https://orcid.org/0000-0003-4294-4133
Figures
Abstract
More than 7 million people are released each year from U.S. jails or prisons, many with chronic diseases that would benefit from primary care in their returning communities. The objective of this study was to provide an in-depth, payer-agnostic description and associational analysis of primary care need and utilization by all individuals ever detained in one county detention facility over a 7-year period. Detention records 2014–2020 were merged with data from an electronic health record with excellent coverage of local primary care, emergency, and hospital services. We found low primary care participation for the group as a whole, with under three quarters of those with serious chronic diseases ever seeing a primary care provider over a 7-year period and less than half ever having a year with more than one visit. Multivariable regression models estimated associations between individual characteristics (demographic, detention-related, and clinical) and ever having access to primary care (logistic) and the number of primary care visits (zero-inflated negative binomial). We found that having more jail bookings was associated with fewer primary care visits, but not one-time access, even controlling for time out of community, age, insurance, and other demographic characteristics. This finding was driven by subgroups with chronic disease such as hypertension, obstructive lung disease, and diabetes, who most need regular primary care. Being Black retained an independent effect, even controlling for bookings and other variables, and was also associated with fewer primary care visits, though not one-time access. To promote primary care utilization among individuals who have the combined challenges of repeated jail involvement and chronic disease, it is crucial to focus on engagement, as much as formal access. Access to health insurance alone will not resolve the problem; Medicaid expansion should be coupled with specialized, tailored support to promote engagement in primary care.
Citation: Easter MM, Schramm-Sapyta NL, Swartz MS, Tackett MA, Greenblatt LH (2024) Primary care need and engagement by people with criminal legal involvement: Descriptive and associational analysis using retrospective data on the entire population ever detained in one southeastern U.S. county jail 2014–2020. PLoS ONE 19(10): e0308798. https://doi.org/10.1371/journal.pone.0308798
Editor: Souparno Mitra, NYU Grossman School of Medicine: New York University School of Medicine, UNITED STATES OF AMERICA
Received: June 10, 2024; Accepted: July 24, 2024; Published: October 7, 2024
Copyright: © 2024 Easter et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All data underlying tables and statistics presented in the study have been provided in Supporting Information files. Non-aggregated data cannot be shared publicly because of agreements with Duke University Health System and Durham County Detention Facility. To obtain access to data, researchers would need to contact Nicole Schramm-Sapyta, Maria Tackett, and Michele Easter to join the research team, which would require appropriate research credentials and a relevant, acceptable analysis plan. Following this initial approval, additional approval of the overall plan would be sought from contributing data partners, and finally institutional approvals including sponsorship as Duke affiliate (requiring a background check), and approvals (human subjects protection, Duke's protected computing environment).
Funding: Funding was provided as part of an undergraduate course series by Duke University Bass Connections Program (NSS and MT, https://bassconnections.duke.edu/) and Rhodes Information Initiative at Duke Data Plus Program (NSS and MT, https://bigdata.duke.edu/participate/data-plus/). Additional funding for Michele Easter's time was provided by the Wilson Center for Science & Justice, but not specifically for this study, (https://wcsj.law.duke.edu/). The sponsors and funders played no role in study design, data collection and analysis, decision to publish, nor preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Introduction
Americans have shorter lives and worse health than those living in comparably developed countries [1], while having the highest rate of incarceration in the world [2]. Those who have experienced incarceration often have significant health and healthcare needs and barriers, including a need for primary care that often goes unmet [3–7]. Individuals from minoritized communities, with adverse childhood experiences, or with other disadvantages are more likely to become involved in the criminal legal system for a number of reasons [8–11]. These disadvantages, in combination with sequelae of confinement (e.g., trauma, infectious disease exposure, life disruption, [12]) and reentry challenges make this population particularly vulnerable to health problems and contribute to health inequities [13–15]. Individuals with criminal legal involvement have a higher prevalence of chronic and infectious disease [16–19], creating needs that could be alleviated by a robust connection to primary care.
A key point of opportunity for connection to primary care is the transition out of incarceration, back into the community. With this in mind, it is important to consider the distinction between jail and prison release. Nearly half a million people were released from U.S. federal and state prisons in 2022 [9], yet jail releases are estimated to be over 7 million (7 million admissions at an average stay of 32 days [11]) Jail-involved people, as compared to prison-involved people, are arguably the most relevant to local community healthcare providers, because they are a much larger population and most return quickly to the community; the vast majority are un-convicted and awaiting court action [11].
Those with primary care needs face significant disruption in continuity across carceral and community systems [20–24]. Community-based primary care providers can play a significant role in improving the health and quality of life for this population. Although people in prison and jail are entitled to and may receive primary care during detention and incarceration, they typically face significant challenges in engagement and continuity of care across the transition. These challenges include lack of insurance, fear of discrimination, lack of coordination across systems, and distrust [6, 23, 25, 26]. Releases can also be unpredictable and therefore difficult to coordinate with community services. Specialized interventions exist to overcome these challenges [6, 27, 28] but are not widely implemented nationwide.
Healthcare systems often lack information about how their local county jail population uses their services. This study uses a unique county-wide dataset to follow the entire population of people detained or incarcerated in one urban county detention facility in North Carolina and to characterize their community healthcare use, using electronic health records that capture nearly all primary care visits in the county. Frequent ED use is more common in this population than in the general population [29] and the general patient population of this specific healthcare system [30]. Ideally, people with frequent ED visits and other acute illness care such as inpatient (IP) stays should become engaged in outpatient care and thereby avoid preventable ED/IP. How far is reality from that ideal? This paper provides an overview of community healthcare use by a jail-involved population with a focus on primary care, and compares factors associated with greater engagement for subgroups defined by chronic disease.
