Adverse Experiences, Mental Health, & Substance Use Disorders as Social Determinants of Incarceration

Abstract

Aims:

This study describes how incarcerated people understand: (1) adverse experiences, mental health, and substance use disorders as determinants of incarceration, (2) the role of gender in impacting this understanding, and (3) strategies to prevent incarceration. Ecosocial theory provides a theoretical framework.

Methods:

Open-ended interviews were conducted (December 2016–January 2017) with recently incarcerated adults in Massachusetts state prisons. Participants described determinants of incarceration and incarceration prevention strategies. Interviews were coded thematically using inductive and deductive approaches.

Results:

Thirty participants, evenly split by gender, reported themes across four levels: society (lacking basic needs, discrimination), community (neighborhood factors), interpersonal (trauma), and individual (social isolation, mental health, substance use). However, there were variations in themes by gender. Proposed prevention strategies included early access to quality individualized cross-system services.

Conclusion:

Findings highlight how investing in social and community building services could prevent incarceration. Policies can support these services by redirecting funding.

Introduction

Background

Improving how the United States (US) addresses mental health and substance use disorders for justice-involved individuals is critical. The US has the highest rate of incarceration in the world which is detrimental to public health (Dumont et al. 2012, Wildeman and Wang 2017). Formerly incarcerated people are at an elevated risk of death from drug overdose, suicide, homicide, and cardiovascular disease (Binswanger et al. 2007). People with mental health (Prins 2014) and substance use disorders (James and Glaze 2006, Mumola and Karberg 2006) are also overrepresented within US prisons and jails. The rate of mental health disorders is as high as 60% among people incarcerated in jails and only slightly lower for people incarcerated in state and federal prisons (Mumola and Karberg 2006). In particular, people with serious mental illness are held in jails at very high rates and account for 15% of men and 31% of women incarcerated in jails (Steadman et al. 2009). Additionally, nearly half of all people incarcerated in the US have problematic substance use prior to their incarceration (Mumola and Karberg 2006).

Given that incarcerated people have disproportionately high rates of mental health and substance use disorders, it is not surprising that these disorders can also play a role in incarceration. For example, people with mental health and/or substance use disorders are at increased risk of arrest (Prince and Wald 2018). Serious mental illness has also been identified as a driver of recidivism (Baillargeon, Hoge, and Penn 2010). Additionally, 33% of people in state and 22% of people in federal prisons reported to be “using drugs at the time” of their offense, while about 18% of incarcerated people committed their crime to obtain money for drugs (Mumola and Karberg 2006). Substance use disorders can also mediate the relationship between mental health disorders and crime (Swartz and Lurigio 2007, Fisher et al. 2014, Fazel et al. 2009).

A potentially important related factor is traumatic or adverse experiences. Adverse experiences have been demonstrated to play a role in mental health, substance use disorders (Dong et al. 2004, Felitti et al. 1998, Henry 2020), and violence (Gillikin et al. 2016). Additionally, rates of exposure to adverse experiences among incarcerated people are higher than the general public (Briere, Agee, and Dietrich 2016). A portion of incarceration is likely explained by the interplay between adverse experiences, mental health, and substance use disorders. Further research is needed to describe the varied pathways by which this may occur, which is a goal of this paper.

Existing research in this area has mostly used quantitative methods and has primarily been conducted with non-incarcerated populations. However, to truly understand how a phenomenon occurs, qualitative methods are needed (Richard 2013). Additionally, given that existing approaches to reducing incarceration have been limited in their effectiveness, it is important to understand how people experiencing the problem of incarceration view potential prevention strategies. Few studies have examined this area. One study (Barnert et al. 2015) of incarcerated juveniles reported that the youth recommended an increase in programs that promote safety and mentorship, both at home and in school. However, similar research with adults is lacking.

Theoretical & Conceptual Framework

This study is guided by ecosocial theory, which is relatively novel within crime studies, outside of domestic violence research (Gil-González et al. 2007). Ecosocial theory is rooted in the field of epidemiology and provides a framework that explains how individuals are impacted by environmental conditions. The theory conceptualizes environments as occurring on multiple overlapping/interconnected levels which include society, community, and interpersonal levels (Krieger 2001). Together, factors relating to these levels are collectively described as social determinants. However, the theory also conceptualizes individual factors, such as race and gender, to be socially constructed. Furthermore, it considers individual health factors to be socially determined, which means that a person’s biology and health is impacted by the social determinants through the process of “embodiment.” Coined by Krieger, the term embodiment describes how people physiologically incorporate experiences with their environments (Krieger 2005). This theory is ideal for explaining how adverse experiences could cause mental health and substance use disorders, for example through shaping brain development.

A study of repeatedly incarcerated adults tested this framework. Results demonstrated that community level factors of neighborhood violence and social capital were associated with drug dependence, while the interpersonal level factor of intimate partner violence mediated the relationship between community, and individual factors. Individual level factors of gender, race, and ethnicity also moderated the relationship (Rogers et al. 2012).

Given that ecosocial theory explains how the interplay between individuals and their environments is further impacted by a person’s past experiences (Krieger 2001), this theory can help to explain how mental health and substance use disorders could go on to shape future behavioral responses to environmental conditions via criminalized behavior. While ecosocial theory has historically been applied to explain the social determinants of health (Krieger 2001), it will be applied in this study to explain the social determinants of incarceration.

