Comorbidities, Health Care Use, and Contact with Health Care Transition Services in Older Veterans After Incarceration

Abstract

Background

The Health Care for Reentry Veterans (HCRV) program was established to support community reintegration for veterans after incarceration. Yet, it is unclear how those with and without HCRV contact differ. We sought to evaluate differences in medical and psychiatric conditions and health care utilization among mid-to late-life reentry veterans who did and did not receive HCRV outreach.

Methods

Study participants were veterans aged ≥50 years who qualified for Medicare fee-for-service, had experienced incarceration for ≥1 year, and were released from incarceration between October 1, 2006, and September 30, 2018 (N=9,733). Using VA and Medicare claims data, we compared prevalence of medical and psychiatric diagnoses, and use of emergency, inpatient, and outpatient medical and mental health services up to 12 months after release between those with and without HCRV contact.

Results

Veterans with HCRV contact (35.5%) had significantly higher rates of psychiatric conditions and medical conditions related to substance use (e.g., liver disease) compared to veterans without HCRV contact. Average time between release and first health care service use was significantly lower for HCRV veterans (36.5 ±SD 59.5 days) versus non-HCRV veterans (58.9 ±SD 77.5 days) and HCRV veterans were more likely to utilize the emergency department, inpatient and outpatient mental health services, and inpatient medical services.

Conclusion

HCRV reaches older reentry veterans with a large burden of mental health and substance use disorders. However, levels of multimorbidity were high among all older reentry veterans, pointing to a need to develop specialized geriatric models of care for this reentry population.

Introduction

Approximately 8% of those incarcerated in U.S. jails and prisons annually are veterans.¹ Upon reentering the community after incarceration, justice-involved veterans are at considerably elevated risk of homelessness², drug overdose³, and suicide^{4, 5} compared with veterans without justice involvement. The Veterans Justice Programs (VJP) were established by the Veterans Administration (VA) in 2007–2009 to facilitate community integration by linking justice-involved veterans to health care and support services. These programs work to identify veterans in the legal system, contact them through outreach, and connect them to VA services. The Veterans Justice Outreach (VJO) program provides outreach at earlier stages of legal system involvement, focusing on jail diversion and case management for veterans being held in local courts and jails. In contrast, the Health Care for Reentry Veterans (HCRV) program coordinates with state and federal prisons to identify incarcerated veterans preparing for community reentry. These transition programs are unique as they are nationwide in scope and partner with courts, jails, and prisons to support veterans in their local communities and provide continuity of care.^4,6,7

While those participating in VJP have been previously described,^6,7 it is unclear if or how these individuals differ from those who have also transitioned from incarceration to the community but without VJP contact. Evaluating differences between reentry veterans with and without VJP contact can help to determine if VJPs are reaching those who are most in need of health care-related reentry assistance (e.g., those with serious mental illness) or those at especially high risk of experiencing homelessness (e.g., those with histories of homelessness). These comparisons can also inform differences in health care utilization following incarceration and can characterize the spectrum of veterans who return to the community after incarceration.

By leveraging a pre-existing cohort of reentry veterans aged 50 and older, we had the unique opportunity to evaluate differences between those who did and did not receive VJP outreach. Focusing on reentry veterans in mid-to-late-life is important because most veterans reentering the community after incarceration are in this age group, reflecting wider trends in the rapidly aging U.S. prison population. In fact, the percentage of adults aged 50 and older sentenced in U.S. prisons more than doubled from 10% in 2012 to 22% in 2020.^{8, 9} Moreover, this population faces additional challenges, including medical comorbidities that are associated with older age but occurring at relatively young ages, in a process called “accelerated aging”.¹⁰ These medical conditions may be exacerbated by long periods of incarceration and contribute to functional impairment, which is often coupled with limited social support upon reentry.¹¹ Thus, connecting this population to health care and social services during the high-risk transitional period of reentry is critical.

In a national sample of reentry veterans aged 50 and older, we sought to compare those who did and did not receive outreach from the HCRV program. We focused on the HCRV program specifically because all study participants had been incarcerated and thus, in theory, were eligible for HCRV outreach. We determined if there were differences in sociodemographic characteristics and health-related conditions. We also determined if there were differences in time to first contact with health services and in utilization of medical and mental health care up to one year after release.

