Health disparities in jail populations: Mixed methods and multi-disciplinary community engagement for justice and health impacts

Abstract

This special issue of Practicing Anthropology presents multidisciplinary and multisectoral views of a community engaged health disparities project titled “Health Disparities in Jail Populations: Converging Epidemics of Infectious Disease, Chronic Illness, Behavioral Health, and Substance Abuse.” The overall project incorporated traditional anthropological mixed-methods approaches with theory and methods from informatics, epidemiology, genomics, evolutionary and computational biology, community engagement, and applied/translational science.

Introduction

Our project to investigate health disparities in jail populations began with an exploratory meeting between one of the authors of this article and the director of the Coconino County Criminal Justice Coordinating Council (CJCC). That director stated, “In my experience, county jail systems throughout the United States should be viewed as more of a broken public health system and less as an effective justice system than anyone assumes.” Subsequent exploration of that opinion, in consultation with key members of the CJCC and Northern Arizona University’s (NAU) Center for Health Equity Research (CHER), created a community engaged research project titled, “Health Disparities in Jail Populations: Converging Epidemics of Infectious Disease, Chronic Illness, Behavioral Health, and Substance Abuse.” The project was subsequently funded as a pilot project by a local charitable trust, the NARBHA Institute. The primary aim of the project is to address the significant and simultaneous public health threats to both county jail prisoners and the general county and state populations.

During the exploratory phase (literature review, community engagement, etc.) of the project, we found research on health issues and incarceration, especially on jail versus prison incarceration, to be relatively scarce and predominantly single disease oriented. Extant research focuses overwhelmingly on relatively stable prison populations as opposed to the much more transient local jail populations, in spite of the potentially greater impact of jail incarceration on public health and population health conditions (Binswanger, Krueger, and Steiner 2009; Massoglia 2008).

In contrast to prison populations, county and municipal jail populations are more transient and significantly more likely to interact regularly with the general population. As an example, Figure 1 shows the annual incarceration rate for Coconino County for 2001 through 2016, when the converging epidemics project started. The county jail population averages between 8 and 10 percent of the total county population in any given year.

Figure 1.

Coconino County Jail Bookings: 2001–2016.

The average length of jail stays is eight to ten days nationally, with a small but significant percentage of the population staying for up to one year or longer while awaiting trial (Boyd et al. 2014; Greenberg and Rosenheck 2008). However, the median stay for Coconino Country is only two days, with nearly half of the booked individuals being released within twenty-four hours. This short-term jail stay and release process becomes a destabilizing public health force for both individuals and communities and is associated with higher rates of substance use relapse, as well as non-adherence to mental and physical health treatment programs (Wakeman, McKinney, and Rich 2009). Jails are also important potential vectors for communicable disease exposure, as particular populations are cycled in and out of jail. Continuity of care is often severely impacted by individuals repeatedly cycling between jail and community and is associated with limited opportunity for stability in health care (Evans Cueller and Cheema 2012). Medicaid provides a disproportionate number of inmates with insurance; however, states are barred from receiving federal Medicaid dollars for health care costs during incarceration with one exception: any care delivered outside of the correctional facility when the individuals has been admitted for 24 hours or more. (Centers for Medicare and Medicaid Services 2017, Department of Health and Human Services 2016) Currently, 31 states suspend or reclassify Medicaid for the duration of incarceration or for a specific amount of time while the remaining 19 states terminate Medicaid while the individual is incarcerated. This creates a barrier to continuity of care for many chronic conditions, treatment regimens for severe mental illness (SMI), and other behavioral health problems due to benefits being temporarily or permanently terminated, as well as significant differences in the formularies offered by the jail as opposed to those offered by Medicaid benefits (Trotter et al. 2018).

In consideration of these general conditions, our research team committed to assessing health disparities in the Coconino County Detention District based on the invitation and the strong potential for community engaged collaboration, the diverse population experiencing incarceration, and the opportunity to take advantage of a relatively simple cultural-ecological environment (one primary detention center, one primary regional health care center, one federally recognized health care system, one regional behavioral health authority).