The conceptual framework for this study is shown in Figure 1. This framework includes four overlapping circles, one for each level. In Figure 1, each circle is labeled with the corresponding level of ecosocial theory and topics under study are listed below. Hypothesized relationships are depicted via boxes and arrows which are imbedded in the circles. As shown, it is hypothesized that 1) adverse childhood experiences will impact both incarceration and mental health/substance use disorders, and 2) having a mental health or substance use disorder will impact incarceration.

Figure 1: Conceptual Framework.

Study Aims

This study aims to describe how incarcerated people understand: (1) the role of adverse experiences, mental health, and substance use disorders as social determinants of their incarceration, (2) the role of gender in impacting this understanding, and (3) potentially effective programs and policies to address social determinants of incarceration. This study will fill a gap in the literature by including the voices of incarcerated people to contextualize their nuanced pathways to prison.

Methods

Sampling

This phenomenological study used open-ended in-depth interviews, followed by a short survey. Participants gave accounts of their pathways to prison and provided suggestions for preventing incarceration. Study participants included both men and women. All participants were recently criminally sentenced (found guilty) adults who were incarcerated in Massachusetts state prisons. Participants were sampled from the two gender specific intake facilities, both maximum security. Within Massachusetts state prisons, 30% of men and 70% of women had an open mental health case, meaning that they were receiving treatment for a mental health diagnosis (Research & Planning Division 2016). This gendered difference is mirrored nationally (James and Glaze 2006).

Due to the gender differences in rates of mental health disorders, stratified purposive sampling was used so that participants of both genders included a portion of participants with and without mental health diagnoses. Strata were divided evenly between men and women, as determined by the gender for which the prison was classified. Individuals with open mental health cases were sampled at a higher rate to provide variation of diagnostic severity within the sample. The prison system’s research department prepared the sample from their internal census.

Participants were excluded if they were incarcerated less than 30 days or more than 6 months. People incarcerated for less than 30-days were excluded to give them time to adjust to their new environment and to allow time for mental health assessments which related to some of the demographic questions. People incarcerated for more than six months were excluded to reduce recall bias. People who could not speak English were also excluded. People incarcerated in a health services unit were excluded so as not to interfere with their medical treatment, as participants would have had to be transported out of the hospital unit to be interviewed.

The overall response rate was 67%. However, it varied by gender with 79% of men and 58% of women participating. All men who declined to participate did so during the first recruitment stage, and therefore did not meet the researcher. Correctional staff reported that men generally gave limited reasons for declining. For example, multiple men reportedly stated that it was “too early” in the day. Women declined to participate at both stages of recruitment. The most common reason that they gave for declining was that they “had something else to do”. This difference between both the rate and reason for declining between men and women was likely related to the fact that the women’s facility had more programming. The women also often had to forgo participating in programming to participate in the study. In contrast, several men reported that they “might as well participate” because they had “nothing else” to do. Information to understand how people who refused to participate might have differed from those who did participate was not available.

The final number of participants was 30 individuals. The number of planned interviews was capped at an upper limit of 30 due to recommendations in the literature regarding a minimum sample size of 25–30 for in depth qualitative interviews (Dworkin 2012). Saturation was determined separately by gender as interviews were conducted and completed first in the men’s facility and then in the women’s facility. Within each gender group saturation began to occur around 12 to 13 interviews, which created a natural endpoint at the cap of 15 interviews per prison.

Recruitment

Standard recruitment protocol per the correctional department was followed. Recruited individuals were first contacted in-person by correctional staff and informed that they were selected to participate a study and given written information about the study. Recruited individuals were then asked if they would like to meet with the researcher where they could learn more about the study. They were advised that they could refuse this meeting and that their decision would have no impact on the conditions of their incarceration. Correctional staff were not provided information about which individuals decided to participate in the study, and which individuals discussed the study but declined to participate. Participants received no compensation for participation.

Data Collection

Data were collected in December 2016–January 2017. The study was approved by the Brandeis University Institutional Review Board and the Massachusetts Department of Correction. Informed consent was obtained from all participants included in the study. Consent capacity was determined by the interviewer per the National Institute of Health’s guidelines (National Institutes of Health 2009). Interviews occurred in the prisons where the participants resided. Interviews generally lasted between 30–60 minutes, which did not include the informed consent process. With consent, interviews were audio recorded. Field notes were taken during all interviews.

Participants were interviewed by the author who was the sole interviewer. Interviews took place in a private room where neither staff, nor other incarcerated people, could overhear the interviews. The rooms were either empty program staff offices or classrooms. Rooms did have a window into the hallway. However, the rooms were also at the end of a corridor, so there were very few people who walked past. Male participants were escorted to and from the room by a staff person who the researcher called on the telephone when needed. Female participants were able to come and go freely from the interview room during scheduled “movement times.”

An interview guide which contained broad topics with example questions was used. Topics comprised of pre-conviction life, pathways to incarceration, and potential programs/policies to prevent incarceration. Pre-conviction life was introduced by asking participants to describe “how things were going in their life” around the time of their crime. They were asked what was helping or harming them at that time and how they thought they ended up in prison. Questions were open-ended and allowed participants to introduce topics of adverse experiences, mental health, and substance use disorders. However, if participants did not bring these topics up, or struggled to answer open-ended questions, then they were asked progressively more specific probing questions.

A variation of the miracle question was utilized to explore crime prevention. The miracle question was originally developed for use in solution-focused psychotherapy where a client is asked to imagine that during their sleep a miracle occurred which solved their problem. The client is then asked to describe how their life would be different (De Jong and Miller 1995). While the purpose of the interviews in this study was not to provide psychotherapy, the modified use of this question stimulated participant creativity to think outside traditional paradigms of social services. An example of how this question was asked in this study is: “Imagine your life had taken a different pathway, and you did not end up in prison. What would have been different in your life for that to have happened?” This phrasing helped stimulate responses that were directed toward the root of incarceration. For example, when asked this question, one young man (Participant A) responded that he would have “been born into a different family.” This response stimulated a description of intergenerational incarceration which created an expectation of incarceration. Had the participant simply been asked what could have prevented his crime, the intergenerational pathway may not have emerged.

