key: cord-0761646-811o8wal
authors: Van Nostrand, Elizabeth; Johnston, Alyssa; Albert, Steven M; Brown, Andre L; Buchanich, Jeanine M; Ray, Brad; Roberts, Mark S
title: Evaluating the effectiveness of problem-solving courts at preventing fatal and non-fatal opioid events: preliminary findings from Indiana
date: 2021-12-15
journal: Drugs Context
DOI: 10.7573/dic.2021-8-1
sha: 73c42f97c18631151fe906e3048a52f0be191272
doc_id: 761646
cord_uid: 811o8wal

BACKGROUND: Fatal and non-fatal events associated with drug misuse are skyrocketing in most United States jurisdictions, including Indiana. Historically, the role of the judiciary is to arrest, impose sanctions and protect society from harm. Adults arrested for drug abuse in Indiana can be sentenced to 1 of 17 correctional facilities. As an alternative, they may be eligible to participate in a problem-solving court (PSC) programme that refers individuals to treatment as a pretrial diversionary strategy. The aim of the study is to determine which interventions offered by PSCs and correctional facilities impact morbidity and mortality. The study began in 2019 and will end in 2023; therefore, the results in this manuscript are preliminary. METHODS: The study cohort included two populations arrested for drug misuse: (1) adults sentenced to Indianan correctional facilities (1 January 2018 to 30 June 2021) and (2) adults participating in an Indiana PSC programme (1 January 2018 to 30 June 2021). The study used a mixed-methods design that integrated qualitative interviews of deputy wardens, PSC team members and service providers with the following quantitative datasets: sentencing information, emergency department visits, inpatient hospitalization admissions, prescription drug monitoring programme data and death records. The individuals will be followed at 2-week, 4-week, 6-month and 1-year intervals post-release. Difference-in-difference and time-to-event analyses will identify impactful interventions. A model will be created to show the effect of impactful interventions in Indiana counties that do not have PSCs. RESULTS: Findings are preliminary. There is variability amongst correctional facilities regarding programme eligibility, provided services and provision of medication-assisted treatment. All correctional facilities were severely impacted by the COVID-19 pandemic. CONCLUSION: It is anticipated that the adoption of impactful interventions will lower opioid-related morbidity and mortality rates.

Fatal events associated with the misuse of drugs are skyrocketing in most United States jurisdictions. Provisional numbers from the Centers for Disease Control and Prevention (CDC) suggest that overdose deaths increased by 30% from 2019 to 2021 in the United States. 1 Opioids, primarily synthetic opioids (other than methadone), were involved in 70.6% of these fatalities. 2 In fact, overdose deaths, along with fatalities from coronavirus, homicides and some chronic diseases, contributed to 2020's single-year life expectancy decline -the largest regression since 1943. 3 Non-fatal overdose events are also on the rise. For example, the annual number of emergency department (ED) visits in the United States involving opioids is increasing across most age groups (Table 1 ). Indiana is one of the many jurisdictions experiencing an opioid crisis. The CDC estimates overdose ISSN: 1740-4398 ORIGINAL RESEARCH -Evaluating the effectiveness of problem-solving courts drugsincontext.com fatalities in Indiana surged from 1704 in 2019 to 2268 in 2020 -a 33.1% increase. 2 With respect to non-fatal events, the number of individuals admitted to Indianan EDs specifically for opioid overdoses escalated from 1460 admissions in 2009 to 3199 in 2016 -a 119% increase. 4 Historically, the role of the judiciary in drug-related crimes is to arrest, impose sanctions and protect society from future harm. Research suggests, however, that there is no significant relationship between state imprisonment rates and selfreported drug use, arrest rates for drug misuse or mortality rates. 5 State prison populations have grown more than 700% since the 1970s, 6 and state correction budgets have nearly quadrupled. 7 The rate of incarceration in Indiana (723 per 100,000) 8 exceeds the national rate (698 per 100,000) 9 by 3.58%.

Individuals of colour are overrepresented in Indiana correctional facilities. Although only 9% of Indianans identify as Black, 10 Black adults comprise 39% of its incarcerated population. 11 Although the rate of incarcerated individuals is higher for all persons of colour (Table 2) , Black adults are about 5.2 times more likely to be incarcerated in Indiana than Whites.