Materials and methods
1. Sample and data
Data were collected from a southeastern mid-sized city in a state that had not expanded Medicaid at the time of the study. Every person who was booked into the county detention facility, 2014–2020, was included in the initial study sample (n = 29,120), i.e., anyone serving a sentence, awaiting trial, or released without spending the night. Their interactions with the healthcare system could have occurred before, during, or after booking. The present analysis includes only the healthcare-using subsample, i.e., those with a 2014–2020 healthcare encounter accompanied by a diagnosis reflecting a provider’s clinical impression (n = 17,974/29,120, 61.7% of the initial sample, S1 Table in S1 File).
Retrospective data on bookings and healthcare encounters January 1, 2014-December 31, 2020 from the detention facility and healthcare system included all county Emergency Departments and hospitals except VA, and most primary care clinics in the county, including all clinics known to serve poor and underinsured clients except one. Retrospective data were accessed for research purposes by informaticians at the detention facility and provided to informaticians at the healthcare system for linkage to data from healthcare records; there was no recruitment or other subject contact for this study. Healthcare system informaticians removed all direct identifiers and adjusted all dates to create a set of linked, limited datasets that were shared with the research team on June 16, 2021. The limited datasets were accessible only via the healthcare system’s secure computing environment and a user-limited study folder. Authors did not have access to information that could identify individual participants. The Duke University School of Medicine IRB included a prisoner representative, waived the requirement of consent for use of these retrospective records, and approved this study (Pro00102050).
2. Measurement
Healthcare encounters were categorized into types (e.g., primary care, Emergency Department (ED), inpatient, urgent care, testing.). Primary care and ED encounters were aggregated into 7-year counts (2014–2020) and calendar-year counts (2014, 2015, etc.). Frequent ED use (FEDU) was defined as 4+ visits in a single calendar year, a commonly used threshold [31–33]. Primary care engagement included visits with any primary care provider and excluded OB/GYN and behavioral health. Encounters were classified as “uninsured” if no payor was listed; “Medicaid/state-funded” (including pending); and “other” if Medicare, private insurance, or another payor was recorded. The first-listed diagnoses at each encounter were assigned to categories according to ICD10 chapters [34]. People were classified into one of three chronic condition statuses: having either (1) a “priority” chronic condition (obstructive lung disease, kidney disease, diabetes, hypertension, heart failure, myocardial infarction, as classified by [35], (2) other chronic condition (as classified by Chronic Condition Indicator Refined [36]), or (3) no record of a chronic condition. The 6 “priority” conditions were chosen because such patients would typically need 2 or more primary care visits annually. In addition, diagnoses were categorized as ambulatory care sensitive conditions [37] and chronic behavioral health conditions (ICD mental disorder chapter excluding tobacco use disorder). Person-level, mutually exclusive categories were created for insurance (ever Medicaid/state insurance, ever other insurance and never Medicaid/state insurance, and never insurance) and chronic condition status (ever “priority” chronic condition, ever other chronic condition, and no chronic condition). Demographic characteristics as recorded by the detention facility were collapsed to facilitate analysis: sex/gender (male, female/other); race (Black, white, other); and ethnicity (Latino, not Latino). Age was measured as of 2021 and categorized into age bands (18–29, 30–39, 40–49, 50+). Bookings and incarceration days were totaled over the 7-year observation period, and bookings categorized as (1, 2, 3–4, and 5 or more bookings).
3. Analysis
Descriptive statistics were used to characterize healthcare utilization and diagnoses. Multivariable logistic regression models were used to estimate the odds of receiving any primary care over the 7-year period for the full sample as well as for three subgroups: those with one or more of the 6 chronic conditions, those with any other chronic condition, and those with no observed chronic condition. Independent variables were gender, race, ethnicity, age, number of jail bookings, insurance type (ever observed), and total incarceration days (to adjust for time out of community). Zero-inflated negative binomial (ZINB) regression models were used to estimate associations between independent variables and the number of primary care visits over 7 years (“count model”), and the “excess zeros”, i.e., the absence of any primary care visits (“zero model”), which could be shaped by different factors than the number of visits. Formal one-sided tests of overdispersion using the Pearson statistic supported use of ZINB. These ZINB models were estimated for the same four groups as the logistic regression and count models included the same independent variables as the logistic regression models for all four groups, as did the zero model for the full set of patients. In zero models for subgroups defined by chronic conditions, several independent variables were collapsed (age from 4 categories to 2) or omitted (Latino, insurance) due to lack of convergence and inflated estimates that were not interpretable for one or more groups. SAS 9.4 was used for all analyses; PROC GENMOD was used to carry out logistic and ZINB models [38].
Results
1. Use of primary care and other healthcare by one county’s jail-involved population
About half (50.6%) of the sample had a primary care visit over the 7-year study period (Table 1, left panel). Consistent with earlier research (Authors 2023), the most common type of encounter with the healthcare system was an ED visit; most (84.1%) had one, half staying overnight at least once. Far fewer visited urgent care (37.3%) or were admitted for inpatient hospitalization (16.1%) during that time period.
Of those who had a primary care visit, the mean number of visits was 8.5 over the 7-year period (11.7 SD), with a range 1–173 (Table 1, right panel) and a median of 4 visits. For those with an ED visit, the mean number of ED visits was only slightly lower at 7.2 (SD 11.8), and ranged 1–309, and the median was also 4. For those who ever stayed overnight at the ED.Those who were hospitalized had a mean of 1.7 admissions, with a median of 1 and range 1–14. Some individuals had more than 800 encounters with the healthcare system over 7 years (S2 Table in S1 File).