Methods

Data and Participants

The study used data from the Centers for Medicare and Medicaid Services (CMS) and the VA National Patient Care Database (NPCD), which includes all VA inpatient and outpatient services. Veterans in the study qualified for Medicare fee-for-service and utilized VA or Medicare-reimbursed health care services. In addition, study participants were age ≥50 years, had been incarcerated for at least 1 year, and were released from incarceration between October 1, 2006, and September 30, 2018 (N=9,733). Incarceration history files through Medicare, which included start and end dates of most recent incarcerations (Medicare turns off when incarcerated and back on when released), were used to identify veterans who had been incarcerated. The ≥1 year length of incarceration was intended to create a group of veterans who were likely to need HCRV services post-release due to loss of community resources such as housing, employment, and health care that may have occurred while they were incarcerated, and who were more likely to be in prison rather than jail.^12,13

Measures

Contact with the Health Care for Reentry Veterans (HCRV) program

Contact with the HCRV program was defined as contact with HCRV specialists during their most recent incarceration or within 6 months after their most recent incarceration release date and was ascertained through stop code 591 in the NPCD or through the VA Homeless Operations Management System (HOMES) database.

Diagnoses and Health Care Utilization

Diagnoses were selected based on their relevance to an aging population and based on prior work^14–16 with additional emphasis on conditions prevalent among incarcerated populations.^11,17 Domains included medical, which was comprised of cardiovascular, musculoskeletal, infectious disease, neurologic and other (i.e., chronic lung disease, cancer, and liver disease), psychiatric, and substance use disorders. NPCD and CMS databases were used to determine diagnoses, identified via ICD-9 and ICD-10 codes, that were linked to health care encounters up to 1 year after the incarceration release date (see Supplemental Material for details). These databases were also used to identify veterans who experienced medical and mental health care visits (i.e., medical and mental health emergency, inpatient, and outpatient encounters) up to 1 year after the incarceration release date and to determine time (in days) to first contact with VA or Medicare health services after release.

Other Measures

Sociodemographics, including age, sex, race/ethnicity (non-Hispanic White, non-Hispanic Black, and Hispanic/Other), U.S. region after release from incarceration, income, and education, were determined using NPCD and CMS databases. Income and education were assessed at the census tract-level, based on zip code of veteran’s residence. History of homelessness prior to the incarceration was indicated in the NCPD data files as homelessness indicator=1, clinic stop codes 28, 37, 522, 528, 529, 530, 590, or ICD-9 code V60.0 and ICD-10 code Z59.0 (“lack of housing”). These same ICD-9/10 codes were also used to identify homelessness in the CMS data files. The NPCD was used to identify those who had a service-connected disability, a disability attributed to a medical or psychiatric illness or injury that occurred during or was exacerbated by military service.

Statistical Analyses

Chi-square tests and t-tests were used to compare veterans with and without HCRV contact in terms of sociodemographics, proportion of medical, psychiatric, and substance use disorder diagnoses, and time to first contact with VA health services and health care utilization. Differences between those with and without HCRV contact were compared using t-tests for continuous variables and χ2 tests for categorical variables.

The study was approved by the San Francisco Veterans Affairs Health Care System and the University of California, San Francisco Institutional Review Boards.

Results

Approximately one-third of our sample had contact with the HCRV program. We identified 3,451 HCRV and 6,282 non-HCRV reentry veterans during the study period. Table 1 presents characteristics of HCRV and non-HCRV veterans. As compared with non-HCRV veterans, veterans in the HCRV program were younger (49.4% were 50–64 years vs. 45.4% non-HCRV), were less likely to be female (1.2% vs. 1.8%), and were more likely to identify as non-Hispanic Black (27.8% vs. 24.1%). Nearly half of HCRV veterans (48.7%) were located in the Midwest or Southeast United States compared to 33.5% of non-HCRV veterans. HCRV veterans also had higher rates of past homelessness (6.1% vs. 4.8%), service-connected disability (42.5% vs. 31.7%), and prior contact with VJP (14.6% vs. 4.4%), and they had longer average lengths of incarceration (4.1±4.0years vs. 3.7±3.5 years), with more than 24% of the HCRV veterans incarcerated for at least 5 years as compared with 20% of the non-HCRV veterans. Each of these comparisons was significant at the p<0.05 level. HCRV veterans were more likely to be low-income but did not differ in education level compared to non-HCRV veterans.

Table 1.

