Links of Previous Incarceration With Geriatric Syndromes and Chronic Health Conditions Among Older Adults in the United States

Abstract

Background

This study investigated the association between previous incarceration and various geriatric and chronic health conditions among adults 50 and older in the United States.

Methods

Data came from the National Longitudinal Study of Adolescent to Adult Health—Parent Study (AHPS) collected in 2015–2017, including 2 007 individuals who participated in the parent study (Parent Sample) and 976 individuals who participated in the spouse/partner study (Spouse/Partner Sample). Multiple logistic regression was used to investigate the relationship between previous incarceration and geriatric syndromes (dementia, difficulty walking, difficulty seeing, difficulty with activities of daily living) and chronic health conditions (self-reported poor/fair health, diagnosis of cancer, hypertension, diabetes, heart disease, stroke, chronic lung disease, depression, and alcohol use [4 or more drinks per week]).

Results

In adjusted analyses, respondents with previous incarceration in the AHPS had significantly higher odds of reporting difficulty walking, activities of daily living difficulty, cancer diagnosis, depression diagnosis, and chronic lung disease (adjusted odds ratios [aORs] = 2.21–2.95). Respondents in the AHPS spouse/partner study reported higher odds of difficulty seeing, cancer, depression, chronic lung disease, and heavy alcohol use (aORs = 1.02–2.15).

Conclusions

Previous incarceration may have an adverse impact on healthy aging. Findings highlight the importance of addressing the enduring health impacts of incarceration, particularly as individual transition into older adulthood.

With nearly 2 million individuals incarcerated on any given day, the United States has the highest incarceration rate of any developed democracy (1). Mass incarceration in the United States is also characterized by stark racial disparities. Disproportionately high rates of incarceration are faced by minoritized populations—particularly Black, Latinx, and Native American persons—a phenomenon rooted in deep systemic inequities and structural racism (1). Notably, the dramatic growth in the size and scope of incarceration during the past 5 decades (1), combined with the use of long prison sentences, has shifted the age structure of prisons over time by increasing the number of older currently and formerly incarcerated persons (2–5).

As of 2021, there were approximately 300 000 persons age 50 or older incarcerated in prisons and jails in the United States (6,7), and estimates suggest that at least 1 in 15 community-dwelling adults over age 50 have been previously incarcerated (3). Despite widespread incarceration exposure among older adults, limited research and data have documented the health profiles of previously incarcerated older adults. This is an important gap in knowledge, considering that incarceration is a social determinant of health with profound repercussions for physical and mental health across the lifespan (8–10). In particular, incarceration is postulated to undermine healthy aging, primarily through the intertwined stressors and stigma of incarceration itself and the subsequent challenges of community reintegration, compounded by inadequate access to quality healthcare and suboptimal health behaviors (3,5,9). Considering the increase in geriatric and chronic health conditions during the transition to older adulthood (11) and evidence that incarceration is a catalyst for worsening health for older adults (3,5,9), the large number of formerly incarcerated older adults makes the potential public health implications of incarceration for healthy aging ever more pressing, emphasizing the critical need for further research (4,12,13).

Existing studies on incarceration and health among older adults have primarily focused on the health profiles of currently incarcerated persons, finding that compared to the general population, older incarcerated adults exhibit higher rates of chronic disease, mental health problems, and challenges with activities of daily living (ADLs) (14–17). Less research has investigated differences in geriatric and chronic health conditions between persons with and without a history of incarceration in the general population. Due to the lack of data on this topic, most knowledge comes from a single, albeit robust, data source—the Health and Retirement Study (HRS) (3,5,17). Of particular note, using national data on community-dwelling adults from the 2012 and 2014 HRS, Garcia-Grossman et al. found that a history of incarceration was associated with an elevated risk of impairment of ADLs, hearing impairment, and certain chronic diseases, including chronic lung disease, mental health conditions, and heavy alcohol use. However, previous incarceration was not associated with other chronic health conditions, including diabetes or cardiovascular outcomes (3).

To expand the evidence base for this important but understudied issue, we aimed to replicate the work of Garcia-Grossman et al. (3) by leveraging a distinct source of national data on incarceration and health among older adults from the National Longitudinal Study of Adolescent to Adult Health—Parent Study (AHPS). Specifically, our core study aim is to assess whether chronic and geriatric health conditions differ among adults aged 50 and older with and without a history of incarceration.

