Abstract
Background:
Treatment with methadone and buprenorphine medications for opioid use disorder (MOUD) during incarceration may lead to better community re-entry, but evidence on these relationships have been mixed. We aimed to identify community re-entry patterns and examine the association between in-jail MOUD and a pattern of successful reentry defined by rare occurrence of reincarceration and preventable healthcare utilization.
Methods:
Data came from a retrospective, observational cohort study of 6 066 adults with opioid use disorder who were incarcerated in New York City jails and released to the community during 2011–14. An outcome was community re-entry patterns identified by sequence analysis of 3-year post-release reincarceration, emergency department visits, and hospitalizations. An exposure was receipt of in-jail MOUD versus out-of-treatment (42% vs. 58%) for the last 3 days before discharge. The study accounted for differences in baseline demographic, clinical, behavioral, housing, and criminal legal characteristics between in-jail MOUD and out-of-treatment groups via propensity score matching.
Results:
This study identified five re-entry patterns: stability (64%), hospitalization (23%), delayed reincarceration (7%), immediate reincarceration (4%), and continuous incarceration (2%). After addressing confounding, 64% and 57% followed the stability pattern among MOUD and out-of-treatment groups who were released from jail in 2011, respectively. In 2012–14, the prevalence of following the stability pattern increased year-by-year while a consistently higher prevalence was observed among those with in-jail MOUD.
Conclusions:
Sequence analysis helped define post-release stability based on health and criminal legal system involvement. Receipt of in-jail MOUD was associated with a marker of successful community re-entry.
1. Introduction
In the United States, individuals with opioid use disorder (OUD) who are incarcerated face many health challenges when transitioning from correctional systems to the community. Being in a facility with a forced abstinence policy which decreases one’s drug tolerance level (Lim et al., 2012; Binswanger at al., 2007; Fox et al., 2015) and low rates of successful linkage to community-based medication for opioid use disorder (MOUD) treatment (i.e., methadone, buprenorphine, naltrexone) (Fox et al., 2015; Krawczyk et al., 2017; Krawczyk et al., 2021; Maradiaga et al., 2016) result in high excess risk of death due to overdose, particularly in the immediate time after release from incarceration. Concomitantly, higher rates of reincarceration, emergency department (ED) visits, and hospitalizations are observed among new community re-entrants (Frank et al., 2013; Westerberg et al., 2016).
In-jail or in-prison MOUD increases odds of continuing MOUD in the community, which in turn improves health and stability (Magura at al., 2009; Moore et al., 2018). Although it has been postulated that in-jail or in-prison MOUD reduces criminal activities after release, studies testing impact of in-prison MOUD on reincarceration have reported mixed results. In a recent New Mexico study, authors found that in-jail MOUD versus detoxification was associated with reduced reincarceration (Westerberg et al., 2016). A recent observational comparison in Massachusetts also showed associations with expanded correctional buprenorphine and lower reincarceration (Evans et al., 2022). On the other hand, a randomized control study found that MOUD during incarceration at Rhode Island jail and prison was not associated with reincarceration (McKenzie et al., 2012). A meta-analysis attributed these mixed results to differential follow-up times and differences in measuring reincarceration across studies (e.g., self-reports vs. administrative records) (Moore et al., 2019). Another explanation may be because commonly used measures of reincarceration (e.g., reincarceration rate, the first episode of incarceration after release) are poor proxies for unsuccessful community re-entry.
This study assesses criminal legal involvement and health care utilization aspects to define successful re-entry into the community following incarceration. We introduce a measure that incorporates both episodes and duration of reincarceration as well as post-release utilization of acute healthcare including ED visits, and hospitalizations using sequence analysis. We then test the hypothesis that jail-based MOUD is associated with increased stability following community re-entry among New York City (NYC) adults with OUD who were incarcerated and released to the community between 2011 and 2014.
