Abstract
Medications for opioid use disorder (MOUD), including buprenorphine, are effective for the treatment of incarcerated persons (IPs) with opioid use disorder (OUD). Racial disparities in such prescribing have been researched in both community and correctional settings. We describe a performance improvement (PI) project to reduce the disparities in prescribing buprenorphine in the New Jersey Department of Corrections that involved public-facing educational efforts directed toward both prescribers and potential patients. Using reports designed to collect baseline and follow-up summary data from the electronic medical record and institutional database, we showed that, from 2021 to 2024, for the entire NJDOC census, the rate of prescribing for Black IPs increased from 3.9 to 6.3 percent (p < .00001) and for Hispanic IPs from 6.6 to 9.6 percent (p = .0009), whereas prescribing for White IPs was unchanged (19.0–18.0%, p = .362). When considering interest in buprenorphine by persons with OUD (either expressing interest or receiving a prescription for it), we noted increases in Black (41.1–56.5%, p < .0001) and Hispanic (54.9–69.7%, p = 0) IPs, but not White (71.6–73.4%, p = .360) IPs. Compared with a study on MOUD prescribing in this setting in 2019, the proportion of IPs prescribed buprenorphine identifying as Black increased (17.0-41.1%, p < .00001). These results support educational efforts for improving access to treatment with MOUD in carceral settings.
Opioid use disorder (OUD) is highly prevalent in incarcerated persons (IPs). Over half of U.S. adults with OUD engage with the criminal justice system in any given year,1 with at least 15 percent of IPs nationwide carrying a diagnosis of OUD.2 Former IPs with OUD often resume using substances, with a substantial risk of overdose death, especially in the early weeks after release.3,–,6 Medications for opioid use disorder (MOUD), including buprenorphine, methadone, and naltrexone, are widely regarded as effective treatments for OUD, with evidence (more consistently for agonist MOUD) supporting reduced drug use, criminal activity, and mortality, including from overdoses.7,–,9 In a study of overdose deaths among 6,932 Medicare beneficiaries by Kuo and colleagues, less than 10 percent were prescribed MOUD at the time of their death.10 Numerous professional organizations, including the National Commission on Correctional Health Care, the American Psychiatric Association, the American Correctional Association, and the American Society for Addiction Medicine support the prescription of MOUD, including during incarceration, although the implementation of these recommendations by jails and prisons in the United States has been inconsistent.11
Rutgers University Correctional Health Care (UCHC) provides medical and mental health treatment for IPs within the New Jersey Department of Corrections (NJDOC). UCHC and NJDOC offer comprehensive MOUD pharmacotherapy services for IPs, and buprenorphine has been prescribed in every facility there since 2017. The NJDOC houses state-sentenced IPs, with most being accepted from county jails. Buprenorphine is the MOUD most preferred by patients within the NJDOC and the one most often prescribed. In fact, consistently, over 90 percent of IPs on MOUD in the NJDOC receive a buprenorphine product, with the remainder on methadone or a naltrexone product. Buprenorphine may be started either prerelease or as long-term maintenance during incarceration when clinically appropriate. It may be prescribed either by psychiatric providers for patients in treatment for a mental health disorder or by general medical providers for any other IP. Based on UCHC’s internal guidelines for the treatment of substance use disorders, the threshold for prescribing prerelease MOUD is lower.
A review by Tamburello and colleagues12 described the practice patterns related to MOUD in the NJDOC, as well as the characteristics of IPs prescribed buprenorphine in calendar year 2019. An unexpected finding was that, out of the cohort of 875 IPs prescribed buprenorphine at the conclusion of 2019, the majority were White (67.3%). Black (17.0%) and Hispanic (13.9%) IPs were prescribed buprenorphine less frequently than expected based on their representation in the overall NJDOC census (21.5% White, 61.6% Black, and 15.8% Hispanic).12 Moreover, almost double the number of Black IPs as White IPs had a diagnosis of OUD but were not prescribed buprenorphine.12 The terms Black and Hispanic are hereafter used in accordance with the language of the NJDOC census, although these categories may denote a wide range of cultural diversity among Black, African-American, Hispanic, and Latino peoples.