At the end of the in-depth interview, participants answered a series of closed-ended demographic questions (e.g., age, gender, race, ethnicity, educational history, employment status) and completed the Adverse Childhood Experiences (ACE) Questionnaire (Felitti et al. 1998). This questionnaire is a 10-item measure of adverse childhood experiences, including abuse and neglect. It is a validated instrument for assessing adverse experiences that has demonstrated an increased risk of multiple negative health outcomes, based on the measured adverse childhood experiences (Felitti et al. 1998, Dube et al. 2004). The measure has also been used with incarcerated populations (Stensrud, Gilbride, and Bruinekool 2018). Participants were given this questionnaire to better understand the demographic profile of the sample and to capture information about all adverse experiences of participants, including those experiences that they did not identify as relating to their incarceration.

Qualitative Analysis

Preliminary data analysis was done via writing memos during data collection. Writing memos is a reflexive research technique used in qualitative data analysis where ongoing writing is done throughout the process of data collection and analysis to document conceptual links in the data (Birks, Chapman, and Francis 2008). Additional qualitative analysis was conducted using both inductive and deductive approaches. Inductive analysis allows for the use of a theoretical framework to shape the interpretation of the data, while a deductive approach allows the researcher to describe themes that are not pre-determined by the theoretical framework (Braun and Clarke 2006). Inductive analysis was used to develop codes generated by the ecosocial theoretical framework, and deductive analysis was used to identify thematic codes unrelated to ecosocial theory (Braun and Clarke 2006). Finally, the inductive and deductive codes were compared and condensed to a single set of codes to streamline analysis. This hybrid approach allows for the description of both expected and unexpected themes (Kristoffer et al. 2016, Fereday and Muir-Cochrane 2006).

The author was the sole coder; however, multiple approaches were used to ensure validity or trustworthiness of results (Meadows and Morse 2001). For example, regular debriefing meetings were conducted between the author and another consulting researcher with expertise in this area. During these meetings data were discussed in relation to emerging themes. Appropriate codes were also selected based on the data. The use of debriefing meetings also facilitated analytic reflexivity which is the iterative and reflexive process of revisiting data to make new connections with emerging themes to refine codes and meaning (Srivastava and Hopwood 2009).

An audit trail was kept as a record of all decisions made during this process. Decisions were repeatedly revisited during coding to assure that themes best fit the data. Coding was done manually, which does not detract from the rigor of the analysis (Basit 2003, Meadows and Morse 2001). Data were reviewed multiple times during both inductive and deductive coding. The application of both inductive and deductive coding approaches also allowed for the use of multiple methods through the combination of analytic approaches, which is also a mechanism of ensuring trustworthiness of results (Meadows and Morse 2001). During the coding multiple diagrams were generated to represent the social determinants of incarceration. These diagrams were condensed based on similarity, until only a single diagram remained. Participant quotes were identified in the text with a letter for example “Participant A.” This was done to indicate that quotes came from different, or the same, participants. The participant letter is otherwise unconnected to any analysis or data storage. While participant gender/age category were listed alongside quotes, additional demographic identifiers were not provided to prevent revealing the identity of participants.

Quantitative Analysis

Descriptive (frequencies and means) and bivariate (t-tests) quantitative analysis of the data was conducted via Stata 15 software (StataCorp 2017). Bivariate analysis using t-tests was conducted to identify differences by gender and number of ACEs. Reported ACEs were categorized into two groups which included “high ACEs” and “low ACEs”. The high ACEs group included people who reported four or more ACEs, while the low ACEs group included people who had an ACE score of 3 or fewer. The cutoff of four was selected based on previous literature which identifies an ACE score of four or more as high risk for negative health outcomes (Hughes et al. 2017).

Results

Participants

As shown in Table 1, participants were evenly split between men (N=15) and women (N=15); no participants identified as transgender. Nearly 40% of participants had an ACE score of four or greater. T-tests revealed that men and women were equally likely to have high ACEs. Stratified of sampling resulted in 76.7% of participants having a mental health diagnosis. Notably, nearly all people with four or more reported ACEs (high ACEs) had a mental health disorder (94.7%). According to t-test results, this figure significantly differed (p = 0.01) from the rate of mental health disorders of people with less than four reported ACEs (low ACEs), which was 54.5%.

Table 1:

