Association between mental health outcomes and lifetime criminal justice involvement in U.S. adults with diabetes

Abstract

Background:

Diabetes is a leading cause of death in the United States, and comorbid mental illness is associated with worse diabetes outcomes. Those with criminal justice involvement (CJI) have high rates of mental illness and diabetes prevalence. However, little is known about the relationship between CJI and mental illness among those with diabetes.

Methods:

Using a nationally representative sample of U.S. adults with diabetes from the National Survey of Drug Use and Health (2015-2018), we investigated the relationship between CJI and mental health outcomes (depression, serious psychologic distress, serious mental illness, moderate mental illness, any mental illness, functional status). Multiple linear and logistic regression models were used to assess the relationship between CJI and each mental health outcome adjusting for multiple socio-demographic and comorbidity variables.

Results:

Of 11,594 respondents, representing 25,834,422 adults with diabetes, 17.1% reported prior CJI. In fully adjusted models, CJI was independently associated with all mental health outcomes: depression (aOR 1.80, 95% CI: 1.41, 2.30), serious psychologic distress (aOR 1.53, 95% CI: 1.23, 1.90), serious mental illness (aOR 2.00, 95% CI: 1.58, 2.52), moderate mental illness (aOR 1.72, 95% CI 1.30, 2.26), any mental illness (aOR 1.92, 95% CI: 1.56, 2.35) and functional status (regression coefficient 3.6, 95% CI: 3.53, 3.79).

Conclusion:

Those with diabetes and criminal justice involvement experience concentrated risk for poor mental health outcomes. Our findings suggest that mental health interventions may be imperative to achieving control of diabetes in the justice-involved population.

Introduction

Diabetes mellitus was the seventh leading cause of death in the United States in 2017, affecting over 30 million adults¹. The burden of poorly-controlled diabetes on the United States health care system is great, accounting for over 400,000 emergency department visits or hospitalizations for hyperglycemic crisis, 130,000 lower-extremity amputations, and nearly 300,000 cases of end-stage renal disease¹. Achieving optimal glycemic control reduces risk of complications such as hyperglycemic crisis, end-stage renal disease, and neuropathy, but requires a high level of self-management². Higher rates of diabetes prevalence and complications are experienced by members of American minority groups, particularly Black Americans^1,3.

Comorbid mental illnesses including depression and serious mental illness (SMI) occur in higher rates among those with diabetes and are known to worsen diabetes outcomes^4,5. Studies examining the prevalence of depression among those with diabetes in the U.S. vary between 8.7% and 19.1%, the latter finding being twice that of the general population^6,7. Adults with serious mental illness experience 70% increased risk of diabetes than the general population⁸. Studies suggest these relationships are bidirectional. For example, insulin-resistance may be exacerbated by anti-psychotic medications in patients with SMI⁹, whereas diabetes distress, overwhelming feelings brought on by the burden of the chronic disease, can worsen depression¹⁰. A substantial amount of prior work has shown the relationship between mental illness and poor diabetes outcomes is mediated by impaired self-care behavior^11,12. While research shows that these conditions can be effectively treated in patients with diabetes, its impact on glycemic control is unclear¹³.

Criminal justice involvement (CJI) is a social risk with significant overlap with populations at highest risk for poor diabetes outcomes. While Black Americans make up 13% of the U.S. population, they account for over 16% of diabetes diagnoses and 40% of those incarcerated^1,14. Black Americans are dramatically more likely to be exposed to CJI at any level, from arrests to prolonged sentences^15,16. Like those with diabetes, populations with CJI experience high rates of mental illness – among persons currently incarcerated, over half will have any mental illness and a quarter will have SMI¹⁷. CJI can worsen mental health, with even minimal contact such as an arrest or overnight stay in jail causing deterioration of overall mental well-being^18,19. The mechanisms by which CJI influences mental health are hypothesized to be associated with chronic stress as well as traumatic experiences²⁰. Thus, it is plausible given the extreme disparities in CJI that undertreated mental illness brought on or exacerbated by CJI may undermine efforts to achieve health equity in diabetes outcomes.