Figure 2 depicts the identified interconnections between the primary care institutions in Coconino County and the detention center. Since all of these institutions regularly interact, we felt it should be possible to model the key interactions among these entities and incarcerated individuals. Approximately 10,000 to 12,000 individuals move in and out of the Coconino County Detention Center in an average year. This relatively rapid cycling of incarcerated individuals shapes the social and cultural ecology of the jail and constitutes a very fluid condition in relation to general population health. The multiplex health conditions that are closely concentrated in time and place with the county jail system produce a dynamic array of exposure to infectious diseases, difficult access to treatment for chronic health conditions, interruption in continuity of care for serious behavioral health conditions, as well as ongoing issues for the prevention and treatment of substance abuse disorders.

Figure 2.

Inter-institutional flow patterns for health care access and utilization.

Complementary Approaches

Our research process involved interdisciplinary collaboration between anthropologists, statisticians, psychologists, epidemiologists, public health researchers, criminal justice researchers, and molecular biologists, as well as extensive support from local Criminal Justice Coordinating Council, Sheriff, and corrections officials within the County Detention Center. The design of the project encompasses multi-disciplinary theoretical frameworks and mixed methods designs (from ethnography, to epidemiology, to informatics, to genomics, and beyond). We have also incorporated community engaged analyses and nascent policy actions as a desired endpoint.

The protocol and design for the project (Trotter et. al 2018) includes three data collection and analytic frameworks including: (1) a comprehensive targeted synthesis and analysis of existing databases on the health status of county jail populations; along with (2) primary data collection through a cross-sectional health and health care services survey of incarcerated individuals; coupled with (3) collection of biological samples to model infectious disease transmission in a county jail population. This “converging epidemics” framework (see Figure 3) is further explored, expanded, and evaluated in our recent publications and the accompanying articles in this PA issue.

Figure 3.

Converging Epidemics Databases.

Comprehensive Analysis of Existing Databases

Our secondary data and epidemiological approach is captured, in part, in a published paper, the Epidemiology of Incarceration (Camplain et al. 2019). From January 2001 to May 2018, over 75,000 unique individuals (78% men, 47% White, 33% American Indian/Alaska Native, 16% Latino/Latina or Hispanic, and 4% Black) were booked into the Coconino County Detention Facility almost 200,000 times, when counting multiple incarcerations of individuals (Camplain et al. 2019). Further exploration identified differences by race/ethnicity and sex stratified by length of stay.

Individuals are granted an initial appearance before a judge within twenty-four hours of booking and subsequently are (1) released without being formally charged, (2) released on recognizance, (3) released on bond, or (4) fully booked into the jail. When comparing racial/ethnic makeup of those who are released (within twenty-four hours) and those who stay in jail (greater than twenty-four hours), the difference is stark. A lower proportion of individuals who stay in jail were White compared to those who were released within twenty-four hours (37%, 54% respectively), and a higher proportion of those who stay in jail were American Indian/Alaska Native or Latino/ Latina or Hispanic compared to those who were released within twenty-four hours (Figure 5). Additionally, a higher proportion of males stayed in jail (83%) versus being released within twenty-four hours (74%, Figure 6).

Figure 5.

Racial/ethnic differences between those who are incarcerated ≤ 24 hours vs. > 24 hours.

Figure 6.

Sex differences between those who are incarcerated ≤ 24 hours vs. >24 hours to hospital discharge and emergency department core databases.

The secondary data and epidemiologic approach is further explored in Organizational and Institutional Compartmentalization as a Barrier to Population Health (Hepp, Fofanov, and Trotter this issue). The Hepp article notes that the Centers for Disease Control and Prevention have defined population health as “an approach that allows health departments to connect practice to policy.” One intention of the approach is that institutions from several different sectors, including academia, non-profits, public health, and local governments, should collaborate towards a holistic view of local health outcomes. In principal, the holistic view should allow for the recognition of substantial health concerns such that resources can be re-allocated to address them. In reality, each institution is bound by unique financial and personnel constraints, and there is little incentive to motivate actions that extend beyond institutional missions. This means that data sharing is a significant impediment to the type of research proposed in our project. In order to assess an inmate’s criminal and health care history, numerous sources of data have to be linked, including incarceration records, jail medical records, primary care patient records, and general information on access to health care services.

Each data set has its own restrictions about exporting data and cross-institutional database linkage. The purpose of the Hepp article is to highlight barriers we have encountered while attempting to integrate a minimum of five large data sets owned by both criminal justice and health care entities, in order to construct an understanding of the complex interconnections between population health characteristics, incarceration, and environmental justice (including data on different forms of health care provided to individuals who have entered the criminal justice system).