2. Methods
2.1. Study population
For this analysis, we used data from a retrospective, observational cohort study where NYC Department of Health and Mental Hygiene, Correctional Health Services at NYC Health + Hospitals, and academic institutions (New York University and Weill Cornell Medical College) partnered to match multiple administrative data and evaluate the impact of jail-based MOUD on mortality and other health outcomes post jail release. Specifically, electronic medical records of individuals aged 18 years or older with OUD diagnosis who were admitted and released from NYC jails during 2011–2017 (n=29 566) were linked at the person level with NYC death certificate data and NY Statewide Planning and Research Cooperative System (SPARCS) hospitalization and ED records. Details about data matching processes and quality assessment have been reported elsewhere (Lim et al., 2023). In this study, we included 6 066 individuals who met two eligibility criteria: 1) no incarceration in NYC jails for at least 1 year prior to the index incarceration, which is the first incarceration event during 2011 to 2014, to address potential confounding of recent incarceration, and 2) at least 3-years of follow-up data without death events following the index incarceration, to strengthen validity of the stability measure (Fig S1 in Supplementary Appendix). Overall, there was no systematic difference between the study sample (n=6 066) and individuals who were excluded due to death during the follow-up (n=207; Table S1 in Supplementary Appendix).
2.2. Measures
Upon admission to NYC jails, Correctional Health Services, the NYC jails health provider, offer all individuals with active OUD short course methadone or buprenorphine for the treatment of acute opioid withdrawal or bridge medication until reported treatment program participation and dosing prior to incarceration can be confirmed. They are then offered voluntary enrollment in the MOUD program, which includes induction or continuation, and maintenance on methadone or buprenorphine, as well as linkage to community-based MOUD treatment upon release. Until late 2017, individuals with felony charges or parole violations were ineligible for MOUD, while those with misdemeanor-level charges were eligible. The details of MOUD eligibility has been described in a previous report (Lim et al., 2023). While incarcerated, individuals with OUD had their OUD treatment options (withdrawal-only, maintenance or out-of-treatment/no MOUD offered) modified based on their decision or, less often, involuntarily. Various treatment patterns (e.g., withdrawal followed by maintenance, maintenance followed by out-of-treatment) were categorized into two groups (in-jail MOUD and out-of-treatment) based on OUD treatments that they received during the last 3 days before discharge. Of the 6 066 study individuals, 2 565 received in-jail MOUD (42%) and 3 501 did not (58%) during the index incarceration.
The study outcome was pattern of community re-entry, incorporating episodes and duration of incarceration, ED visits (not culminating in an inpatient stay), and hospitalizations, as well as their dynamic interactions. This study created the outcome variable utilizing sequence analysis (described below) and using distinct patterns of community re-entry. Of the patterns identified, we defined stability as having a pattern of community re-entry with rare occurrence of incarceration, ED visits, and hospitalizations, and short duration of these events when they occurred.
In addition to the exposure and outcome, we included covariates related to demographic (age at release from the index incarceration, sex, race, Hispanic ethnicity, marital status, education), clinical (HIV diagnosis), behavioral (lifetime injection drug use, current smoking, alcohol use disorder, cocaine use disorder, serious mental illness, bipolar and related disorders, schizophrenia spectrum and other psychotic disorders, personality disorders), housing (self-reported homelessness or transient shelter stay at the time of incarceration), and criminal legal characteristics (incarceration length, highest charge severity at/near jail admission) to describe the sample and adjust for baseline differences between MOUD and out-of-treatment groups. Diagnosis of serious mental illness is based on criteria from the New York State (NYS) Office of Mental Health (until mid-2017) and Diagnostic and Statistical Manual of Mental Disorders (DSM) 5th edition (after mid-2017).
The Institutional Review Boards of DOHMH (18–106), two academic institutions (i18–00445 and 1811019740), and Biomedical Research Alliance of New York (19‐PRS‐156‐419(HHC)) reviewed and approved this study. The Office for Human Research Protections of the US Department of Health and Human Services also provided certification.
2.3. Statistical Analysis
We performed sequence analysis to identify patterns of community re-entry for three years after release from the index incarceration. Sequence analysis is a statistical method that generates nonoverlapping groups of individuals who share similar duration and sequencing of events (Abbott and Tsay, 2000; McAllister, Lennon and Kuang, 2011). We used three types of events for this analysis: jail incarceration, hospitalizations, and ED visit. The three events were considered mutually-exclusive on a given day; if no event of interest occurred, that day was considered to be a community stay.
