How Does Timing Affect Trauma Treatment for Women Who Are Incarcerated? An Empirical Analysis

Trauma Exposure in Incarcerated Women

Although women make up only 7% to 10% of incarcerated people in the United States (Federal Bureau of Prisons, 2018), women have been the fastest growing demographic segment of prison and jail populations throughout the past several decades (Carson, 2018). In fact, the population of women in state prisons has increased by over 700% since the 1970s (Carson, 2018). Rising rates of incarceration have largely been driven by increased enforcement and prosecution of drug-related offenses, especially low-level nonviolent offenses (Fellner, 2000). Because many survivors of sexual assault, who are most often women, commonly use drugs and/or alcohol to cope with posttraumatic symptoms (Ullman et al., 2013), greater enforcement of drug laws disproportionately increases women’s justice involvement. Indeed, a recent review found that 56% to 82% of incarcerated women have experienced lifetime sexual victimization, with 50% to 66% having experienced sexual abuse in childhood (Karlsson & Zielinski, 2018).

Not surprisingly, sexual victimization is associated with worse mental health. For instance, 30% of women who were raped met criteria for a depressive disorder (vs. 10% for non-victims), and 31% reported symptoms consistent with posttraumatic stress disorder (PTSD) (vs. 5% of non-victims; Kilpatrick, 2000). In a nationally representative sample of adult women, child sexual abuse survivors were significantly more likely than non-victims to have alcohol use problems (33.9% vs. 18.2%), other drug problems (27.6% vs. 10.1%), depression (39.3% vs. 19.2%), and PTSD (39.1% vs. 5.7%) during their lifetime (Molnar et al., 2001). Among women with rape-related PTSD, 20.1% had alcohol problems and 7.8% had problems related to other drug abuse (Kilpatrick, 2000). Few studies have investigated the association between trauma exposure and mental health in incarcerated women; however, those that have done so found that childhood and adult sexual assault are associated with higher rates of mental illness including depression, PTSD, and substance use disorders (Karlsson & Zielinski, 2018).

Mental illness is not the sole negative emotional consequence of sexual victimization; maladaptive emotions such as shame often emerge following trauma (Saraiya & Lopez-Castro, 2016). Shame is an emotion tied to perceptions that the self is defective, inadequate, or “bad,” and tends to lead to negative outcomes such as self-criticism, poor self-esteem, and maladaptive thought patterns (Gilbert & Miles, 2000). Interpersonal traumas like sexual assault likely have unique relationships with shame: perpetrators of sexual assault are often family members, friends, or acquaintances of the victim, increasing the likelihood victims will view the attack as a reflection on themselves and impacting victims to a greater extent than other forms of trauma (Green et al., 2000). Shame may also be part of “prison culture,” where shame and guilt around crimes might be encouraged (Markel, 2016). In our groups, we have also treated many women who were blamed for their victimization and/or taught that it was shameful.

Trauma-Focused Therapy in Prison

In response to the high rates of trauma exposure in incarcerated women, there have been calls to make jails and prisons more trauma-informed (Miller & Najavits, 2012) such that care is delivered in ways that enhance safety, trust, collaboration, and empowerment (Substance Abuse and Mental Health Services Agency, 2014). If adopted, this transformation would result in the provision of both trauma-informed practices (e.g., same-sex medical exams or describing pat-down procedure prior to physical contact) and trauma-focused therapy. Most trauma-focused therapies emphasize providing psychoeducation and teaching coping skills (e.g., distress tolerance, cognitive restructuring, or problem solving) within a supportive and safe environment; they can be further delineated into those that use exposure (e.g., the current treatment; Karlsson et al., 2015) and those that do not (e.g., Helping Women Recover & Beyond Trauma, Seeking Safety; reviewed in King, 2017). Beginning trauma therapy, including those involving exposure, can be temporarily unsettling and difficult (Hayes et al., 2007). Therefore, some researchers have argued that specialized treatments for trauma sequelae and/or PTSD should occur only after clients have been provided with time and assistance in adjusting to significantly daily stressors (Miller & Rasmussen, 2010).