Data

Data are from the AHPS, a national study that gathered social, behavioral, and health data in 2015–2017 from a probability sample of the parents of children selected to participate in the prospective cohort study of the National Longitudinal Study of Adolescent to Adult Health (Add Health), who were originally interviewed in 1994–1995. Parents eligible for participation in this study were the biological parents, adoptive parents, or stepparents of an Add Health respondent in 1994–1995; not deceased or incarcerated at the time of AHPS sampling; and had at least 1 Add Health child who was not deceased at the time of AHPS sampling (18). Because most parent respondents were mothers, the sample is more than 96% female. In addition, the AHPS also gathered separate data on the current cohabiting spouse or partner of AHPS parents (Spouse/Partner sample). Approximately 95% of the spouses/partners were male. In both samples, the focal respondents (either the parent or spouse/partner) reported their own history of incarceration as well as a variety of health conditions. To focus on older adults, we restricted both samples to those aged 50 and older, which removed 6 cases from the parent data (n = 2 007) and 12 cases from the spouse/partner data (n = 976).

Dependent Variables

The dependent variables selected to mirror those in prior research using the HRS by Garcia-Grossman et al. (3) are categorized as geriatric syndromes and chronic health conditions based on self-reported physical health outcomes and health conditions any healthcare provider diagnosed (3). Supplementary Appendix A provides details on the measurement of each variable.

Independent Variable

Previous incarceration is measured from the question, “Have you ever been incarcerated, that is, spent time in a jail, prison, juvenile detention center or other correctional facility?” (yes or no).

Control Variables

We drew from available measures in the AHPS to mirror the control variables in prior research (6), including respondent age in years, sex (female or male), race/ethnicity (non-Hispanic White, non-Hispanic Black, Hispanic, or other race), highest educational attainment (less than high school, high school graduate, college graduate), whether a respondent received public assistance benefits in the prior 12 months including Unemployment Compensation, Workers’ Compensation, veterans benefits or cash assistance from a state or county welfare program (yes or no)¸ usual source of care when a respondent is sick or needs health care (private doctor, emergency department, none, or other), if a respondent had previously been abused by a parent (yes or no), whether a respondent had an overnight stay in the hospital in the past 12 months (yes or no), health insurance status (uninsured, Medicare, Medicaid, private, other, or multiple sources), and the year of data collection (2015, 2016, or 2017).

Analytic Approach

We assessed the association between previous incarceration and geriatric syndrome and chronic health condition outcome variables using multiple logistic regression. Following the approach of Garcia-Grossman et al. (3), we first estimated the bivariate association (Model 1) using 2-tailed t-tests. Next, we perform a series of multiple logistic regression analyses, first including covariates for age and sex (Model 2), and finally including a fully adjusted model with all covariates (Model 3).

Missing data were addressed using multiple imputation with chained equations using 20 multiply imputed data sets, resulting in an analytic sample of 2 007 respondents aged 50 and older in the parent data and 976 respondents aged 50 and older in the spouse/partner data. Patterns of missing data across both samples are reported in Supplementary Appendix B. Analyses were performed in Stata v. 17.0 (StataCorp LLC, College Station, TX); all analyses include robust standard errors. Survey weights are included in the parent sample because parents were drawn from the same nationally representative sampling frame as the initial Add Health study. However, because the spouse/partner data were not part of this nationally representative sampling frame, there are no survey weights for the spouse/partner sample. The use of the AHPS for this study received approval from the University of Texas Health Science Center at Houston institutional review board.

Results

Table 1 summarizes the descriptive statistics of the parent and spouse/partner data stratified by incarceration history. In the largely female parent study, 4.2% reported previous incarceration, whereas in the mostly male spouse/partner study, 15.4% reported previous incarceration.

Table 1.