We assessed the degree of dissimilarity among all possible pairs of sequences using an algorithm known as Levenshtein distance and summarized the results as a distance matrix (Abbott and Tsay, 2000). We then performed a hierarchical cluster analysis with the Ward method, resulting in five nonoverlapping clusters representing patterns of community re-entry for three years post-release (Fig 1) (Maechler et al., 2022). Individual‐level sequences of daily events (represented as horizontal lines) for three years after release were stacked together and divided into five distinct clusters based on their similarities. The Calinski and Harabasz Index, which is based on ratio of between-cluster variation and within-cluster variation, suggested four or five clusters as an optimal cluster solution (Milligan and Cooper, 1985), and we decided that an optimal number of clusters was five after reviewing descriptive statistics from four versus five clusters and a dendrogram (Fig S2 in Supplementary Appendix).
Fig 1.
Community re-entry patterns for 3 years after release from New York City jails among 6066 adults who were incarcerated in 2011–14 and had opioid use disorder
We performed propensity score matching using an optimal full matching algorithm in order to account for potential confounding. We evaluated covariate balancing between MOUD and out-of-treatment groups by comparing absolute standardized mean difference in each of aforementioned baseline covariates before and after propensity score matching, and tested if the difference was < 0.1, a threshold at which covariate balancing is achieved (Normand et al., 2001). We then ran marginal multinomial logistic regression with three-year community re-entry patterns as the outcome and jail-based MOUD as the exposure and computed predicted probabilities of following each community re-entry pattern, while accounting for propensity score matching. To test if jail-based MOUD was associated with stability patterns, we repeated propensity score matching and regression analysis using 500 bootstrap replicates and computed 95% confidence intervals around predicted probabilities. To assess whether estimated treatment impacts differed by index years, we performed propensity score matching and regression analysis per each index year (2011, 2012, 2013, 2014).
All statistical analysis was conducted using SAS version 9.4 (Cary, NC) and R studio version 3.5.2 (Vienna, Austria). Sequence analysis was performed using TraMineR (Gabadinho et al., 2011) and cluster packages (Maechler et al., 2022). Statistical significance was determined by two-sided p-values <0.05.
3. Results
The majority of the 6 066 study cohort members were male (81%) and non-Hispanic Black or Hispanic (70%) (Table 1). On average they were 43 years old (standard deviation (SD)= 11 years) when they were released from the index incarceration. The highest highest legal charges near admission were felony (52%), followed by misdemeanor (40%). Five percent met criteria for serious mental illness, and 12% experienced homelessness. While 43% reported lifetime injection drug use, 25% had cocaine use disorder and 24% had alcohol use disorder. On average, they made 0.26 ED visits (SD= 0.83) during the three months prior to the index incarceration. Table 1 shows that receipt of MOUD versus out-of-treatment during the index incarceration was associated with being charged with misdemeanor at admission (63% vs. 24%), and shorter jail stay (46 days vs. 84 days, on average), reflecting in-jail MOUD eligibility. Additionally, the receipt of MOUD was associated with being diagnosed with personality disorder (9% vs. 15%) and being female (28% vs. 13%) although females comprised a small proportion of the study cohort.
Table 1.
