁵ But the trend in the aggregate masks stark differences in death rates by race and ethnicity. A study by the federal Centers for Disease Control and Prevention looked at overdose fatalities in 25 states and the District of Columbia and found that in these jurisdictions, the death rate increased the most among Black and American Indian/Alaska Native people (44% and 39%, respectively).⁵⁶ By comparison, the increase for White people was 22%.⁵⁷And because of systemic barriers to treatment, Black and Hispanic people are less likely to initiate and stay on medication than their White peers.⁵⁸

These inequities demonstrate that states must disaggregate data by factors such as race and ethnicity, geographic location, gender, and age. Doing so can provide decision-makers in government and health care with rich information about where resources can best be directed to help ensure that all people have equitable access to services at each stage of the cascade of care. As noted in a recent Commonwealth Foundation report on the inadequacy of race/ethnicity data in government data systems, “We cannot improve what we cannot see or fail to measure. The path to health equity starts with collecting complete, accurate, and reliable race and ethnicity data.”⁵⁹

Addressing data challenges

States can use data from insurers to track Core OUD Treatment Measures, but the quality of race and ethnicity data varies. Many commercial health plans do not collect race and ethnicity data.⁶³ Although state Medicaid agencies are required to collect race/ethnicity data, and all collect self-reported data (a best practice) through the application process, their completeness and accuracy varies among states.⁶⁴ Mismatches between responses on Medicaid applications and Census Bureau surveys and missing data have prompted researchers to raise “high concerns” about Medicaid race/ethnicity quality in 19 states.⁶⁵

Experts point to a number of steps that Medicaid agencies could take to improve their reporting of race and ethnicity, including widespread use of intake workers to help applicants fill out forms and the use of CMSrecommended uniform reporting forms that feature more granularity in the reporting on Hispanic and Asian ethnicity and race categories.⁶⁶ In lieu of these steps, to identify potential inequities among communities with different racial and ethnic profiles, states can compare performance on core measures. These techniques could be applied to both commercial and Medicaid claims.

Conclusion

The year 2021 was the worst on record for fatal opioid overdoses. Yet when states use data to address this crisis, there are signs of hope. In Missouri, state officials used data to pinpoint and address racial disparities in access to treatment. In Rhode Island, metrics informed an effort to decrease the stigma of using lifesaving medications for opioid use disorder. And in Kentucky, data made the case for removing barriers to these medications, helping people to get care when they were ready.

These examples show what’s possible when states collect and act on data. By using the Core OUD Treatment Measures to inform policy and practice, states can help transform their treatment systems to save lives.

Appendix A: Expert panel participants

Pew selected the expert panel to provide a diversity of perspectives on measuring the OUD cascade of care. By including the recovery community and harm reduction perspectives, treatment providers and quality measurement experts, federal officials, and the state agency staff who would ultimately use the data, Pew convened an expert panel with a range of points of view, all of whom engaged in a process to go from constructive disagreement to consensus. 

Appendix B: Full metrics inventory and methodology

To identify the most appropriate metrics for tracking the cascade of care, Pew first created an inventory of all existing SUD and OUD measures, drawing from the following sources:

• National Quality Forum (NQF) Quality Positioning System.
• Centers for Medicare and Medicaid Services Measures Inventory Tool (CMIT) (which includes measures from the Behavioral Health Core Set).
• Medicaid 1115 substance use disorder (SUD) waiver demonstration monitoring metrics.
• Treatment-related metrics found in state opioid dashboards. If multiple states reported a similar measure, the panel treated it as a single measure. Pew identified state dashboards from a review of the Carolina Center for Health Informatics at UNC Chapel Hill’s “Opioid Dashboards and Data” page supplemented by a Google search for all 50 states and the District of Columbia in the format of “[state] opioid dashboard.” The research team only examined interactive dashboards for which metrics appeared to be updated regularly by a state or local government; the team did not review documents such as PDFs or PowerPoints for which metrics may have been reported at a point in time.

This inventory was developed in the summer of 2021; these sources may have had changes made to them since then.

To ensure that the inventory metrics were relevant to evidence-based OUD treatment, Pew excluded any measures that met the following criteria:

• Mental health only.
• A specific SUD other than opioid use (e.g., tobacco or alcohol use only).
• Controlled substance prescribing.
• Acute care utilization.
• Overdoses.
• Spending.
• Specific to inpatient psychiatric treatment care setting.
• Non-MOUD treatment utilization.
• Duplicative of other measures.

Pew then classified the metrics according to the stage of the care cascade that they measured or impacted. Some measures could be applied to multiple stages and were therefore assigned to multiple groups. Several measures for related health conditions that did not meet the exclusion criteria were considered even though they could not be placed in the care cascade. However, none of these measures was ultimately selected.

Complete metrics inventory