Separate Roles for State and Trait Anxiety in the Formation of SHU Syndrome: Testing a Moderated Mediation Hypothesis

Abstract

It has been proposed that state anxiety, aroused when an inmate is initially placed in restrictive housing, interacts with the sequestering Special Housing Unit (SHU) environment to overtax the individual's already limited coping resources, and promotes later emotional problems and psychological deterioration. This study tested a SHU syndrome hypothesis with a moderated mediation path analysis of 69 male inmates. Results revealed that group status (general population vs. restrictive housing) interacted with state anxiety to increase trait anxiety and ineffective coping, which then gave rise to higher staff ratings of psychological disturbance in inmates with no prior history of mental illness.

Keywords

SHU syndrome coping skills state and trait anxiety

Introduction

The use of special housing to manage correctional populations presents both opportunities and problems. Special housing is frequently employed as punishment for institutional rule violations and as a way of managing the more violent inmates in the prison population while protecting the more vulnerable ones. As a general rule, special or restrictive housing has been found to be effective in reducing prisoner misconduct, particularly when hard-to-manage inmates like gang members are targeted (Fischer, 2002; Ralph & Marquart, 1991). There are concerns, however, that special housing may be overused and that keeping an inmate in segregation over an extended period of time may create serious psychological problems and issues, even in inmates without a history of prior mental health disorder (Grassian, 1983). One such problem is a constellation of behaviors, abnormalities, and symptoms known as SHU (Special Housing Unit) syndrome. The purpose of the current investigation was to test a conceptual model of SHU syndrome with a moderated mediation research design in a group of inmates with no apparent mental health needs.

SHU Syndrome

Grassian (1983, p. 2006) developed the concept of SHU syndrome from interviews held with 15 male inmates living in a restrictive housing unit at the Massachusetts Correctional Institution-Walpole. All 15 were plaintiffs in a class action suit alleging Eighth Amendment violations emanating from their exposure to long-term solitary confinement. The symptoms of this syndrome, as outlined by Grassian, were generalized hyperresponsivity to external stimuli, perceptual distortions and hallucinations, affective disturbances, concentration and memory difficulties, disturbances of thought content, and problems with impulse control. Of particular note, these symptoms were said to remit or significantly diminish within hours of the inmate's removal from isolation. This led Grassian to conclude that restrictive or segregation housing was the cause of these symptoms. Early cross-sectional studies comparing inmates in restrictive housing to inmates in general population produced mixed and inconclusive results (Hodgins & Côté, 1991; Suedfeld et al., 1982), and exposed the existence of large individual differences between prisoners in restrictive housing and prisoners in general population in terms of their premorbid mental status (Motiuk & Blanchette, 2001). Researchers soon realized that longitudinal investigations were required to properly evaluate the validity and prevalence of SHU syndrome.

The first longitudinal study conducted on psychological adjustment in restrictive housing was performed by Zinger et al. (2001). Using a sample of 60 Canadian prisoners evaluated at three points in time, with 30 days between each evaluation, Zinger et al., failed to observe psychological deterioration in inmates from the restrictive housing unit, although two months may not be sufficient enough time to observe a noticeable drop in functioning. Andersen et al. (2003) compared 113 Danish inmates in restrictive housing to 95 general population inmates, and unearthed modest support for SHU syndrome to the extent that cross-sectional comparisons identified more adjustment problems in the restrictive housing inmates. As well, longitudinal analyses disclosed reduced symptomatology in general population inmates but no change in restrictive housing inmates. It was further noted that transfer from restrictive housing to general population led to reductions in psychological distress and disturbance. Chadick et al. (2018) also observed that psychological functioning in general population inmates improved over time, whereas administrative detention inmates showed little to no change. In a study of inmates housed in high security facilities in Colorado, O’Keefe et al. (2013) failed to observe any major differences between restrictive housing and general population inmates, with both groups exhibiting reduced symptomatology over time.

It should be noted that SHU syndrome remains steeped in controversy. Despite efforts by Grassian (1983, p. 2006) and others (Haney, 2003; Kupers, 2016; Muller, 2018) to promote SHU syndrome as a major correctional concern, many researchers and correctional practitioners are skeptical (Gendreau & Bonta, 1984; Hanson, 2011; Morgan et al., 2016). Criticisms that have been leveled against the original Walpole Prison study include a strong possiblity of response bias, given that participants were engaged in a class action suit against the institution for alleged cruel and unusual punishment, absence of a control group, and use of a cross-sectional design that did not consider pre-existing mental health conditions (Labrecque et al., 2020). In a study where pre-existing mental health conditions were taken into account, Walters (2018) determined that inmates with a history of mental health disorder were more likely to experience a rise in psychiatric symptomatology over the course of their incarceration than inmates with no prior mental health history. Using data from the Colorado State Prison study, Walters (2018) discovered that inmates suffering from significant mental health problems were at increased risk for psychological deterioration over time, although they were just as likely to deteriorate in general population as they were in segregation.

State and Trait Anxiety and Ineffective Coping

Anxiety can be defined as a series of physical and psychological responses designed to prepare the organism for threatening situations, events, and phenomena. In an effort to decipher the mechanisms behind anxiety, some investigators have further separated anxiety into state and trait anxiety. Spielberger et al. (1983), in fact, developed the State-Trait Anxiety Inventory (STAI) for this very purpose. While state anxiety is a transient emotional response to a situational concern, trait anxiety is a more enduring set of beliefs about internal and external stimuli that threaten a person's sense of self and well-being. In this way, state anxiety reflects momentary stresses and strains (Agnew, 1992), whereas trait anxiety represents a developmental pattern in which a person's ability to cope effectively with stress and strain has been compromised (Wadsworth, 2015). One of the principal assumptions upon which the current study was based maintains that maladaptive coping techniques give rise to elevated levels of state and trait anxiety, with state anxiety representing a person's response to being placed in situations not adequately covered by their coping skills and trait anxiety reflecting the longer-term consequences of poor coping. A review of the international literature on the STAI–coping relationship provides support for the notion that ineffective coping and trait anxiety are moderately to strongly correlated and that the relationship permeates many aspects of life, though it appears to find its greatest expression in research on people's efforts to cope with personal and family medical issues.