With regard to insurance (non-mutually exclusive categories), most (83.2%) of the sample had at least one uninsured encounter over the 7-year period. About 41% ever had an encounter paid for by Medicaid or state fund, and nearly half ever had an encounter paid for by another type of payer, including private insurance, Medicare, a carceral facility, or another type of payer. Those with Medicaid or state-funded health care had more visits (mean 30.6 visits) than those who were uninsured (mean 11.1 visits) or had other types of coverage (mean 14.5 visits).
2. Characteristics of healthcare user types: demographic, detention-related, and clinical features
The sample was classified into four groups based on frequency and type of healthcare utilization (Table 2): “Any IP” ever had an inpatient admission; “Frequent ED” ever had a year of frequent ED use (and never an inpatient admission), “Any ED” ever used the ED (but never frequently and never inpatient), and “No ED/IP” never used ED or inpatient. The largest group (“Any ED” 54.2%) visited the ED at least once, but never frequently, and were never hospitalized. The other three groups were approximately equal in size (15–16%): those who never went to the ED nor hospital (“No ED/IP”); those who went to the ED frequently but were never hospitalized (“Frequent ED”); and those who were hospitalized (“Any IP”). Of the “Any IP” group, 42% were also frequent ED users at some point over the study period. All demographic, detention-related, and clinical characteristics were statistically significantly associated with healthcare user type.
The “Frequent ED” group was the only group in which most people (61.5%) had more than one jail booking (combining categories), and 25.2% had 5 or more bookings, compared to 5.6–15.4% for the other groups. The “Frequent ED” group also had the highest percentage of Black individuals (71.8%), the lowest percentage of white individuals (17.6%) and more older individuals, with 28.1% 50 or older (compared to 20.2–22.6% for the other groups) The “Frequent ED” group and the “Any IP” group had the highest proportions with one or more of the 6 “priority” chronic conditions (54.6% and 50.2%), ambulatory care sensitive conditions (63.1% and 64.6%), and behavioral health conditions (64.3% and 70.7%).
By contrast, the “No ED/IP” group had the fewest jail bookings; 72.9% had one jail booking only, the minimum to be included in the study (vs. 38.5–54.9% of other groups). The “No ED/IP” group also had highest percentage of white individuals (32.4% vs. 17.6–21.2%) and the lowest of Black (45.7% vs. 63.4–71.8%), as well as the highest % Latino (7.9% vs. 1.9–4.9%). This group also had the lowest proportion ever having public insurance and the highest proportion ever having a private or other non-public payer. The “No ED/IP” group had the fewest diagnosed chronic conditions, ambulatory care sensitive conditions, and behavioral health conditions.
Of the remaining two groups, the “Any IP” group had the highest proportion female (63.5% vs. 22.0–31.7% for other groups), and the highest proportion ever having public insurance (79.0% vs. 17.1–57.3%). Characteristics of the “Any ED” group usually fell between or were similar to those of other groups.
With regard to chronic conditions, most of the sample had a chronic condition, either one or more of the 6 “priority” conditions (34.3%) or another chronic condition (48.2%); 17.5% had no record of a chronic condition. Virtually all people in the “Frequent ED” and “Ever IP” groups were ever diagnosed with a chronic condition, and half or more had one of the “priority” conditions (54.6% and 50.2% respectively). Most people in “Frequent ED” and “Ever IP” had an ambulatory care sensitive condition (63.1% or 64.6% respectively), and/or a chronic behavioral health condition (64.3% and 70.7% respectively).
3. Common diagnoses at ED or IP by type of healthcare user
To better understand the health needs of healthcare user types, we classified all diagnoses received at ED or IP for the three groups that used these services (“Any ED”, “Frequent ED”, “Any IP”). The two ED groups had the same five diagnosis types among their top 5 (at ED): Injury/poisoning, mental/behavioral, musculoskeletal/connective, digestive, respiratory. Mental/behavioral diagnoses were most common among the “Frequent ED” group, and injury/poisoning diagnoses (which includes drug or alcohol intoxication or overdose) among the “Any ED” group. For the “Any IP” group, mental/behavioral diagnoses were the most common, but the next 4 most common diagnoses were not shared with the other two groups: pregnancy/childbirth, metabolic diseases, neoplasms, and nervous system disorders.
4. Primary care visits by healthcare user groups and chronic condition groups
Out of a total of 514,725 encounters (S3 Table in S1 File), 15.0% (n = 77,224) were for primary care. About a third of all visits were uninsured (32.3% of all encounters, 32.6% of primary care visits), and about 40% were covered by Medicaid or state insurance (43.7% of all encounters, 39.8% of primary care visits). The remaining quarter were covered by another insurance plan (24.0% of all encounters, 28.6% of primary care visits).
The four healthcare user groups varied in their use of primary care (Table 3, upper panel). The “Any IP” group used primary care the most, with the highest proportion ever using it (69.4%) and highest mean visits (12.1). The “Frequent ED” group had the next highest use, with 62.6% ever using it and a mean of 10.5 visits for those who did. Consistency of use, measured as years with at least one primary care visit, was also highest for “Any IP” followed by the “Frequent ED” group (S4 Table in S1 File).