In 2019, approximately 15% of all United States arrests involved at least one drug abuse violation. 12 As of January 2021, 29 .25% of all adults incarcerated in Indiana were convicted of one or more drug offenses. 13 Adults arrested for drug-related crimes in Indiana can be sentenced to 1 of 17 correctional facilities. Characteristics of these facilities are found in Table 3 .

Prisoners are the only population in the United States constitutionally entitled to healthcare. Under the Eighth Amendment, 14 the denial of medical services to the incarcerated is tantamount to "cruel and unusual punishment". A prisoner's accessibility to healthcare, however, is not limitless. "[S]ociety does not expect that prisoners will have unqualified access to health care." 15 Rather, incarcerated individuals have the right to "adequate medical care" for "serious" medical conditions. 16 The courts have defined "adequate medical care" as services that are "at a level reasonably commensurate with modern medical science and of a quality acceptable within prudent professional standards". 17 The onus is placed upon correctional facilities to determine what level and types of services fulfil a prisoner's constitutional right to medical care. Indiana prisoners post-release are 129 times more likely to die of a drug overdose than the general population, especially if their addiction was not adequately treated whilst incarcerated. 18 As an alternative to incarceration, individuals arrested for drug misuse may be eligible to participate in a problemsolving court (PSC) programme. PSC programmes refer arrested individuals to treatment as a pretrial diversionary strategy and address the underlying issues associated with criminal offenses. Participation in PSC programmes is not automatic; rather, individuals have to apply to participate and can be denied enrolment. PSCs use a phased programme model that is typically longer in duration than traditional prison sentences (correctional facility interviews, unpublished data, October 2020 to July 2021). For example, the advisory sentence for an individual arrested in Indiana for possession of less than 5 g of heroin is 1 year. 19 It would take that same individual 12-36 months to complete a PSC programme. 20 Indiana first began using PSCs in 1996. 21 Oversight for Indiana PSCs falls to the Indiana Office of Court Services (IOCS). 22 There are currently 117 Indiana PSCs serving various populations ( Table 4 ; the types of PSCs participating in this study are bolded).

Although there are recognized 'best practices' for each type of PSC, the policies and treatments offered by individual programmes vary tremendously. For example, each PSC has its own eligibility and exclusionary criteria. The roles and responsibilities of the judges differ, including whether professional training is required. The length of the programmes and availability of medical therapies, cognitive therapies and social services fluctuate. Sanctions for participant non-compliance and incentives for programme completion are not standardized. 23 Although previous studies have examined the effect of PSCs on recidivism, 24 health and wellness of individuals and their communities by impacting subsequent morbidity and mortality rates?

The aims of the study are to (1) provide an overview of correctional facility services and PSC programmes and individual characteristics of persons arrested for drug misuse;

(2) identify the strategies, or combination of strategies, that keep individuals alive and healthy after they are released from a correctional facility or complete a PSC programme; and (3) model the impactful strategies to show their effect in Indiana communities not currently utilizing them. The study began in 2019 and will end in 2023; therefore, the results presented in this manuscript are preliminary.

Legal epidemiology is a discipline that includes the study and application of law as an intervention to prevent disease and injury. 25 As with other disciplines rooted in science, ISSN: 1740-4398 ORIGINAL RESEARCH -Evaluating the effectiveness of problem-solving courts drugsincontext.com 

Adult drug court 42 Provides specialized long-term drug treatment programmes for individuals suffering from substance use disorder 44 7 Domestic violence court 43 Provides social services and interventions to domestic violence offenders and support to their victims 1 2

Family recovery court 44 Provides support to children and long-term treatment programmes for individuals suffering from substance use disorder 19 2 Juvenile drug court 45 Provides support services and intensive, continuous judicial supervision for youth 1 1

Juvenile problem-solving court 46 Provides immediate and highly structured judicial intervention for youth 2 0

Juvenile mental health court 47 Provides services and support that allow youth to remain in or re-enter their communities 0 1

Mental health court 48 Provides community-based treatment programmes for individuals with serious mental illnesses 8 4 Operating vehicle whilst intoxicated court 49

Provides community-based programmes for individuals charged with drunk driving as a Level 6 felony 1 1

Problem-solving court 50 Provides immediate and highly structured judicial intervention for eligible individuals 1 0

Re-entry court 51 Provides support to individuals transitioning into the community by coordinating mental health treatment and other services 11 0

Truancy court 52 Provides a partnership between schools, law enforcement, social service providers, mental health providers and clinicians to stabilize families and reengage youth in education legal epidemiological methods are scientific, systematic, transparent, replicable, precise and measurable. 26 The application of legal coding methodology allows law and policy to become data that can be measured and analysed to show impact and effectiveness on health outcomes.