Perhaps the results from this study should have been anticipated. Black patients in the community are less likely than White patients to be prescribed opioids, even for obviously painful injuries like fractures.13–14 Other research suggests that non-White persons are offered buprenorphine less often in the community compared with White persons, even among those experiencing an opioid overdose.15,–,18 A related concern raised by Abraham and colleagues19 was a disparity observed in terms of less access to buprenorphine prescribers for Black and Hispanic recipients of Medicare Part D. Hirchak et al.20 described a lack of waivered buprenorphine providers in zip codes with greater numbers of Hispanic persons as an “obstacle.” Dong and colleagues21 reviewed over 11 million buprenorphine prescriptions, finding that 84.1 percent were prescribed to White persons, 8.1 percent to Black persons, and 6.3 percent to Hispanic persons, with fewer Black persons receiving extended treatment. News coverage of the opioid public health emergency primarily depicts young White males as the most affected group.22 Although OUD may be more common in White Americans, it is increasingly common in Black Americans over the age of 35,23 and the median age of NJDOC IPs is 37.24 As overdose deaths in Black and Hispanic Americans are increasing,25,–,30 along with the well-described disproportionate effects of mass incarceration by race,31 identified gaps in offering effective treatments require urgent attention.
According to the Substance Abuse and Mental Health Services Administration’s (SAMHSA) 2020 resource documents on race and MOUDs,29,30 there are numerous factors to consider in addressing these disparities, including not only inherent bias in prescribers but also culture, history, and both real and perceived concerns about access to care in the community. In terms of culture, Black Americans with substance use disorders must cope with two stigmatized statuses.29 Many non-White persons do not subscribe to the disease model of substance use disorders.29 One respondent told SAMHSA that “addiction is seen as a weakness not a disease” (Ref. 29, p 8). Political and law enforcement responses to widespread drug problems beginning in the 1970s led to more incarceration and disruption for Black families. The legacy of the “War on Drugs” is a fear, especially among Black persons, that seeking care for a drug problem will expose them to harsh punishment rather than treatment.29 Persons in minority communities may lack information about when treatment is necessary and the effectiveness of treatment options, and may lack access to culturally competent providers.29–30 Hispanic persons may also be concerned about a language barrier or a fear of their immigration status being exposed while seeking treatment.30 Historical distrust of the justice and health care systems is a key contextual factor in terms of providing health care services for non-White persons within prison settings.32
Methods
The investigators initiated this work as a performance improvement (PI) project. The workgroup has changed over time but has included health care staff with diversity in backgrounds, demographics, credentials, and medical specialties. We began by doing a literature review (much of which is summarized here) on evidence-based reasons for disparities in access to MOUD for non-White persons and approaches to addressing these disparities. We hypothesized that the discrepancies in the prescribing of buprenorphine previously observed could be mitigated through education, both of ourselves and of other stakeholders.
We developed an in-service training and provided this education via live or electronic means in April 2021 to UCHC medical and mental health leadership and prescribers of buprenorphine about the observations from the 2022 study on our buprenorphine prescribing practices in advance of its publication.12 The content of the training is reflected in the introduction and included the current literature about social determinants of health, known cultural factors related to the treatment of OUD, existing research about disparities in prescribing MOUD, and potential contributing factors, including inherent bias. We invited discussions on ideas to address these disparities, including partnering with the community (e.g., aftercare providers) to improve equitable access to care, education for providers, continuous quality improvement, and education of incarcerated persons with a goal to reduce stigma and encourage help-seeking.
We then designed content for distribution to IPs, informed by SAMHSA’s strategies to address opioid misuse and OUD in Black and Hispanic communities.29–30 For example, we avoided lecturing or using terms like “opioid crisis” that non-White persons may find threatening. According to the SAMSHA resource document, Black community leaders expressed concern that terms like “epidemic” or “crisis” put “residents on high alert and [trigger] fears of incarceration” (Ref. 29, p 7). Flyers were distributed to prescribers for posting in health care areas to encourage clinical conversations about MOUD with staff best able to provide accurate information (see Fig. 1). Our PI team collaborated with the NJDOC Public Information department to create a video for internal distribution via closed-circuit television and tablet computers readily available to IPs and may be seen here: https://www.youtube.com/watch?v=Q2Wligf8vuc.33 This was also made available to the public via a Rutgers University press release.34
Figure 1. Educational flyer on medications for opioid use disorder posted in medical areas.