On the contrary, some have argued trauma sequelae can be addressed concurrently with daily stressors; trauma symptoms may even exacerbate the stress felt by daily hassles (Neuner, 2010), suggesting offering treatments earlier in the context of prison adjustment may be beneficial. Trauma treatments (both exposure-based and non-exposure-based) offered in times of crisis or disruption such as in war zones, refugee camps, and domestic violence shelters show benefits (e.g., Adler et al., 2009; Johnson et al., 2011; Neuner et al., 2004). Therefore, it is possible that even within potentially traumatic and/or stressful environments, such as prisons, people could see significant improvements from trauma-focused treatment.

Adjustment to Incarceration

Although for some women prison can offer more stability and safety than their lives in the community (e.g., Bradley & Davino, 2002), prison entry periods are often characterized by feelings of uncertainty, fear, and isolation (Yardley & Wilson, 2013). Incarceration is a stressful life event that requires modifications of basic life patterns. People who are incarcerated are required to live in close proximity to one another, adhere to strict rules and regulations, and lose the ability to make many of their own decisions (Yardley & Wilson, 2013). Female gender, younger age, having children, lower socioeconomic status, and preexisting psychopathology and/or victimization experiences are each associated with greater difficulties adjusting to life in prison (DiCataldo et al., 1995; Islam-Zwart & Vik, 2004). Women often have less access to social support due to being incarcerated further distances from their families in comparison to men (Johnston & Gabel, 1995), and often have to cope with losing custody of dependent children. Women with sexual assault histories tend to take longer to adjust to incarceration and have slower rates of improvement over time (Islam-Zwart & Vik, 2004). These factors’ contributions to prison maladjustment are consistent with both deprivation theory, which suggests that poor adjustment stems from the prison environment itself, and importation theory, which emphasizes the role of individual differences (e.g., previously held beliefs, “street attitudes,” and personal desires) in adjustment to prison (Irwin & Cressey, 1962).

Adjustment problems often present in the form of internalizing symptoms (e.g., discomfort, insomnia, anger), externalizing behaviors (e.g., arguments, physical altercations), and an increased risk of physical illness and suicidal behavior (e.g., Liebling, 1999; O’Donnell & Edgar, 1999). Successful adaptation often involves establishing feelings of safety, psychological well-being, a sense of autonomy, making meaning of one’s experience, improved functioning (e.g., meeting the demands of the environment), and having a positive outlook on the future (van der Laan & Eichelsheim, 2013; Van Ginneken, 2015).

Previous research suggests that incarceration can exacerbate symptoms of emotional distress and behavioral dysregulation in people with trauma histories (e.g., Listwan et al., 2010); however, some people’s symptoms improve during incarceration, and researchers have underscored the importance of reassessing trauma symptoms after an initial period of adjustment (Hochstetler et al., 2004). Indeed, prior work has found various adjustment problems change over time (Van Ginneken, 2015). One study found that adjustment continued to improve throughout incarceration for men; however, incarcerated women experienced periods of both increased and decreased adjustment (Hassan et al., 2011). Fluctuations in symptoms are often most pronounced at the beginning of prison sentences. Improved adjustment is reflected by less frequent rule violations, less time spent in solitary confinement, fewer interpersonal conflicts, and a greater ability to cope with changes over time (Bonta & Gendreau, 1990; Harding & Zimmerman, 1989; Toch et al., 1989). The classification of prison itself may also contribute to the adaptation process. Adjustment to prison may be more difficult in maximum-security facilities, as people in custody are often subject to greater restrictions, fewer therapeutic opportunities, and an increased likelihood of exposure to violence compared with people in minimum- or medium-security facilities (Camp & Gaes, 2005).

The fluctuations in symptoms and potential exacerbation of symptoms when adjusting to life in prison make sense when considering Maslow’s (1943) hierarchy of needs, which assumes that people are not easily able to focus attention on complex needs (e.g., self-betterment in preparation for the future) when basic needs have not been met (e.g., when experiencing mental health problems, having difficulty accepting one’s experience, low ability to cope and meet the demands of the environment, low subjective well-being and hopelessness for the future; Maslow, 1943; Van Ginneken, 2015). Van Ginneken (2015) proposed a model of adjustment where self-actualization or the movement toward reaching one’s full potential is a function of adjustment. It is therefore possible that people who become incarcerated must first get their basic needs met, improve their ability to manage existing psychological symptoms, gain a sense of autonomy, identify social groups, and gain an understanding of institutional rules, procedures, and policies (Crawley & Sparks, 2006). People may be better able to participate in prison rehabilitative programs that promote working on more long-standing, self-oriented factors to enhance quality of life and reduce recidivism (e.g., treatment for mental health disorders, interpersonal skills) once these other tasks have been accomplished. Indeed, research suggests people who have not yet adapted to incarceration are less likely to participate in activities and programming than those who have had a period of adaptation (Goodstein et al., 1984). Overall, research on adaptation suggests early incarceration can be a turbulent and uncertain time and suggests adjustment to prison likely influences both motivation for and involvement in rehabilitative treatment programs.