There have been several studies published on the relationship between coping skills and state/trait anxiety within the past decade. One disclosed that Russian patients coping with chronic leukemia displayed moderate levels of both state and trait anxiety, and that while both forms of anxiety correlated negatively with positive coping strategies, trait anxiety also achieved a strong positive correlation with the negative coping strategy of repression (Stepanchuk et al., 2013).A Polish investigation showed that trait and state anxiety correlated positively with emotion-oriented coping and negatively with task-oriented coping in adults diagnosed with cancer (Michalowska et al., 2019). In research from India, Bhattacharya et al. (2016) discovered that positive coping strategies in mothers of children diagnosed with cancer correlated inversely with scores on the state and traits scales of the STAI. French patients with major depressive disorder and their primary caregiver displayed correlations of −.27 to −.47 between scores on the STAI trait anxiety scale and positive coping and of .31 to .38 between scores on the STAI trait anxiety score and negative coping (Marguerite et al., 2017). In a study of Pakistani adolescents, Fatima and Tahir (2013) likewise discovered that the STAI state and trait anxiety scores correlated positively with coping styles marked by avoidance and denial, and negatively with positive or problem-focused coping styles. Finally, among German adults diagnosed with an alcohol use disorder, Wedekind et al. (2013) determined that poor interpersonal coping, as measured by an insecure attachment style, correlated with the STAI trait anxiety scale but not STAI state anxiety scale.

A Conceptual Model

The conceptual model proposed in this article rests on two assumptions. The first is that the STAI trait anxiety scale reflects failed or inadequate coping. This assumption was previously described in the state and trait anxiety section. The second assumption is that offenders, as a group, favor externalizing coping strategies. This is described in the current section. The definition of an externalizing coping strategy is one in which the person interacts with the external environment in order to distract themselves from current issues and problems. It is further assumed that externalizing tendencies increase criminal risk, the probability of incarceration, and the likelihood of placement in segregation housing. Childhood and adolescent problems have traditionally been divided into internalizing and externalizing disorders (Hopwood & Grilo, 2010). Internalizing disorders are directed inward, toward the person (depression, anxiety, and withdrawal); externalizing disorders are directed outward, toward the environment (aggression, bullying, and conduct disorder). In recent work by van der Ende et al. (2020), externalizing disorders of childhood were reported to correlate with externalizing problems of adulthood assessed some 20 years later. Externalizing and internalizing disorders of childhood also rely on different types of coping mechanisms, with problem avoidance being a core feature of an externalizing coping strategy. In fact, one study found that an avoidant coping style appeared to connect childhood externalizing disorder to later adult criminality (Aebi et al., 2014).

The conceptual model tested here is depicted in Figure 1. According to this model, externalizing tendencies are the basis of the first two elements of the model. First, externalization is believed to be responsible for the proposed interaction between group status (restrictive housing vs. general population) and state anxiety. Given the importance of environmental interaction to someone with strong externalizing tendencies, a significant reduction in environmental stimulation, created by placement in restricted housing, will increase situational distress or state anxiety. Second, someone with externalizing tendencies should experience more chronic or trait anxiety as a result of weak or underdeveloped internal coping skills and strategies. Therefore, entering restrictive housing and the sudden increase in state anxiety to which this gives rise, burdens and overwhelms the individual's habitually weak internal coping skills, thus highlighting and augmenting the trait anxiety that serves as the second step of this sequential model. Augmented trait anxiety, a consequence of an overtaxed and inefficient internal coping system, then leads to psychological disturbances that are eventually diagnosed as SHU syndrome. This explains how SHU syndrome can occur in persons with no formal history of mental health disorder and why it is the exception rather than the rule in prison studies on restrictive housing because it is reserved for those with the strongest externalizing tendencies and weakest internal coping skills.

Figure 1.

Overview Of the model tested in the current study with the externalizing tendencies that overlay the two processes (strain created by placement in restrictive housing and ineffective coping) that lead to psychological deterioration in the top row, the two processes and psychological deterioration in the middle row, and the measures used to assess each element in the model in the bottom row.

Present Study

The data for the current investigation comes from the previously reviewed O’Keefe et al. (2013) study. In an analysis conducted using these same data, Walters (2018) uncovered minimal support for SHU syndrome, noting instead that psychological decompensation was primarily a function of mental illness in both general population and restrictive housing inmates. Still, there is the possibility that a relatively small group of individuals with no prior history of mental health disorder experience genuine signs of SHU syndrome and that mental illness in studies such as the one conducted by Walters (2018) obscure this effect. Consequently, the current study confined itself to individuals rated by staff as having no major mental health needs. The one control variable included in this investigation was a prior measure of the dependent or outcome variable designed to assess psychological adjustment so that deterioration in psychological adjustment could be assessed as a lagged dependent variable.

The moderated mediation hypothesis tested in this study predicted that:

H1: Time 2 trait anxiety would mediate the relationship between the Time 1 state anxiety and Time 3 psychological maladjustment, controlling for Time 1 psychological maladjustment, but only in inmates assigned to restricted housing.