In November 2021, we requested reports from the UCHC Quality Improvement Department for baseline data. These reports accessed the AthenaHealth Centricity electronic medical record (EMR) and NJDOC iTAG (“i” is for institutional, and TAG is not an acronym) databases, providing summary information about Department of Corrections (DOC) census by race, buprenorphine prescribing by race, diagnosis of OUD by race, and expressed interest in buprenorphine by race. IPs are asked at intake and at other key points in their period of incarceration about OUD and interest in MOUD.
Since 2019, the UCHC quality improvement (QI) department has distributed a biweekly report listing persons with a diagnosis of OUD who are not presently prescribed MOUD and are anticipated to be released to the community within the next six months. Providers are directed to proactively follow up with these individuals and ask about their interest in MOUD. One hypothesis to explain the racial disparities in prescribing buprenorphine was that such follow-up was not happening for non-White patients (a theoretical consequence of inherent bias), so we also looked at whether there were differences by race in terms of whether this question was documented as asked.
Additional sets of these reports were run in 2022, 2023, and 2024. Given the substantial social importance of our findings so far from this PI project, we wish to report them for the benefit of other carceral systems. As the analysis involved only reports and no protected health information, this study was determined to be exempt from review by the Rutgers-Robert Wood Johnson Medical School Institutional Review Board. The New Jersey Department of Corrections Departmental Research Review Board approved it as described above.
Data, when available, were carried forward from the prior study on buprenorphine prescribing in the NJDOC at the end of 2019.11 Racial identifications occurring at less than one percent in the NJDOC were not included in the analysis. Statistical analysis on categorical variables was performed using chi-square tests for three by two tables and Fisher’s exact test for two by two tables, with significance set a priori at p < .05.
Results
Table 1 and Figure 2 compare the changes in the percentage of prescribing buprenorphine over time to persons who identify as Black, White, or Hispanic. Using the numbers of persons prescribed buprenorphine by race on December 31, 201912 and the publicly available information about the numbers of NJDOC IPs (total census, irrespective of OUD diagnosis) on January 1, 2020,24 1.3 percent of Black, 14.8 percent of White, and 4.2 percent of Hispanic IPs were then prescribed buprenorphine. As of May 9, 2024, this increased for each group in that 6.3 percent of Black (p < .00001), 18.0 percent of White (p = .0004), and 9.6 percent of Hispanic (p < .00001) IPs were prescribed buprenorphine. When limiting the analysis to the time frame of the study (2021 – 2024), the rate of prescribing for Black IPs increased from 3.9 to 6.3 percent (p < .00001) and for Hispanic IPs from 6.6 to 9.6 percent (p = .0009), whereas prescribing for White IPs leveled off from 19.0 percent to 18.0 percent (p = .362).
Figure 2. Buprenorphine prescription by race (among all IPs) in the NJDOC prior to and following PI project. IP = incarcerated person; NJDOC = New Jersey Department of Corrections; PI = performance improvement; Bup = buprenorphine. ap < .00001. bp = .0004.
Percentage of All NJDOC IPs Prescribed Buprenorphine, By Race
In terms of interest in MOUD, the percentage of IPs with OUD expressing no interest in MOUD on November 16, 2021 (the first report collecting this information) was far higher in Black (67.9%) compared with White (18.2%) or Hispanic (12.8%) IPs (p < .00001). Also, there were higher numbers of Black IPs with OUD missing documentation of being asked about their interest in MOUD in 2021 (55.2% Black, 30.9% White, 12.6% Hispanic, p < .00001). In 2024, these missing documentation observations changed in various directions (37.2% Black, 72.0% White, 54.0% Hispanic, p < .00001, with all comparisons over time from 2021 to 2024 also with a p < .00001).