Purpose and Hypotheses

To our knowledge, no studies have examined the role of timing of therapeutic interventions relative to incarceration dates. Morgan and Flora (2002) conducted a meta-analysis where they attempted to examine the effects of group therapy timing in prison on anxiety, depression, anger, self-esteem, interpersonal relations, and locus of control; however, they were unable to answer the question of optimal timing for treatment due to lack of research studies and missing data. In non-carceral settings, some researchers have argued that it is beneficial to prepare clients for the difficult work of trauma therapy by providing a baseline period to adjust to life stressors, learn to manage distress, and develop coping skills (Campbell et al., 2016; Miller & Rasmussen, 2010).

Given the transition to prison involves significant disruptions in daily activities and has been described as highly stressful (even somewhat traumatic; Yardley & Wilson, 2013), we evaluated whether trauma treatments might be more effective if delivered after an initial adjustment period to prison by analyzing data from women in a minimum-security prison who participated in our exposure-based group trauma treatment. In addition to our group, the women are required by the prison to participate in various therapeutic programming (e.g., mindfulness meditation, Celebrate Recovery, spirituality-based groups, jobs, individual counseling). While these programs likely help with rehabilitation and increase coping skills, our treatment group is unique in its focus on treating trauma, specifically through exposure. Having more time to adjust to the prison setting while learning these foundational skills may increase women’s ability to manage the emotional intensity of exposure-based trauma therapy. Therefore, consistent with the notion that people may need time to acclimate to prison, we hypothesized a replication of prior findings suggesting greater time since incarceration would be associated with lower internalizing symptoms (depression, posttraumatic distress, and shame; H1). Consistent with prior evaluations of this treatment (Karlsson et al., 2015), we hypothesized women would report significant declines in internalizing symptoms from pretreatment to posttreatment (H2). Finally, we hypothesized the more time between initial incarceration date and the date of beginning group trauma treatment participants had, the greater the magnitude of change in psychiatric symptoms they would report at posttreatment, controlling for pretreatment symptoms (H3).

Method

Participants

Participants in this study were women in a minimum-security prison who enrolled in a voluntary eight-session weekly therapy group aimed at treating symptoms related to pre-incarceration sexual assault and/or abuse. All had a self-reported history of sexual victimization; 74.7% reported both childhood sexual abuse and sexual victimization in adulthood, 2.3% reported only child sexual abuse, and 18.4% reported only sexual victimization in adulthood. A total of 128 women across 27 groups participated in treatment. Of these, 116 provided consent to participate in the study (six did not consent to have their data used for research purposes at pretreatment, and an additional six did not consent at posttreatment). Of those who consented, 93 (80.2%) completed treatment (defined as having attended six of the eight sessions). Three participants did not provide data on their incarceration start date, so they were excluded from the analyses. The final sample included 90 women who completed group treatment between May 2015 and August 2018 and consented to have their data used at both time points. Demographic characteristics of these participants are provided in Table 1.

OPEN IN VIEWER

Table 1:

Participant Descriptive Statistics

Variable	M (SD)	n (%)
Age (n = 88, 2 missing data)	33.3 (9.2)
Maternal status
Mother		82 (90.1)
Average number of children	2.6 (1.7)
Race (n = 89, 1 missing data)
White		76 (84.4)
African American		3 (3.3)
Native American		7 (7.8)
Other		3 (3.3)
Latinx ethnicity (n = 87, 3 missing data)		2 (2.2)
Marital status
Married		27 (30.0)
Single		35 (38.9)
Divorced or separated		24 (26.7)
Other (e.g., engaged, widowed)		4 (4.4)
Sexual trauma history (n = 87, 3 missing data)
Sexual assault by family member or acquaintance		74 (82.2)
Sexual assault by a stranger		47 (52.2)
Child sexual abuse		67 (74.4)