Method

Participants

The sample for the current study consisted of 69 non-mental health participants from the Colorado Department of Corrections Longitudinal Study of Psychological Effects of Administrative Segregation (CDOC-PEAS: O’Keefe et al., 2011). All participants were male, although other individual demographic data were unavailable as these data were masked in the version of the dataset accessed by the author through the Inter-University Consortium for Political and Social Research (ICPSR).The average participant in the full CDOC-PEAS sample (N = 270) was 31.8 years of age (SD = 9.1, range = 17–59), and the racial/ethnic breakdown was 40% White, 36% Hispanic, 18% African American, 4% Native American, and 1% Asian.

Sample Construction

Participation was confined to CDOC-PEAS members with no apparent history of mental health disorder or treatment as evaluated by a committee of staff members with access to mental health reports on each inmate. It was reasoned that because advocates of SHU syndrome propose that segregation housing creates mental health symptoms in inmates without a documented history of mental illness, it would be best to restrict the sample to individuals who entered the study without a documented history of mental health need. A prior study using these same data, in fact, determined that mentally ill inmates were just as likely to deteriorate in general population as they were in segregation housing (Walters, 2018). There were 102 participants in the CDOC-PEAS study who satisfied the first criterion (i.e., no documented history of mental illness), 60 of whom were housed in segregation, and 42 of whom were housed in general population.

Listwise deletion was used to select a sample of inmates with complete data on all five study variables. Discarding data on the 33 inmates with missing data on one or more variables resulted in a sample of 69 participants with complete data and no history of mental illness (42 were housed in segregation and 27 were housed in general population). In implementing listwise deletion, it was discovered that 85% of the cases with missing data had not been evaluated on the outcome measure (BPRS-3), either because they withdrew their consent to participate or were released from prison on parole. Listwise deletion was viewed as justified based on the results of Little's MCAR test which showed that data were missing completely at random: χ² (10) = 9.60, p = .48.

Measures

Group Status

Participants belonged to one of two groups. One group was housed in general population (1) and the other group was housed in administrative segregation (AS) (2). Both groups had participated in an AS hearing and, as a result, were either assigned to the general population of a regular institution or transferred to a locked facility referred to as the segregation housing unit (SHU) or restrictive housing in the current study. Inmates in AS were single-celled and spent 23 to 24 hours a day in their cells. Group status served as the moderator variable in this study, and was included in an interaction with Time 1 state anxiety after both variables had been centered.

State-Trait Anxiety Inventory (STAI)

The STAI (Spielberger et al., 1983) is a 40-item self-report measure designed to assess state or situational anxiety (20 items: e.g., “I am tense” “I feel confused,” “I feel self-confident” [reverse coded]) and trait or dispositional anxiety (20 items: e.g., “I feel that difficulties are piling up so that I cannot overcome them,” “I worry too much over something that really doesn't matter,” “I make decisions easily” [reverse coded]). Each item on the STAI is rated on a four-point scale—1 = almost never, 2 = sometimes, 3 = often, 4 = almost always—after which, item scores are summed to create two scores, a state anxiety score and a trait anxiety score. Internal consistency estimates (Cronbach alpha: α) for the Time 1 State and Time 2 Trait scales of the STAI were excellent in the current sample of participants: .91 and.92, respectively.

Brief Psychiatric Rating Scale (BPRS)

The BPRS (Overall & Gorham, 1962; Ventura et al., 1993) is a 24-item rating scale designed to assess rapidly changing symptoms of serious psychiatric disorder like “somatic concern,” “anxiety,” “depression,” “suicidality,” “guilt,” “hostility,” “elevated mood,” “grandiosity,” “suspiciousness,” “hallucinations,” “unusual thought content,”“bizarre behavior,” “self-neglect,” “disorientation,” “conceptual disorganization,” “blunted affect,” “emotional withdrawal,” “motor retardation,” “tension,” “uncooperativeness,” “excitement,” “distractibility,” “motor hyperactivity,” and “mannerisms and posturing.” Each item on the BPRS is rated on a 7-point scale (1 = not present, 7 = extremely severe), and the ratings summed to create a score that can range from 24 to 168. There is research to suggest that the BPRS is sensitive to the psychological status of incarcerated forensic patients (Greenwood & Burt, 2001; van Beek et al., 2015). Internal consistency for the BPRS was adequate in the current sample of participants during Time 1 (α = .71) and Time 3 (α = .70) of the CDOC-PEAS.

Research Design and Statistical Analyses

The mental health status of participants was evaluated with the BPRS every six months by mental health staff affiliated with the CDOC. Several self-report measures were also completed during each time period, the STAI being one of them. The model tested in this investigation began with the interaction between state anxiety (independent variable) and group status (moderator variable) at Time 1, moved to trait anxiety at Time 2 (mediator variable), and ended with the BPRS at Time 3 (dependent variable). A longitudinal panel design administered across three time periods, with a six-month gap between periods, was implemented with all 69 non-mental health participants from the CDOC study who had complete data on all five study variables.

Analyses were conducted with SPSS Version 26 (IBM, 2019). This included both descriptive/correlational and inferential (regression) analyses. With respect to the regression analysis, an ordinary least squares (OLS) moderated mediation path analysis was performed with Hayes (2018) Process program, Model 7. Besides assessing the results with normal theory procedures (i.e. OLS regression), significance was also determined by constructing percentile bootstrapped 95% confidence intervals, which research indicates are more reliable and less subject to bias than normal theory procedures (MacKinnon et al., 2007; Preacher, 2015).A sensitivity analysis was performed using Kenny’s (2013) “failsafe ef” procedure. Considering the fact that missing data were handled with listwise deletion, all participants had complete data on all five variables.