Considering the potential confounder that being prescribed buprenorphine would obviate the need to ask about interest (and that persons on MOUD were not included in that analysis), we looked at the percentage of IPs with OUD who were either prescribed buprenorphine or expressing an interest in MOUD, as shown in Figure 3. In 2021, 41.1 percent of Black IPs with OUD were either prescribed buprenorphine or expressing an interest in it, whereas in 2024, this increased to 56.5 percent (p < .00001). In 2021, 54.9 percent of Hispanic IPs with OUD were either prescribed buprenorphine or expressed an interest in it, which increased to 69.7 percent in 2024 (p = 0). In 2021, 71.6 percent of White IPs with OUD were either prescribed buprenorphine or expressing an interest in it, and this was similar in 2024 (73.4%, p = .36). Although the comparison by race of OUD on buprenorphine or interested versus not interested or asked was statistically significant in 2021 (Black 41.1%, White 71.6%, Hispanic 54.9%, p < .00001), in 2024, the differences were not statistically significant (Black 56.5%, White 73.4%, Hispanic 69.7%, p = .201).
Figure 3. NJDOC IPs with OUD prescribed or interested in Bup prior to and following PI project. IP = incarcerated person; OUD = opioid use disorder; Bup = buprenorphine; PI = performance improvement; NS = not significant. ap < .00001. bNS.
Table 2 and Figure 4 reflect the relative or proportional prescribing of buprenorphine by race, including the results from the study of 2019 data.12 As of the most recent reports collected in May 2024, of the total number of IPs prescribed buprenorphine, 41.1 percent were Black, 43.0 percent were White, and 15.9 percent were Hispanic. This was a statistically significant change when compared with the relative prescribing of buprenorphine by race as reported from the 2019 dataset (17.0% Black, 67.3% White, 13.9% Hispanic, p < .00001).
Figure 4. Proportional prescription of buprenorphine by race in the NJDOC prior to and following PI project. IP = incarcerated person; NJDOC = New Jersey Department of Corrections; PI = performance improvement; Bup = buprenorphine. ap < .00001, proportion of prescribing compared with December 31, 2019.
Relative Racial Identification among NJDOC IPs Prescribed Buprenorphine
Tables 3 and 4 show data, including absolute numbers of patients as well as percentages, for the prevalences by race of OUD and buprenorphine from the beginning to the end of the PI project, including the census, diagnosis of OUD, prescription of buprenorphine, and expressed interest among those who were not currently prescribed buprenorphine. The overall rate of OUD diagnosis in the NJDOC increased from 25.3 percent to 28.6 percent (p < .00001), although changes by race were only statistically significant for Black IPs (22.5–27.6%, p < .00001). Changes in OUD diagnosis for White IPs (36.7–36.0%, p = .590) and for Hispanic IPs (20.7–21.8%, p = .381) were not statistically significant. Limiting the analysis to patients with OUD prescribed buprenorphine from November 2021 to May 2024, the rates increased from 17.1 to 22.8 percent for Black IPs (p = 0) and from 32.1 to 43.9 percent for Hispanic IPs (p = .0005). The rates for White IPs were similar across this time period (51.7–50.1%, p = .462). The overall percentage of prescribing buprenorphine for a diagnosed OUD increased from 30.0 to 33.1 percent (p = .009).
NJDOC OUD Diagnosis and Bup Prescribing by Race 11/16/2021
NJDOC OUD Diagnosis and Prescribing by Race 5/9/2024
Discussion
Following a multipronged PI effort, we observed reduced racial disparities in the diagnosis of OUD and the prescribing of buprenorphine in the NJDOC. This appears to be driven by an increase in diagnosing OUD in Black IPs and an increase in prescribing of buprenorphine (and interest in being prescribed buprenorphine) for Black and Hispanic IPs with an OUD diagnosis.