Description of Sample

		Total	Gender			ACE Score
			Men	Women	p	High ACEs	Low ACEs	p
N (%)		30.0 (100.0%)	15.0	15.0		19.0 (63.3%)	11.0 (36.7%)
ACE score Mean (std)		3.7 (2.1)	4.0 (2.3)	3.4 (2.0)	0.45	5.0 (1.4)	1.7 (1.3)	*<0.001
Age Mean (std)		38.7 (12.4)	42.0 (12.2)	35.3 (12.0)	0.14	36.2 (12.0)	43.0 (12.3)	0.14
Mental health diagnosis N (%)		23.0 (76.7%)	11.0 (73.0%)	12.0 (80.0%)	0.68	18.0 (94.7%)	6.0 (54.5%)	*0.01
Substance use disorder N (%)		18.0 (60.0%)	10.0 (66.6%)	8.0 (53.3%)	0.47	14.0 (73.7%)	4.0 (36.4%)	0.05
Use for self-medication N (%)		19.0 (63.3%)	8.0 (73.3%)	11.0 (53.3%)	0.27	16.0 (84.2%)	3.0 (27.3%)	*0.001
Regular drug user ever N (%)		19.0 (63.3%)	13.0 (86.6%)	6.0 (40.0%)	*0.01	13.0 (68.4%)	6.0 (54.5%)	0.46
Regular alcohol user ever N (%)		12.0 (40.0%)	7.0 (46.6%)	5.0 (33.3%)	0.47	9.0 (47.4%)	3.0 (27.3%)	0.30
Regular user at time of crime N (%)		17.0 (56.6%)	8.0 (53.3%)	9.0 (60.0%)	0.72	13.0 (68.4%)	4.0 (36.4%)	0.09
Using at time of crime N (%)		14.0 (46.6%)	9.0 (60.0%)	5.0 (33.3%)	0.15	11.0 (57.9%)	3.0 (27.3%)	0.11
Race	White N (%)	14.0 (46.7%)	6.0 (40.0%)	8.0 (53.3%)		9.0 (47.4%)	5.0 (45.5%)
	Black N (%)	6.0 (20.0%)	4.0 (26.7%)	2.0 (13.3%)		3.0 (15.8%)	3.0 (27.3%)
	Hispanic N (%)	10.0 (33.3%)	5.0 (33.3%)	5.0 (33.3%)		7.0 (36.8%)	3.0 (27.3%)
Employment	Fully employed N (%)	15.0 (50.0%)	7.0 (46.7%)	8.0 (53.3%)		9.0 (47.4%)	6.0 (54.6%)
	Under-employed N (%)	3.0 (10.0%)	3.0 (20.0%)	0.0 (0.0%)		3.0 (15.8%)	0.0 (0.0%)
	Unemployed N (%)	12.0 (40.0%)	5.0 (33.3%)	7.0 (46.7%)		7.0 (36.8%)	5.0 (45.5%)

Note:

^*significant at p<0.05; T-tests not conducted when n<15

Over half (60.0%) of participants reported having a substance use disorder, including alcohol use disorder. More than half of participants also reported using substances to self-medicate mental health or trauma symptoms. According to t-test results, people with high ACEs were significantly more likely (p = 0.001) to report using drugs to self-medicate (84.2% versus 27.3%). Over half of participants were also using substances regularly around the time of the crime, with just under half reporting that they were under the influence of substances at the time of the crime. Men and women did not differ significantly except by history of ever using drugs, with men being more likely (p = 0.01) to have ever used drugs regularly. Among all participants, 47% identified as primarily white (non-Hispanic), while 20% identified as black, and 33% identified primarily as Hispanic. Half of participants were under or unemployed at the time of their crime. Under employment was defined as having a job where they did not earn enough money to meet their basic needs.

Overview of Social Determinants

Figure 2 depicts themes relating to how social determinants of incarceration operate. Themes emerged through inductive and deductive strategies. Inductive coding, based on ecosocial theory, provided a skeleton to map participant reported themes. Participant description of how the social determinants operated verified that this structure was appropriate. Pathways detailed in the individual level of the social determinants were also derived deductively. On the left side of Figure 2 is a box containing the social determinants of incarceration. Across the top of this box are the four levels of the theoretical framework relating to ecosocial theory: society, community, interpersonal, and individual. Social determinants reported by participants are shown within light colored boxes that are nested within a darker colored box that is below a subheading which describes the appropriate level for the social determinant. Themes related to society level determinants included lack of opportunity to meet basic needs and discrimination. The primary community level theme was neighborhood factors, while the primary interpersonal level theme was trauma. Individual level social determinants included social isolation, substance use, and mental health symptoms.

Figure 2: Overview of Social Determinants.

Note: Two participants reported being innocent of their crime and are not included in the types of crime. Social determinants categories are adapted from ecosocial theory (Krieger, 2001).

The figure depicts the boxes as nested within each other. This nesting conveys the relationship between the levels of the social determinants as being interconnected. The box containing the society level determinants contains all boxes for other levels of determinants. Nesting demonstrates how societal level determinants, like discrimination, impact all other levels of determinants, such as neighborhood factors, trauma, and social isolation. Nesting also indicates how individual level determinants, like mental health and substance use, are characteristics that are socially determined, for example through traumatic experiences. Nesting in the figure also provides a visual representation for how these individual level characteristics can act as social determinants by impacting social processes. For example, having a substance use disorder may impact how other people perceive a person through a societal determinant of discrimination. Discrimination could then impact the type of neighborhood that a person might live in or directly lead to a trauma. The individual level determinants box depicts the relationships between individual level determinants, as described by participants. The arrows in this section depict how these individual level determinants mediate the impact of the other social determinants to contribute to the criminalized behaviors.

Themes relating to crime and incarceration are shown in another box on the right side of the figure. Themes that participants reported are organized into four types of crime which include: 1) crimes to gain acceptance (18%), 2) crimes to obtain money (36%), 3) crimes due to impaired judgement (39%), and 4) crimes to release emotion (7%). Men and women described engaging in all types of crimes at about the same frequencies. There was variation in how men and women were impacted by the determinants. For example, in crimes to gain acceptance, men sought acceptance from a peer group that was gang involved. For women, this acceptance was more often sought from a romantic partner, but sometimes from a peer group or family member. Both men and women also engaged in crime to obtain money to meet basic needs, particularly if they also had a substance use disorder. For some people, mental health symptoms and or the use of substances (including alcohol) impaired their judgement which led to a crime. However, some other participants reported engaging in crime to regulate emotions, sometimes related to untreated symptoms of a mental health disorder.