Because people with “Any IP” or “Frequent ED” use patterns are likely to have needs that would benefit from greater engagement in primary care, we compared groups by whether they ever had 2 or more primary care visits in any calendar year. The “Any IP” group had the highest engagement, with 42.6% ever having 2+ visits, followed by the “Frequent ED” group, with 35.0% having 2+ visits (consistency of use followed the same pattern).
We also compared use of primary care by chronic condition status Those with “priority” conditions were more connected to primary care than the other two groups: 73.1% ever had a primary care visit (vs. 45.6% and 20.4%); the highest number of visits (mean of 12.2 vs. 5.3 and 2.0), and more years with 2+ primary care visits (2.6 vs. 1.8 and 1.1).
5. Correlates of primary care engagement
What factors were associated with primary care engagement? Demographic characteristics, number of bookings, and insurance factors were considered together in models predicting any primary care visit over 7 years (Table 4 upper panel: logistic regression models) and the number of primary care visits over 7 years (Table 4 lower panel: zero-inflated negative binomial regression count model, see S5 Table in S1 File for zero model). In addition to the full sample, separate models were estimated for three subsamples: people with “priority” chronic conditions, other chronic conditions, and no recorded chronic conditions.
In general, across most models and chronic disease subgroups, being female, Latino, older, or insured was associated with having primary care. Insurance (primarily Medicaid/state insurance) and age (primarily age 50+) had the largest effect sizes for people with chronic conditions. However, for the subgroup with no chronic conditions, being Latino was associated with the largest effect sizes (ever having access and having more visits), more so than age, and being female was associated with a lower odds of ever having primary care. Being Black and having a higher number of bookings each had mixed results across models and are described separately.
In full group models (logistic, count, and zero), the largest effect sizes were associated with having public insurance compared to no payer listed (a 4.9 times greater odds of ever having primary care and an average of 3.6 additional visits) and age, particularly being 50+ years old compared to 18–29 (e.g., a 4.0 times greater odds of ever having a visit, and an average of 4.0 additional encounters). In the zero model for the full group, being Latino and ever having “other” insurance were additionally among the largest effect sizes, in the direction of greater primary care engagement.
Similar results were found in the two subsamples with chronic conditions; the largest effect sizes were also with older age, Medicaid/state insurance, and other insurance in both logistic and count models. In the model for the subsample with no chronic conditions, the largest effect sizes were associated with public insurance, other insurance, and being Latino.
Less consistency was seen across models in estimates related to being Black and the number of jail bookings. Compared to being white, being Black was associated with a greater odds of ever having primary care for those with “priority” chronic conditions, but a lower odds for those with other chronic conditions. Being Black was also associated with fewer visits in the count models for the full group and both subsamples with chronic conditions.
The number of jail bookings had an impact on the number of primary care encounters but mixed results for ever having access. Regarding one-time access, having more jail bookings usually was not statistically significantly associated. By contrast, in count models, having a high number of jail bookings was consistently associated with fewer visits. This pattern was seen in the full group. (3–4 bookings and 5+ bookings vs. 1), the subgroup with “priority” chronic conditions (same) and those with other chronic conditions (5+ bookings vs. 1). Specifically, for those with the priority chronic diseases, people with 3–4 bookings (or 5+ bookings) have an average of 80% as many visits compared to those with only one booking, i.e., a 20% decrease in visits. These associations with bookings were independent of total time detained in jail.
Discussion
This study found that about half of people who were detained or incarcerated in one county jail 2014–2020 and interacted with the local healthcare system had seen a primary care provider during that time period. Rates were higher for those whose healthcare use or diagnoses indicated greater need, such as those with inpatient stays, frequent ED use, and diagnosis of one or more “priority” chronic conditions. Typically, these individuals still underutilized primary care for important health conditions, with only a third or fewer of those groups having a visit most years and under half ever having a year with two primary care visits, the recommended minimum for people with the priority diagnoses. In 2019, 67% of Americans 18 or older had a primary care-like visit in the past year (i.e., a wellness visit, physical, or check-up was had by 78.9% of the 84.9% who saw a physician [39]).
This jail-involved sample was enriched with characteristics associated with lower engagement in care; in the general U.S. population, non-White, younger, male, or uninsured individuals tend to have lower access to a “usual source of care” compared to White, older, female, or insured individuals [40]. More than three quarters of the sample (83%) ever had a healthcare encounter that was uninsured, and when viewed across 7 years, the modal category had received Medicaid/state insurance (40.9%), and the next largest group had never had a payer listed (26.9%).
Women, who are underrepresented in the jail population, were overrepresented among inpatients compared to men. This was consistent with pregnancy/childbirth being one of the top 5 diagnostic families for inpatient encounters but not in the other healthcare user groups, and lower rates of primary care engagement. Women of reproductive age may get Medicaid when pregnant and be admitted for childbirth. Their involvement in the healthcare system may have been driven by healthy childbearing, and their primary care needs may have been minimal or addressed well by OBGYN specialists at the time of care.
There was evidence of feedback loop between insurance and health service type. Nearly 80% of the inpatient group had a public payer, likely due to the hospital assisting eligible pregnant women and others to get care covered by Medicaid or other local resources. There may also be selection effects in which those who cannot pay are referred elsewhere for non-emergent care, e.g., a state psychiatric hospital rather than a private academic hospital. Only 3.6% of the inpatient group never had a payer listed.