Characteristics of the study population
| Total (N = 6,066) | MOUD (N = 2,565) | Out-of-treatment (N = 3,501) | Standardized difference | |
|---|---|---|---|---|
| Demographic characteristics | ||||
| Male | 81% | 72% | 87% | 0.381 |
| Race and ethnicity | 0.068 | |||
| Hispanic | 47% | 46% | 48% | |
| Non-Hispanic Black | 23% | 25% | 22% | |
| Non-Hispanic White | 28% | 28% | 28% | |
| Othersa | 2% | 1% | 2% | |
| Education | 0.087 | |||
| Less than high school degree | 45% | 47% | 43% | |
| High school degree or equivalent | 34% | 33% | 34% | |
| Some college/trade school | 18% | 17% | 19% | |
| Unknown | 3% | 3% | 3% | |
| Marital status | 0.013 | |||
| Single, never married | 76% | 75% | 76% | |
| Married/partnered | 18% | 19% | 18% | |
| Others | 6% | 6% | 6% | |
| Average age at admission, years (SD) | 43 (11) | 44 (10) | 42 (11) | 0.207 |
| Behavioral characteristics | ||||
| Self-reported injection drug use | 43% | 49% | 39% | 0.208 |
| Current smoking | 80% | 83% | 79% | 0.114 |
| Clinical characteristics | ||||
| Alcohol misuse/dependence/use disorder | 24% | 24% | 24% | 0.009 |
| Cocaine use/dependence/use disorder | 25% | 28% | 24% | 0.091 |
| Serious mental illnessb | 5% | 5% | 5% | 0.017 |
| Personality disorders diagnosis | 13% | 9% | 15% | 0.186 |
| Bipolar and related disorders diagnosis | 8% | 8% | 7% | 0.047 |
| Schizophrenia and psychotic disorders diagnosis | 6% | 5% | 6% | 0.062 |
| HIV diagnosis | 9% | 10% | 8% | 0.102 |
| Housing characteristics | ||||
| Homelessnessc | 12% | 13% | 11% | 0.075 |
| Legal characteristics | ||||
| Highest charge severity at/near admission (≤ 15 days) | 0.875 | |||
| Felony | 52% | 34% | 65% | |
| Misdemeanor | 40% | 63% | 24% | |
| Other | 7% | 3% | 11% | |
| Average length of incarceration (SD), in days | 68 (85) | 46 (62) | 84 (96) | 0.464 |
| Healthcare utilization | ||||
| Average number of emergency room visits for 3 months prior to discharge date (SD) | 0.26 (0.83) | 0.34 (0.95) | 0.21 (0.72) | 0.153 |
Notes:
Others include individuals who reported no or unknown Hispanic ethnicity and Asian or Pacific Islander, American Indian/Alaskan, other-unspecified race, or unknown race.
Based on criteria from the New York State Office of Mental Health and Diagnostic and Statistical Manual of Mental Disorders V edition
Self-reported homelessness or transient shelter stay at the time of incarceration
According to Fig 1, for 3 862 individuals (64%), their time in the community was rarely disrupted by subsequent incarcerations, ED visits, or hospitalizations (stability pattern). On the other hand, 122 (2%) experienced almost consistent incarceration with long jail stays (continuous incarceration pattern). Four percent experienced reincarceration and then transitioned to stability (immediate return-to-jail pattern) whereas 7% followed the opposite pattern, experiencing stability initially and then being re-incarcerated (late return-to-jail pattern). Lastly, 1 424 individuals (23%) experienced higher numbers and longer duration of hospitalizations along with shorter duration of reincarcerations with no distinct pattern than the others (hospitalization pattern). Table 2 confirms that characteristics of these five patterns were well reflected in numbers of episodes and duration of three events. For example, individuals with the hospitalization pattern were hospitalized for an average of 10 days (SD = 29), which was the highest across the five patterns. Individuals with the continuous incarceration pattern were incarcerated for 596 days on average (SD = 145) for three years, which was substantially higher than those in the other patterns. Lastly, while the baseline demographic, behavioral, clinical, and housing characteristics were similar across community re-entry patterns, a few notable differences stood out (Table S2 in Supplementary Appendix). Percents of male and personality disorders diagnosis among individuals with the stability pattern were the lowest (78% and 11%, respectively) compared with the other patterns. Those with the continuous incarceration pattern had the highest percents of homelessness (16%) and misdemeanor charges (46%).
Table 2.