This study utilized a mixed-methods design. The investigators created quantitative data from correctional facility and PSC records and written policies describing provided services. Qualitative interviews were also coded. An 11-person Stakeholder Committee, comprised of a student with lived experience as well as nationally recognized experts in the judiciary, corrections, public health, medicine, modelling and legal epidemiology, was convened to advise the researchers on the project's overall aims, design and methods, contribute to the analysis and interpretation of research results, and assist with dissemination strategies. ISSN The IOCS also suggested that the investigators limit eligibility to individuals who participated in PSCs that used an electronic case management system, the Supervised Release System (SRS), to track programme involvement. Amongst other data, SRS records include chronological case summaries and hearing information. PSCs that collect data on paper records lack uniformity and granularity and are difficult to access. 

Qualitative interviews were critical to understand more fully and accurately the types of interventions offered to individuals sentenced for drug misuse. The interviews captured inclusionary and exclusionary criteria, substance abuse treatment programming, and barriers to treatment. Four types of individuals were identified for the qualitative interviews: deputy wardens in Indiana's 17 adult correctional facilities; Addiction Recovery Services (ARS) representatives who provided services in Indiana's correctional facilities; members of PSC teams (i.e. judges, court administrators, coordinators, prosecutors and defence attorneys); and PSC treatment providers. The University of Pittsburgh Institutional Review Board (IRB) issued an 'Exempt' designation for the qualitative interviews. As future interviews may also take place at Temple University, the qualitative interviews were also submitted to Temple's IRB for review and received an 'Exempt' designation.

Each qualitative interview was structured in two parts. First, a brief written Qualtrics survey collected background information about the person being interviewed. Next, a semi-structured virtual in-person qualitative interview was conducted and audio recorded. As of 13 October 2021, 11 deputy wardens and 6 ARS providers have been interviewed across 82% of Indiana's adult correctional facilities. Written protocols that also inform eligibility criteria and provided services were obtained from five correctional facilities. The audio recordings were uploaded to a transcription service website. To ensure confidentiality, the audio was segmented so that multiple individuals transcribed each interview. The transcription service's platform was built on an Amazon Web Services cloud solution that provides for additional security.

Codebooks were developed to create data for the qualitative interviews and written protocols. The conceptual evaluation framework developed by the US Department of Justice in its Multi-Site Adult Drug Court Evaluation study 29 served as a model for the codebooks. Multiple collaborative sessions were conducted amongst the investigators to discuss common definitions, rules, assumptions and questions. Weekly meetings were held to discuss issues and inconsistencies in coding approaches.

Inter-rater reliability (IRR) was calculated to ensure the validity and reliability of the coding process. An IRR methodology similar to that developed by McAlister et al. 30 was used. For the IRR exercise, coded responses for three interview transcripts and three written protocols were exported from Qualtrics as Excel documents. Responses to each question were examined. Agreements and disagreements between the triplet of coders were totalled and IRR scores were calculated by dividing the total number of agreements (i.e. consistent ding amongst all the three coders) by the total number of responses given by each individual coder. Once each coder had their own IRR score for each document, the number of consistent answers was averaged. The three transcripts of recorded deputy warden interviews yielded IRR scores of 71%, 76% and 77%. The three written protocols obtained from correctional facilities yielded IRR scores of 84%, 89% and 95%.

The COVID-19 public health emergency significantly impacted operations in Indiana's courts. Because of unforeseen issues associated with the pandemic, the IOCS requested that the investigators suspend their collaboration with PSCs in March 2020. Collaboration resumed in December 2020. Qualitative interviews of PSC team members and service providers have begun. It is anticipated that these interviews will be concluded by 31 March 2022. As of 13 October 2021, the initial interviews have not been coded or analysed.