The following sub-sections of the results describe social determinants of incarceration by level and provide examples from participants of how these determinants contributed to their incarceration via the various types of crime. Given the nested nature of the levels of social determinants, themes within the more overarching level of society are discussed both within their own section, to provide examples of how they directly contribute to incarceration, and within the other nested sections, to describe how they contribute to incarceration indirectly.

Society Level Social Determinants

Multiple societal level social determinants of incarceration were identified. Overarching themes across gender were lack of opportunity to meet basic needs and discrimination. Basic needs that participants lacked were: transportation, employment, housing, dental, and healthcare including treatment for mental health, and substance use disorders. Participants highlighted a lack of ability to access services due to limited financial resources. They also described how lack of ability to meet basic needs sometimes led directly to their incarceration where for example they might sell drugs to earn money. Other participants reported an indirect relationship, whereby their inability to meet basic needs reduced their ability to obtain employment which led to incarceration. For example, one participant who was a young woman reported the following:

I was born without adult teeth and it’s hard because right now [Medicaid] doesn’t pay for that anymore. I got my first partial [dentures] when I was 12…Say you get [job training] but you don’t have teeth…How a person looks on the outside makes them feel confident…Are you going to be confident, are they going to give you the job? People think it’s all about programs, but it’s all interconnected.

(Participant B)

The above participant also described selling drugs and sex to meet basic needs. This quote highlights a theme where participants reported having complex and interconnected needs which required integrated services and cross-system communication. They reported that a lack of coordination sometimes also indirectly led to incarceration. For example, one man reported that he was incarcerated because he missed an appointment with his probation officer due to being hospitalized.

Stigma based discrimination was highlighted as contributing to social isolation and heightened structural barriers. For example, participants described how stigma was perpetuated on multiple levels and impacted their ability to meet basic needs for acceptance and employment, as highlighted by the following quote from a young man.

It’s the situation where you’re a convict and no one wants to hire you…It took me [9 months] to get a job, because they asked for my background and after that they see you differently. That’s the big issue on how people be coming back. They expect you to be doing good, but society…doesn’t give you a chance to prove yourself…One job he told me honestly, he said I’m not hiring you because of…your record…You see a lot of people mumbling because they see you with tattoos, different. Some people…know they were jail tattoos, so they look at you with a different point of view so like, there’s a criminal right there…The area where I was staying at…was like a little white neighborhood, so they was looking at me with tattoos and looking a different way and it hurt because I seen them pushing they kids to the side, or holding they kids or holding the purse, like at the end of the day I look at it, I’m just like you. I feel, I cry, I bleed, just like you. Maybe I did mistakes in the past but please…come and talk to me. So, I had to go back to what I know, selling drugs and scheming around…the life that I’ve been living since I was 13.

(Participant C)

This theme was echoed by many participants who also reported struggling with multiple intersectional stigmatized or oppressed identities. Participants described the compounding effects of racism and stigma. Stigmatized identities included: having a criminal record, living in poverty, and having a mental health or substance use disorder. For example, participants reported being repeatedly rejected by employers, programs, and service providers. They also experienced daily micro-aggressions, such as those highlighted in the above quote. While these experiences were often implicitly connected to their marginalized identities, sometimes they were overtly told that their existence in a space was unwelcome. These experiences led participants to return to spaces where they felt welcome, such as old neighborhoods with prevalent opportunities for crime.

Community Level Social Determinants

The primary community level social determinant of incarceration that participants described was growing up in neighborhoods with many opportunities to participate in gang activity and a culture impacted by gangs. This was described as a direct pathway to incarceration, as highlighted by the following quote from a young woman.

When I came out the womb I was born gangsta and that’s my lifestyle. It’s something that I will never change. For example, the reason why I am here is that I had a little altercation with somebody on the bus…He axed me where I was from and that’s something I didn’t reveal at the time, because I don’t gang-bang, but I do keep a banger on me, which is also known as a gun or a weapon. He said I’m from [this gang] and what I said was, if I was still from [another gang] I would have killed your ass by now.

(Participant D)

This quote highlights how a subgroup of participants embraced what they referred to as the “lifestyle.” This “lifestyle” was often marked by high rates of violence and or easy access to money, drugs and or sexual partners. Participants described the “lifestyle” as both exciting and in some ways addictive, as one young woman (Participant E) summarized by saying: “I wasn’t addicted to drugs; I didn’t even use them. I was addicted to selling them.” Some older, primarily male, participants reported having lived the “lifestyle” in their youth but having grown out of it. However, they reported feeling lucky, as many of their peers died young. Other participants who reported having grown up in these neighborhoods described their pathway to incarceration as unrelated to this exciting “lifestyle” and instead related to a lack of access to education and employment.

Interpersonal Level Social Determinants

Trauma was the primary interpersonal level social determinant and operated through individual level determinants of social isolation, mental health, and substance use disorders. Both men and women frequently reported that the recent precipitant to their crime was an traumatic event which triggered a relapse of mental health and or substance use symptoms. These traumas were interpersonal in nature, and therefore occurred at the interpersonal level. However, the trauma impacted incarceration by acting at the individual level, primarily through mental health or substance use symptoms. Two exemplars, both from middle-aged men, are shown below:

I wasn’t with my son’s mother anymore, so that was a tough situation. Then I got into a car accident, which was a pretty bad one, one person died…he was a good friend of mine. To see him lying there, it was tough. He was probably my best friend out there. I kind of put the burden on my own shoulders. [Then] I had surgery and I got addicted to Percocet and that’s when I lost my job, because of the Percocet. Once the doctor whittled me off, that’s when I started buying them from the streets and it got real bad. I tried to get help from a psychiatrist and I ended up abusing Suboxone and Klonopin on top of [Percocet]…I went to selling drugs to get the money to buy the drugs. I was selling heroin. I made some mistakes, not spending time with my loved ones and family, they would have helped me get off…but the drugs took over.