Average numbers of episodes and duration of post-release reincarceration, ED visits, and hospitalizations by community re-entry patterns among the study population (N=6,066)
| Hospitalization | Stability | Immediate return to jail | Late return to jail | Continuous incarceration | |
|---|---|---|---|---|---|
| Mean (SD) | Mean (SD) | Mean (SD) | Mean (SD) | Mean (SD) | |
| Number of reincarcerations | 3.8 (2.2) | 1.6 (1.1) | 3.8 (2.1) | 4.2 (2.2) | 4.9 (3.0) |
| Number of transfers to the federal/state custody | 0.4 (0.6) | 0.05 (0.2) | 0.6 (0.6) | 0.5 (0.7) | 1.0 (1.0) |
| Number of ED visits | 4.8 (10.2) | 2.0 (4.5) | 4.1 (14.6) | 4.9 (10.0) | 4.1 (6.6) |
| Number of hospitalizations | 1.7 (4.5) | 0.6 (1.6) | 0.8 (1.9) | 1.5 (4.5) | 1.1 (2.5) |
| Length of incarceration, in days | 122.6 (66.7) | 7.1 (14.6) | 347.4 (10.1.9) | 266.7 (101.6) | 595.5 (144.8) |
| Length of hospitalization, in days | 9.8 (28.8) | 2.7 (8.4) | 4.6 (16.0) | 7.0 (19.5) | 6.5 (16.4) |
Propensity score matching successfully reduced differences in baseline demographic, clinical, behavioral, housing, and legal characteristics between MOUD and out-of-treatment groups. The balance plots confirmed that standardized mean differences in all covariates were less than 0.1 or at around 0.1 (Fig S3 in Supplementary Appendix). After accounting for propensity score matching, 6 to 7 higher percent of individuals who received MOUD during incarceration, compared with those who were out-of-treatment, experienced stability after release (Table 3). On the other hand, differences in the likelihood of following the other patterns between MOUD and out-of-treatment groups were small and not statistically significant. Fig 2 shows that year-by-year percent of individuals who experienced stability post-release in both groups increased while the impact of MOUD persisted.
Table 3.
Predicted probabilities of community re-entry patterns by MOUD versus out-of-treatment from marginal multinomial regression analysis of propensity score matched data among the study population (N=6 066)
| 2011 | 2012 | 2013 | 2014 | |||||
|---|---|---|---|---|---|---|---|---|
| MOUD | Out-of-treatment | MOUD | Out-of-treatment | MOUD | Out-of-treatment | MOUD | Out-of-treatment | |
| Estimate (95% CI) | Estimate (95% CI) | Estimate (95% CI) | Estimate (95% CI) | Estimate (95% CI) | Estimate (95% CI) | Estimate (95% CI) | Estimate (95% CI) | |
| Hospitalization | 0.25 (0.22,0.28) | 0.28 (0.25,0.32) | 0.23 (0.21,0.25) | 0.26 (0.24,0.29) | 0.21 (0.19,0.24) | 0.24 (0.22,0.26) | 0.18 (0.16,0.21) | 0.21 (0.19,0.24) |
| Stability | 0.64 (0.60, 0.67) | 0.57 (0.53,0.60) | 0.66 (0.63,0.68) | 0.59 (0.57,0.61) | 0.68 (0.65,0.71) | 0.62 (0.59,0.64) | 0.71 (0.68,0.74) | 0.65 (0.62,0.68) |
| Immediate return to jail | 0.03 (0.02,0.04) | 0.04 (0.03,0.06) | 0.03 (0.02,0.04) | 0.04 (0.03,0.05) | 0.03 (0.02,0.04) | 0.05 (0.04,0.06) | 0.03 (0.02,0.03) | 0.04 (0.03,0.05) |
| Late return to jail | 0.07 (0.05,0.08) | 0.08 (0.06,0.10) | 0.06 (0.05,0.08) | 0.08 (0.07,0.09) | 0.06 (0.05,0.07) | 0.08 (0.06,0.09) | 0.06 (0.05,0.08) | 0.08 (0.06,0.09) |
| Continuous incarceration | 0.02 (0.01,0.03) | 0.02 (0.01,0.04) | 0.02 (0.01,0.03) | 0.02 (0.01,0.03) | 0.02 (0.01,0.03) | 0.02 (0.01,0.03) | 0.02 (0.01,0.02) | 0.02 (0.01,0.03) |
Notes: Propensity score matching balanced differences between MOUD and out-of-treatment groups in the following variables (age, sex, race and ethnicity, marital status, the highest level of education, serious mental illness, personality disorder diagnosis, schizophrenia diagnosis, bipolar disorder diagnosis, alcohol use disorder, cocaine use disorder, HIV, injection drug use, current smoking, homelessness, total numbers of ED visits during 3 months prior to discharge date, length of incarceration, and highest charge severity). Estimates were obtained from marginal multinomial regression analysis of propensity score matched data and 95% CI were estimated via 500 bootstrap replicates.
