Abstract
This study examines the association between psychological distress and two aggravating factors (childhood adversity and substance use) and two mitigating factors (social support and resilience) in a correctional sample of 943 men. Participants completed a questionnaire probing psychiatric distress using the DASS (Depression, Anxiety, and Stress Scale) and substance use behaviors using the ASSIST (Alcohol, Smoking, and Substance Involvement Screening Test). Adverse childhood experiences (ACEs) were measured incrementally and cumulatively. ACEs were reported by a majority (57.6%) of the sample. The number of ACEs and childhood emotional, physical, and sexual abuse experiences were strongly and consistently associated with prison-based psychological distress. The graded relationship found between ACE and psychological distress among incarcerated men supports the cumulative risk hypothesis. Prison-based substance use was positively associated with psychological distress, whereas resilience and, to a lesser extent, social support were negatively associated with distress. These findings can be used to inform the design of trauma-sensitive integrated interventions in correctional settings for men.
Keywords
Introduction
History of childhood trauma, such as experiences of neglect and abuse during childhood, heightens the risk of adverse adult health, behavioral health, and psychosocial outcomes (Anda et al., 2006; Shonkoff et al., 2012). Research conducted in the United States and other countries consistently finds strong associations between childhood adversity and psychiatric distress (Merrick et al., 2017), substance abuse (Dube, Anda, Felitti, Edwards, & Croft, 2002; Dube et al., 2003), and incarceration (Elklit, Karstoft, Armour, Feddern, & Christoffersen, 2013; Mersky, Topitzes, & Reynolds, 2012; Roos et al., 2016). Hence, it should not be surprising that dozens of studies consistently document the overrepresentation of mental health (Baranyi et al., 2019; Bronson & Berzofsky, 2017; Fazel, Hayes, Bartellas, Clerici, & Trestman, 2016; Fazel & Seewald, 2012; Gottfried & Christopher, 2017) and substance abuse disorders (Baranyi et al., 2019; Fazel, Yoon, & Hayes, 2017), as well as trauma exposure (Drury et al., 2017; Wolff, Shi, & Siegel, 2009), among incarcerated men and women.
Although prison conditions and services can promote healing (Morgan et al., 2012), for some, prison can be an unsafe place (Gibbons & deBelleville Katzenbach, 2006). Mental illness, in particular, is a risk factor to several adverse outcomes in prison. Incarcerated people with mental illnesses are at higher risk for adverse outcomes, such as suicide (Rivlin, Hawton, Marzano, & Fazel, 2010; Sánchez, Fearn, & Vaughn, 2018), self-harm (Fotiadou, Livaditis, Manou, Kaniotou, & Xenitidis, 2006; Hawton, Linsell, Adeniji, Sariaslan, & Fazel, 2014), victimization (Blitz, Wolff, & Shi, 2008; Steiner, Ellison, Butler, & Cain, 2017; Wolff, Blitz, & Shi, 2007), and placement in segregation (Labrecque, 2018).
Childhood trauma also has been linked to the mental health problems reported by incarcerated people (Briere, Agee, & Dietrich, 2016; Debowska & Boduszek, 2017; Driessen, Schroeder, Widmann, von Schonfeld, & Schneider, 2006; Wolff & Shi, 2012). In a study based on a sample of approximately 7,000 male residents in state prisons, more than half (57%) reported childhood physical abuse and one in 10 reported childhood sexual abuse (Wolff et al., 2009). Similarly, Caravaca-Sánchez and Wolff (2018), based on a sample of 2,484 male prison residents, found emotional abuse was the most frequently reported type of prison-based abuse, with more than one third of male respondents reporting emotional abuse during the past 6 months of their incarceration. This same study found the risk of experiencing multiple types of victimization in prison increased significantly with childhood emotional abuse. Post-traumatic stress disorder (PTSD) is a psychiatric morbidity predicated on prior trauma. Baranyi, Cassidy, Fazel, Priebe, and Mundt (2018), in a meta-analysis of 36 studies, estimated the prevalence rate of PTSD among incarcerated men to be 6.1%, compared with 3.5% to 5% for community samples (Breslau, 2009; Kessler et al., 2005).
Substance use disorder and mental illness often co-occur (Peters, Wexler, & Lurigio, 2015; Sung, Mellow, & Mahoney, 2010). Vaeroy (2011) found strong correlations among depression, anxiety disorder, and substance use disorders using an incarcerated male sample drawn from Norway prisons. Rates of community- and prison-based substance use were also found to be positively associated with depression, anxiety, and stress symptoms during incarceration for men residing in Spanish prisons (Caravaca-Sánchez & Wolff, in press). More generally, based on a survey of residents in state and federal prisons in the United States, James and Glaze (2006) found that nearly three quarters of state prison residents reporting a mental health problem also met the criteria for substance dependence or abuse.