The frequent ED-using subgroup had significant health and other challenges: a quarter had 5 or more bookings during the study period, more than half had at least one of the 6 priority chronic diseases, nearly two thirds ever had an ambulatory care sensitive condition, and similar numbers ever had a behavioral health disorder. The inpatient group was similar in these respects to the frequent ED group with the notable exception of having fewer jail bookings. In this sample, inpatients had better access and engagement with primary care than frequent ED-users despite similar diagnoses. This difference is consistent with system-level factors, such as requirements surrounding discharge and referral for inpatients [41], as well as person-level factors that may be associated with both frequent ED-use and multiple jail bookings.
Results showed that jail bookings were associated with a reduced number of visits, though not one-time access. This finding was driven by the two subgroups with chronic disease, who most need regular primary care. Having more jail bookings was independently associated with reduced repeat primary care visits, even controlling for age, insurance, other demographic characteristics, and the number of days detained. This suggests that much more needs to be done for this vulnerable group to promote and support ongoing engagement in primary care; a person may go once but to have a real impact there should be repeated visits.
A similar pattern of associations was found for being Black as for jail bookings; being Black was associated with reduced visits but not one-time access, suggesting a similar need for support to sustain engagement. Being Black was independently associated with reduced visits even when adjusted for insurance, number of bookings, time incarcerated and other characteristics. Sustained engagement in primary care is likely shaped by structural and institutional racism, and likely to be worse for Black men [42, 43].
More robust linkage efforts following ED and IP are needed, especially for frequent users. Healthcare systems may provide appointments or referrals to primary care or other providers, but some people lack transportation and also need help to remember appointments; communicate by phone, email, or patient portal; navigate and coordinate across systems; pay for care; and overcome distrust, stigma, and other barriers [23, 44]. One successful intervention model, the Transitions Clinic Network, integrates community health workers with lived experience of incarceration to serve as a liaison between patient and provider, provide person-centered case management to address needs of the patient beyond healthcare, such as housing [24]. An RCT of TCN compared to standard primary care found the program reduced overall emergency room visits by 51% [45]. A propensity-matched comparison also found that TCN reduced preventable hospitalizations, shortened hospitalizations, and reduced incarceration days [46] and costs [47].
Adopting the principles used by the Transitions Clinic Network would reduce barriers to community primary care for jail- and prison-involved individuals. One such program, NC Formerly Incarcerated Transition Program (FIT) [48], is operating in the region and their patients and FIT-supported encounters would be among those in this dataset. FIT’s eligibility criteria are recent release from incarceration (prison or jail) and a chronic disease diagnosis (including mental illness and/or substance use disorder), which in this sample would certainly include 6,162 people with one of the priority chronic conditions and many others as well. Considering only those with the six priority conditions, 26.9% never had primary care, which would equate to 1,657 who could use FIT services. But even those who have had primary care also show signs of not being well connected, with 67% (n = 4,128) going most years without a primary care visit, and 52.8% (n = 3,253) never having a year with 2 or more visits, which would be indicated for people with these high-priority chronic diseases. FIT’s capacity is primarily defined by the number of community health workers it can employ to manage a caseload of 30–60 clients each. Even taking care of just the 1,657 with serious chronic disease and no evidence of primary care would require at least 27 community health workers, fewer as people stabilize.
There are also specialized models for people with serious mental illness, such as Assertive Community Treatment (ACT), that could support greater engagement in primary care for those with serious mental illness who are eligible [49]. Collaborative care models wherein patients with mild or moderate psychiatric conditions are managed by a consultation team of care manager and consulting psychiatrist are associated with lower readmission rates to IP and ED, and a lower costs of care overall [50], and may use principles from ACT to engage and redirect frequent users toward other community resources, including outpatient primary care [51]. Initiatives to reduce frequent use of 911, jail, and other city and county resources [52, 53] should include partnerships with primary care providers, along with community paramedics and others.
Greater investment and prioritization to meet the needs of jail-involved frequent ED users with highly complex care needs would benefit patients and providers alike. Healthcare policy makers have argued that improving the U.S. healthcare system requires maximizing the Triple Aim, namely enhancing patient experience, improving population health, and reducing costs, simultaneously [54, 55]. Not only would shifting healthcare use by this patient population from preventable ED/IP to primary care support the Triple Aim, it would also address the additional need articulated in the Quadruple Aim [56]: improving the work life of health care providers, including clinicians and staff. People with preventable conditions or social needs contribute to overcrowding and longer wait times at the ED. There is a subset of jail-involved people who exhibit violent behavior at the ED, whether due to substance or alcohol use disorders, mental health disorders, difficulty communicating, distrust and fear, or frustration by their needs that exceed ED treatment capacity. Healthcare staff who experience violence and/or are not able to give patients what they need to be healthy and stable may experience physical injury, moral injury, and associated stress and burnout. Interventions such as medical-legal partnerships that effectively address the complex needs of specific populations can improve provider wellbeing because there is some way to help their patients. Further, prioritizing needs of CL-involved people, who are disproportionately from minoritized groups, would support the Quintuple Aim that includes racial equity [57].
Study limitations include measurement of disease based on encounters rather than universal screening; people who come for care are more likely to be diagnosed, particularly if they come frequently. In addition, this analysis excludes those diagnosed by healthcare providers outside the system. Because most people encountered the health system by going to the E.D. and only half ever had a primary care visit, there were nevertheless many opportunities to be diagnosed outside of a primary care context and our approach was able to detect many with chronic disease apart from the primary care context. Moreover, the medical records capture almost all primary care providers likely to serve this county population. Behavioral health-related enhanced services (such as ACT teams) and some routine visits are not as well captured. Our measures of primary care may or may not represent access to a usual source of care; individuals with apparently good engagement actually may be visiting different primary care providers and lack a stable long-term connection to care. This dataset does not support a direct comparison between those who are and are not jail-involved. Last, the data are from just one geographic area. However, this is a payer-agnostic dataset that includes uninsured people, and thus trades geographic coverage for comprehensive inclusion of everyone detained at the county detention facility and nearly everyone who would have sought care nearby.