A Cross-Sectional Health and Health Care Services Survey of Incarcerated Individuals

Our prospective mixed-methods data collection effort focused on a combination of in-jail observations and surveys combined with biological data collection to establish bio-markers for health conditions. We successfully enrolled, interviewed, and collected bio-specimens from 199 incarcerated individuals. Some of our preliminary results from direct contact with incarcerated individuals is documented in A Survey of Health Disparities, Social Determinants of Health, and Converging Morbidities in a County Jail Population (Trotter et al. 2018), and we have presented at regional and national conferences and both scholarly meetings and community forums. We clearly document the ways the environmental health status of jail populations in the United States constitutes a significant public health and population health threat.

Each respondent was asked to identify if they had been informed (by a health care professional) that they had one or more of the twenty-eight conditions most frequently identified in the literature as being critical to the health of incarcerated populations (see Figure 7). Respondents reported a higher number of those twenty-eight conditions than is evidenced in the general population. Additionally, our research documents a study population with a high prevalence of mental and physical health conditions, in addition to substance abuse. Incarcerated populations are disproportionately burdened by higher rates of chronic and infectious diseases, with significant rates of comorbidities (multiple health problems). For example, the rates of many chronic and infectious diseases in United States incarcerated populations are more than double of those in the general population, including diabetes, emphysema, liver disease, hepatitis C, HIV, and tuberculosis.

Figure 7.

Self-report of twenty-eight primary health conditions in an incarcerated population (N = 199).

We subsequently approached our data on health conditions from a multiple response perspective (i.e., for each individual we indicated all of the conditions reported, rather than simply summing individual conditions). Approximately three-fourths (78.5%) of the individuals surveyed reported a minimum of two and up to ten or more co-morbid conditions (Figure 8). We also found that co-morbidities present a special challenge to county detention centers and incarcerated individuals in terms of treatment priorities, health policy, and general population health characteristics (Trotter et al. 2018).

Figure 8.

Number of self-reported comorbidities in a county jail population

Our exploration of these co-morbidities included treating the links between health conditions as a map for potential multi-sectoral intervention, so that we could look at disease clusters through the lens of a social network visualization process. In Figure 9, each of the twenty-eight health conditions is represented by a node in the network diagram. The connections between conditions identify the self-report matches the various respondents identified and are represented by the lines between the nodes. The thickness of the lines between nodes represents the number of times the dyadic co-morbidity is identified by multiple respondents. The size of the node represents the number of times that the condition is listed as a comorbidity. We plan on looking at individual (ego-centric) clusters to provide policy and procedural information for treating the most prominent clusters of diseases at the same time, rather than sequentially.

Figure 9.

Network visualization of interacting comorbidities in a county jail population.

Our survey population also shows a much lower self-report of global health status (excellent, very good, good, fair, poor) than the general population. Since this global health indicator correlates with projected five-year mortality rates, this finding is an important projective measure for the intersection of the social determinants of health and incarcerated populations (Barger, Cribbet, and Muldoon 2015). Combined, these findings have a number of important policy implications for both morbidity and mortality in incarcerated populations.

Infectious Disease Information from the Project

The third focus of our research project has been to create a model for understanding communicable disease transmission. We accomplished this through biological and genomic testing for Staphylococcus aureaus in the jail population. We successfully collected bio-specimens from 198 incarcerated individuals (see Figure 10 for training picture). The primary elements of the bio-data collection (including concurrent survey data collection) included: (1) recruitment; (2) full informed consent; and (3) survey and bio-specimen collection, transport, processing, and subsequent analysis. The individuals who elected to participate in the study were interviewed and had bio-specimens collected in the activity rooms of their living pod, in groups of two to five. The consent process was thoroughly addressed, and informed consent forms were signed. Only one out of 199 participants elected to complete the survey but opted out of the bio-testing. The sampling process did not require trained personnel and was performed by participants themselves. The study coordinator first demonstrated the use of the sample collection equipment. The participant then used a sterile, single-tipped cotton swab and inserted the swab head less than 1 cm into one of their nostrils. Next, they gently rubbed the inside of their nose with the swab head for one to two seconds, removed the swab from their nostril, then placed the same swab head in their other nostril, repeating the procedure. After they removed the swab head from their second nostril, they carefully returned the swab head to their holding tube and gave the tube back to the project team member. The tube was placed into a Ziplock storage container for transportation to the lab and further analysis.

Figure 10.

Training picture for collection of nasal swabs.