Fig 2.
Predicted probabilities of community re-entry patterns and 95% confidence intervals by MOUD versus out-of-treatment
Notes: Predicted probabilities were estimated from multinomial regression analysis with propensity score matched data and 95% confidence intervals were estimated using the bootstrap method.
4. Discussion
In this study of adults with histories of incarceration and OUD, we found that transition to the community after jail release and challenges due to reincarceration, ED visits, and hospitalizations were characterized by five unique trajectories. We identified a pattern where time in community was rarely disrupted by reincarceration, ED visits, and hospitalizations, representing a relatively smoother community transition. Individuals who received in-jail MOUD were more likely to experience this stability pattern. For example, for those who were released at their index incarceration in 2011, the percentage having the stability pattern was 64% (in-jail MOUD) versus 57%, and this 7% difference was consistently observed in the next three index incarceration years (2012–14).
This finding provides important evidence that in-jail MOUD leads to more stability post-release, which is a marker of a successful transition of individuals with OUD from jail to community. One possible mechanism to explain this positive impact is increased community MOUD enrollment and maintenance among individuals who received MOUD during incarceration and linkage-to-care contacts. Several studies found that in-prison MOUD was significantly associated with continuing MOUD post-release (Moore et al., 2019; Jalali et al., 2022), which could reduce risk of accidental overdose and acquisitive crime associated with illicit substance use (Samples et al., 2020).
About 23% of the cohort members followed the hospitalization pattern characterized by on average two hospitalization events with an average of 10 cumulative days in hospital for 3 years, a much higher rate than the other patterns. However, total hospital stays were shorter than those from the previous Rhode Island study (12 hospital days per year) (Howell et al., 2021). The main reason for this discrepancy was that the post-release event analysis was restricted to individuals who were not re-incarcerated after release to the community in the Rhode Island study. With frequently occurring reincarceration in an urban setting, the chance to be hospitalized might decrease in the current study cohort. Similar to the Rhode Island study (Howell et al., 2021), we did not find an association between receipt of in-jail MOUD and subsequent hospitalization, suggesting that there may be a particular group of people who suffer from underlying chronic physical and behavioral health issues, which could lead to hospitalization regardless of receipt of in-jail MOUD.
We observed year-by-year increase of individuals following the stability pattern and decrease of those following the hospitalization pattern. This positive result may be attributed to improvement of re-entry services in jail (referrals to community care) or other changes in the environment of re-entry from jail, including court-mandated participation in treatment programs and overall decreased contacts with the criminal justice system.
In addition to the hospitalization pattern, the sequence analysis discovered two distinct trajectories: immediate return to jail and later return to jail. The former points to individuals who experienced early struggles during the community re-entry, but ultimately transitioned into stability. The latter includes individuals who experienced some stability followed by increased reincarceration. Although represented by a small percent of individuals and not associated with receipt of in-jail MOUD, these are novel trends and require more study to understand their impacts on community re-entry experiences.
Although the majority were rarely engaged with the NYC criminal legal system consistently throughout the three years, 2% were entering a cycle of reincarceration despite no record of incarceration for 1 year prior to the index incarceration. Combining all incarceration events after the index incarceration, 54% of their three-year follow-up time on average was spent in jail. A similar finding was observed in a recent study of the NYC jail data (Chan et al., 2020). Chan and colleagues found that 5% of 3 114 young adults with limited experiences of incarceration had a rapid succession of subsequent incarcerations. These individuals were more likely to be homeless, have mental illness, use substances, and be charged with misdemeanors (Chan et al., 2020). Although their cohort was much younger and not restricted to individuals with OUD, both studies suggest that there is a group of individuals who are likely to experience continuous incarcerations even with in-jail MOUD.