The MPH is governed by a variety of statutory and regulatory provisions, policies and agreements regarding privacy and sharing of data received from its partners. 31 Each agency identified in Table 5 i has agreed to have its datasets linked.

Data dictionaries have been shared and meaningful variables identified. As of 13 October 2021, the agencies, MPH and the researchers are in the process of finalizing a charter agreement. After the charter agreement is perfected, the MPH will assign a unique identifier known as a 'global ID' to each individual in the study cohort. University of Pittsburgh i As of October 13, 2021, permission to link the SRS data is pending. researchers will access data through MPH's Enhanced Research Environment, which is a scalable environment that uses a Windows-based desktop application containing Linux-based open-source tools supported on multi-cloud environments (e.g. Jupyter Notebooks and R Studio with capability for 'bringyour-own-license' for SAS, STATA, IBM, Power BI, Tableau, and Microsoft Office products). Researchers and analysts will be able to 'meet' collaboratively in the virtual environment to share and analyse data. A code repository will allow analysts to iterate on and keep track of their code. The Enhanced Research Environment is also a hardened environment in which data can flow in but cannot flow out without approval from MPH. The MPH data review team will ensure that privacy and suppression laws, contractual requirements, and security protocols are met.

After the MPH links the datasets identified in Table 5 , the investigators will use the coded qualitative interviews and written protocols to determine the specific therapies, treatments, and programmes offered to each member of the study cohort. As individuals are at higher risk for overdose death following their release from incarceration, 32 the study cohort will be followed for morbidity (measured by ED visits, emergency medical services (EMS) calls and in-patient hospitalizations) and mortality outcomes (death) at 2 weeks, 4 weeks, 6 months and 1-year post-programme intervals. The data analysis portion of the study received a 'Not Human Subjects' research designation from Temple University IRB, University of Pittsburgh IRB, and Wayne State University IRB.

Individual-level covariates will include race/ethnicity, gender, age (categorical) and time at risk (days). All outcome variables will be measured dichotomously (yes, no) and include any nonfatal overdose or death. The outcome of primary interest will be death, which will come from death certificates submitted by county coroners to the Indiana State Department of Health Division of Vital Records. Three mortality outcomes of particular public health importance will be evaluated: allcause (non-drug related); drug-related causes (ICD10 X40-44, X60-X64, X85, Y10-Y15) and opioid-related events (any mention of T40.0-T40.4; T40.6).

Non-fatal overdose EMS events will be operationalized as EMS contacts where naloxone was administered, the patient was resuscitated and the patient survived at least 1 day following resuscitation. These data are submitted by provider agencies to the National Emergency Medical Services Information System (NEMSIS), which is managed by the Indiana Department of Homeland Security. Data on ED visits are compiled in the 'Uniform Billing 2004 form' and include inpatient data that will be operationalized using primary or secondary diagnosis codes for overdose (ICD-10-CM T40.0-T40.4; T40.6).

Descriptive analyses on all study variables will be conducted. Outcomes amongst the cohort will then be evaluated in ISSN: 1740-4398 ORIGINAL RESEARCH -Evaluating the effectiveness of problem-solving courts drugsincontext.com county. 'What-if' scenarios, such as forecasting the impact of increased access to medical assisted treatment or the provision of naloxone concurrent with release, will be tested.

All findings and conclusions are preliminary; however, some themes have surfaced from the interviews of assistant wardens and ARS providers.

Although the Department of Correction determines the policies for the individual correctional facilities, the ARS Director determines which inmates can participate in programmatic activities. It is easier for inmates to be accepted into programmes if they are ordered to by the court. GAIN scores (i.e. a screening tool used to discern problems and symptoms, such as substance use, behavioural health issues, and violent tendencies) are frequently used to determine programme eligibility but acceptable scores are not uniform. Gang activity and a history of violent behaviour are often exclusionary criteria. One facility excludes any individual who is in restricted housing (i.e. individuals who have markedly reduced out-of-cell time, reduced privileges and limited activities) from participating in any programmes.

The impact of the pandemic COVID-19 has significantly impacted the types and frequency of programmes. In-person interactions have been extremely limited, and it has been difficult to keep inmates engaged. Several facilities have provided programmes in the virtual space; however, this necessitates providing inmates with tablets or other interactive devices. In all but one facility, the cost of the device was borne by the inmate or their family. If they cannot afford the purchase, they are not afforded the services.