Characteristics of Reentry Veterans

Characteristic	Overall Reentry (N=9,733)	No HCRV (N=6,282)	HCRV (N=3,451)	t value or χ2	p value
Sociodemographics, n (%) unless otherwise specified
Age at release in yrs, mean (SD)	64.8 (8.4)	65.0 (8.6)	64.4 (8.1)	3.33	<.001
Age at release				23.15	<.001
50–64	4,555 (46.8)	2,851 (45.4)	1,704 (49.4)
65–74	3,949 (40.6)	2,573 (41.0)	1,376 (39.9)
75+	1,229 (12.6)	858 (13.7)	371 (10.8)
Female	151 (1.6)	110 (1.8)	41 (1.2)	4.62	.03
Race				16.36	<.001
Non-Hispanic White	6,847 (70.4)	4,491 (71.5)	2,356 (68.3)
Non-Hispanic Black	2,473 (25.4)	1514 (24.1)	959 (27.8)
Hispanic/Others	413 (4.2)	277 (4.4)	136 (3.9)
Median income tertile #				9.41	.02
Low tertile (<$37,734)	3,108 (31.9)	1,971 (31.4)	1,137 (33.0)
Middle tertile	3,095 (31.8)	1,965 (31.3)	1,130 (32.7)
High tertile (>$49,388)	3,138 (32.2)	2,079 (33.1)	1,059 (30.7)
Education: live in area >25% college	2,833 (29.1)	1,810 (28.8)	1,023 (29.6)	3.63	.16
Region *				538.60	<.001
West	1,282 (13.2)	785 (12.5)	497 (14.4)
Southwest	1,307 (13.4)	838 (13.3)	469 (13.6)
Midwest	1,499 (15.4)	746 (11.9)	753 (21.8)
Northeast	854 (8.8)	492 (7.8)	362 (10.5)
Southeast	2,283 (23.5)	1,357 (21.6)	926 (26.8)
History of homelessness	516 (5.3)	304 (4.8)	212 (6.1)	7.54	.01
History of service-related disability	3,460 (35.6)	1,994 (31.7)	1,466 (42.5)	112.11	<.001
Duration of most recent incarceration, years, mean (SD)	3.83 (3.7)	3.67 (3.5)	4.12 (4.0)	-5.74	<.001
Duration of most recent incarceration				35.79	<.001
1 year - ≤2 years	3,665 (37.7)	2,484 (39.5)	1,181 (34.2)
>2 years - ≤5 years	3,971 (40.8)	2,537 (40.4)	1,434 (41.6)
>5 years - ≤10 years	1,552 (16.0)	940 (15.0)	612 (17.7)
>10 years	545 (5.6)	321 (5.1)	224 (6.5)
CCI, mean (SD)	2.32 (2.7)	2.35 (2.8)	2.26 (2.6)	1.55	.12
CCI category				2.65	.85
0	2,745 (28.2)	1,774 (28.2)	971 (28.1)
1	2,234 (23.0)	1,415 (22.5)	819 (23.7)
2	1,475 (15.2)	949 (15.1)	526 (15.2)
3	1,065 (10.9)	695 (11.1)	370 (10.7)
4	625 (6.4)	405 (6.5)	220 (6.4)
5	449 (4.6)	296 (4.7)	153 (4.4)
6	1,140 (11.7)	748 (11.9)	392 (11.4)

Abbreviations: HCRV- Healthcare for Reentry Veterans; SD-standard deviation; VJO-Veterans Justice Outreach; CCI-Charlson Comorbidity Index

^#Based on median income from 2013 census

^*For this variable, 25.8% values are missing

Figure 1 presents the distribution of medical diagnoses in our sample of HCRV and non-HCRV veterans and Figure 2 presents psychiatric and substance use disorders. Comparing medical diagnoses between HCRV and non-HCRV veterans, HCRV veterans had a significantly higher proportion of liver disease (10.3% vs. 8.2%) and hepatitis C (10.8% vs. 8.1%), a significantly lower proportion of stroke (9.9% vs. 11.4%) and chronic lung disease (18.2% vs. 22.0%), and a higher proportion of mild cognitive impairment (1.4% vs. 0.9%) but lower proportion of dementia (3.7% vs. 5.1%), with no significant differences among the other medical conditions. HCRV veterans had higher prevalence of depression (35.9% vs. 26.5%), bipolar disorder (11.3% vs. 9.0%), posttraumatic stress disorder (PTSD) (22.3% vs. 13.7%), and generalized anxiety disorder (GAD) (5.2% vs. 3.4%). HCRV veterans also had higher prevalence of all substance use disorders, including alcohol (26.3% vs. 17.7%), opioid (7.9% vs. 4.6%), amphetamine (4.2% vs. 1.9%), cannabis (8.6% vs. 4.0%), cocaine (13.7% vs. 7.2%), and tobacco (33.4% vs. 29.0%) compared to non-HCRV veterans. The average Charlson Comorbidity Index was not significantly different between the two groups.