The primary results of this feasibility and pilot study are more fully presented in Overcoming Institutional, Scientific, and Cross-disciplinary Barriers for Health Care and Disease Transmission Research in County Jail Settings: Integrating Genomics, Survey Data, and Biological Data Collection Strategies (Fofanov et al. this issue). Figure 11 identifies the high rates of S. aureus carriage in the detention center population, compared with the general county population. These findings are being further explored through phylogenetic testing of the various strains of S. aureus to determine the relative contributions of community acquired versus institutionally acquired transmission of the bacteria, both of which have significant policy and procedure implications for the detention system, in terms of dealing with infectious disease.

Figure 11.

S. aureus carriage rates in general and incarcerated populations of Northern Arizona.

Discussion: Challenges for a Multi-Disciplinary Approach to Health in Incarcerated Populations

The combined data from the project allows for a nuanced picture of both the public health and the environmental health conditions within a county detention center and the intense interactions between the incarceration environment and local communities. The primary questions we attempt to answer include: (1) What is the distribution of the number of incarcerations and lengths of stay for the population cycling through the Coconino County Jail system? Who are the “super users” of the system, and how do they compare to the rest of the jail population in terms of overall health status, co-morbidity conditions, and carriage of infectious diseases. (2) What are the demographics and socioeconomic characteristics (or social determinants of health) of persons in the jail system (both super users and single stay individuals), and how do these individuals compare with the population(s) utilizing county emergency services, and primary health care systems? (3) What are the incidence and prevalence rates for specifically targeted diseases and conditions in the jail population and general health care populations in Northern Arizona (priority conditions: hepatitis C, HIV/AIDS, sexually transmitted diseases (STI), MRSA, TB, cardiovascular, oral health, substance abuse, diabetes, and SMI). (4) Are incarceration variables (length of stay, type of criminal activity, ability to post bail, number of incarcerations, etc.) predictive values for infectious disease, chronic illness, or behavioral health status in jail populations? (5) Is there a statistically significant difference in prevalence of disease (contagious/infectious/chronic/behavioral) between the different stratifications of the number of incarcerations and other criminal justice variables in the jail populations?

We continue to wrestle with both logistical and ethical questions that necessarily emerge when working with this vulnerable population. Challenges in recruiting jail inmates without undue coercion, while simultaneously addressing proper consent procedures, are significant barriers. Language barriers and mental health issues can make it difficult to explain complex concepts like disease transmission, genomics, and the need for access to health care records. Even simple things like participant compensation are complicated in an environment when participants do not have direct access to cash or choices for places to spend it. Finally, simply carrying out questionnaires and biological specimen collection is complicated by the requirement of not introducing contraband into controlled jail environments—something as simple as paper clips, ball-point pens, heavy-duty test tube labels, etc. are among the “contraband” articles that had to be negotiated within the scope of the research project.

Incarcerated individuals are both ethically and legally considered vulnerable populations and are eligible for numerous additional protections from coercion or harm. As a consequence, the Health Disparities in Jail Populations project explored a number of collaborative options with our Institutional Review Board and with the appropriate county legal offices and the judicial system, in collaboration with the Coconino County Criminal Justice Coordinating Council. The ethical conditions framing the “converging epidemics” project are articulated in the article Ethics of Working with Incarcerated Populations: A View from the IRB (Traustadóttir, Hanabury, and Peoples this issue). That article presents our experiences and issues when working with an incarcerated population, from the perspective of the university Institutional Review Board (IRB). The goal of the article is to provide information to other IRBs and researchers that might help the review process for research protocols involving prisoners and mitigate potential problems before they arise. Vulnerable populations such as prisoners require special conditions and protections for recruitment, consent, compensation data collection, data protection, and analysis, as well as compliance with detention center procedures, especially when the overall research project includes analysis of secondary databases containing personal identifying information (PII) and personal health information (PHI), prospective surveys of prison conditions, and biological data collection. The three main issues the article addresses are (1) the balance between privacy and confidentiality of the subjects and researchers’ safety within the facility (researcher safety not IRB purview), (2) payment to subjects, and (3) the ability of subjects to communicate with the IRB. The authors also present some recommendations for research “best practices.”