Results

Preliminary Analyses

Descriptive statistics and correlations for the five variables included in this study are listed in Table 1.As these results indicate, the temporal relationship between state and trait anxiety over a period of six months was strong, as was the test-retest reliability of the BPRS over a period of one year. If SHU syndrome is defined as a BPRS-3 score greater than or equal to one standard deviation above the mean (raw score ≥ 31) in an individual with a BPRS-1 score less than one standard deviation above the mean (raw score < 34), nearly four times as many inmates in segregation (6/37 or 16.2%) as compared to general population (1/24 or 4.4%) satisfied the criteria for SHU syndrome, although the difference was not statistically significant, χ²(1) = 2.08, p = .15. Collinearity diagnostics conducted on the two regression equations revealed no evidence of multicollinearity between predictor variables (tolerance = .698–.919; variance inflation factor = 1.088–1.432).

Table 1.

Descriptive Statistics and Correlations for the 5 Variables Included in the Current Investigation.

Variable	M	SD	Range	2	3	4	5
1. Group Assignment	1.61	-	1–2	.06	.28	.16	.07
2. State Anxiety-1	2.04	0.54	1.15–3.80		.47	.20	.14
3. Trait Anxiety-2	1.86	0.52	1.00–3.20			.17	.44
4. BPRS-1	28.81	4.65	24–44				.58
5. BPRS-3	27.22	4.26	24–48

Note. Variable = variable name; M = mean, SD = standard deviation; Range = range of scores in the current sample; Group Assignment = general population (1) vs. restrictive housing (2), State Anxiety-1 = STAI state anxiety score at Time 1, Trait Anxiety-2 = STAI trait anxiety score at Time 2, BPRS-1 = Brief Psychiatric Rating Scale score at Time 1, BPRS-3 = Brief Psychiatric Rating Scale score at Time 3, N = 69.

Main Analyses

A path analysis was performed, the results of which showed evidence of moderated mediation (see Table 2 and Figure 2)¹. The independent variable (state anxiety), moderator variable (group), and their interaction (state anxiety x group), all measured at Time 1, correlated significantly with trait anxiety, measured at Time 2. Trait anxiety, in turn, mediated the relationship between the state anxiety x group interaction at Time 1 and the total BPRS score at Wave 3, controlling for the direct effect of Time 1 state anxiety on Time 3 BPRS and prior maladjustment (BPRS-1).Also, the use of a lagged dependent variable (BRPS-3, controlling for BPRS-1) revealed that the indirect effect predicted a change in adjustment status, but that the direct effect of State Anxiety-1 on BPRS was non-significant.²

Figure 2.

Ordinary Least squares (OLS) path analysis of trait anxiety as a mediator of the relationship between group assignment, state anxiety, and the group x state anxiety interaction, on the one hand, and time 3 BPRS total score, on the other hand. Note. Standardized beta coefficients are reported; the BPRS control and precursor estimates are not shown; N = 69. *p < .05, **p < .001.

Table 2.

Results of a Least Squares Moderated Mediation Path Analysis Composed of Two Regression Equations.

Predictor	β	SE	t	p	95% PBCI
Trait Anxiety-2 (Outcome)
Constant	−0.336	.116	−2.89	.005	[−0.564, −0.114]
Group Assignment	0.296	.088	3.36	.001	[0.120, 0.461]
State Anxiety-1	0.474	.098	4.84	<.001	[0.296, 0.680]
Group × State Anxiety-1	0.259	.094	2.76	.008	[0.057, 0.429]
BPRS-1	0.071	.145	0.49	.625	[−0.195, 0.329]
BPRS-3 (Outcome)
Constant	−0.082	.085	−0.97	.336	[−0.418, 0.183]
Trait Anxiety-2	0.336	.080	4.20	<.001	[0.156, 0.512]
State Anxiety-1	−0.129	.077	−1.68	.097	[−0.278, 0.006]
BPRS-1	0.589	.101	5.82	<.001	[0.186, 0.868]

Note. Outcome = dependent or outcome measure for the regression equation, with the variables under each outcome measure, the predictors for that equation; Constant = constant or intercept; Group Assignment = general population (1) vs. restrictive housing (2), State Anxiety-1 = STAI state anxiety score at Time 1, Trait Anxiety-2 = STAI trait anxiety score at Time 2, BPRS-1 = Brief Psychiatric Rating Scale at Time 1, BPRS-3 = Brief Psychiatric Rating Scale at Time 3; β = standardized coefficient; SE = standard error; t = asymptotic t-test; p = significance level of the asymptotic t-test; 95% PBCI = 95% percentile bootstrapped confidence interval; N = 69.

The results of the moderated mediation analysis, as summarized in Table 3, indicated that the indirect effect of state anxiety at Time 1 on BPRS at Time 3, via trait anxiety at Time 2, was significant in participants housed in SHU but not in inmates housed in general population (Conditional Indirect Effects) and that the difference between indirect effects was statistically significant (Index of Moderated Mediation). A diagram of the significant interaction along the a path of the indirect effect in Figure 3 illustrates how this relationship worked. Higher levels of state anxiety contributed significantly more to higher subsequent levels of trait anxiety in participants housed in SHU than in participants housed in the general population facility, as evidenced by the difference in the slope between the two lines.³

Figure 3.

Interaction between state anxiety and group assignment in predicting time 2 trait anxiety. Note. GP = general population, SHU = special housing unit/restrictive housing.

Table 3.

Results of the Conditional Indirect Effects and Index of Moderated Mediation.