(Participant F)

I was sober for many years, but [after my father died] I started using drugs for three or four days…to try and avoid that pain that I was going through…[Because I was high at the time] everything happened so fast that I started doing stuff like robbing people, even though I had money. Which is weird right? But, I didn’t want to go home because I didn’t want my family to see me high like that. So, I started going around and robbing people to buy more drugs and keep going…but I’ve never committed a crime when I wasn’t high.

(Participant G)

These examples highlight how traumatic loss of a key support person led to a relapse of mental health symptoms and substance use. Neither of these participants had ever committed crimes when not using, and both only committed crimes to obtain money to access drugs. Both also highlighted how shame, stemming from internalized stigma, kept them from seeking help from their remaining support systems. This pathway to crime represented about a fifth of participants.

Individual Level Social Determinants

Primary themes reported as individual level social determinants included social isolation, poor mental health, and or substance use. Participants primarily described social isolation as contributing to crime either by leading to crimes to gain acceptance, or by contributing to substance use and mental health symptoms. Substance use was reported to contribute to crimes to obtain money, and to access drugs. Both substance use and mental health symptoms contributed to impaired judgement. For example, one middle-aged male participant stated:

[The problem is] just trying to keep my connection with the people in my life. Because it’s when I start isolating it’s when I start feeling alone and I get depressed and then my anxiety starts running away with everything and I don’t make the right decisions and then I lean toward drugs and as soon as I start doing drugs all else is off the table. I don’t feel anything. I don’t think about consequences. I just want to stay in that state of mind. So, then I do whatever I have to do to keep getting access to more drugs.

(Participant H)

Symptoms of mental health disorders impacted participants by impairing their judgement and through the direct effects of symptoms. Two participants with serious mental illness reported limited insight into crime occurring around them, by family members or peers. Their peripheral involvement in these activities eventually led to their incarceration. Another described difficulty managing rigid obligations of probation. Finally, another group engaged in crime to regulate emotions related to symptoms. For example, a male participant with a serious mental illness reported engaging in crime to self-regulate distressing emotions and symptoms of his serious mental illness. He reported that engaging in risky, violent, or dangerous behaviors was a form of psychological release. He also reported being comfortable with incarceration and feeling a sense of belonging in prison, where he was well known due to multiple incarcerations. He described the world outside prison as overwhelming, isolating, and confusing. Within the prison he reported feeling at ease and less isolated due to the rigid structure and predictable human contact.

Another participant who was a middle-aged woman described how her crime was an act that discharged her anger and frustration, but also was a mechanism of resistance in the face of feeling powerlessness within an oppressive system. This participant stated:

I’m a well-educated person. I have my undergrad in accounting and my MBA so I’ve spent my whole career as an accounting manager. I had a great life, still have a great life, very blessed and so I’ve been going back thinking why did I make such a horrible choice, because it’s just not like me. My charge is [felony larceny], I knew when I did it I would go to jail. Thinking about what led me down that path, it was just a really frustrating place to work, safety violations, sexual harassment, people not getting paid overtime that they were entitled to and in my position, it was my job to fix those things. So, going to the company president and having him tell me to mind my own business…I was so frustrated that it was my way of getting back at him.

(Participant I)

While these two narratives are similar in that the crime allowed them to release emotion, the characteristics of the participants were different. The first had significant history of trauma, a serious mental illness, and multiple incarcerations, while the second had almost no childhood trauma, no diagnosed mental health or substance use disorder, and no previous legal involvement. However, both experienced the same relief in emotional distress through committing a crime.

Gender Variation

While there were many similar themes across gender, there were also some themes that were reported more by gender. For example, men more frequently reported coping with social isolation by joining gangs and acting violently to affirm their manhood. Women more often reported maintaining unhealthy relationships, frequently characterized by domestic violence by male romantic partners. However, unhealthy relationships with male family members and female partners were also reported. Both men and women experienced high rates of mental health symptoms and using substances to cope with both symptoms and social isolation. Additionally, men were more likely to report having been formerly and repeatedly incarcerated. They described how this process eroded their social networks and led to social isolation upon release. Additionally, men were more likely to have been homeless, prior to their current incarceration. Women also reported unstable housing as a substantial problem. Women frequently reported meeting their housing needs through maintaining an unhealthy relationship usually with a male partner. Consequently, women were less likely to report social isolation, but were more likely to report unhealthy relationships. For women, these relationships, reportedly, led directly to their incarceration. For example, one woman who had a serious mental illness was charged in conjunction with her male relative who had used her apartment for human trafficking. The woman appeared to have limited insight to what had been happening in her apartment and her male relative testified against her in trial.