We compared completers and non-completers on demographic and pretreatment symptom measures. There were no significant differences in ethnicity, race, marital status, age, number of children, or pretreatment depression, PTSD, and shame scores (all p values > .19). Although data regarding reasons for discontinued participation were not systematically recorded, typical reasons included unanticipated schedule conflicts (e.g., work or other groups), release from the facility or transfer to another facility, conflict with other group members, violating the two-absence policy for unforeseen reasons (e.g., illness, segregation), and perceptions that personal traumas had already been addressed.

Results

Descriptive Statistics

Table 2 presents bivariate correlations and Table 3 presents descriptive statistics for all study variables. On average, participants were in the moderate range for depression, well above the clinical cutoff for posttraumatic stress symptoms, and about one standard deviation above measure norms for feelings of shame. In contrast, at posttreatment, on average participants were in the mild range for symptoms of depression, at the clinical cutoff for posttraumatic stress symptoms, and at the normative sample mean for feelings of shame. The mean number of days participants had been incarcerated prior to starting SHARE was 160.7 (SD = 104.2), or about 5.4 months. Days incarcerated before beginning SHARE did not significantly differ between participants who did and did not complete the group, t(113) = −0.47, p = .640.

OPEN IN VIEWER

Table 2:

Correlations and Descriptive Statistics for Study Variables

Variable	1.	2.	3.	4.	5.	6.	7.
Pretreatment
1. Days incarcerated before beginning SHARE	—	−.34**	−.18	−.10	−.23*	−.14	−.16
2. PHQ-9		—	.70***	.55**	.30**	.26*	.12
3. PCL-5			—	.59***	.29**	.25*	.12
4. PFQ-2 Shame				—	.27*	.23*	.39***
Posttreatment
5. PHQ-9					—	.68***	.48***
6. PCL-5						—	.54***
7. PFQ-2 Shame							—
Cronbach’s α		.86	.94	.83	.83	.93	.81

Note. PHQ-9 = Patient Health Questionnaire–9; PCL = Posttraumatic Checklist; PFQ-2 = Personal Feelings Questionnaire–2.

*

p < .05. **p < .01. ***p < .001.

OPEN IN VIEWER

Table 3:

Change in Symptoms From Pretreatment to Posttreatment

Variable	PreM (SD)	PostM (SD)	Test statistic	Cohen’s d
PHQ-9	12.8 (6.3)	5.9 (4.9)	t(89) = 9.8***	1.04
PCL-5	39.0 (19.6)	15.0 (12.6)	t(89) = 11.1***	1.17
PFQ-2 Shame	2.1 (0.8)	1.5 (0.7)	t(87) = 6.8***	0.72

Note. PHQ-9 = Patient Health Questionnaire–9; PCL = Posttraumatic Checklist; PFQ-2 = Personal Feelings Questionnaire–2.

***

p < .001.

Hypothesis Testing

H1. An examination of the bivariate correlations in Table 2 revealed that number of days incarcerated before beginning SHARE was negatively associated with pretreatment PHQ-9 scores (r = −.34), but not significantly associated with pretreatment PCL-5 (r = −.18) or PFQ-2 Shame (r = −.10) scores. Therefore, our first hypothesis was not fully supported.

H2. Table 3 provides results of three dependent-sample t tests comparing pretreatment to posttreatment internalizing symptom scores. Consistent with our second hypothesis, declines in symptoms were significant for all three measures (all p values for dependent-sample t tests <.001). These declines were large in magnitude for PHQ-9 and PCL-5 scores, and medium to large in magnitude for PFQ-2 Shame scores.

To determine how many participants evidenced clinically significant improvements, we calculated RCIs. RCI values of 1.96 or greater are suggestive of improvements beyond chance alone. A total of 53.3% of participants (n = 48) obtained RCIs of 1.96 or greater on the PHQ-9; 65.6% obtained RCI values ≥1.96 on the PCL-5 (n = 59), as did 32.6% on the PFQ-2 Shame (n = 28; 4 were missing data).