Figure 1. Prevalence of Diagnoses of Medical Conditions in Older Previously Incarcerated Veterans With and Without HCRV Contact.

Abbreviations: MI- myocardial infarction; CHF-congestive heart failure; HIV/AIDS-human immunodeficiency virus/acquired immunodeficiency syndrome; STI-sexually transmitted infection; TBI-traumatic brain injury; MCI-mild cognitive impairment; MSK-musculoskeletal

Figure 2: Prevalence of Diagnoses of Psychiatric and Substance Use Conditions in Older Previously Incarcerated Veterans With and Without HCRV Contact.

Abbreviations: PTSD-posttraumatic stress disorder; GAD-generalized anxiety disorder; SCZ-schizophrenia, schizoaffective disorder and other psychotic disorders; AUD-alcohol use disorder; OUD-opioid use disorder; UD-use disorder; SUD-substance use disorder

Average time from release from incarceration to first VA health care service use was significantly lower for HCRV veterans (36.5±59.5 days) versus non-HCRV veterans (58.9±77.5 days). In the 12 months following release from incarceration, a significantly higher proportion of HCRV veterans versus non-HCRV veterans used emergency mental health (2.1% vs. 0.8%), inpatient mental health (10.1% vs. 4.5%), outpatient mental health (66.2% vs. 35.8%), emergency medical (53.1% vs. 48.5%), and inpatient medical services (37.5% vs. 35.3%), with all comparisons significant at the p<0.05 level. Utilization of outpatient medical health services did not differ between the two groups (Figure 3).

Figure 3: Mental Health and Medical Care Utilization among Older Previously Incarcerated Veterans With and Without HCRV Contact.

Abbreviations: MH-mental health; ED-emergency department

Discussion

This is the first national study to compare the characteristics and health services use of older reentry veterans who did versus did not have HCRV contact. We found that those with HCRV contact had a higher burden of psychiatric disorders, substance use disorders (SUDs), and medical diagnoses commonly linked to SUDs, and were more likely to have histories of homelessness. We also found that those with HCRV contact had fewer days between release date and first date of health care services use and had higher mental health care and acute medical care utilization rates. Together these findings demonstrate that while all older reentry veterans have substantial medical morbidity, HCRV is reaching those with greater mental health and substance use needs.

As compared with older reentry veterans with no HCRV contact, those with HCRV contact had substantially higher rates of most psychiatric disorders. Notably, the rates of depression and PTSD, which are strong risk factors for suicide in both veterans and civilians,^18–21 were nearly 10 percentage points higher in the group that had HCRV contact. Similarly, rates of SUDs were also 10 percentage points higher in those with HCRV contact; more than 35% of those with HCRV contact had a SUD. SUDs among persons age ≥50 are among the fastest growing health problems in the U.S. Between 2000 and 2020, those age ≥50 with SUDs increased from 1.7 million to 5.7 million.²² Moreover, justice involvement has been found to be associated with increased risk of overdose among veterans³ and substance use among justice-involved veterans puts them at greater risk of additional negative outcomes such as homelessness and reincarceration.²³ Consequently, reaching many older reentry veterans with SUDs and linking them with treatment at the VA or in their local communities is an important role for HCRV and other reentry programs.

Given our finding that a high proportion of older reentry veterans with HCRV contact had SUDs, it is perhaps not surprising that these veterans also had higher proportions of substance use-related comorbidities including liver disease and hepatitis C. In contrast, those with HCRV contact had lower proportions of stroke, chronic lung disease, and dementia as compared to those without HCRV contact. Further research is needed to determine if those veterans who are reentering the community with conditions such as stroke, chronic lung disease, and dementia have high rates of disability and if they are likely to have discharge plans that include home care or transfers to skilled nursing facilities. Thus, they may require a different set of transition services than those offered by HCRV.

Another key goal of HCRV is to reduce rates of homelessness among reentry veterans. We found that a significantly greater proportion of the older reentry veterans who received HCRV outreach had a history of homelessness, again indicating that HCRV is making targeted connections. Interestingly, the proportion of those with a service-related disability was also high among those with HCRV contact, potentially mirroring the high rates of psychiatric disorders. It is also possible that when trying to reinstate their VA benefits at the time of discharge, persons with service-related disability are made aware of or are made apparent to the HCRV program and so they may be more likely to have contact with HCRV.