A subset of women (four out of 15) repeatedly referred to themselves as “midlife mavericks,” which they reported was the name of a group that was offered at the prison. This subgroup was comprised of women in midlife who were incarcerated for the first time. They were also typically middle income, and half were serving long sentences for serious crimes. These women reported lower levels of mental health and substance use disorders. Their pathways to prison were often characterized by either a recent significant adverse experience, or an exploitative relationship. One middle-aged woman described her situation as follows:

I’m not in here for drugs. I have never done drugs. I am in here because I took money from my employer…I had finally gotten rid of my husband. It was 28 years of abuse, mental and physical…After that I was by myself…I was on every dating site imaginable and thought I was talking to two different people…and then they started asking for things…and I ended up taking money from my employer and giving it to them.

(Participant J)

This woman’s story highlights how some of the participants had significant adverse experiences, but no subsequent development of a mental health disorder. However, this adverse experience, combined with social isolation, created a vulnerability to exploitation.

Men also more frequently discussed the experience of incarceration as inherently traumatic, with one young man stating, “the motto inside the jail is you eat, or you get eaten” (Participant C). Men described the role of toxic masculinity, in creating a dangerous prison environment and as a pathway to prison. For example, men described how they saw their worth as inherently tied to their ability to financially support a family. When presented with limited legal job opportunities, almost half of men reported turning to a “criminal lifestyle,” to obtain access to basic resources and develop a sense of manhood or self. However, women who were not impacted by the previously described neighborhood factors also reported being quite traumatized by their incarceration, primarily due to difficulty adjusting to the deprivation of incarceration.

Interrupting Pathways to Incarceration

Proposed prevention policies and programs focused on early intervention for substance use and adverse childhood experiences. Additionally, participants called for increased access to quality comprehensive healthcare, more individualized cross-system services, and increased services for domestic violence victims especially in rural areas. Services inside prisons were reported to be both helpful and sometimes not enough, as the following quote from a middle-aged man describes.

People with drug issues should have priority to go into treatment, especially people that have shorter sentences…work with those people first, give them the help they need, so when they go out [they can] be a better person. [When I was in prison before] they never asked me about drug problems, but this time they did ask me…and they’ve got some programs now, so they referred me. I even asked the judge for help [with my drug problem] and she said I’d get help. What they are doing now is different…they realize there is an epidemic going on, that people are committing crimes not because they want to but because they need it to get the drugs and they can’t control it…They should have even more programs that are even longer…and keep you learning the whole time.

(Participant G)

Another participant called for an in-prison program to teach men how to be fathers and expose them to healthy male role models to help them understand how taking care of others is what makes a “real man.” Another participant called for a mentoring model for women, that would similarly allow participants to feel connection and purpose. Multiple participants reported a deep need for prisons to undergo a culture shift to be more humane and trauma informed. For example, one participant, who was a young woman, reported a need to “help people feel like they are something in the world; like you are important. If someone did that for me I would be like, wow am I a person?” (Participant B)

Participants also reported a need for continuous services, before, during, and after prison. They reported a need for services that could help them build long term relationships with service providers and services that support their other relationships. Access to religious services and strong community connections were cited by multiple participants as factors that had helped them both manage their symptoms and stay away from crime. Participants reported that having a supportive community buffered them against the effects of social isolation. Participants also reported a need for job training and pre-release job programs in prison. About a fifth of participants reported having had caregiving or parental duties as a child. They reported that more emotional and financial support for their family when they were a child could have been helpful. For example, additional tutoring, mental health services, childcare for themselves and siblings, and additional positive adult role models were called for specifically. Addressing racism and stigma against those with a criminal record and or mental health or substance use disorder was also highlighted as something that contributed to social isolation and limited access to resources particularly employment opportunities.

Discussion

This study provides guidance on how to address incarceration based on people’s own lived experience. Results of this study also improve knowledge of how to address adverse experiences within this population. This knowledge may help providers develop trauma informed programs or strategies to prevent further victimization and promote rehabilitation. Findings may also inform policies that address the root causes of crime and improve public health outcomes.

Theoretical Implications

The study’s conceptual framework was supported by the data and described how ecosocial theory can be applied to explain determinants of incarceration. For example, in comparing the a priori conceptual framework (Figure 1) to the figure summarizing the findings (Figure 2), the themes that emerged deductively both mapped to the levels of social determinants, and further explained how the social determinants operated. Findings also built on previous research which identified environmental factors that impact crime, such as physical/architectural design, zoning boundaries, and available green space. Previous research in this area has documented how features of the built and community environment-such as abandoned properties and presence of intoxicated people-are associated with increased rates of violent crime (Furr-Holden et al. 2008). This study also further described how previously identified pathways function and connected environmental determinants to social determinants within the conceptual framework. For example, by describing how racism and stigma contributed to crime by exacerbating social isolation.

Social Determinants of Incarceration

This study also provides context to situate previous research linking poverty and crime. While previous studies have documented a correlation between incarceration rates and neighborhood disadvantage (Friedson and Sharkey 2015), this study describes how individuals navigated from these environments into prison. A question raised by this study is how social isolation might be connected to neighborhood disadvantage. Previous studies have described how concentrated police oversight and high neighborhood rates of incarceration can contribute to the erosion of social cohesion within neighborhoods (Friedson and Sharkey 2015). However, this study is unable to describe how this relationship might operate. Future research is needed.

The link between social isolation, mental health, substance use disorders, and crime has been documented by others. A quantitative study of the neighborhood context of youth alcohol and drug problems found that within neighborhoods, lower levels of social cohesion were associated with higher levels of perceived alcohol and drug use by youth which in turn was associated with youth drug and alcohol arrests (Duncan, Duncan, and Strycker 2002). This study both supports previous research and adds to it by describing how these relationships operate. Participants described how social isolation leads to depression and anxiety that contributed to substance use as an escape. Participants in this study also described how mental health and substance use disorders were connected to their crimes, either through inhibited judgement or emotional release.