There are several limitations in this study. Although percent of federal or state custody transfer was small, time spent in prison may be misclassified as time in the community. Only three types of events were recorded, and other major events that could disrupt transition to community, such as homelessness, or events outside of NYS were not captured. Although propensity score matching accounted for a large number of covariates to reduce observed associations between in-jail MOUD and out-of-treatment groups (e.g., higher % of females and lower % of individuals with personality disorder among those with in-jail MOUD), unobserved differences between two groups (e.g., arrests, prior and current history of conviction, jail diversion, probation, chronic health conditions) could have introduced bias due to confounding. We did not have data about dosing and duration of in-jail MOUD, preventing us from testing a dose-response relationship between in-jail MOUD and stability post-release. Lastly, generalizability of the study findings might be limited because data came from a large urban setting. Despite these limitations, large matched administrative data allowed for capturing more objectively three distinct events that were strongly linked to life disruption over three years, which could have been very challenging if these data were collected via conventional survey instruments. Another strength was use of sequence analysis, a data-driven method to help identify patterns of community re-entry from duration and episodes of reincarceration, ED visits, and hospitalizations and their dynamic relationships.
5. Conclusion
Receipt of MOUD during incarceration was associated with more stable transition to the community among NYC adults with OUD. Those who received MOUD during incarceration, compared with those who were out of treatment, were more likely to have their post-release time in the community rarely disrupted by reincarceration, ED visits, and hospitalizations. This finding provides important evidence that providing MOUD during incarceration has potential to break a vicious cycle of reincarceration and addiction and supports expanding access to MOUD in criminal legal settings.
Supplementary Material
Highlights.
Sequence analysis identified five community re-entry patterns.
64% of the cohort members followed the stability pattern.
Receipt of in-jail MOUD contributed to a positive outcome of community re-entry.
In-jail MOUD has potential to break a vicious cycle of reincarceration and addiction.
Footnotes
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
References
- Lim S, Seligson AL, Parvez FM, Luther CW, Mavinkurve MP, Binswanger IA, Kerker BD, 2012. Risks of drug-related death, suicide, and homicide during the immediate post-release period among people released from New York City jails, 2001–2005. Am J Epidemiol 175(6), 519–526. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Binswanger IA, Stern MF, Deyo RA, Heagerty PJ, Cheadle A, Elmore JG, Koepsell TD, 2007. Release from prison—a high risk of death for former inmates. NEJM 356(2), 157–165. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Fox AD, Maradiaga J, Weiss L, Sanchez J, Starrels JL, Cunningham CO, 2015. Release from incarceration, relapse to opioid use and the potential for buprenorphine maintenance treatment: a qualitative study of the perceptions of former inmates with opioid use disorder. Addict Sci Clin Pract 10, 1–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Krawczyk N, Picher CE, Feder KA, Saloner B, 2017. Only one in twenty justice-referred adults in specialty treatment for opioid use receive methadone or buprenorphine. Health Aff 36(12), 2046–2053. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Krawczyk N, Mojtabai R, Stuart EA, Fingerhood MI, Agus D, Lyons BC, Weiner JP, Saloner B, 2021. Opioid agonist treatment is highly protective against overdose death among a US statewide population of justice-involved adults. Am J Drug Alcohol Abuse 47(1), 117–126. [Google Scholar]
- Maradiaga JA, Nahvi S, Cunningham CO, Sanchez J, Fox AD, 2016. “I kicked the hard way. I got incarcerated.” Withdrawal from methadone during incarceration and subsequent aversion to medication assisted treatments. J Subst Abuse Treat 62, 49–54. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Frank JW, Andrews CM, Green TC, Samuels AM, Trinh TT and Friedmann PD, 2013. Emergency department utilization among recently released prisoners: a retrospective cohort study. BMC Emerg Med 13, 1–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Westerberg VS, McCrady BS, Owens M, Guerin P, 2016. Community-based methadone maintenance in a large detention center is associated with decreases in inmate recidivism. J Subst Abuse Treat 70, 1–6. [DOI] [PubMed] [Google Scholar]