Although the epidemiological evidence on behavioral health disorder and trauma exposure among incarcerated persons is growing worldwide, this literature has focused more on documenting the presence of behavioral health disorder within correctional populations and less on the factors that foster or inhibit adverse behavioral health outcomes of people while they are incarcerated (Wolff & Shi, 2012). The paucity of knowledge about behavioral health symptoms inside prison, in light of robust findings that link childhood adversity and maladaptive coping, warrants careful investigation. If exposure to childhood adversity increases adult sensitivity to psychological distress, then incarceration, as a potentially stressful and harmful experience, may also increase the expression of anxiety and depression symptoms, heightening, in some cases, already preexisting symptoms.
Given that trauma and chronic stress are antecedents to psychiatric disorder and that conditions of privation and isolation associated with incarceration are expected to elevate stress and the risk of trauma (Aneshensel, Rutter, & Lachenbruch, 1991; Murali & Oyebodde, 2004), the ability of prison residents to cope with these conditions may be mitigated by the support they receive from friends and families in the community. Social support has been identified as a key environmental resource promoting positive adaptive change to life crises and transitions (Schaefer & Moos, 1998) and traumatic events (Tedeschi & Calhoun, 1996). In a meta-analytic review of 113 studies of change following a traumatic event, Prati and Pietrantoni (2009) found that social support moderately predicted positive change in posttraumatic growth. Research conducted by the World Health Organization (2012) found a strong association between mental health and immediate social context, with social support from family and friends having an abating effect on mental health determinants. More recently, using a sample of college students, Huang, Costeines, Kaufman, and Ayala (2014) found that social support moderated depression and anxiety.
In terms of prison adjustment, social support, inclusive of financial and emotional resources, may serve as a mitigating factor if generosity and engagement provided through social connections help the resident cope with and adjust to prison life. The impact of social support on prison adjustment has only recently been explored in the correctional literature. Steiner and colleagues (2017) examined the relationship between social support and segregation placement and found that prison residents with more family/friend visits were less likely to experience disciplinary segregation. Caravaca-Sánchez and Wolff (in press) examined the association between social support and substance use during incarcerated. In their study, residents reporting substance use in prison also reported lower levels of social support. Relatedly, Listwan-Johnson, Colvin, Hanley, and Flannery (2010), studying 1,600 formerly incarcerated men, found that social support was positively associated with psychological well-being. Similar results were reported by Lee, Guilamo-Ramos, Munoz-Laboy, Lotz, and Borrnheimer (2016). They found that social support prevented recidivism and promoted desistance after release from prison. These findings suggest that social support may provide benefits through emotional and filial connections that abet people while incarcerated to endure and manage prison stress. Having people in the community who care enough to stay connected may be the “port in the storm” that brings relief or the “lighthouse” that keeps them away from hazards.
Social support focuses on external resources, whereas resilience emphasizes internal reserves. The American Psychological Association (2014) defines resilience as “the process of adapting well in the face of adversity, trauma, tragedy, threats or even significant sources of stress” (p. 1). In this sense, resilience acts like a defense mechanism or an adaptive immune system that helps people cope with life’s threats and challenges (Patel & Goodman, 2007). Resilience research concentrates on an individual’s ability to draw on internal reserves, such as positive attitude, optimism, self-esteem, to “rise up” and deal with adversity in healthy ways (Rutter, 1987). Community-based research on resilience and psychological distress shows that resilience decreases the risk of being depressed or stressed or anxious (Färber & Rosendahl, 2018; Komiti et al., 2003; Ryden, Karlsson, Sullivan, Torgerson, & Taft, 2003; Southwick, Litz, Charney, & Friedman, 2011). Resilience, a measure of personal coping resources that facilitate the ability to “bounce back,” also has been found to moderate psychological distress. For example, among intimate partner victims, resilience was inversely and strongly associated mood symptoms and perceived stress, controlling for social support (Jose & Novaco, 2016). In addition, researchers studying cancer patients found that patients with higher resilience scores had lower levels of psychological distress, again controlling for social support (Matzka et al., 2016). It is expected that endowments of resilience will provide similar benefits to incarcerated people as they attempt to manage the difficulties of privation concomitant with incarceration.