Conclusions
In a single county population of individuals ever booked into the detention center, utilization of primary care was inconsistent and overall low. Individuals too often relied on the Emergency Department for care. To promote primary care utilization among individuals who have the combined challenges of repeated jail involvement and chronic disease, it is crucial to focus on engagement, as much as access. This study demonstrated that among jail-involved individuals with high-priority chronic diseases such as hypertension, obstructive lung disease, and diabetes, having repeated bookings was associated with less ongoing engagement in primary care—not one-time access, but repeated, sustained engagement. This barrier was in the context of low primary care participation for the group as a whole, with under three quarters of those with serious chronic diseases ever seeing a primary care provider and less than half ever having a year with more than one visit. Access to health insurance alone will not resolve the problem; Medicaid expansion should be coupled with specialized, tailored support to promote engagement in primary care. Programs such as Transitions Clinic Network can address barriers to accessing primary care thereby addressing acute and chronic physical health conditions and coordinating or providing mental health and substance abuse care. There should be sufficient support to expand such programs so they can reach every jail-detained individual with chronic disease, along with others who need but do not engage in primary care.
Acknowledgments
This work could not have been completed without the cooperation of Sheriff Clarence Birkhead, Major Elijah Bazemore, and Caroline Andrews of the Durham County Sheriff’s Office; Gudrun Parmer of the Durham County Criminal Justice Resource Center; Roshanna Parker of the Durham County Justice Services Department; and members of Durham County Stepping Up Initiative. Funding to support data collection and analysis was provided by Duke University Bass Connections Program, Rhodes Information Initiative at Duke Data Plus Program, and the Wilson Center for Science & Justice.
References
- 1.
National Academies National Research Council, Institute of Medicine. U.S. Health in International Perspective: Shorter Lives, Poorer Health. Woolf SH, Aron L, editors. Washington, DC: The National Academies Press; 2013.
- 2.
Fair H, Walmsley R. World Prison Population List, 13th edition. 2021.
- 3. Frank JW, Linder JA, Becker WC, Fiellin DA, Wang EA. Increased hospital and emergency department utilization by individuals with recent criminal justice involvement: Results of a national survey. J Gen Intern Med. 2014;29(9):1226–33. Epub 2014/05/13. pmid:24817280; PubMed Central PMCID: PMC4139534.
- 4. Khatri UG, Bilger A, Zeidan A, Meisel Z, Delgado MK, Samuels-Kalow M, et al. "A Catch-22": A Firsthand Account of Barriers and Facilitators to Accessing Health Care During Reentry. Journal of Correctional Health Care. 2022;28(2):109–16. WOS:000766225200001. pmid:35244474
- 5. Lee J, Vlahov D, Freudenberg N. Primary care and health insurance among women released from New York City jails. J Health Care Poor Underserved. 2006;17(1):200–17. pmid:16520527.
- 6. Shavit S, Aminawung JA, Birnbaum N, Greenberg S, Berthold T, Fishman A, et al. Transitions Clinic Network: Challenges And Lessons In Primary Care For People Released From Prison. Health Affairs. 2017;36(6):1006–15. WOS:000404980600007. pmid:28583958
- 7. Wang EA, Redmond N, Himmelfarb CRD, Pettit B, Stern M, Chen J, et al. Cardiovascular Disease in Incarcerated Populations. Journal of the American College of Cardiology. 2017;69(24):2967–76. WOS:000403064700012. pmid:28619198
- 8.
Alexander M, West C. The new Jim Crow: mass incarceration in the age of colorblindness. Revised edition / ed. New York, Jackson, Tennessee: New Press, Perseus, Distribution; 2012.
- 9. Carson EA, Kluckow R. Prisoners in 2022—Statistical Tables. In: Bureau of Justice Statistics Office of Justice Programs, U.S. Department of Justice, editor. 2023.
- 10. Fox BH, Perez N, Cass E, Baglivio MT, Epps N. Trauma changes everything: Examining the relationship between adverse childhood experiences and serious, violent and chronic juvenile offenders. Child Abuse & Neglect. 2015;46:163–73. WOS:000359508100017. pmid:25703485
- 11.
Zeng Z. Jail Inmates in 2022-Statistical Tables. U.S. Department of Justice; Office of Justice Programs; Bureau of Justice Statistics: 2023 December 2023. Report No.: Contract No.: NCJ 307086.
- 12. Massoglia M, Pridemore WA. Incarceration and Health. Annu Rev Sociol. 2015;41:291–310. Epub 2015/08/01. pmid:30197467; PubMed Central PMCID: PMC6124689.
- 13. Cloud DH, Garcia-Grossman IR, Armstrong A, Williams B. Public Health and Prisons: Priorities in the Age of Mass Incarceration. Annual Review of Public Health. 2023;44:407–28. WOS:000958817200022. pmid:36542770
- 14. Sugiea NF, Turneya K. Beyond Incarceration: Criminal Justice Contact and Mental Health. American Sociological Review. 2017;82(4):719–43. WOS:000406179200003.
- 15. Wildeman C, Wang EA. Mass incarceration, public health, and widening inequality in the USA. Lancet. 2017;389(10077):1464–74. WOS:000398345100038. pmid:28402828
- 16.