The article Measuring Health in Jails: Limitations and Opportunities in Measurement of Health-Related Constructs (Arazan et al. this issue) provides a helpful “how to” approach for researchers who would like to become engaged in social justice research among jailed populations. The article stresses the importance of expanding on the health literature among jailed populations and focuses on how to overcome some of the challenges that are specific to doing health research in jails including: issues related to selecting appropriate self-report items on questionnaires and the need for formal testing of measurement properties of commonly-used scales in jailed populations. The opportunities and issues for dealing with combined social and biological conditions in these populations are further described in Overcoming Institutional, Scientific, and Cross-disciplinary Barriers for Health Care and Disease Transmission Research in County Jail Settings (Fofanov et al. this issue). The article focuses on a number of institutional, scientific, and disciplinary barriers to conducting research in this setting. The challenges can take the form of core scientific reproducibility—for example, national surveys (like NHANES, PROMIS, BRFSS, and Healthy People 2020) are explicitly administered to non-incarcerated populations and avoid incarcerated populations, making comparison between those populations very difficult.

The article Reflexive Challenges in Community Engaged Research in a County Jail (Eaves, Kohlbeck, and Camplain this issue) continues in the vein of “best practices” approaches to studying incarcerated populations: traditionally, anthropology has been conducted far away from home and researchers are able to develop friendships and relationships with their participants and then leave them to go back to their other lives as academics, students, or practicing anthropologists. As local, participatory research becomes more common, the logistics of this local research can challenge our traditional sense of disconnection from research participants. The Health Disparities in Jail Populations project provides an example of a project where a major barrier to conducting research in the jail stemmed from concern about getting to know inmates at the local jail in a small community. Concern that researchers or students would run into participants at local bars or in other locations was a source of concern. Practicing anthropology in our own local communities and overcoming fears about getting to know people representing demographics other than our own are considered here through qualitative interviews with the researchers and students involved in conducting research.

We are also incorporating an article on the “spin-off” effect of the pilot project. One of our recently graduated MA students, Bailey Kohlbeck, used her experience and connections developed by the pilot project to successfully conduct an ethnographic study of the safety concerns of women engaged in the Coconino County probation and parole department. Her article Lost to Found: The Creation of a Master’s Thesis provides important insights into the evolution of procedures and ideas from the parent project and provides important framing for addressing the need for policy changes in the existing probation and parole system.

We have also incorporated a set of community-oriented commentaries. The overall project was embedded in NAU’s Center for Health Equities Research (CHER), which provided a strong framework for supporting a substantial multi-disciplinary approach to health equities. The Search for Health Equity among Individuals Incarcerated in Jail (Camplain and Baldwin this issue) focuses on the opportunities and challenges for incorporating and hopefully sustaining community-based, multidisciplinary efforts for the integration of multi-faceted projects within the context of a formal university Center for Health Equity Research. From the community perspective, we have invited commentary on the utility and impact of the Converging Epidemics project from the perspective of our primary funder, the NARBHA Institute (Carroll), the Coconino County Administration (Peoples), and county probation and parole (Douthit). These commentaries provide very important framing for both future research as well as program and policy development. As can be seen from those commentaries, Coconino County has a number of highly innovative programs and a strong orientation toward multi-sectoral structural innovations. Many our findings are not a surprise to our partners, but they do provide evidence-based information for supporting existing programs and identifying areas of policy and program innovation. Hopefully, some of our unexpected findings will also provide strong support going forward.

Conclusions and Observations

Dealing simultaneously with all the health and health care access conditions for jailed populations, given the average time of incarceration for each individual, creates serious issues that must be dealt with on a system-wide basis, rather than by detention facilities alone. We are addressing the primary social and environmental impacts on the health of incarcerated individuals by taking the position that a simple categorical approach (single disease, single solution) ultimately fails to appropriately address the public health ecology of incarceration and its impact on population health. One of our working assumptions is that jails may not be an appropriate venue to conduct all prevention and intervention programs in all four areas of the converging morbidities in our study (chronic illness, behavioral health, infectious disease, and substance abuse). Rather, some form of multi-sectoral approach is needed to comprehensively address public health needs of this population. At the same time, jails may play a pivotal role in addressing the overall impact of incarceration on public health. We believe a multi-sectoral or collective impact framework (Trotter et al. 2018) will be necessary to reduce the “broken public health” system in jails. Because of intersecting conditions, we aim to produce a model for prevention and intervention programs in all four areas of health, potentially resulting in significant cost reduction for health care among incarcerated individuals.

Figure 4.

Ethnicity of incarceration by year and type of crime (misdemeanor or felony) in Coconino County from 2008–2016.