Study Focus and Hypotheses
This study examines psychological distress among a sample of men incarcerated in Spain. Our focus is on men (exclusive of women) because men are disproportionately represented in prison (more than 90% of the correctional population) and because of important gender differences related to the nature of trauma and its prevalence and impact (Covington, Burke, Keaton, & Norcott, 2008; Dembo, Williams, Wothke, Schmeidler, & Brown, 1992). The aim is to test three hypotheses:
Method
Participants and Procedures
Cross-sectional survey data were collected from January through March 2017 in three prisons located in the southeast of Spain. The security levels of the selected prisons were medium (2) and minimum (1), and collectively they housed 1,800 men (Spanish Prison System, 2018). Of those eligible, 1,150 residents met all three of the following inclusion criteria: (a) resided in general population, (b) were incarcerated for at least 3 months, and (c) literate in Spanish. In total, 188 (16.3%) residents refused to participate in the survey, leaving 962 residents completing the questionnaires (2.0% of the completed questionnaires were dropped because of missing items on the psychological distress and social survey instruments).
The final study sample consisted of 943 male residents ranging from 19 to 83 years of age (M = 37.2, SD ±12.3). The sample was representative of the overall male Spanish prison population at the end of 2018 (N = 54,786; Spanish Prison System, 2018). Study participants were most likely to be Spanish (78.2% and 73.1%, study sample and prison system, respectively), incarcerated for a property offense (36.5% and 34.7%, study sample and prison system, respectively), and incarcerated for the first time (67.1% and 63.8%, study sample and prison system, respectively). In addition, participants (N = 943) were most likely not to be in an intimate relationship (52.1%) and more likely to be poorly educated (64.7%) and unemployed prior to prison (58.0%). In terms of prison-based substance use, nearly one third (31.1%) of respondents reported using drugs (including cannabis, cocaine, amphetamine-type stimulants, inhalants, sedatives, hallucinogens, or opioids) and more than one fifth (21.5%) reported using alcohol.
Survey questionnaires were administered in a private room and completed in groups of approximately 15 residents; the administration process was observed by research staff (social workers). All participants answered questions arranged in the same order. Surveys were rapidly administered by unit to minimize contamination bias. The questionnaire order began with the adapted Childhood Trauma Questionnaire, followed by the ASSIST (Alcohol, Smoking, and Substance Involvement Screening Test), DASS-21 (Depression, Anxiety, and Stress Scale-21), MOS (Medical Outcomes Study) Social Support Survey, and Brief Resilience Scale (BRS) instruments. It took participants approximately 45 to 60 min to complete the questionnaire. Completed questionnaires were returned directly to research staff. The Spanish Prison System and the University of Murcia review boards and committees approved the protocols for the current study.
Measures
Psychological Distress (Criterion Variable)
Psychological distress during the past week while incarcerated was measured using the DASS-21, developed by Lovibond and Lovibond (1995). The DASS-21 instrument is divided into three scales: Depression (DASS-21D), Anxiety (DASS-21A), and Stress (DASS-21S), with unique scores constructed for each scale. There are seven items for each scale and each item response ranges from 0 (“did not apply to me at all”) to 3 (“applied to me very much or most of the time”). Each item score within a scale (Depression, Anxiety, and Stress) is doubled to yield a scale range of 0 to 42. The DASS-21, combined Depression, Anxiety, and Stress, ranges from 0 to 126. The psychometric properties of the DASS-21 are strong based on studies of adults with anxiety and/or mood disorders (Antony, Bieling, Cox, Enns, & Swinson, 1998; Brown, Chorpita, Korotitsch, & Barlow, 1997; Clara, Cox, & Enns, 2001). The Spanish translation of the DASS-21 has been validated (Daza, Novy, Stanley, & Averill, 2002).
ACE (Predictor, Aggravating Factor)
The childhood adversity items were adapted from the Childhood Trauma Questionnaire (Bernstein & Fink, 1998). Participants were asked about the following seven ACEs (answering yes or no to each experience that occurred prior to age 18): (a) a parent was often too drunk to take care of the family, (b) a parent was often too high to take care of the family, (c) a parent had died, (d) a parent had been incarcerated; and they had experienced, (e) emotional abuse, (f) physical abuse, and (g) sexual abuse. An affirmative response to an ACE was assigned a score of 1 (otherwise 0) and were added together to create various ACE variables. The cumulative ACE variable ranged from 0 to 7.