Conditional Indirect Effects	Effect	SE	LLCI	ULCI
General Population	0.056	0.055	−0.029	0.189
Restrictive Housing	0.232	0.070	0.103	0.375
Index of Moderated Mediation	Index	SE	LLCI	ULCI
Group Assignment	0.176	0.076	0.030	0.328

Note. Effect = value of the percentile bootstrapped conditional indirect effects, Index = value produced by the index of moderated mediation, SE = standard error of the percentile bootstrapped conditional indirect effects or index of moderated mediation, LLCI = lower limit of the percentile bootstrapped 95% confidence interval, ULCI = upper limit of the percentile bootstrapped 95% confidence interval, N = 69.

Sensitivity testing was performed with the aid of the “failsafe ef” procedure to assess for omitted variable bias. Using Kenny’s (2013) formula— (r_my.x) × (sd_m.x) × (sd_y.x) / (sd_m) × (sd_y)—the “failsafe ef” was calculated for the purpose of determining how well an unobserved covariate confounder would need to comport with the mediator and dependent variables, controlling for the independent and mediator variables in the case of the latter, to completely eliminate the b path coefficient of the significant indirect effect. The results of the current sensitivity analysis revealed that a confounding covariate would need to correlate .38 with Trait Anxiety-2 and .38 with BPRS-3, controlling for State Anxiety-1 and Trait Anxiety-2, to fully eliminate the significant indirect effect along the b path that runs from Trait Anxiety-2 to BPRS-3.

Inclusion of Mental Health Cases

When the 131 participants from the CDOC-PEAS study with a history of mental health disorder were added to the analysis (N = 200), the mediating effect of trait anxiety at Time 2 on the relationship between Time 1 state anxiety and Time 3 BPRS remained significant, whereas the moderating effect of group status on this same relationship was no longer significant. This outcome was not unexpected. First, prior research has shown that mentally ill inmates from the CDOC-PEAS are just as likely to decompensate in general population as they are in segregation (Walters, 2018). Second, mentally ill inmates were nearly twice as prevalent in the general population group as they were in the segregation group.

Discussion

SHU syndrome is one of the more controversial and contentious topics in the field of corrections. There are those who view it and its alleged cause, restrictive housing, as an ongoing threat to correctional clients and a common occurrence in jails and prisons (Muller, 2018). This is countered by an equally strong belief that support for SHU syndrome is weak and based almost exclusively on anecdotal and poorly controlled cross-sectional studies (Hanson, 2011). The truth, it would seem, lies somewhere in-between these two extremes. Longitudinal studies, currently the best way to test SHU syndrome outside of a true experiment, have produced mixed results (Andersen et al., 2003; Chadick et al., 2018; O’Keefe et al., 2013; Zinger et al., 2001), yet several general conclusions can be offered. First, most inmates in restrictive housing, even those who spend a great deal of time in such housing, do not decompensate psychologically, although the fact that a minority of inmates do experience such deterioration is reason for concern. Second, SHU syndrome is not the sole result of exposure to restrictive housing. Unlike the sensory deprivation experiments of the 1950s (Hebb, 1980), inmates in restrictive housing are not completely shut off from sensory experience and most find ways to cope with the diminished sensory input found in these environments. Nonetheless, some inmates, even those without a prior history of mental illness, display signs of psychological deterioration when placed in restrictive housing. The purpose of the current study was to account for these two general conclusions while testing a model of SHU syndrome development, with results showing the presence of a relatively weak effect that seems to stem from ineffective inmate coping.

According to the model tested in this study, certain characteristics of the individual interact with the decreased sensory input instilled by restrictive housing to put some inmates at risk for SHU syndrome. A guiding assumption of this model is that inmates are externally oriented and as such, more likely to use external coping strategies to deal with stress than persons who are more internally oriented. External coping strategies, defined as coping behaviors directed at the external environment, are much less effective in combating stress in a locked unit than in the regular prison milieu where there are more opportunities for varied environmental experience. The end result is an elevated level of situational strain upon being confined in restrictive housing (represented in the current study by the state anxiety x group interaction), followed by a dysfunctional coping response marked by chronic strain and negative affect (represented in the current study by trait anxiety), which, in turn, gives rise to symptoms of SHU syndrome (represented in the current study by an increase in the BPRS from Time 1 to Time 3). A test of the model produced results that were specific to the proposed sequence (i.e., the trait anxiety x group interaction was nonsignificant; see Footnote 2) and highly robust to the effects of omitted variable bias (“failsafe ef” results). These findings also help explain two points highlighted in the previous paragraph. First, SHU syndrome is the exception rather than the rule when it comes to inmates in restrictive housing (i.e., only those with the weakest internal coping resources and least amount of environmental support will decompensate), and second, SHU syndrome is not solely a function of being placed in restrictive housing but involves a complex person x situation interaction.

Theoretical Implications

The person x situation interactive model evaluated in the current study found support in a moderated mediation analysis of data from the CDOC-PEAS in which group status interacted with state anxiety to promote trait anxiety, which, in turn, led to an increase in psychological symptomatology. This implies that restrictive housing is a necessary but not sufficient condition for the appearance of symptoms of SHU syndrome and that certain characteristics of the individual are important in determining whether placement in restrictive housing will result in psychological deterioration. Although not all aspects of the conceptual model were tested in this study, the results support the idea that an interaction between placement in restrictive housing and a high degree of situational anxiety set off a sequence of events that can lead to mild psychological deterioration that partially supports the notion of SHU syndrome. Once this sequence is set into motion, the inmate's psychological stability will become more tenuous because their habitual means of coping is no longer able to assist them in warding off the mounting anxiety of confinement in a restricted housing unit. It is possible that the model tested in this study is incomplete and requires additional elements, one of which may be a feedback loop whereby ineffective coping (high trait anxiety) leads to more situational distress (high state anxiety), which then makes coping even less effective. Such a possibility should be tested in future research.