H3. To examine whether number of days incarcerated before beginning SHARE predicted therapeutic benefits of participating in SHARE group, we conducted two sets of hierarchical linear regressions. Number of days incarcerated before beginning SHARE was entered at the first step, and pretreatment scores were entered on the second step. The criterion variables for the first set of three regressions were PHQ-9, PCL-5, and PFQ-2 Shame posttreatment scores, respectively, while in the second set, change scores served as the criterion variables. All regressions met assumptions of normality of residuals, with no standardized residual with an absolute value >3. Results are displayed in Table 4. Contrary to our hypothesis, there was no association between days since incarceration and improvements in internalizing symptoms for any of the measures with one exception. Number of days incarcerated before beginning SHARE was a significant predictor of posttreatment PHQ-9 scores. However, when controlling for pretreatment symptoms, it was no longer significant. We reran the regressions on the full sample in an intent-to-treat analysis using a last-observation-carried-forward procedure; results did not change.

OPEN IN VIEWER

Table 4:

Hierarchical Linear Regressions Predicting Posttreatment Internalizing Symptoms

Model	Posttreatment score			Change score
	Model F	R 2	β	Model F	R 2	β
Depression
Model 1	4.91*	.05		1.97	.02
Days since incarcerated			−.23*			−.15
Model 2	5.22**	.11		49.43***	.53
Days since incarcerated			−.15			.11
PHQ-9 pre			.25*			.76***
PTSD
Model 1	1.73	.02		0.63	.01
Days since incarcerated			−.14			−.08
Model 2	3.38*	.07		79.65***	.65
Days since incarcerated			−.10			.06
PCL-5 pre			.23*			.81***
Shame
Model 1	2.21	.03		0.17	.00
Days since incarcerated			−.16			.04
Model 2	8.64***	.17		27.26***	.39
Days since incarcerated			−.12			.10
PFQ-2 Shame pre			.38***			.63***

Note. PHQ-9 = Patient Health Questionnaire–9; PTSD = posttraumatic stress disorder; PCL = Posttraumatic Checklist; PFQ-2 = Personal Feelings Questionnaire–2.

*

p < .05. **p < .01. ***p < .001.

Finally, we conducted independent-sample t tests to examine whether participants who evinced reliable change (RCI values ≥ 1.96) and those who did not (RCI values below 1.96) differed in mean number of days incarcerated before beginning SHARE (Figure 1). Results revealed no significant differences in PHQ-9 [t(88) = 1.36, p = .177], PCL-5 [t(88) = 1.55, p = .126], or PFQ-2 Shame [t(84) = −0.78, p = .437] RCI groups.

OPEN IN VIEWER

Figure 1:

Mean Days Between Incarceration and Group Therapy Start Dates by Reliable Change Groups

Discussion

Adjusting to being incarcerated can be stressful, possibly even traumatic, and women with histories of sexual assault struggle with adjustment more so than do women without sexual assault histories (Islam-Zwart & Vik, 2004 ; Yardley & Wilson, 2013), although the concern that prisons are necessarily chronically stressful is likely overstated (Bonta & Gendreau, 1990). Because most women who are incarcerated have a history of prior trauma exposure (Karlsson & Zielinski, 2018), incorporating trauma treatments into therapeutic programming in prison settings is a priority (Hall et al., 2013). However, there are questions about whether it is appropriate to offer such treatment to women at all, especially when the treatments use exposure techniques which necessitate the retelling of traumatic, intrusive memories at a time and in a place where women may be feeling especially vulnerable (Miller & Najavits, 2012). Here, we investigated whether the amount of time women had been incarcerated related to their ability to benefit from trauma treatment. We utilized an open trial of 90 women incarcerated primarily for drug-related offenses who had all had a history of sexual violence victimization, most since early childhood. All participated in an 8-week therapy group that included imaginal exposure to trauma-specific memories. We evaluated changes in depression, posttraumatic stress symptoms, and shame scores from pre- to post-treatment. Because this was an open trial and did not include a comparison group, our findings are limited. Nevertheless, they help inform whether there may be reason to wait a period of time to allow women to adjust to prison before asking them to engage in trauma work.