Substance Use (Predictor, Aggravating Factor)
Adapted from the Spanish version of the ASSIST (V3.0; Organización Mundial de la Salud, 2011), participants were asked about their substance use in the past 3 months while incarcerated. Substances were grouped into two categories: (a) any alcohol and (b) any drug including cannabis, cocaine, amphetamine-type stimulants, inhalants, sedatives, hallucinogens, or opioids. Use of any drugs was coded as 1 (0, if “no”), indicating prison-based substance use. Any use of alcohol during the past 3 months was coded as 1 (0, if “no”), indicating prison-based alcohol use. The ASSIST instrument has strong psychometric properties with high internal consistency in English (Humeniuk et al., 2008; World Health Organization ASSIST Working Group, 2002) and Spanish (Valladolid et al., 2014).
External and Internal Support (Predictor, Mitigating Factor)
External social support was measured using the MOS Social Support Survey (SSS; Sherbourne & Stewart, 1991). The SSS has 19 items scored using a 5-point Likert-type scale (1 = “none of the time,” 2 = “a little of the time,” 3 = “some of the time,” 4 = “most of the time,” and 5 = “all of the time”) and aggregates to an overall functional Social Support Scale (range = 19-95) that can be decomposed into four dimensions of support: emotional (range = 8-40), tangible (range = 4-20), interactional (range = 4-20), and affectionate (range = 3-15). Higher scores indicate more social support. The SSS English (Giangrasso & Casale, 2014) and Spanish (Requena, Salamero, & Gil, 2007) versions have strong psychometric properties. This measure has been previously administered to a Spanish prison sample (Rodríguez-Martínez et al., 2010). Resilience, internal support, was measured by the BRS (Smith et al., 2008). This instrument comprises six items assessed using a 5-point scale (1 = “strongly disagree,” 2 = “disagree,” 3 = “neutral,” 4 = “agree,” and 5 = “strongly agree”), with a range of 6 to 30. The English (Smith et al., 2008) and Spanish (Rodríguez-Rey, Alonso-Tapia, & Hernansaiz-Garrido, 2016) versions of the BRS have strong psychometric properties.
Sociodemographic (Control Variables)
Data were collected on age (continuous variable), nationality (Spanish or foreigner), and marital status (married or in a relationship or single, divorced, or widowed).
Statistical Analyses
Univariate, bivariate, and multivariate analyses were conducted using SPSS statistical software Version 22 with a minimum significant level set at 95% (p ≤ .05). Univariate and bivariate analyses were conducted to examine the distribution of ACEs (Table 1) and the intercorrelations among predictor and criterion variables (Table 2). Ordinary least square (OLS) linear multiple regression analysis was used to examine the association between psychological distress, the criterion variable, and the following predictor variables: ACE (Hypothesis 1), any drug use in prison, any alcohol in prison (Hypothesis 2), social support, and resilience (Hypothesis 3). The estimated models differ in their specification of the criterion variable, which is a continuous measure of depression (Model I, range = 0-42), anxiety (Model II, range = 0-42), stress (Model III, range = 0-42), and overall psychological distress, DASS-21 (Model IV, range = 0-126). Each model is estimated with a unique specification of the ACE and social support variable: ACE and social support are aggregated in Table 3 (Models I-IV), ACE and social support are disaggregated into their different types (testing the effect of unique types of experiences) in Table 4 (Models I-IV), and ACE is disaggregated into the number of different types (testing the dosage effect) and social support is disaggregated into its different types in Table 5 (Models I-IV). Estimated beta coefficients are reported unstandardized and standardized. For the unstandardized coefficients, the binary variables are interpreted as the expected response difference between the defined category and the reference category, ceteris paribus. All models control for age (younger participants in an institutional environment may experience higher levels of distress), nationality (being a noncitizen may contribute to distress), and marital status (being isolated from a spouse/partner may elevate distress), characteristics that vary within the sample and might be related to the criterion variables. Controlling for these variables better isolates the relationship between the criterion and predictor variables. The mean differences in the criterion variables among the prisons was significant only for the stress variable (p = .047). For this reason, no dummy variables were added for facility in the linear regression models. In terms of the regression analysis, each model is estimated separately, has a unique dependent variable, and all independent variables are entered into the regression equation at once.
Distribution of Different Types of ACEs and Cumulative Number of Different Types of ACEs Reported by Male Participants (N = 943)
Note. ACE = adverse childhood experience.