Maruschak LM, Berzofsky M, Unangst J. Medical problems of state and federal prisoners and jail inmates, 2011–12. In: Statistics BoJ, editor.: US Department of Justice; 2015.
- 17. Puglisi LB, Brinkley-Rubinstein L, Wang EA. COVID-19 in Carceral Systems: A Review. Annual Review of Criminology. 2023;6:399–422. WOS:000920255800018.
- 18. Trotter RT 2nd, Lininger MR, Camplain R, Fofanov VY, Camplain C, Baldwin JA. A Survey of Health Disparities, Social Determinants of Health, and Converging Morbidities in a County Jail: A Cultural-Ecological Assessment of Health Conditions in Jail Populations. Int J Environ Res Public Health. 2018;15(11). Epub 20181108. pmid:30413098; PubMed Central PMCID: PMC6267226.
- 19. Binswanger IA, Krueger PM, Steiner JF. Prevalence of chronic medical conditions among jail and prison inmates in the USA compared with the general population. J Epidemiol Community Health. 2009;63(11):912–9. Epub 20090730. pmid:19648129.
- 20. Dauria EF, Kulkarni P, Clemenzi-Allen A, Brinkley-Rubinstein L, Beckwith CG. Interventions Designed to Improve HIV Continuum of Care Outcomes for Persons with HIV in Contact with the Carceral System in the USA. Curr HIV/AIDS Rep. 2022;19(4):281–91. Epub 20220608. pmid:35674879; PubMed Central PMCID: PMC9175158.
- 21. Freudenberg N, Heller D. A Review of Opportunities to Improve the Health of People Involved in the Criminal Justice System in the United States. Annu Rev Public Health. 2016;37:313–33. Epub 20160118. pmid:26789388.
- 22. Hicks DL, Comartin EB, Kubiak S. Transition Planning from Jail; Treatment Engagement, Continuity of Care, and Rearrest. Community Ment Health J. 2022;58(2):288–99. Epub 20210409. pmid:33835278.
- 23. Jennings L, Branson CF, Maxwell AM, Winkelman TNA, Shlafer RJ. Physicians’ perspectives on continuity of care for patients involved in the criminal justice system: A qualitative study. PLoS One. 2021;16(7):e0254578. Epub 20210714. pmid:34260620; PubMed Central PMCID: PMC8279398.
- 24. Wang EA, Hong CS, Samuels L, Shavit S, Sanders R, Kushel M. Transitions Clinic: Creating a Community-Based Model of Health Care for Recently Released California Prisoners. Public Health Reports. 2010;125(2):171–7. WOS:000274599900004. pmid:20297743
- 25. Redmond N, Aminawung JA, Morse DS, Zaller N, Shavit S, Wang EA. Perceived Discrimination Based on Criminal Record in Healthcare Settings and Self-Reported Health Status among Formerly Incarcerated Individuals. Journal of Urban Health-Bulletin of the New York Academy of Medicine. 2020;97(1):105–11. WOS:000491404100001. pmid:31628588
- 26. Albertson EM, Scannell C, Ashtari N, Barnert E. Eliminating Gaps in Medicaid Coverage During Reentry After Incarceration. Am J Public Health. 2020;110(3):317–21. Epub 2020/01/17. pmid:31944846; PubMed Central PMCID: PMC7002937.
- 27. Goulet MH, Dellazizzo L, Lessard-Deschenes C, Lesage A, Crocker AG, Dumais A. Effectiveness of Forensic Assertive Community Treatment on Forensic and Health Outcomes: A Systematic Review and Meta-Analysis. Criminal Justice and Behavior. 2022;49(6):838–52. WOS:000730392200001.
- 28. Spaulding AC, Drobeniuc A, Frew PM, Lemon TL, Anderson EJ, Cerwonka C, et al. Jail, an unappreciated medical home: Assessing the feasibility of a strengths-based case management intervention to improve the care retention of HIV-infected persons once released from jail. PLoS One. 2018;13(3):e0191643. Epub 2018/03/31. pmid:29601591
- 29. Frank JW, Andrews CM, Green TC, Samuels AM, Trinh TT, Friedmann PD. Emergency department utilization among recently released prisoners: A retrospective cohort study. BMC Emerg Med. 2013;13:16. Epub 2013/11/06. pmid:24188513; PubMed Central PMCID: PMC3818565.
- 30. Easter MM, Schramm-Sapyta NL, Tackett MA, Larsen IG, Tang B, Ralph MA, et al. Reliance on Community Emergency Departments by People Ever Detained in Jail: Retrospective Cross-Sectional Study. J Correct Health Care. 2023;29(4):282–92. Epub 20230619. pmid:37335965.
- 31. Bieler G, Paroz S, Faouzi M, Trueb L, Vaucher P, Althaus F, et al. Social and medical vulnerability factors of emergency department frequent users in a universal health insurance system. Acad Emerg Med. 2012;19(1):63–8. Epub 2012/01/10. pmid:22221292.
- 32. Butler A, Love AD, Young JT, Kinner SA. Frequent Attendance to the Emergency Department after Release from Prison: A Prospective Data Linkage Study. J Behav Health Serv Res. 2020;47(4):544–59. Epub 2019/12/11. pmid:31820327; PubMed Central PMCID: PMC7578130.
- 33. LaCalle E, Rabin E. Frequent users of emergency departments: the myths, the data, and the policy implications. Ann Emerg Med. 2010;56(1):42–8. Epub 2010/03/30. pmid:20346540.
- 34. World Health Organization. ICD-10: International statistical classification of diseases and related health problems: Tenth revision, 2nd ed2004.