Summary Statistics of Analytic Sample Stratified by Incarceration

	Parent Sample		Parent’s Spouse/Partner Sample
Variable	Never Incarcerated (n = 1 923)	Previous Incarceration (n = 84)	Never Incarcerated (n = 826)	Previous Incarceration (n = 150)
Age (mean [SD])	62.5 [5.6]	59.7 [5.7]	64.7 [5.9]	63.0 [6.3]
Sex
Female	96.7%	87.4%	4.9%	2.1%
Male	3.3%	12.6%	95.1%	97.9%
Race/ethnicity
Non-Hispanic White	69.8%	69.4%	78.5%	74.1%
Non-Hispanic Black	12.6%	14.6%	6.0%	10.8%
Hispanic	10.4%	7.1%	8.5%	6.9%
Non-Hispanic other race	7.2%	8.8%	7.0%	8.2%
Educational attainment
Less than high school	10.4%	12.3%	10.4%	16.2%
High school graduate	63.4%	69.5%	55.9%	71.0%
College graduate	26.2%	18.2%	33.7%	12.8%
Public assistance recipient
No	95.9%	91.5%	91.6%	86.7%
Yes	4.1%	8.5%	8.4%	13.3%
Usual source of care
Private doctor	69.5%	65.0%	66.7%	58.8%
Emergency department	4.0%	7.1%	3.6%	6.0%
None or other	26.5%	27.8%	29.7%	35.2%
Abused by parent
No	86.2%	79.6%	88.2%	79.5%
Yes	13.8%	20.4%	11.8%	20.5%
Overnight hospital stay
No	85.9%	73.4%	84.3%	77.6%
Yes	14.1%	26.6%	15.7%	22.4%
Health insurance
Uninsured	4.6%	10.4%	2.4%	12.3%
Medicare	15.3%	12.6%	16.0%	4.8%
Medicaid	5.1%	10.5%	1.3%	39.8%
Private	43.8%	36.0%	40.0%	7.3%
Other	6.0%	6.2%	4.4%	27.5%
>1	25.2%	24.3%	35.9%	12.3%
Year
2015	24.4%	29.4%	24.1%	15.4%
2016	39.2%	41.8%	41.9%	50.8%
2017	36.4%	28.7%	34.1%	33.8%

Notes: Summary statistics are weighted in the parent sample. However, because the partner/spouse sample was not part of the original Add Health study sampling frame, no survey weights are available for these data.

Table 2, Panel A, shows the results assessing geriatric syndromes and chronic health conditions for the parent data. Several bivariate differences emerged in Model 1, as those with previous incarceration report a higher prevalence of difficulty walking (18.5% vs 6.9%, p ≤ .001), ADL difficulty (48.2% vs 23.1%, p ≤ .001), poor/fair self-rated health (46.3% vs 28.1%, p = .001) cancer (22.8% vs 12.1%, p = .032), heart disease (22.2% vs 13.2%, p = .043), stroke (11.0% vs 3.6%, p = .012), depression (45.3% vs 23.7%, p = .001), and chronic lung disease (37.6% vs 17.5% p ≤ .001). In the fully adjusted model (Model 3), findings show that after adjusting for covariates, previous incarceration is associated with significantly higher odds of reporting difficulty walking (adjusted odds ratio [aOR] = 2.391, 95% confidence interval [CI] = 1.132, 5.051), ADL difficulty (aOR = 2.952, 95% CI = 1.1774, 4.913), cancer diagnosis (aOR = 2.206, 95% CI = 1.205, 4.040), depression diagnosis (aOR = 2.432, 95% CI = 1.416, 4.179), and chronic lung disease (aOR = 2.417, 95% CI = 1.415, 4.129).

Table 2.

Multiple Logistic Regression of Geriatric Syndromes and Chronic Health Conditions on Previous Incarceration

Panel A: Parent Sample (N = 2 007)
	Model 1: Bivariate			Model 2: Control for Age and Sex		Model 3: Fully Specified
	No Incarceration	Previous Incarceration	p Value	OR	95% CI	OR	95% CI
Geriatric syndromes
Dementiaa	1.2%	2.5%	.733	1.108	(0.224–5.493)	0.752	(0.134–4.208)
Difficulty walking	6.9%	18.5%	<.001***	3.352***	(1.678–6.697)	2.391*	(1.132–5.051)
Difficulty seeing	16.7%	20.5%	.213	1.496	(0.802–2.790)	1.226	(0.655–2.293)
ADL difficulties	23.1%	48.2%	<.001***	3.709***	(2.253–6.104)	2.952***	(1.774–4.913)
Chronic health conditions
Poor/fair health	28.1%	46.3%	.001**	2.189**	(1.336–3.588)	1.564	(0.932–2.625)
Cancer	12.1%	22.8%	.032*	2.285**	(1.258–4.149)	2.206*	(1.205–4.040)
Hypertension	47.8%	54.3%	.409	1.415	(0.864–2.318)	1.226	(0.751–2.001)
Heart disease	13.2%	22.2%	.043*	1.834*	(1.004–3.348)	1.523	(0.798–2.907)
Stroke	3.6%	11.0%	.012*	2.974*	(1.228–7.200)	2.172	(0.884–5.334)
Depression	23.7%	45.3%	<.001***	3.043***	(1.830–5.061)	2.432**	(1.416–4.179)
Diabetes	21.0%	27.1%	.369	1.385	(0.802–2.389)	1.199	(0.673–2.138)
Chronic lung disease	17.5%	37.6%	<.001***	2.908***	(1.720–4.916)	2.417**	(1.415–4.129)
Heavy alcohol use	7.2%	5.9%	.837	0.872	(0.299–2.539)	0.700	(0.202–2.424)