There has been increasing attention on interventions targeting improved health outcomes among those with CJI^21,22. While better understanding the relationship between CJI and mental health for patients with diabetes may help us to improve treatment options for this population, little is known about the intersection between these two risk factors among U.S. adults with diabetes. To better understand this relationship, we used nationally representative survey data to study the association between mental health outcomes including depression, serious psychologic distress, any, moderate or severe mental illness, and mental health-related functional impairment and lifetime CJI among U.S. adults with diabetes. We hypothesized that CJI would be significantly associated with worse mental health outcomes in this population, even after controlling for significant socio-demographic and medical comorbidity confounders.

Methods

Data for this study included respondents to the National Survey of Drug Use and Health (NSDUH) between the years 2015-2018. The Substance Abuse and Mental Health Services Administration’s (SAMHSA) administered this nationally representative, cross-sectional survey on an annual basis to establish the prevalence of substance use and mental health conditions among adolescents and adults in the United States. The survey data excludes respondents living in institutionalized settings, including those who are in long-term care facilities or currently incarcerated in prison or detention centers. In addition, it collects important associated information, including mental health and addiction treatment utilization, some measures of chronic disease and criminal justice involvement. While interviews are conducted in person, the majority of responses are provided directly and anonymously into a computer. The survey response rate was between 66-69% during included years²³.

Population

Our sample included responses from all adults who reported a diagnosis of diabetes (n=11,594, weighted n=25,834,422). We included all respondents who provided a “yes” response to the question, “Have you ever been told by a doctor or other healthcare professional that you have diabetes or sugar diabetes?”

Main Measures

Our main exposure variable was lifetime involvement in the criminal justice system. We considered respondents to have CJI if they responded “yes” to the question, “Not counting minor traffic violations, have you ever been arrested and booked for breaking the law? Being ‘booked’ means that you were taken into custody and processed by the police or by someone connected with the courts, even if you were then released.” Missing data were excluded from the analysis.

Outcomes

Our dependent variables included five non-exclusive dichotomous variables for past year mental illness and one continuous measure of functional impairment due to poor mental health.

1. Depression: measured by questions containing DSM-IV diagnostic criteria to categorize participants by history of past year major depressive episodes²⁴.

2. Serious psychological distress: measured using the Kessler Psychological Distress Scale (K6). The K6 contains six scaled questions, which when scored predict serious psychological distress at a validated cutoff²⁵.

3. Serious mental illness (SMI): SAMHSA defines serious mental illness as having combined past year mental illness combined with serious functional impairtment²⁶. The NSDUH is designed to estimate three degrees of mental illness (mild, moderate, severe) using responses to short scales of psychological distress and functional impairment to create a scalar prediction model with validated cut-points²⁷. Respondents scoring above a certain cut-point were considered to have SMI.

4. Moderate mental illness (MMI): Respondents scoring above a certain cut-point but below the estimated criteria for SMI were considered to have MMI.

5. Any mental illness (AMI): Respondents estimated to have mild, moderate, or serious mental illness was considered to have AMI.

6. Mental health-related functional impairment: measured using a 13-question abbreviated version of the World Health Organization disability assessment score 2.0 (WHO-DAS), a validated self-reported instrument based which quantifies functional impairment due to poor mental health²⁸. Questions about daily functioning relative to “emotions, nerves, or mental health” are scored to arrive at a value between 0 and 24²⁹. Higher scores indicate higher degrees for functional impairment.