Adjustment theories suggest skills-focused therapeutic programming can be helpful to men and women who enter prison (Bonta & Gendreau, 1990), recognizing that concerns about safety, increases in environmental stressors such as noise, and limitations in the ability of people in custody to engage with social supports can all increase psychological distress (Jiang & Winfree, 2006; MacKenzie et al., 1989). We therefore expected SHARE group, a trauma-focused treatment that requires participants to disclose often-avoided emotions and memories of traumatic experiences with other group participants, to be most helpful to women who had been incarcerated for longer periods of time, presumably after they had successfully navigated some of the early concerns associated with adjusting to prison. Contrary to our expectations, however, there was not a significant association between days incarcerated before beginning SHARE and treatment completion, suggesting no specific risk of dropout associated with early engagement in SHARE.

Bivariate correlations between days incarcerated before beginning SHARE and pretreatment symptoms revealed a significant association between time and levels of depression (but not shame or PTSD); the longer women had been in the prison facility, the lower their pretreatment levels of depression were. This is consistent with the idea that adjusting to prison may involve some uptick in symptoms associated with self-regulation, including sleep, appetite, and mood. Perhaps depression symptoms improved as women adjusted to life in prison, including sharing sleeping quarters, a change in diet and daily routines, and acclimating to being temporarily separated from natural social supports. Prisons must also provide medical care, which often includes prescribing medications to treat psychological symptoms. A portion of the women in our study were likely prescribed psychotropic medications, which may have contributed to observed changes in psychological symptoms throughout treatment. Unfortunately, we were unable to obtain information on our participants’ use of antidepressants or other psychotropic medications.

Also contrary to our expectations, number of days incarcerated before beginning SHARE was not associated with reductions in shame or PTSD symptoms at the bivariate level. These symptoms remained elevated in women even as depression appears to declined with longer times since incarceration began. Such elevations are consistent with research suggesting PTSD-related symptoms do not self-resolve in the absence of treatment (Morina et al., 2014), and point to the high need to provide trauma treatments for women in prisons. Furthermore, because women at this particular facility begin to engage with therapeutic rehabilitation programming as soon as they arrive, the lack of an association between time since incarceration and trauma symptoms suggests indirect targeting of symptoms (e.g., through other programming such as developing assertiveness skills or enhancing sobriety) is insufficient to treat the unique sequelae associated with trauma exposure.

The large declines in symptoms from pretreatment to posttreatment seen in this study are consistent with what other evaluations of SHARE have shown (e.g., Karlsson et al., 2015): Treatment participation was associated with declines in depression, PTSD, and shame symptoms. Although analyses revealed reductions in internalizing symptoms, declines were not associated with number of days women had been incarcerated. Women who participated in trauma treatment benefited, whether their participation occurred early in their incarceration period or many months later. On average, women initiated treatment 5 months after being incarcerated, and while the range included 4 days to over 500 days, the majority of women in the study had been incarcerated for at least 3 months before beginning treatment. We suggest that cautionary notes of the benefits of waiting to begin trauma treatment or avoiding treatments that focus on past trauma experiences are well-meaning but perhaps misguided (cf. Campbell et al., 2016; Miller & Rasmussen, 2010; Miller & Najavits, 2012). Indeed, the ability of women to benefit from trauma treatment across the spectrum is consistent with an increasing body of evidence suggesting PTSD treatment can be beneficial even under the most adverse of circumstances (Adler et al., 2009; Johnson et al., 2011; Neuner et al., 2004). Importantly, some of the concerns others have expressed about the possibility that dealing with past traumas while in prison can result in emotional destabilization of people in custody (e.g., Miller & Najavits, 2012) are not consistent with our findings.

Our anecdotal observation, having conducted these groups for more than 8 years with hundreds of women, is that the degree to which women struggle to adjust to life in prison is variable. We have certainly noticed that some women who are recently incarcerated or who are very near their release date may be overwhelmed by all that they have to do. Nevertheless, the ability to process stressful situations and the self-regulation and coping strategies they learn as part of the group may in fact be helpful to women as they are preparing for or undergoing such transitions. We have further seen the benefits of group cohesion that forms quickly as a result of treatment. Forming such strong bonds with others in the prison may be helpful to women as they are adjusting to their new circumstances. It is also possible that women who complete such treatment earlier during their time at prison might be better able to engage in other programs and opportunities while incarcerated. If they are effectively able to process their traumatic experiences, then maybe they are less triggered by them and more able to fully engage in other programming. Because our groups are voluntary, women chose to participate, indicating at minimum they perceived this group would be beneficial at that particular time, whenever that may have been. Our results certainly might have been different if women were either mandated to treatment at a particular time or if we had employed a multiple baseline design with women at the time they reported being interested in signing up for the group. On the other hand, we also believe one benefit of waiting to participate in group is that women often hear from prior group participants about the benefits of treatment; a longer period of time in prison before participating in group may allow women to learn of the group through positive word-of-mouth and may motivate people who might otherwise not be interested in recounting traumatic memories to a group of peers.