Panel B: Spouse/Partner Sample (N = 976)
	Model 1: Bivariate			Model 2: Control for Age and Sex		Model 3: Fully Specified
	No Incarceration	Previous Incarceration	p Value	aOR	95% CI	aOR	95% CI
Geriatric syndromes
Dementiab	1.1%	0.7%	.639	0.668	(0.087–5.118)	0.427	(0.048–3.814)
Difficulty walkingc	5.6%	12.7%	.002**	2.514**	(1.421–4.449)	1.823	(0.976–3.405)
Difficulty seeing	13.1%	23.3%	<.001***	2.067***	(1.342–3.182)	1.634*	(1.022–2.611)
ADL difficulties	19.2%	28.3%	.012*	1.633*	(1.094–2.438)	1.297	(0.826–2.037)
Chronic health conditions
Poor/fair health	22.4%	29.1%	.080	1.369	(0.930–2.016)	0.937	(0.607–1.448)
Cancer	12.7%	18.9%	.044*	1.883**	(1.173–3.024)	1.919*	(1.165–3.162)
Hypertension	61.0%	54.9%	.170	0.803	(0.559–1.152)	0.696	(0.474–1.023)
Heart disease	20.0%	23.1%	.392	1.359	(0.883–2.094)	1.311	(0.814–2.113)
Stroked	3.2%	6.1%	.095	2.176	(0.979–4.836)	1.813	(0.770–4.268)
Depression	10.9%	20.9%	.001**	2.207***	(1.395–3.491)	1.864*	(1.144–3.036)
Diabetes	26.2%	22.2%	.301	0.822	(0.537–1.257)	0.665	(0.418–1.059)
Chronic lung disease	11.2%	21.2%	.001**	2.194***	(1.386–3.472)	1.691*	(1.019–2.806)
Heavy alcohol use	13.2%	20.5%	.022*	1.731*	(1.099–2.724)	2.146**	(1.335–3.451)

Notes: All estimates in Table 2 are weighted in the parent sample. However, because the partner/spouse sample was not part of the original Add Health study sampling frame, no survey weights are available for these data. aOR = adjusted odds ratio; ADL = activities of daily living; CI = confidence interval; OR = odds ratio.

^aVariables for sex and health insurance omitted due to perfect prediction of the outcome.

^bVariables for sex, race, and health insurance omitted due to perfect prediction of the outcome.

^cVariable for sex omitted due to perfect prediction of the outcome.

^dCategorical variable for health insurance replaced with a binary variable for uninsured status due to perfect prediction of several health insurance groups with the outcome.

***p < .001. **p < .01. *p < .05.

The results from spouse/partner reports of their own incarceration history and geriatric and chronic health conditions are presented in Table 2, Panel B. Across the bivariate differences reported in Model 1, those with previous incarceration have a higher prevalence of difficulty walking (12.7% vs 5.6%, p = .002), difficulty seeing (23.3% vs 13.1%, p < .001), ADL difficulty (28.3% vs 19.2%, p = .012), cancer diagnosis (18.9% vs 12.7%, p = .044), depression diagnosis (20.9% vs 10.9%, p = .001), chronic lung disease (21.2% vs 11.2% p = .001), and heavy alcohol use (20.5% vs 13.2%, p = .022). The findings from the fully adjusted model (Model 3) show that formerly incarcerated persons have higher odds of difficulty seeing (aOR = 1.634, 95% CI = 1.022, 2.611), cancer diagnosis (aOR = 1.919, 95% CI = 1.165, 3.162), depression diagnosis (aOR = 1.864, 95% CI = 1.144, 3.036), chronic lung disease (aOR = 1.691, 95% = 1.019, 2.806), and alcohol use 4 or more times per week (aOR = 2.146, 95% CI = 1.335, 3.451).