Pearson Correlation Coefficients Among Predictor and Criterion Variables for Full Sample (N = 943)
Note. D-21 = Combined Depression, Anxiety, and Stress DASS-21 scores; DEP = Depression; ANX = Anxiety; STR = Stress; ACE = adverse childhood experience; ACE 1 = parent drank alcohol excessively; ACE 2 = parent used illegal drugs excessively; ACE 3 = a parent died; ACE 4 = a parent was incarcerated; ACE 5 = experienced emotional abuse; ACE 6 = experienced physical abuse; ACE 7 = experienced sexual abuse; 1 ACE = one type of ACE; 2 ACEs = two different types of ACEs; 3 ACEs = three different types of ACEs; 4+ ACEs = four to seven different types of ACEs; SS, Agg = Social Support Aggregate; SS-Emo = Social Support Emotional; SS, Tan = Social Support Tangible; SS-Int = Social Support Interactional; SS-Aff = Social Support Affectionate; RES = resilience; P-B Drg = Prison-based Drug use; P-B Alc = Prison-based Alcohol use.
p < .05. **p < .01.
OLS Linear Multiple Regression by Psychological Distress and ACE, Prison-Based Substance Use, Social Support, and Resilience (N = 943)
Note. OLS = ordinary least square; ACE = adverse childhood experience; DASS-21 = Depression, Anxiety, and Stress Scale-21.
Any drug use in prison includes cannabis, cocaine, amphetamine-type stimulants, inhalants, sedatives, hallucinogens, or opioids. bCompared with no use of drugs in past 3 months while incarcerated. cCompared with no use of alcohol in past 3 months while incarcerated. dSpanish compared with foreigner. eMarried or in an intimate relationship compared with single/divorced/widow.
p < .05. **p < .01. ***p < .001.
OLS Linear Multiple Regression by Psychological Distress and ACEs, Prison-Based Substance Use, Social Support, and Resilience (N = 943)
Note. OLS = ordinary least square; ACE = adverse childhood experience; DASS-21 = Depression, Anxiety, and Stress Scale-21.
Any drug use in prison includes cannabis, cocaine, amphetamine-type stimulants, inhalants, sedatives, hallucinogens, or opioids. bCompared with no use of drugs in past 3 months while incarcerated. cCompared with no use of alcohol in past 3 months while incarcerated. dSpanish compared with foreigner. eMarried or in an intimate relationship compared with single/divorced/widow.
p < .05. **p < .01. ***p < .001.
OLS Linear Multiple Regression by Psychological Distress and ACEs, Prison-Based Substance Use, Social Support, and Resilience (N = 943)
Note. OLS = ordinary least square; ACE = adverse childhood experience; DASS-21 = Depression, Anxiety, and Stress Scale-21.
1 ACE = one type of adverse childhood experience; 2 ACEs = two different types of adverse childhood experiences; 3 ACEs = three different types of adverse childhood experiences; 4+ ACEs = four to seven different types of adverse childhood experiences, each compared with 0 ACE. bAny drug use in prison includes cannabis, cocaine, amphetamine-type stimulants, inhalants, sedatives, hallucinogens, or opioids. cCompared with no use of drugs in past 3 months while incarcerated. dCompared with no use of alcohol in past 3 months while incarcerated. eSpanish compared with foreigner. fMarried or in an intimate relationship compared with single/divorced/widow.
p < .05. **p < .01. ***p < .001.
Results
Descriptive Analyses
Table 1 shows the distribution of ACE in terms of the different types of adverse experiences and the cumulative number of separate types for the full sample of 943. Overall, well, more than half of the male respondents (57.6%) reported at least one ACE and the mean number of different types of ACEs was 1.2. More than one in four men reported experiencing physical abuse (25.0%), emotional abuse (24.5%), and the death of a parent (20.8%) prior to age 18. Approximately four in 10 men (38.2%) reported some form of emotional, physical, or sexual abuse during childhood. A plurality (28.6%) of all male respondents reported one type of ACE, with another 28.9% reporting two or more different types of ACEs.
The intercorrelations among psychological distress measures, ACE measures, prison-based drug and alcohol use, social support, and resilience are shown in Table 2. According to Cohen’s (1988) classification for effect size, associations between two continuous variables with a value of .10 are considered a weak/small association, .30 moderate, and .50 strong/large. Strong correlations were found between DASS-21 and depression, anxiety, and stress; depression and anxiety; depression and stress; and anxiety and stress. Intercorrelations were the strongest between the number of ACEs and any ACE and different types of ACEs and higher correlations except for ACE 3 (a parent died) and ACE 4 (a parent incarcerated), number of ACE and four or more ACEs. Four or more ACEs were also strongly correlated with any ACE, childhood physical abuse, and childhood sexual abuse. The overall social support variable was strongly correlated with its constitute dimensions, emotional, tangible, interactional, and affectionate, and there were strong correlations between interactional support and emotional support and tangible support. Cronbach’s alpha for the resilience scale was .87 and for the social support scale ranged from .87 (full scale) to .62 (affectionate).