Covariates

We included additional covariates in our models to account for potential socio-demographic and clinical confounders. Demographic covariates included: age category (18-25, 26-34, 35-49, 50+); race/ethnicity (white, non-Hispanic; Black, non-Hispanic; Hispanic, other, including non-Hispanic Native American/Alaskan native, non-Hispanic native Hawaiian or Pacific Islander, non-Hispanic Asian, or more than one race); sex (male, female), poverty level (<100% federal poverty level as determined by the Census Bureau, 100-200% poverty level, >200% poverty level), educational status (less than high school, high school degree or equivalent, some college, or college graduate); marital status (married, widowed/divorced/separated, never married); current employment status (full time, part time, unemployed, other, including disabled, keeping house full-time; in school/training; retired; some other reason); and insurance coverage (Medicare, Medicaid, VA healthcare, private, none).

To account for clinical comorbidities, we captured all self-reported physical medical conditions included in the survey, including high blood pressure, heart condition, kidney disease, asthma, COPD, cirrhosis, hepatitis B/C, HIV/AIDS, and cancer. All comorbidities were included as independent dichotomous variables.

Statistical Analyses

First, descriptive statistics were used to summarize the sociodemographic characteristics and health profiles of the sample and compare these variables by CJI status. Second, we calculated proportion of the total population with each mental health outcome, as well as by CJI status. Third, we ran unadjusted and adjusted multiple logistic regression models to investigate the relationship between each mental health outcome coded as a binary variable (depression, serious psychologic distress, serious mental illness, moderate mental illness, and any mental illness) and CJI status as the primary independent variable. Similarly, we ran unadjusted and adjusted multiple linear regression model to investigate the relationship between functional impairment due to mental health and CJI status as the primary independent variable. All models were adjusted analyses for age, sex, race, poverty level, education, marital status, employment, health insurance coverage type, and comorbid medical conditions.

We conducted all analyses with Stata version 16 (Stata Corp LP, College Station, TX) and used weights provided by SAMHSA that account for complex survey design and allow extrapolation for the US population as a whole. We used 2-sided p-values of less than 0.05 to determine statistical significance.

Results

Unweighted sample included 11,594 respondents, representing 25,834,422 U.S. adults. In this sample of U.S. adults with diabetes, 17.1% reported a history of CJI (95% confidence interval (CI), 16.2, 18.0). Demographic characteristics by CJI status are displayed in Table 1. Those with CJI were more likely to be in the middle age ranges compared to those without CJI (26-34: 5.2 versus 4.1%; 35-49 23.5 versus 15.5%; p<0.001). Compared to those without CJI, those with CJI were more likely to be male (72.2 versus 42.8% p<0.001) and Black (18.6 versus 14.8% p=0.02) but less likely to be Hispanic or of “other” race/ethnicity (14.9 versus 16.7%; 6.7 versus 7.9% p=0.002). They were also more likely to be of lower socioeconomic status, with 47.3% reporting income <200% FPL (versus 39.8%, p<0.001); and 54.3% having the equivalent HS graduate degree or less (versus 47.0%, p<0.001). Those with lifetime CJI were more likely to have Medicaid, VA health or be uninsured, and less likely to have private coverage or Medicare. Rates of common diabetes comorbidities were high in both groups; those with CJI were marginally more likely to report a heart condition (26.1% versus 23.4% p=0.09) but slightly less likely to report hypertension (42.5% versus 45.8% p=0.02). Rates of kidney disease did not differ (6.3% versus 6.7%, p=0.53).

Table 1:

Sample demographic characteristics of U.S. adults with diabetes by lifetime criminal justice involvement (CJI), 2015-2018