Although there was an increase in overall prescribing of buprenorphine over the course of our work, our results are unlikely to be explained by an increase in general prescribing of buprenorphine, as during the study period, increases in prescribing (or interest in) buprenorphine were only observed for non-White IPs. It is an empirical question whether the apparent plateauing of buprenorphine prescribing to White IPs at around 50 percent of those with OUD is an actual ceiling effect reflecting maximal interest. There are other options (including nonpharmacologic treatments) besides buprenorphine for IPs in the NJDOC, and some may prefer to pursue MOUD in the community after release. Not all persons with OUD are ready for treatment; according to 2021 data on adults with OUD, 27.8 percent had received MOUD in the past year.35
Our baseline data suggested that, more often than not, we were not asking Black IPs about an OUD or their interest in MOUD. Rather than a confirmation of inherent bias, these data could reflect legal or institutional factors. For example, persons who have been in the system since before screening procedures were established may have been missed. Although patients may request an evaluation for substance use disorder diagnosis and treatment at any time, this is more likely to occur through the action of a provider when the IP is prerelease, so persons with longer sentences may come to the attention of MOUD prescribers later. By 2024, these numbers flipped so that it appeared that health care staff were more often documenting interest in MOUD for Black IPs with OUD and usually missing them for similarly diagnosed White and Hispanic IPs. It is possible that this reflected a focused effort to correct for inherent bias in the context of the staff education provided. Further research may be needed to monitor these changes and to consider other factors, such as offense and sentencing data not included in our protocol, that could be explanatory. Including persons prescribed buprenorphine with those expressing interest results in a more relevant measure of acceptance of the treatment.
There are several limitations to the interpretation of our results. As multiple interventions were undertaken at the same time, it is not possible to determine the relative contribution of each, and there are likely other factors we did not consider or measure. Our protocol did not include surveys of staff or patients asking them whether they were aware of or utilized any of the modalities of the PI project, although this would be operationally challenging and would require far more robust institutional review and approvals. We did not record the county of commitment, which may be relevant, as the prevalence of OUD may vary regionally and county jails may have different policies related to OUD screening, education, and treatment.36 The biweekly continuous quality improvement (CQI) report of IPs with OUD who are soon to be released and not yet prescribed buprenorphine started in 2019. UCHC does not measure the usage of that report (assuming this is practicable), although our PI training intervention for providers encouraged it. We did not assess the diversity of the workforce, which could have potentially influenced the acceptance rates of MOUD by a diverse patient population. The timescale of our study, though, limited the effects of hiring in a relatively stable workforce, and diversity of hiring is a longstanding value of Rutgers University.37
It is reasonable to question the value of relatively modest actions (like a single lecture or a flyer on the wall in the clinic), although research suggests that overly aggressive campaigns to reduce prejudice may be counterproductive.38 The effort required was modest enough to be easily reproducible in other systems. Our intention was to get the attention of staff and potential patients without being heavy handed. Although the promotional video was widely and freely distributed to IPs as well as made available to the public (and thus family members) by the NJDOC and Rutgers, we do not know how many IPs viewed it. Outside factors may have influenced our results, such as greater public awareness of OUD and MOUD. Arguing against this, a recent survey showed that, despite public education efforts in the community, Black respondents were more likely than others to believe that a primary care physician was unable to prescribe MOUD to them.39
The changes observed could simply be a natural product of program maturity. Over time, potential patients will naturally interact with others who are participating in and benefiting from treatment. Based on limited access to buprenorphine in the community,15,–,21 non-White IPs may have been less familiar with this medicine on arrival at the NJDOC. More experience and acceptance of the program would reasonably lead to more comfort and confidence in both prescribers and patients, regardless of their backgrounds. Then again, in terms of systemic culture changes, what we are observing is relatively and refreshingly brisk. It is possible that providers’ and patients’ awareness of our multifaceted (and hard to miss) efforts to foster interest in MOUD was part of a positive feedback loop that promoted such growth that was already in progress. Whatever the source of these changes, we are confident that improved access to MOUD will save lives both during incarceration and in the community.
Acknowledgements
The authors acknowledge nonauthor members of the performance improvement team: Kerri Edelman, PsyD and Paulo Juan Verdeflor, DNP, MSN, BSRT, APN, RN; report production by Patti Ford; consultation on communication with Dan Strasser, PhD of Rowan University; and video production by Chris Carden of the New Jersey Department of Corrections.
Footnotes
Disclosures of financial or other potential conflicts of interest: None.
- © 2025 American Academy of Psychiatry and the Law