Research Implications

There are several potentially important research implications stemming from the results of this study. These implications range from how SHU syndrome should be studied to who should serve as the focus of investigation. As the present findings and prior research indicate, restrictive housing is but one component of a multi-component process responsible for SHU syndrome. It is therefore imperative that person factors and the person x situation interaction be included in future research on psychological decompensation in response to restrictive housing. The notion that restrictive housing alone is responsible for SHU syndrome is outdated and inconsistent with the extant literature. In the current study, for instance, the direct effect of group status on psychological adjustment was non-significant and negative (see Footnote 1) and therefore incapable of explaining SHU syndrome in isolation. Research indicates that inmates with mental health problems are more subject to psychological deterioration in both restrictive housing and general population (Walters, 2018). Hence, a second research implication centers around the question of who should serve as participants in research on SHU syndrome. The answer, I would argue, is that inmates without significant mental health problems should be the focus of investigation in research on SHU syndrome. Given that mentally ill inmates are equally likely to decompensate in restrictive and general population housing, the true test of SHU syndrome is when participants are free of mental illness, as was the case in the current study. This is not meant to suggest that research on mental illness in prison is unimportant, just that including the mentally ill in research on SHU syndrome may obscure a small but meaningful effect.

Practical Implications

There are several practical implications that can also be drawn from the current results. First, correctional programs could start at the source by teaching inmates to focus more on themselves and less on their environment. Given that the externalizing tendencies of criminal offenders are long-standing and engrained, altering these tendencies will require more time and manpower than most correctional systems can afford, with no guarantee that long-lasting change will occur. A more realistic alternative would be to engage inmates in cognitive-behavior therapy as a way of enhancing their internal coping skills through stress and anger management and positive self-talk (Tafrate & Mitchell, 2014).Two programs developed specifically for inmates in restrictive housing, Taking a Chance on Change (Folk et al., 2016) and Stepping Up and Stepping Out (Batastini et al., 2021), have produced promising initial results using a cognitive-behavioral framework. The symptoms of SHU syndrome can also be managed by avoiding single-person cells in restrictive housing, as long as this does not put an inmate in jeopardy. There are indications that inmates assigned to two-person restrictive housing cells are less likely to commit suicide than inmates housed alone (Tartaro & Lester, 2010). A reasonable assumption, then, is that celling inmates together could potentially reduce the stress and isolation that contribute to SHU syndrome. Allowing inmates to spend more than the allotted 30 to 60 minutes a day outside their cells, providing them with more recreational opportunities, offering incentives for positive behavior, and placing a limit on how much time inmates spend in segregation are other policy initiatives than could potentially alleviate some of the factors that contribute to psychological decompensation in restrictive housing.

Limitations

Two of the principal strengths of this study are, first, that it employed a longitudinal research design with no overlap between waves and, second, that it minimized mono-operational bias (Shadish et al., 2002) by relying on separate sources of data—self-report measures for the independent, moderator, and mediator variables, and staff ratings for the dependent variable. Weaknesses, on the other hand, are also evident. First, the sample, as has been the case with most studies on SHU syndrome, was minute. In addition, basic demographic information (i.e., age and race) were unavailable from the version of the dataset accessed through the ICPSR. Second, the effects attributed to SHU syndrome in this study were small, with the cutoff for “SHU syndrome” being only 7 points above the minimum score possible on the BPRS. Third, externalization, which is a key component of the model used to explain SHU syndrome in this study, was not directly assessed. Instead, it was simply assumed that given the externalizing nature of criminal behavior that many of the inmates in this study would employ externalizing coping techniques to deal with the stress and strain of restrictive housing. Fourth, as with externalization, coping strategies were not directly assessed either. Instead, weak internalizing coping strategies were estimated from their putative consequences, i.e., elevated levels of both state and trait anxiety. In the future, it will be important to directly assess both externalizing tendencies and coping strategies when conducting research on the model presented in this study as a means of explaining SHU syndrome. Such research should also be of assistance in illustrating why SHU syndrome does not occur in all inmates placed in long-term restrictive housing.

Conclusion

Based on the current findings and prior research, it is concluded that SHU syndrome exists, although it exists in only a minority of inmates and then only as part of an ongoing person x situation interaction. Furthermore, the effect, while measurable and potentially meaningful, is rather small. Most offenders, it would seem, have sufficient coping skills and resources to deal effectively with the deprivations (Sykes, 1958) presented by restrictive housing. A relatively small group of offenders, however, because of some combination of deficit internal coping skills, weak social support, and extreme environmental stress suffer psychological deterioration after spending time in restrictive housing, even without a documented history of mental illness. Hence, SHU syndrome is seated neither exclusively in the environment nor in the individual, but, rather, is a complex expression of an ongoing person x situation interaction. As a consequence, the psychological symptoms associated with SHU syndrome may occur in one inmate but not another, at one point in time but not at another point in time, or under one set of circumstances but not under another set of circumstances. Helping offenders develop their internalized coping skills and strategies, while working to make restrictive housing less stressful and isolating, could go a long way toward alleviating the problem of SHU syndrome for inmates who are at risk of experiencing this potentially debilitating disorder.

Footnotes

Notes

Author Biography

Glenn D. Walters, PhD, is professor of criminal justice at Kutztown University. His background includes 29 years working as a clinical psychologist in the military and Federal Bureau of Prisons. Current research interests focus on causal mediation analysis and the development of an overarching psychological theory of criminal behavior.