Multivariate Analyses
Table 3 shows results from the OLS linear multiple regression models (each estimated separately) exploring the association between different continuous measures of prison-based psychological distress and aggravating predictor factors (total number of different types of ACEs ranging from 0 to 7 and prison-based drug and alcohol use) and mitigating predictor factors (overall social support and resilience). Significantly positive associations were found between Psychological Distress (for Models I [Depression], II [Anxiety], III [Stress], and IV [DASS-21]) and all aggravating factors (except for alcohol use in Model II). Higher levels of mitigating factors (except for Social Support in Model II) were significantly associated with lower levels of Psychological Distress.
In the next set of separately estimated regression models, the ACE (aggravating) and social support (mitigating) factors were disaggregated by type of adverse experience and social support. Table 4 shows the association between Psychological Distress measures and different types of ACEs and prison-based drug and alcohol use, resilience, and external social support subclassified into four types: emotional, tangible, interactional, and affectionate. With a minority of exceptions, the different types of ACEs were significantly and positively associated with Depression and Anxiety symptoms (Models I and II). The two exceptions were as follows: parents who used illegal drugs (compared with no use of illegal drugs) excessively were not associated with depression symptoms and parents incarcerated (compared with no parent incarcerated) were not correlated with anxiety symptoms. These two types of ACEs, in addition to a parent died (compared with a parent did not die), were also not significantly correlated with stress (Model III) or DASS-21 (Model IV) symptoms. The other four types of ACEs were significantly and positively associated with Stress symptoms and total Psychological Distress (DASS-21) symptoms. The largest associations in all four models were between Psychological Distress and experienced emotional abuse (ACE 5, compared with no emotional abuse) and sexual abuse (ACE 7, compared with no sexual abuse). Prison-based drug and alcohol use (compared with no use) was significantly and positively associated with all four measures of Psychological Distress (Models I, II, III, and IV). Social Support and Resilience were negatively associated with Psychological Distress but only Emotional Support was significantly associated with symptoms of Depression (Model I) and Stress (Model III), whereas Affectionate Support was significantly associated with symptoms of Stress (Model III) and total Psychological Distress (Model IV). Resilience was significantly and negatively associated with all four measures of Psychological Distress (Models I, II, III, and IV).
The models that were separately estimated in Table 5 show the associations between psychological distress and the cumulative number of different types of ACEs (compared with 0 ACE). Because only a small number of residents reported experiencing four or more different types of ACEs, these responses were combined together into one cumulative variable (4+ ACEs). Across all four models, ACE variables (compared with no ACE) are significantly associated with Psychological Distress and the magnitude of the association increases monotonically with the number of different types of ACEs. More specifically, compared with no ACE, the association triples in size between Stress symptoms and one ACE versus four or more ACEs, and more than doubles between Depression symptoms and one ACE compared with four or more ACEs. Any drug or alcohol use (compared with no use) in prison remains positively associated with all measures of Psychological Distress. There is no significant association between mitigating factors and Depression symptoms when ACE factors are measured cumulatively. By contrast, some significant associations between mitigating factors and Anxiety, Stress, and total Psychological Distress symptoms persist in this specification of the model. Resilience and Emotional Support remain negatively associated with Anxiety, Stress, and total Psychological Distress symptoms, as does Affectionate Support and Stress and total Psychological Distress symptoms.
Discussion
Experiences of neglect and abuse prior to age 18 were reported by a majority of the incarcerated men in our sample. Our findings support Hypothesis 1. Having an ACE was significantly and consistently associated with prison-based psychological distress and some types of childhood adversity, particularly childhood emotional, physical, and sexual experiences, compared with household adversities had more robust associations with psychological distress. In addition, distress symptoms, including depression, anxiety, and stress, marginally increased with the accumulation of different types of adverse experiences. Finding a graded relationship between ACE and psychological distress among incarcerated men supports the cumulative risk hypothesis within the ACE literature (Chartier et al., 2010; Dube et al., 2003; Felitti et al., 1998). It is also consistent with the finding that cumulative trauma increased the likelihood of current PTSD among incarcerated men (Briere et al., 2016). Not only do individuals with the most severe ACE have the highest risk of incarceration (Roos et al., 2016) but they are, according to these findings, also most likely to experience psychological distress during incarceration.