Although our results are promising, they are preliminary and should be taken with the study’s limitations in mind. For instance, the data we present on the timing of trauma treatments were with women in a minimum-security correctional facility that is highly focused on rehabilitation and therapeutic programming; it is not clear whether results would generalize to other facilities, especially those of medium or maximum security, or to men. Furthermore, women received a great deal of prior and/or concurrent therapeutic programming. Most participants had thus already learned about cognitive techniques such as thought stopping, behavioral skills such as assertiveness and communication, and emotion regulation skills such as grounding or diaphragmatic breathing. A related limitation is we did not capture what other programs group members attended; future studies should include this information. We also did not document systematically the reasons why women did not complete treatment, and the women on average had been incarcerated 5 months. Some research suggests critical periods of adjustment are much shorter than this—often just a few weeks (Bonta & Gendreau, 1990). Indeed, the recognition that distress is a normal and temporary response to potentially traumatic events is built into the definition of PTSD (American Psychiatric Association, 2013). Therefore, it is possible our sample was entirely or almost entirely composed of women who had already navigated the most critical adjustment period of life in prison. Only four women in our sample had started group within the first 30 days of their incarceration term.

Our choice to index adjustment with measures assessing depression, PTSD, and shame was also limiting (cf. Dhami et al., 2007). A broader definition of adjustment that focuses on components such as engagement in rehabilitation activities, earning privileges, avoiding sanctions, and an enhanced sense of well-being should be included in future studies. In addition, our measure of shame asked about experiences of shame broadly and was not specific to trauma. For example, items measuring shame asked participants questions like, how often do you “feel you deserve criticism for what you did” and “feel disgusting to others.”

The design choice (pre–post without random assignment and no control group) was dictated by the applied nature of this study; the primary purpose of offering groups was to provide therapeutic programming to women in the facility. A secondary goal was to evaluate the group. Although applied work has the potential to be highly informative, the scientific limitations are many and include a lack of ability to establish cause–effect relations or to rule out alternative explanations for observed changes in symptoms across time. Addressing these design limitations would be critical for drawing stronger conclusions about the findings. Some additional methodological limitations to the study include our lack of baseline symptom measurement at the time of entry into prison, a lack of posttreatment data for participants who discontinued participation in group, and use of a self-reported incarceration start date. The latter limitation is especially important because women could have had varying interpretations of the item, with some women indicating how long they had been incarcerated across multiple facilities (e.g., jail prior to prison) and others indicating how long they had been incarcerated in the current facility. We also did not ask about prior incarceration experiences. It would also have been helpful to provide baseline symptom measures to all women entering the facility as this would allow for a clearer evaluation of what symptoms may be declining as a function of adjusting to prison versus those that may be declining as a function of specific therapeutic programming. Future studies should consider gathering data from all group participants, even those who chose to discontinue participation, so intent-to-treat analyses can be conducted. In this study, we used time between incarceration start date and the start of trauma treatment group to index acclimation to life in prison; however, obtaining baseline symptom measures or more direct measures of adjustment would be important to verify these findings. Participants’ previous experiences with being incarcerated could also have an effect on their ability to adjust, but we lacked data on prior incarceration experiences.

Nevertheless, the findings of our study contribute to a growing body of research suggesting trauma treatments can be effective with people even under adverse circumstances and can help address aspects of mental health that may be relatively impervious to change unless directly targeted. Specifically, regardless of the length of time women had spent in prison before beginning trauma treatment, rates of shame and PTSD were similar at pretreatment and declined significantly with trauma treatment. Waiting to offer trauma treatment until other aspects of a person’s life are less chaotic, while a well-intentioned and reasonable recommendation on its surface, may not be warranted.