It is difficult to maintain a sober prison environment. Illicit drugs are pervasive and are brought into by inmates' families and friends as well as correctional facility staff.

Reducing the length of the sentence is used to encourage inmate participation in offered programmes and services.

Commonly offered services Some of the most frequently provided services include job training, life skills training (i.e. tools to assist with independent living as well as inmate jobs and mentor programmes), support groups including Alcoholics Anonymous, Narcotics Anonymous and the Self-Management And Recovery Training two main ways: difference-in-difference (DID) analysis and time-to-event analysis. DID is a quasiexperimental design that uses longitudinal data from treatment (PSC) and control (correctional facility) groups to obtain an appropriate counterfactual to estimate a causal effect. DID will be used to estimate the impact of PSCs by comparing the changes in outcomes over time between PSC and correctional facility populations by including an interaction term between the treatment and dummy groups.

Time-to-event analysis for morbidity and mortality outcomes will also be examined. Cox proportional hazards regression 33 will be used. Whether survival is associated with demographic factors (gender, race/ethnicity and age) and programme and sentencing characteristics will be examined. These will be modelled as separate variables because they are applications of distinct policies and practices. Because it is anticipated that study members may be at risk for multiple adverse events, order and timing of events will be considered. This type of conditional approach assumes that a subject is not at risk for a subsequent event until a prior event occurs, and hence takes the order of events into account. The Prentice-Williams-Peterson model 34 on the calendar time scale will be used to define time intervals for all recurrent events (i.e. EMS presentations, ED visits and in-patient hospital admissions) and/or terminal event (death).

Under a separate CDC contract, the Public Health Dynamics Laboratory at the University of Pittsburgh is developing a model representing the transition amongst various stages of opioid misuse. Typically, this type of transition model is used to simulate the 'natural history' of disease progressions and is a cornerstone for informing various legal and policy interventions. 35 The use of modelling has been limited, however, in the opioid epidemic due to the variability of laws and policies. In this study, the investigators will (1) reconstruct the natural history of the opioid epidemic in Indiana in jurisdictions without PSCs and (2) simulate the potential effect of impactful treatments where PSCs have not been utilized.

To avoid overfitting of the opioid transition model, it will be validated using a combination of cross-validation and back-casting simulations. In cross-validation, a random set of counties will be excluded from the estimation phase. The remaining counties will be used to predict the impact of interventions in these excluded counties. The best models will then be chosen based upon their performance in the excluded counties. Back-casting simulation exercises will also be used to predict the performance of the model in each county whilst excluding the last few years from the calibration.

After the model is calibrated, a simulation will project the expected impact of various impactful interventions in each ISSN: 1740-4398 ORIGINAL RESEARCH -Evaluating the effectiveness of problem-solving courts drugsincontext.com

Whether individuals should be incarcerated for drug misuse is a debatable issue; however, it is clear that, until we can figure out which interventions actually work, tens of thousands of lives will be lost, family units will be inextricably broken and communities will be decimated by the opioid crisis. The COVID-19 pandemic exacerbated drug misuse and created barriers to treatment, especially for those who are incarcerated. 37 Prison costs are escalating, leaving fewer available resources to provide constitutionally protected treatments, therapies and services.

We hope that the proximal outcomes of this project will increase awareness for public health practitioners, health officials, law enforcement and members of the judiciary across the United States regarding which treatments, services and therapies positively impact health outcomes. We hope that the distal outcomes of the project will lead to the adoption of identified impactful recommendations in both correctional facilities and PSCs and that, like precision medicine, services will be targeted in populations where they are likely to have the greatest positive results. Finally, by adopting evidence-based practices, we hope that the long-term outcomes of this project will result in lowered opioid-related morbidity and mortality rates for individuals sentenced for drug-related charges.

(SMART) programme (the SMART Recovery programme is a science-based addiction recovery programme; unlike Alcoholics Anonymous or Narcotics Anonymous, there is no spiritual component in this therapy), and Thinking for Change (an integrated cognitive behavioural change programme comprising a series of lessons that build upon each other).