	Overall (n=11,594)	Prior CJI (n=9,384)	No Prior CJI (n=2,178)	p-valuea
		%	%
Variables
Age				<0.001
18-25	2.0	1.7	2.1
26-34	4.3	5.2	4.1
35-49	16.9	23.5	15.5
50+	76.8	69.5	78.3
Sex				<0.001
Male	47.9	72.2	42.8
Female	52.1	27.8	57.2
Race				0.02
White NH	60.5	59.9	60.7
Black NH	15.4	18.6	14.8
Hispanic	16.4	14.9	16.7
Other	7.7	6.7	7.9
Poverty Level				<0.001
<100%	15.2	20.0	14.2
100-200%	25.1	27.3	24.7
>200%	59.7	52.7	61.1
Education				<0.001
< High School	18.9	20.9	18.5
HS Grad	29.4	33.4	28.5
Some College	29.6	31.4	29.4
College Grad	22.1	14.3	23.7
Marital Status				<0.001
Married	57.5	48.8	59.3
Divorced/Widowed	29.4	32.3	28.8
Never Married	13.1	18.9	11.9
Employment				<0.001
Full Time	32.9	34.2	31.5
Part Time	9.9	8.4	10.2
Unemployed	3.0	4.4	2.7
Otherb	55.2	53.0	55.7
Health Insurance
Private	60.5	49.1	62.9	<0.001
Medicaid	19.7	27.5	18.1	<0.001
Medicare	46.8	39.3	48.4	<0.001
VA Health	9.0	11.7	8.4	0.002
Uninsured	5.0	7.3	4.6	<0.001
Comorbidities
Heart Condition	23.8	26.1	23.4	0.09
Hypertension	45.3	42.5	45.8	0.02
Kidney Disease	6.6	6.3	6.7	0.53
COPD	10.2	12.5	9.8	0.01
Asthma	11.3	11.2	11.3	0.94
Cirrhosis	1.2	1.8	1.1	0.08
Hepatitis B/C	2.2	4.8	1.7	<0.001
HIV/AIDS	0.2	0.5	0.2	0.06
Cancer	10.7	9.5	10.9	0.18

^ap-value: comparing those with prior CJI to those with no prior CJI

^bOther employment status includes disabled, keeping house full-time; in school/training; retired; some other reason

Prevalence of mental health outcomes in the overall sample and by prior CJI status are displayed in Table 2. Per reported mental health symptoms in the past 12 months, 11.67% of those with prior CJI experienced depression, compared to 6.49% of those without prior CJI (p<0.001). 14.98% of those with prior CJI met criteria for serious psychologic distress, compared to 5.94% of those without prior CJI (p<0.001), whereas 30.23% of those with prior CJI had any mental illness, compared to 19.23% among those without prior CJI (p<0.001). Those with prior CJI also experienced greater mental health-related functional impairment (4.97 versus 3.40, p<0.001).

Table 2:

Prevalence of mental health outcomes among adults with diabetes by lifetime criminal justice involvement (CJI), 2015-2018

	Overall	Prior CJI	No Prior CJI	p-valuea
Mental Health Outcome	%	%	%
Major Depressive Episode	7.39	11.67	6.49	<0.001
Serious Psychologic Distress	8.95	14.98	9.54	<0.001
Serious Mental Illness	10.78	17.34	9.28	<0.001
Moderate Mental Illness	5.77	8.85	5.09	<0.001
Any Mental Illness	21.11	30.23	19.23	<0.001
	Mean (95% CI)	Mean (95% CI)	Mean (95% CI)
WHO-DAS 2.0b	3.66 (3.53-3.79)	4.97 (4.58-5.36)	3.40 (3.25-3.54)	<0.001

^ap-value: comparing those with prior CJI to those with no prior CJI

^bWHO-DAS 2.0 – World Health Organization disability assessment score; a measure of mental health-related functional impairment

Lifetime CJI was significantly associated with all mental health outcomes in both unadjusted and adjusted models (see Table 3). In adjusted models, lifetime CJI was associated with depression (aOR 1.80, 95% CI 1.41-2.30), serious psychologic distress (aOR 1.53, 95% CI 1.23-1.90), serious mental illness (aOR 2.00, 95% CI 1.58-2.52), moderate mental illness (aOR 1.72, 95% CI 1.30-2.26), and any mental illness (aOR 1.92, 95% CI 1.56-2.35). Lifetime CJI was also associated with a higher mental health-related functional impairment score (adjusted regression coefficient 1.50, 95% CI 1.05-1.96.