- 35. Nemani K, Li C, Olfson M, Blessing EM, Razavian N, Chen J, et al. Association of Psychiatric Disorders With Mortality Among Patients With COVID-19. JAMA Psychiatry. 2021;78(4):380–6. pmid:33502436; PubMed Central PMCID: PMC7841576.
- 36.
Agency for Healthcare Research and Quality. Chronic Condition Indicator Refined for ICD-10-CM, Healthcare Cost and Utilization Project (HCUP). Rockville, MD2024.
- 37.
Agency for Healthcare Research and Quality. AHRQ PQI Technical Documentation, Version v2023. Rockville, MD2023.
- 38.
SAS Institute Inc. SAS 9.4. 2023.
- 39. Percentage of having a wellness visit in past 12 months for adults aged 18 and over, United States, 2019–2022. National Health Interview Survey. Generated interactively: May 28 2024 [Internet]. 2024. Available from: https://wwwn.cdc.gov/NHISDataQueryTool/SHS_adult/index.html.
- 40. Davis KE. Access to Health Care of Adult Men and Women, Ages 18–64, 2012 (MEPS Medical Expenditure Panel Survey,). AHRQ Agency for Healthcare Research and Quality, 2014.
- 41.
Patel PR, Bechmann S. Discharge Planning. 2024. In: StatPearls [Internet]. StatPearls Publishing, LLC. Available from: https://www.ncbi.nlm.nih.gov/books/NBK557819/.
- 42. Mitchell JA, Williams EG, Li Y, Tarraf W. Identifying disparities in patient-centered care experiences between non-Latino white and black men: results from the 2008–2016 Medical Expenditure Panel Survey. BMC Health Serv Res. 2020;20(1):495. Epub 20200603. pmid:32493469; PubMed Central PMCID: PMC7268709.
- 43. Hedden BJ, Comartin E, Hambrick N, Kubiak S. Racial Disparities in Access to and Utilization of Jail- and Community-Based Mental Health Treatment in 8 US Midwestern Jails in 2017. Am J Public Health. 2021;111(2):277–85. Epub 20201222. pmid:33351663; PubMed Central PMCID: PMC7811091.
- 44. Fahmy N, Kouyoumdjian FG, Berkowitz J, Fahmy S, Neves CM, Hwang SW, et al. Access to Primary Care for Persons Recently Released From Prison. Ann Fam Med. 2018;16(6):549–51. pmid:30420371; PubMed Central PMCID: PMC6231934.
- 45. Wang EA, Hong CS, Shavit S, Sanders R, Kessell E, Kushel MB. Engaging individuals recently released from prison into primary care: a randomized trial. Am J Public Health. 2012;102(9):e22–9. Epub 20120719. pmid:22813476; PubMed Central PMCID: PMC3482056.
- 46. Wang EA, Lin HJ, Aminawung JA, Busch SH, Gallagher C, Maurer K, et al. Propensity-matched study of enhanced primary care on contact with the criminal justice system among individuals recently released from prison to New Haven. BMJ Open. 2019;9(5). WOS:000471192800328. pmid:31048315
- 47. Harvey TD, Busch SH, Lin HJ, Aminawung JA, Puglisi L, Shavit S, et al. Cost savings of a primary care program for individuals recently released from prison: a propensity-matched study. BMC Health Services Research. 2022;22(1). WOS:000789831900001. pmid:35501855
- 48. Clark EE, Ashkin E. Spotlight on the safety net: connecting people released from incarceration to essential health services: the North Carolina Formerly Incarcerated Transition (FIT) Program. North Carolina Medical Journal. 2019;80(6):377–9. pmid:31685578
- 49. Vanderlip ER, Henwood BF, Hrouda DR, Meyer PS, Monroe-DeVita M, Studer LM, et al. Systematic literature review of general health care interventions within programs of assertive community treatment. Psychiatric Services. 2017;68(3):218–24. pmid:27903142
- 50. AIMS Center University of Washington. Evidence base for Collaborative Care (CoCM) 2024. Available from: https://aims.uw.edu/evidence-base-for-cocm/.
- 51. Cunningham NT, Sendak M, Bratcher M, Gamble J, Dillard J, Bowlby L, et al. Homebase: successfully using the principles of assertive community treatment to decrease emergency department and hospital utilization among super-utilizers. Mental illness. 2016;19:76.
- 52. National Association of Counties (NACo). Familiar Faces Initiative: Improving outcomes through coordinated health and justice systems 2022 [cited 2022 July 13]. Available from: https://familiarfaces.naco.org/.
- 53.
Corporation for Supportive Housing (CSH). FUSE [Frequent User System Engagement] 2022. Available from: https://www.csh.org/fuse/.
- 54. Berwick DM, Beckman AL, Gondi S. The Triple Aim Applied to Correctional Health Systems. JAMA. 2021;325(10):935–6. pmid:33555303.
- 55. Berwick DM, Nolan TW, Whittington J. The triple aim: care, health, and cost. Health Aff (Millwood). 2008;27(3):759–69. pmid:18474969.
- 56. Bodenheimer T, Sinsky C. From triple to quadruple aim: care of the patient requires care of the provider. Ann Fam Med. 2014;12(6):573–6. pmid:25384822; PubMed Central PMCID: PMC4226781.
- 57. Nundy S, Cooper LA, Mate KS. The Quintuple Aim for Health Care Improvement: A New Imperative to Advance Health Equity. JAMA. 2022;327(6):521–2. pmid:35061006.