Prior studies have been unable to determine what percent of veterans would connect with the health care system without HCRV contact.^6,7 By linking Medicare and VA data, we were able to evaluate differences in health care utilization among older reentry veterans between those with and without HCRV contact. We found that the HCRV group accessed health care services an average of 22 days earlier than the non-HCRV group. This shorter time to health care utilization and increased engagement with health care services in the HCRV group suggests that HCRV services may facilitate connecting reentry veterans with the health care system. In addition, compared to those without HCRV contact, a higher proportion of those with HCRV contact utilized all types of mental health care (ED, inpatient, outpatient) and acute medical care (ED, inpatient). Access to health care is an important protective factor that promotes successful reintegration into the community for reentry adults.²⁴ Given that reentry to the community after incarceration is considered a high-risk transition time, the three-week shorter presentation to care as well as greater health care utilization could be meaningful for multiple health and safety outcomes. Future research will be needed to assess how health care access and the time between release and accessing health care services affects adverse outcomes such as suicide attempts, overdose, and reincarceration.

Our findings also highlighted that nearly two-thirds of our sample did not have HCRV contact and that considerable proportions of these individuals had high-risk factors such as psychiatric disorders, SUDs, and homelessness. Therefore, there may be opportunity for HCRV to reach more veterans who may benefit from these services. There are currently 44 HCRV specialists working nationwide and serving 1,061 state and federal prisons.²⁴ Additional funding for HCRV staff may be indicated to expand outreach to more reentry veterans. Given the regional variation in how HCRV operates and how many specialists are available, further work is needed to understand best practices among HCRV specialists that could be tailored for a range of settings nationwide.

This study has several limitations. For inclusion, veterans had to be eligible for both Medicare and VA benefits, so the cohort does not include the full population of reentry veterans ages 50–64 years. However, disability due to serious mental illness can be basis for Medicare eligibility, and individuals released from incarceration are more likely to be Medicare-eligible owing to disability than persons in the general population.²⁵ Character of discharge from the military (such as dishonorable discharge) was not available, which could affect eligibility for VA health care benefits.

Whereas a strength of this study was the ability to quantify time incarcerated, we were unable to distinguish between jail and prison incarceration. Because prisons are long-term institutions that typically house incarcerated persons who are convicted of a crime and sentenced, the 1-year minimum period of incarceration likely selected for those in prisons and who had greater need of reentry services due to extended time removed from the community. However, jail sentence length can vary substantially by state, so this approach does not ensure exclusion of a jail population. Jail vs. prison information was not available in VA and Medicare claims data and will be an important distinction to make in future studies. Furthermore, the reason for incarceration was not available. In addition, information on probation and parole, which could involve mandated mental health and substance use treatment, is important to consider but was not available for this study. This is an observational study and so cannot be used to make claims about causality; it is possible that reentry veterans who are more likely to seek out health care are also more likely to engage with HCRV. It is unknown whether reentry veterans who did not access any health care services post-release did not need services or could not access services. Finally, generalizability may be limited for women and non-veterans who reentered the community after incarceration. Still, those in our sample who are eligible for Medicare fee-for-service based on age are likely representative of those in the general population who are aged 65 and older and reenter the community after incarceration. This study offers a more complete picture of reentry veterans by including those who do not have HCRV contact.²⁶

This study identifies veterans who enter the health care system after incarceration, but additional research is needed to understand whether reentry veterans are receiving appropriate services to support their quality of life and long-term health outcomes. These findings point to the need to develop better geriatric models of care for this specialized patient population.

Conclusion

Accessing health care services may be especially important for veterans in later age who are reentering the community after incarceration. By comparing those with and without HCRV contact, this study is a first step towards understanding the impact of the HCRV program in this vulnerable population. While our findings suggest that HCRV is appropriately identifying older reentry veterans who may be at especially high risk for negative health-related outcomes, older reentry veterans without HCRV contact also carry a significant burden of medical and psychiatric illness and may benefit from outreach.

Key Points:

• Among older veterans reentering the community after incarceration, those who had contact with the Health Care for Reentry Veterans (HCRV) program had a higher burden of psychiatric disorders and medical conditions commonly linked to substance use disorders as compared to older reentry veterans with no HCRV contact.

• Older reentry veterans with HCRV contact had fewer days between release date and first date of health care services use and had higher health care utilization rates.

• Together these findings reflect the clinical and psychosocial needs of older reentry veterans that may inform specialized geriatric models of care for the aging prison population.

Why does this paper matter?

Our findings indicate that the Health Care for Reentry Veterans program is reaching older reentry veterans with high prevalence of psychiatric and substance use disorders who have high utilization rates of acute medical and mental health services in the 12 months post-release.