Table 3:

Association between mental health outcomes and lifetime criminal justice involvement (CJI) status among U.S. adults with diabetes, 2015-2018

Mental Health Outcome	Odds Ratio (95% CI)	Adjusted Odds Ratioa (95% CI)
Major Depressive Episode	1.90 (1.54-2.35)	1.80 (1.41-2.30)
Serious Psychologic Distress	1.67 (1.40-2.00)	1.53 (1.23-1.90)
Serious Mental Illness	2.05 (1.67-2.51)	2.00 (1.58-2.52)
Moderate Mental Illness	1.81 (1.46-2.24)	1.72 (1.30-2.26)
Any Mental Illness	1.82 (1.54-2.14)	1.92 (1.56-2.35)
	Coefficient (95% CI)	Adjusted Coefficient (95% CI)
WHO-DAS 2.0b	1.57b (1.14-2.00)	1.50b (1.05-1.96)

Reference group: no lifetime CJI

^aadjusted for age, sex, race, poverty level, education, marital status, employment, health insurance coverage type, and comorbid medical conditions

^{b –}WHO-DAS 2.0: World Health Organization disability assessment score; a measure of mental health-related functional impairment

Discussion

In this nationally representative sample of adults with diabetes, lifetime exposure to the criminal justice system was high, reported by one in six respondents. A large proportion of participants also reported mental health outcomes, with one in five experiencing any mental illness. Those with prior CJI were more likely to be Black men, live below poverty level, and have a high school diploma or less. They were also more likely to be insured by Medicaid or to lack health insurance altogether. Baseline rates of all mental health outcomes were high in our entire study population, but they were much higher among those with prior CJI. Our multivariate analyses found a strong, independent association with prior CJI and all mental health outcomes. After adjustment for socio-demographic confounders and comorbid medical conditions, those with CJI experienced 80% greater odds of having past year depression, 50% likelier to have serious psychologic distress, and were twice as likely to have serious mental illness. Moreover, functional impairment scores were 50% higher among those with CJI. Those with CJI and diabetes experience particularly concentrated risk of mental illness.

While it is well known that those with diabetes are at increased risk for mental illness and those with a history of CJI are at an increased risk for mental illness, this, to our knowledge, is the first study to investigate the overlap of these two risks. Our findings suggest that those with diabetes and CJI are at high risk for a diverse range of negative mental health conditions from depression to serious psychological distress. The strong association with poor mental health-related functional status reveal the potential impacts of these findings on quality of life for those with diabetes and prior CJI. Moreover, our results suggest that screening for and providing tailored mental health care to this population may be a crucial component to achieving optimal glycemic control. The findings underscore the need for additional research into the intersection between mental illness and CJI among those with chronic disease to identify and implement effective interventions.

Serious mental illness was the mental health outcome most strongly associated with criminal justice involvement, with those with prior CJI experiencing two times the adjusted odds of SMI than those without. Comorbid serious mental illness, which the World Health Organization defines as mental disorders that are “severe in degree, persistent in duration, and result in significant functional impairment” is linked to worse quality of healthcare³⁰ and worse health outcomes in those with diabetes³¹. When combined with the social risk of CJI, individuals are likely to face additional barriers to high quality care as well as barriers to self-management of their chronic disease^12,32. Similarly, major depression was strongly associated with prior CJI, with an 80% greater adjusted odd of diagnosis among the exposed. Those with diabetes, our baseline sample, experience depression at rates about twice that of the general U.S. population; thus, this strong association underscore that depression is particularly common among those with diabetes and CJI. Extensive research into the intersection of diabetes and depression has causally linked the mental illness to worse glycemic control, higher rates of complications, and substantially increased healthcare costs^4,33. Taken in sum, these results suggest that additional resources should be focused on providing care to those with both CJI and SMI or depression.