References

Aebi

Giger

Plattner

Metzke

C. W.

Steinhausen

H.-C.

(2014). Problem coping skills, psychosocial adversities and mental health problems in children and adolescents as predictors of criminal outcomes in young adulthood. European Child and Adolescent Psychiatry, 23(5), 283–293. https://doi.org/10.1007/s00787-013-0458-y

Agnew

(1992). Foundation for a general strain theory. Criminology; An Interdisciplinary Journal, 30(1), 47–87. https://doi.org/10.1111/j.1745-9125.1992.tb01093.x

Andersen

H. S.

Sestoft

Lillebæk

Gabrielsen

Hemmingsen

(2003). A longitudinal study of prisoners on remand: Repeated measures of psychopathology in the initial phase of solitary versus nonsolitary confinement. International Journal of Law and Psychiatry, 26(2), 165–177. https://doi.org/10.1016/S0160-2527(03)00015-3

Batastini

A. B.

Lester

M. E.

Morgan

R. D.

Atterberry

(2021). Stepping up, stepping out: A program description and preliminary findings. Psychological Services, 18(4), 679–688. https://doi.org/10.1037/ser0000430

Bhattacharya

Pal

Acharyya

Dasgupta

Guha

Datta

(2016). Depression and anxiety in mothers of children with cancer and how they cope with it: A cross-sectional study in eastern India. ASEAN Journal of Psychiatry, 17(1), 1–8.

Chadick

C. D.

Batastini

A. B.

Levulis

S. J.

Morgan

R. D.

(2018). The psychological impact of solitary: A longitudinal comparison of general population and long-term administratively segregated male inmates. Legal and Criminological Psychology, 23(2), 101–116. https://doi.org/10.1111/lcrp.12125

Fatima

Tahir

(2013). Comparison of coping strategies used by adolescents on state and trait anxiety. FWU Journal of Social Sciences, 7(2), 165–174.

Fischer

D. R.

(2002). Arizona department of corrections: Security Threat Group (STG) program evaluation. National Institute of Justice, U.S. Department of Justice.

Folk

J. B.

Disabato

D. J.

Daylor

J. M.

Tangney

J. P.

Barboza

Wilson

J. S.

(2016). Effectiveness of a self-administered intervention for criminal thinking: Taking a chance on change. Psychological Services, 13(3), 272–282. https://doi.org/10.1037/ser0000079

10.

Gendreau

Bonta

(1984). Solitary confinement is not cruel and unusual punishment: People sometimes are!. Canadian Journal of Criminology, 26(4), 467–478. https://doi.org/10.3138/cjcrim.26.4.467

11.

Grassian

(1983). Psychopathological effects of solitary confinement. American Journal of Psychiatry, 140(11), 1450–1454. https://doi.org/10.1176/ajp.140.11.1450

12.

Grassian

(2006). Psychiatric effects of solitary confinement. Washington University Journal of Law and Policy, 22, 324–383.

13.

Greenwood

Burt

(2001). Validity of the brief psychiatric rating scale within a forensic inpatient hospital. Expert Evidence, 8(1), 15–30. https://doi.org/10.1023/A:1012285717450

14.

Haney

(2003). Mental health issues in long-term solitary and “supermax” confinement. Crime & Delinquency, 49(1), 124–156. https://doi.org/10.1177/0011128702239239

15.

Hanson

(2011, August). Solitary confinement: Rumor and reality. Psychology Today . https://www.psychologytoday.com/us/blog/shrink-rap-today/201108/solitary-confinement-rumor-and-reality

16.

Hayes

A. F.

(2018). Introduction to mediation, moderation, and conditional process analysis: A regression-based approach (2nd ed.). Guilford.

17.

Hebb

D. O.

(1980). Essays on mind. Psychology Press.

18.

Hodgins

Côté

(1991). The mental health of penitentiary inmates in isolation. Canadian Journal of Criminology, 33(2), 177–182. https://doi.org/10.3138/cjcrim.33.2.175

19.

Hopwood

C. J.

Grilo

C. M.

(2010). Internalizing and externalizing personality dimensions and clinical problems in adolescents. Child Psychiatry and Human Development, 41(4), 398–408. https://doi.org/10.1007/s10578-010-0175-4

20.

IBM. (2019). IBM SPSS statistics for Windows (Version 26.0) . IBM Corp.

21.

Kenny

D. A.

(2013). Mediation: Sensitivity analysis [webinar] . http://davidakenny.net/webinars/Mediation/Sensitivity/Sensitivity.html

22.

Kupers

T. A.

(2016). The SHU post-release syndrome: A preliminary report. Correctional Mental Health Report, 17(6), 81–95.

23.

Labrecque

R. M.

Gendreau

Morgan

R. D.

King

M. M.

(2020). Revisiting the Walpole Prison Solitary Confinement Study (WPSCS): A content analysis of the studies citing Grassian (1983). Psychology, Public Policy, and Law, 26(3), 378–391. https://doi.org/10.1037/law0000247

24.

MacKinnon

D. P.

Fairchild

A. J.

Fritz

M. S.

(2010). Mediation analysis. Annual Review of Psychology, 58, 593–614. https://doi.org/10.1146/annurev.psych.58.110405.085542

25.

Marguerite

Laurent

Marine

Tanguy

Karine

Pascal

Xavier

(2017). Actor-partner interdependence analysis in depressed patient-caregiver dyads: Influence of emotional intelligence and coping strategies on anxiety and depression. Psychiatry Research, 258, 396–401. https://doi.org/10.1016/j.psychres.2017.08.082

26.