- Magura S, Lee JD, Hershberger J, Joseph H, Marsch L, Shropshire C Rosenblum A, 2009. Buprenorphine and methadone maintenance in jail and post-release: a randomized clinical trial. Drug Alcohol Depend 99(1–3), 222–230. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Moore KE, Oberleitner L, Smith KM, Maurer K McKee SA, 2018. Feasibility and effectiveness of continuing methadone maintenance treatment during incarceration compared to forced withdrawal. J Addict Med 12(2), 156. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Evans EA, Wilson D Friedmann PD, 2022. Recidivism and mortality after in-jail buprenorphine treatment for opioid use disorder. Drug Alcohol Depend 231, 109254. [DOI] [PMC free article] [PubMed] [Google Scholar]
- McKenzie M, Zaller N, Dickman SL, Green TC, Parihk A, Friedmann PD Rich JD, 2012. A randomized trial of methadone initiation prior to release from incarceration. Subst Abus 33(1), 19–29. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Moore KE, Roberts W, Reid HH, Smith KM, Oberleitner LM, McKee SA, 2019. Effectiveness of medication assisted treatment for opioid use in prison and jail settings: A meta-analysis and systematic review. J Subst Abuse Treat 99, 32–43. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Lim S, Cherian T, Katyal M, Goldfeld KS, McDonald R, Wiewel E, Khan M, Krawczyk N, Braunstein S, Murphy SM, Jalali A, Jeng PJ, MacDonald R, Lee JD 2023. Association between jail‐based methadone or buprenorphine treatment for opioid use disorder and overdose mortality after release from New York City jails 2011–17. Addiction 118(3), 459–467. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Abbott A and Tsay A, 2000. Sequence analysis and optimal matching methods in sociology: Review and prospect. Sociol Methods Res 29(1), 3–33. [Google Scholar]
- McAllister W, Lennon MC, Kuang L, 2011. Rethinking research on forming typologies of homelessness. Am J Public Health 101(4), 596–601. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Maechler M, Rousseeuw P, Struyf A, Hubert M, Hornik K, 2022. Cluster: cluster analysis basics and extensions. R package, Version 2.1. 3. R Foundation for Statistical Computing, Vienna, Austria. [Google Scholar]
- Milligan GW, Cooper MC, 1985. An examination of procedures for determining the number of clusters in a data set. Psychometrika 50, 159–179. [Google Scholar]
- Normand SLT, Landrum MB, Guadagnoli E, Ayanian JZ, Ryan TJ, Cleary PD, McNeil BJ, 2001. Validating recommendations for coronary angiography following acute myocardial infarction in the elderly: a matched analysis using propensity scores. J Clin Epidemiol 54(4), 387–398. [DOI] [PubMed] [Google Scholar]
- Gabadinho A, Ritschard G, Müller NS, Studer M, 2011. Analyzing and visualizing state sequences in R with TraMineR. J Stat Softw, 40, 1–37. [Google Scholar]
- Jalali A, Jeng PJ, Polsky D, Poole S, Ku YC, Woody GE, Murphy SM, 2022. Cost-effectiveness of extended-release injectable naltrexone among incarcerated persons with opioid use disorder before release from prison versus after release. J Subst Abuse Treat 141, 108835. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Samples H, Williams AR, Crystal S, Olfson M, 2020. Impact Of Long-Term Buprenorphine Treatment On Adverse Health Care Outcomes In Medicaid: The impact of longer treatment on health care outcomes for opioid use disorder within a key population of Medicaid enrollees. Health Aff 39(5), 747–755. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Howell BA, Martin RA, Lebeau R, Truong AQ, Wang EA, Rich JD, Clarke JG, 2021. Changes In Health Services Use After Receipt Of Medications For Opioid Use Disorder In A Statewide Correctional System: Study examines changes in postrelease health care use among formerly incarcerated people who received treatment for opioid use disorder while in the Rhode Island Department of Corrections system. Health Aff 40(8), 1304–1311. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Chan PY, Kaba F, Lim S, Katyal M, MacDonald R, 2020. Identifying demographic and health profiles of young adults with frequent jail incarceration in New York City during 2011– 2017. Ann Epidemiol 46, 41–48. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.