The significant association between prior CJI and mental-health related functional impairment among those with diabetes is a novel finding. In this sample mental health-related functional impairment was 50% greater among those with CJI. Prior studies have found that mental health conditions lead to functional impairment in adults with diabetes and may undermine self-care behavior, but factors such as social support may ameliorate this association³⁴. These findings suggests that interventions which aim to improve self-care for those with diabetes may need to consider optimizing mental-health related function first.

Access to mental health care is poor in the United States, partially due to stigma surrounding mental health conditions and a societal divestment in community-based mental health services^35,36. Studies have found that those with CJI have particularly high barriers to mental health treatment³⁷. Barriers may include decreased access to employment-sponsored health insurance as well as prior negative healthcare encounters and perceived discrimination experienced by those with CJI^32,38. Though the mechanism by which CJI undermines mental health treatment is unclear, evidence suggests it extends beyond decreased access to health insurance coverage. One study found that despite impressive gains in insurance coverage among those with CJI after the implementation of the Affordable Care Act, there was no increase receipt of mental health treatment during the same time period³⁹. Additional research is needed to better understand how those with CJI experience barriers to mental health care, particularly for those with comorbid chronic conditions such as diabetes.

Prior work has documented high rates of co-occurring substance use disorders and mental illness in justice-involved populations⁴⁰. In our models, we did not control for substance use disorders in order to focus on the large but under-studied issue of the intersection of mental illness and chronic disease among those with criminal justice involvement. However, additional study into the overlapping effects of substance use disorder, criminal justice involvement, and chronic disease is warranted.

Our findings have important implications for researchers and policy makers pursuing health equity. Due to previously described disparities in diabetes prevalence¹, our baseline study population was disproportionately comprised of minorities and those of low socioeconomic status; our results found these disparities were additionally concentrated among those with prior CJI. Given these profound disparities in the study population and the strong and diffuse association of CJI with mental health outcomes, incorporating mental health treatment into diabetes care, especially for those with justice-involvement, may be one crucial component among many in the effort to eliminate differential rates of diabetes outcomes between individuals of different races.

We acknowledge several limitations to our study. First, due to the cross-sectional study design, causal association between CJI and mental health outcomes cannot be established and should not be assumed. Secondly, the self-reported survey design may underestimate those with diabetes, particularly those who are undiagnosed. Those missed are likely overrepresented in the CJI-exposed group, as they were more likely to be uninsured which is a predictor for undiagnosed diabetes⁴¹. Therefore, this limitation is most likely to render our findings more conservative. While criminal justice exposure is also self-reported, prior work has shown this method of information gathering equally accurate as administrative data.⁴² Finally, because this survey data excludes those currently incarcerated, a substantial proportion of those with CJI in the United States are missed in this otherwise nationally representative survey.

In conclusion, those with diabetes and criminal justice involvement experience concentrated risk for mental illness. Taken in sum with prior work which has demonstrated that mental illness often undermines diabetes care and results in worse outcomes, our findings underscore the importance of improving treatment options for those with complex layers of health and social risks and incorporating mental health interventions into treatment options for adults with diabetes exposed to the justice system.

Highlights.

• Using a nationally representative sample of U.S. adults with diabetes from the National Survey of Drug Use and Health (2015-2018), this study investigated the relationship between criminal justice involvement (CJI) and mental health outcomes (depression, serious psychologic distress, serious mental illness, moderate mental illness, any mental illness, functional status).

• CJI was independently associated with all mental health outcomes: depression (OR 1.80), serious psychologic distress (OR 1.53), serious mental illness (OR 2.00), moderate mental illness (OR 1.72), any mental illness (OR 1.92) and functional status (beta 3.6).

• Those with diabetes and criminal justice involvement experience concentrated risk for poor mental health outcomes.