Discussion

This study replicated research by Garcia-Grossman et al. (3) on the effects of incarceration history on geriatric syndromes and chronic health issues using AHPS data. Our results corroborate the findings that formerly incarcerated individuals are more likely to suffer from various geriatric and chronic health conditions such as ADL difficulties, difficulty walking and seeing, depression, chronic lung disease, and heavy alcohol use. Likewise, both our study and that of Garcia-Grossman et al. (3) found no significant association between incarceration history and cardiovascular health issues. Notably, our data showed an elevated prevalence of cancer in formerly incarcerated older adults compared to nonformerly incarcerated persons, an outcome that was not included in Garcia-Grossman’s study and adds to the scant literature on previous incarceration and cancer diagnosis (19). Furthermore, it is important to highlight fundamental differences across the majority-female parent AHPS sample and the majority-male spouse/partner AHPS sample. The relationship between incarceration and mobility issues was more prominent in the female sample, possibly due to gender differences in aging and chronic health conditions (11). The relationship between incarceration and heavy alcohol use was notably higher among formerly incarcerated individuals in the majority-male spouse/partner data, consistent with broader trends of male drinking patterns (20).

Considering the large number of older adults with an incarceration history, the growing population of aging adults in the country, and limited research on this topic, our findings emphasize the need to understand better the impact of incarceration on healthy aging trajectories and the pathways between incarceration and health among older adults. Given the elevated geriatric and chronic health issues found among formerly incarcerated persons, as highlighted by our study and prior research (3), there is a need to develop and test the feasibility, acceptability, and efficacy of interventions tailored to improve geriatric and chronic health conditions among previously incarcerated older adults. These steps may include policy enhancements within the penal system, such as better access to preventive health screenings and improved quality of life and medical services during incarceration (eg, specialty referrals and diagnostics, physical therapy). Additionally, enhanced, streamlined coordination of healthcare and social services during the reentry process that promotes social determinants of health postrelease may improve outcomes. Considering the differences in outcomes observed between males and females in the sample, prioritization of gender-responsive care is also warranted. Ultimately, additional research on incarceration and healthy aging will be essential to deepening understanding of how incarceration experiences affect morbidity and mortality among older adults. Results will inform targeted health interventions and policy changes to address the unique needs of this vulnerable population.

Limitations and Future Directions

Despite its strengths, this study has limitations. The measure of incarceration lacks details on the duration, frequency, and experiences during incarceration. Further, the sample size of the AHPS limited the ability to assess heterogeneity in the findings across racial and ethnic groups, and the ability to examine the intersectionality of incarceration, race/ethnicity, and sex on outcomes. Certain conditions like dementia were infrequently observed in the AHPS data, possibly affecting association detection. We also could not identify mechanisms that might link incarceration to our outcomes, which may be biological, psychosocial, or both. Potential pathways to be explored in future longitudinal studies include psychological stress, lack of social support, loneliness, mental health challenges, unemployment and financial hardships, the stigma of incarceration, and poor health behaviors over the lifespan (5,9,10). Finally, the study could not ascertain whether observed associations between incarceration and health outcomes are causal due to the observational and cross-sectional data and potential unmeasured confounders. Research with larger and longitudinal samples can consider alternative methodologies such as propensity score matching, fixed-effects modeling, or interrupted time series designs to build upon the findings we have presented here (5).

Conclusion

This study illuminates the health challenges experienced by previously incarcerated older persons and adds to the small but growing body of research exploring the enduring effects of incarceration on aging and health issues. Building a deeper understanding of the long-term health repercussions of incarceration is vital to crafting programmatic and policy initiatives that mitigate health disparities and foster the well-being of vulnerable older adults.

Funding

C.K. is supported by the National Institute on Aging (K01AG061239). A.S. is supported by grants from the National Institute on Aging. R.J.T., Jr., is supported by the National Institute of Aging (P30AG059298, K02AG059140) and the National Institute on Minority Health and Health Disparities (U54MD000214)

Conflict of Interest

A.S. is supported by grants from the National Institute on Aging and received payment for serving as a consultant for Merck, received honoraria from Springer Nature Switzerland AG for guest editing special issues of Current Sleep Medicine Reports, and is a paid consultant to Sequoia Neurovitality and BellSant, Inc. The other authors declare no conflict. The views presented are of the authors, alone, and do not represent the views of the U.S. Department of Veterans Affairs, or any other federal agency.

Author Contributions

A.T.: conceptualization, writing, reviewing, editing. D.B.J.: conceptualization, writing, reviewing, editing. M.N.: writing, reviewing, editing. C.K.: writing, data analysis, reviewing. C.G.: writing, reviewing, editing. J.T.: writing, reviewing, supervision. A.P.S.: writing, reviewing, supervision. R.J.T.: writing, reviewing, supervision.