Michalowska

Matusewicz

Samochowiec

(2019). Coping and anxiety in cancer patients. Archives of Psychiatry and Psychotherapy, 21(2), 28–32. https://doi.org/10.12740/APP/103093

27.

Morgan

R. D.

Gendreau

Smith

Gray

Labrecque

R. M.

MacLean

Van Horn

S. A.

Bolanos

A. D.

Batastini

A. B.

Mills

J. F.

(2016). Quantitative syntheses on the effects of administrative segregation on prisoners’ well-being. Psychology, Public Policy, and Law, 22(4), 439–461. https://doi.org/10.1037/law0000089

28.

Motiuk

L. L.

Blanchette

(2001). Characteristics of administratively segregated offenders in federal corrections. Canadian Journal of Criminology, 43(1), 131–143. https://doi.org/10.3138/cjcrim.43.1.131

29.

Muller

R. T.

(2018, May). Solitary confinement is torture: Research reveals numerous adverse effects of solitary confinement on inmates. Psychology Today . https://www.psychologytoday.com/us/blog/talking-about-trauma/201805/solitary-confinement-is-torture

30.

O’Keefe

M. L.

Klebe

K. J.

Metzner

Dvoskin

Fellner

Stucker

(2013). A longitudinal study of administrative segregation. Journal of the American Academy of Psychiatry and Law, 41(1), 49–60.

31.

O’Keefe

M. L.

Klebe

K. J.

Stucker

Sturm

Leggett

(2011). One year longitudinal study of the psychological effects of administrative segregation (NCJRS 232973). U.S. Department of Justice.

32.

Overall

J. E.

Gorham

D. R.

(1962). The brief psychiatric rating scale. Psychological Reports, 10(3), 799–812. https://doi.org/10.2466/pr0.1962.10.3.799

33.

Preacher

K. J.

(2015). Advances in mediation analysis: A survey and synthesis of new developments. Annual Review of Psychology, 66, 825–852. https://doi.org/10.1146/annurev-psych-010814-015258

34.

Ralph

P. H.

Marquart

J. W.

(1991). Gang violence in Texas prisons. Prison Journal, 71(2), 38–49. https://doi.org/10.1177/003288559107100205

35.

Shadish

W. R.

Cook

T. D.

Campbell

D. T.

(2002). Experimental and quasi-experimental designs for generalized causal inference. Mifflin.

36.

Spielberger

C. D.

Gorsuch

R. L.

Lushene

P. R.

Vagg

P. R.

Jacobs

A. G.

(1983). Manual for the State-Trait Anxiety Inventory. Consulting Psychologists Press.

37.

Stepanchuk

Zhirkov

Yakovleva

(2013). The coping strategies, psychological defense mechanisms and emotional response to the disease in Russian patients with chronic leukemia. Procedia: Social and Behavioral Sciences, 86, 248–255. https://doi.org/10.1016/j.sbspro.2013.08.559

38.

Suedfeld

Ramirez

Deaton

Baker-Brown

(1982). Reactions and attributes of prisoners in solitaryconfinement. Criminal Justice and Behavior, 9(3), 303–340. https://doi.org/10.1177/0093854882009003004

39.

Sykes

G. M.

(1958). The society of captives: Study of a maximum security prison. Princeton University Press.

40.

Tafrate

R. C.

Mitchell

(2014). Forensic CBT: A handbook for clinical practice. Wiley.

41.

Tartaro

Lester

(2010). Suicide and self-harm in prisons and jails. Lexington.

42.

van Beek

Vuijk

P. J.

Harte

J. M.

Smit

B. L.

Nijman

Scherder

E. J. A.

(2015). The factor structure of the Brief Psychiatric Rating Scale (expanded version) in a sample of forensic psychiatric patients. International Journal of Offender Therapy and Comparative Criminology, 59(7), 743–756. https://doi.org/10.1177/0306624X14529077

43.

van der Ende

Verhulst

F. C.

Tiemeier

(2020). Multitrait-multimethod analyses of change of internalizing and externalizing problems in adolescence: Predicting internalizing and externalizing DSM disorders of adulthood. Journal of Abnormal Psychology, 129(4), 343–354. https://doi.org/10.1037/abn0000510

44.

Ventura

Lukoff

Nuechterlein

K. H.

Liberman

R. P.

Green

M. F.

Shaner

(1993). Brief Psychiatric Rating Scale (BPRS) expanded version (4.0): Scales, anchor points, and administration manual. International Journal of Methods in Psychiatric Research, 3, 227–244.

45.

Wadsworth

M. E.

(2015). Development of maladaptive coping: A functional adaptation to chronic, uncontrollable stress. Child Development Perspectives, 9(2), 96–100. https://doi.org/10.1111/cdep.12112

46.

Walters

G. D.

(2018). Checking the math: Do restrictive housing and mental health need add up to psychological deterioration? Criminal Justice and Behavior, 45(9), 1347–1362. https://doi.org/10.1177/0093854818780177

47.

Wedekind

Bandelow

Heitmann

Havermann-Reinekce

Engel

K. R.

Huether

(2013). Attachment style, anxiety coping, and personality-styles in withdrawn alcohol addicted inpatients. Substance Abuse Treatment, Prevention, and Policy, 8(1). http://www.substanceabusepolicy.com/content/8/1/1 https://doi.org/10.1186/1747-597X-8-1

48.

Zinger

Wichmann

Andrews

D. A.

(2001). The psychological effects of 60 days in administrative segregation. Canadian Journal of Criminology, 43(1), 47–83. https://doi.org/10.3138/cjcrim.43.1.47