Naltrexone an opioid antagonist is the most frequently offered medication-assisted treatment. Vivitrol is the brand name for an injectable, extended-release form of Naltrexone. ReVia and Depade are brand names for a once-a-day pill form of Naltrexone. A few facilities have started to provide inmates with Naloxone (a medication to counter breathing difficulties during opioid overdose) upon release. Naloxone is sold under the brand name Narcan.

To encourage responsibility and foster kindness, one facility gives inmates animals to care for. One prison uses a geo-tracker to ensure programme participation and meeting attendance. Two facilities offer dialectal behavioural therapy that focuses on four core skills (mindfulness, emotional regulation, distress tolerance and interpersonal effectiveness) to help with problem-solving and challenging issues. 36

Contributions: AJ contributed to the research and data analysis portions of the paper and results. She also edited the final manuscript. SMA contributed to the writing of the qualitative methods portion of the paper. AB contributed to the writing of the qualitative methods portion of the paper. JB contributed to the writing of the data linkage and data analyses portions of the paper. BR contributed to the writing of the data linkage and analyses portions of the paper. He also edited portions of the final manuscript. MSR contributed to the writing of the modelling portion of the paper. All named authors meet the International Committee of Medical Journal Editors (ICMJE) criteria for authorship for this article, take responsibility for the integrity of the work as a whole, and have given their approval for this version to be published. 

Provisional life expectancy estimates for 2020. Vital Statistics Rapid Release

Vital signs: trends in emergency department visits for suspected opioid overdoses -United States

Meta-analysis of drug-related deaths soon after release from prison

The Price of Prisons: What Prisons Cost the Taxpayer

Prison Policy Initiative website

What Percent of the U.S. is Incarcerated? (and other ways to measure mass incarceration). Prison Policy Initiative website

United States Census Bureau website

Overrepresentation of Blacks in Indiana. Prison Policy Initiative website

Fact Cards. Indiana Department of Correction website

The Crisis in Indiana: Addiction Affects Every Aspect of Hoosier Life

Code § 35-50-2-7 (b)

Indiana drug courts: a summary of evaluation findings in five adult programs

What is "treatment" for opioid addiction in problem-solving courts: a study of 20 Indiana drug and veterans courts

Do race and ethnicity matter? An examination of racial/ethnic differences in perceptions of procedural justice and recidivism among problem-solving court clients. Race Justice

A transdisciplinary approach to public health law: the emerging practice of legal epidemiology

The growing field of legal epidemiology

Opioid overdose mortality among former North Carolina inmates: 2000-2015

Management Performance Hub website

The multi-site adult drug court evaluation: the drug court experience

Qualitative coding: an approach to assess inter-rater reliability

Code § § 4-1-6-2, 4-1-6-8.6(b)

A conceptual model understanding post-release opioid-related overdose risk

Regression models and life-tables

On the regression analysis of multivariate failure time data

Evaluating test strategies for colorectal cancer screening: a decision analysis for the U.S. Preventive Services Task Force

What is Dialectical Behavior Therapy (DBT)? Verywellmind website

Care for incarcerated patients hospitalized with COVID-19

Nonfatal drug overdoses treated in emergency departments -United States

Prison Policy Initiative website

Adult Correctional Facilities. Indiana Department of Correction website

Indiana Problem Solving Court Program Directory

What are Drug Courts? National Drug Court Resource Center website

Family recovery court helps with addiction and stabilizes home life. National Public Radio website

Juvenile Drug Court Programs. US Department of Justice

Problem-Solving Courts. The Unified Judicial System of Pennsylvania

An Overview of Juvenile Mental Health Courts. American Bar Association website

Practice Profile: Adult Mental Health Courts. National Institute of Justice Crime Solutions website

Allen Circuit Court website

Tracking Problem-Solving Issues across the Country. National Center for State Courts website

Re-entry Court Law and Legal Definition

What is a Veterans Treatment Court? Justice for Vets website

Acknowledgements: It takes a talented, multidisciplinary team to tackle a project as complex and multifaceted as this study. The students at the University of Pittsburgh Graduate School of Public Health are enthusiastic, dedicated contributors who make working together so rewarding: Ozge Burget, Adam Kruchen, Patrick O'Toole, Jordan Perkins, Ethan Rosentel and Devin Strynkowski. Also, the study's stakeholder committee members are nationally recognized innovators and we greatly appreciate their advice regarding the project's overall aims, design,