Pathways to Prevent Incarceration

The viability and importance of investing in up-stream prevention strategies that are often considered outside the traditional scope of the criminal justice system was also described by participants in this study. For example, addressing health and educational needs were frequently described as having enormous potential for disrupting early pathways to prison. This builds on previous research which demonstrates the effectiveness of specific models in preventing or reducing crime, such as after school programming (Gottfredson et al. 2004). Given that an overarching theme from this study was social isolation particularly related to poverty additional community building and economic opportunities could be helpful. Another potentially viable pathway to prevent incarceration that was raise by this study is addressing childhood adversity through early intervention to strengthen families and communities. Future research should incorporate these strategies for prevention into the tested conceptual framework described in this study. Such research could provide guidance on how investment in public health could prevent crime. Furthermore, findings demonstrate the need for researchers to collaborate across the fields of criminal justice and public health. On a policy level it also provides context for how criminal law reform that acts at the root cause of crime could also improve health. However, additional studies should continue to investigate this area.

Gender Variation

While this study was qualitative in nature with a small sample size, it still described variation by gender. For example, women described the role that unhealthy relationships played as a pathway to their incarceration. These findings support previous research which described the role that male partners play in the incarceration of women (Kennedy and Mennicke 2018). However, this study also builds on those findings by describing the role of other relationships in women’s pathway to incarceration. While findings generally supported previous research, which documents gendered pathways to prison (Salisbury and Van Voorhis 2009), they also suggest that the role adverse experiences play for incarcerated men may be underestimated in previous literature. Additionally, within this study men were also able to identify how toxic masculinity contributed to social isolation and caused harm to them by disrupting their ability to form nurturing relationships with others. Findings build on previous research which has documented the “hypermasculine” culture in men’s prisons (Jewkes 2005). Men in this study specifically called for both a cultural shift and increased programming to assist men with their need to build interpersonal capital. Additional research which examines specific programs in this area is needed.

Limitations

Despite this study’s strengths it is not without limitations. For example, interviews were conducted in one state at a single point in time. Given this, factors relating to the historical point in time or regional variation could have impacted participant responses. For example, given the disproportionate impact of the opioid epidemic in Massachusetts during this time (Kaafarani et al. 2017) there is some potential for a slight inflation in the role of opioids in crime. However, only a few participants described opioids as playing a role in their crime.

This study also relies on self-reported data which can be subject to bias. Limiting participants to only those people within 6-months of admission should have mitigated recall bias as much as possible. Additionally, recall bias in relation to remembering adverse childhood experiences, actually leads to an under reporting of experiences (Hardt and Rutter 2004) which makes results a conservative estimate. Other forms of response bias could have also impacted study findings. For example, given the retrospective nature of the questions, and the targeting of a population with mental illness, mood congruency bias, could be an issue. Mood congruency bias is the situation where respondents experiencing negative moods are more likely to recall negative experiences (Roxburgh and MacArthur 2014). Bias from this factor would likely influence the results, but in an opposite direction from the recall bias, and therefore would likely not override the overall effects, but may influence the magnitude. Finally, social desirability bias might have also influenced the responses. Given the prison setting, respondents may have felt stronger than usual pressure to appear tough, thereby underreporting experiences of victimization, or abuse. However, the additive impact of the Hawthorne effect may have caused respondents to feel a pressure to overinflate adverse experiences, which could nullify the effects of social desirability bias.

Reporting biases are inherent in nearly all qualitative and survey-based research. Therefore, given that this method is the most appropriate way to collect this information, some amount of response bias is to be expected. The degree of response bias in this study is not anticipated to be so great as to skew the direction of the results. Given these many strengths and the unavoidable nature of many of the limitations, the strengths of this study outweigh the limitations. However, policy action based on these results should also consider other evidence. In particular, other evidence from jail-based populations should be considered, as there are fundamental differences between the populations related to psychological symptoms and functioning (Bronson, Maruschak, and Berzofsky 2015). Finally, future research should seek to explore these relationships by using quantitative modeling with a larger and more regionally diverse samples to triangulate findings.

Conclusion

This study provides evidence of how incarcerated people understand the role that adverse experiences, mental health, and substance use disorders play as social determinants of their incarceration. Additionally, this study described what incarcerated people understand to be potentially effective programs and policies to address these social determinants. For example, an overarching theme in this area was the importance of enhancing cross-system preventive and rehabilitative policies and programs to reduce rates of incarceration. These activities could be targeted to the social determinants of incarceration identified by the participants, including increased access to quality, affordable, individualized medical, dental, mental health, and substance use services. Other social services to improve housing and employment opportunities were identified as potential pathways to prevention of crime. Looking forward, policy efforts to create incentives to improve inter-system collaboration could also be helpful in addressing these individual needs on a systemic level.

Broader cultural implications include a need to build communities to alleviate social isolation and create support systems that act as a safety net for individuals experiencing adversity. Community building should also include efforts to break down stigma and racism that served as barriers to participants’ community integration. Policies and programs to address community building should be considered as viable alternatives to direct justice reinvestment funding. Justice reinvestment is a relatively new policy approach that aims to reinvest spending on the drivers of crime to reduce costs, increase evidence-based services, expand the types of programs aimed at addressing crime, and reduce harm caused by both crime and the criminal justice system (Taxman, Pattavina, and Caudy 2014). Results from this study offer evidence regarding social determinants of crime that could inform justice reinvestment policies. Such reinvestment could also serve as a preventive measure to address the negative downstream health consequences of legal involvement.