As expected, the childhood adversities experienced by incarcerated men were not equal in their prevalence or association with psychological distress. Emotional and physical abuse and death of a parent were the most commonly reported childhood adversities, whereas the adversities of emotional and sexual abuse had the largest marginal associations with psychological distress. These abuse types appear to have a more enduring injurious quality. This set of findings suggest a personal vulnerability to risk perceptions and conditions in some prison environments. For example, sexual victimization inside prison, while relatively rare (Beck & Harrison, 2007; Wolff, 2018), is part of the prison lore that is sensationalized through popular culture (Levan, Polzer, & Downing, 2011). This outside view of prison informs the perceptions of those entering prisons, heightening safety concerns and catalyzing fears. Although the fear of sexual assault during incarceration may be democratized, the actual risk of being sexually assaulted in prison significantly increases for men with mental illness and experiences of childhood sexual abuse (Caravaca-Sánchez & Wolff, 2016; Wolff et al., 2007). In this context, the positive association between exposure to childhood sexual abuse and psychological distress during incarceration makes rational sense.
Psychological sensitivities developed as a consequence to childhood emotional abuse may be triggered by the prison environment itself, as well as by the elevated risk of predation inside prison. Studies examining polyvictimization in prisons have found high levels of emotional abuse among incarcerated men (Caravaca-Sánchez & Wolff, 2018; Listwan, Daigle, Hartman, & Guastaferro, 2014). High rates of reported emotional abuse inside prison would suggest that the prison environment has the potential to both heighten preexisting psychological distress and provoke new symptoms of distress.
Prison-based substance use was the second aggravating factor examined in this study. Nearly one in three male respondents reported prison-based drug use, with one in five reporting prison-based alcohol use in the past 3 months. This finding is consistent with the extant literature on substance use during incarceration and the overrepresentation of substance use disorder among the prison population, with more than half of incarcerated individuals estimated to have a significant substance abuse disorder (Baranyi et al., 2019; Fazel et al., 2017; Peters, Bartoi, & Sherman, 2008). Furthermore, our findings support Hypothesis 2: Prison-based drug use and alcohol use, independently, were positively associated with psychological distress, controlling for ACE. A similar finding was reported in a study of incarcerated men in Norway, except the use of drugs and alcohol pertained to a time prior to incarceration (Væroy, 2011). Apropos to this study, the marginal impact of prison-based substance use was approximately equivalent to two ACEs for depression and anxiety symptoms and three ACEs for stress. This strong positive association persisted independent of whether childhood adversities were measured separately or cumulatively. Both the independence and magnitude of the association between prison-based substance use and psychological distress reaffirm the comorbidity of substance use disorders and mental illnesses (Kessler et al., 1996; Regier, Burke, & Burke, 1990). Illicit substances may be used to manage distress associated with prison and/or the use of illicit substance in prison contributes to distress cannot be disentangled with cross-sectional data, but this association does suggest the importance of integrating treatments for both behavioral health problems in correctional settings (Wolff, Huening, Shi, & Frueh, 2015).
Social support and resilience were examined as potential mitigating factors to prison adversity. We hypothesized that both social support and resilience would be negatively associated with psychological distress. Here, the results are mixed and unstable, primarily for social support. Resilience, with one notable exception, was consistently and negatively associated with psychological distress. To our knowledge, our study is the first to examine the connection between resilience and psychological distress in a prison sample. Whether more resilient residents have developed stronger “immune” systems to prison-based adversity or have more impenetrable defense systems remains unclear. This finding, however, does suggest that in an effort to help residents deal with the vicissitudes of prison life and perhaps attenuate psychological distress, resilience promotion programs may have potential as part of the correctional behavioral health care package. These types of programs have demonstrated modest effectiveness (Kim, Lim, Kim, & Park, 2018; Macedo et al., 2014; McMurray, Connolly, Preston-Shoot, & Wigley, 2008; Vanhove, Herian, Perez, Harms, & Lester, 2016; Yoon, Slade, & Fazel, 2017).
The findings regarding social support are less robust. Social support, in the aggregate, was found to be negatively associated with stress and symptoms of depression (not anxiety symptoms). When ACE were controlled for separately and cumulatively, and social support was disaggregated into subscales, Emotional, Informational, Tangible, and Affectionate, only Emotional and Affectionate subscales were consistently and negatively associated with Stress, whereas only the Emotional Support subscale was negatively associated with Depression (ACE separated only model) and Anxiety (ACE cumulated only model). Consistent with other prison studies (Caravaca-Sánchez & Wolff, in press; Lee et al., 2016; Listwan-Johnson, Colvin, Hanley, & Flannery, 2010), higher levels of Social Support, in general, were associated with improved outcomes. Yet, in our study, while the sign of the association was consistently negative, the significance of the association varied with the specification of the ACE and social support predictor variables.
Limitations
A limitation of this study is its cross-sectional study design, which limits the analysis to isolating associations without the ability to draw causal inferences. Although these associations suggest a number of pathways, without knowing temporal sequences, it is impossible to isolate the causal pathway producing these associations. It is, however, very likely that there are multiple pathways, not one, that are linking psychological distress and ACEs, current substance use, social support, and resilience. Having longitudinal survey data, in addition to qualitative data, would be ideal. But, in the absence of the ideal, these associations suggest a fecund area for clinical and correctional intervention, as well as research investigation. Also, that our measure of ACEs includes adverse experiences prior to age 18, it limits our ability to examine the association between age of abuse (childhood vs. adolescence) and psychological distress. Although it is customary to define ACEs using the age 17 years or younger cut-off, there is growing evidence suggesting that neglect and abuse occurring in early life are more detrimental to formative development (Hildyard & Wolfe, 2002; Stirling & Amaya-Jackson, 2008). Another possible limitation is the inclusion of male residents drawn from only three prisons located in the southeast of Spain. These three prisons represent approximately 1.8% of the total Spanish male prison population (Spanish Prison System, 2018). For this reason, findings based on this sample may not be representative of the whole Spanish Prison System, especially with respect to prisons with higher security levels or specialty populations (residents with serious mental illness are housed in two specialty prisons that were not included in our sampling frame and our sample inclusion criteria excluded residents housed in protective custody or specialized psychiatric sections of the three prisons included in the sample). Moreover, because participants were required to read and write in Spanish to complete the survey, residents with literacy deficits were excluded from the sample; hence, our findings do not generalize to residents with literacy deficits or who were not literate in Spanish. This limits the generalizability of our results to a class of prisons within a single country. Given the design of our sample, our findings are only representative of incarcerated men held in medium and minimum secure prisons. Similar research is needed using samples from prison systems in other countries, inclusive of high security prisons. Finally, the predictor and control variables included in our models explain less than 30% of the variance in the psychological distress measures. Other aggravating (e.g., prison-based victimization, housing location, recent relocation, loss of a family member, loss of health, other family stressors) and mitigating (e.g., number of recent visits, friends in prison, involvement in church or school or employment) factors may improve the performance of the models in ways that could affect the stability of the estimated associations found in our models. Alternative model specifications need to be pursued with a broader data set.
Implications
The corpus of ACE literature developed over the past two decades clearly and consistently shows the high prevalence of ACE among the general population in the United States and other countries and the enduring detrimental impact of childhood adversity on adult health, behavioral health, and psychosocial outcomes (Anda et al., 2006; Hughes et al., 2017). These findings are now being replicated with prison populations (Elklit et al., 2013; Mersky et al., 2012; Roos et al., 2016). Together, they suggest the need for a “universal precautions” practice when working with children and adults in institutional settings; a practice supported by trauma-sensitive interventions and environments that are guided by the goal of preventing retraumatization (Grossman, Spinozzola, Zucker, & Hopper, 2017; Substance Abuse and Mental Health Services Administration [SAMHSA], 2014). In terms of prisons and jails, the prevalence of behavioral health disorder and trauma sensitivity among those at risk of incarcerated is extremely high and, to prevent triggering or accentuating psychological distress, adding dual diagnosis treatment that is trauma-sensitive and resilience-building programming into the correctional behavioral health care package, as well as correctional staff trauma-sensitivity training, enhances efforts to make prisons safer and more rehabilitative environments (Bronson & Berzofsky, 2017). Such an integrated approach may reap benefits that include the reduction of psychopathology among prison residents, as well as less self-harm and fewer suicides and suicidal attempts, victimizations, and placements in segregation, especially among those with un- or undertreated psychiatric symptoms.
Conclusion
This is the first study providing evidence on the association between childhood adversity among incarcerated men and prison-based psychological distress and that each marginal addition to cumulative childhood adversity is associated with a larger increase in psychological distress during incarceration. Prison-based substance use also was an aggravating factor, whereas resilience and, to a lesser extent, social support were mitigating factors. Most studies of psychopathology among prison populations document the overrepresentation of mental illness, substance use disorder, and trauma exposure in carceral settings. Other studies, focusing on the risk of bodily harm inside prison, have provided compelling evidence that mental illness is a risk factor in prison. This study identifies potential aggravating and mitigating factors that could inform strategies for identifying psychological distress among incarcerated men and ways to build and provide integrated programming for those affected by childhood adversity.
Footnotes
Authors’ Note:
The authors would like to thank the three external reviewers and the Managing Editor, Jaime S. Henderson, for constructive feedback that guided the revision of the article and ultimately improved its quality. The opinions expressed herein are those of the authors. The authors contributed equally to the development of the article.
