|Home | About | Journals | Submit | Contact Us | Français|
Individuals with severe mental illness (SMI) are at greatly increased risk for trauma exposure and for the development of posttraumatic stress disorder (PTSD). This study reports findings from a large, comprehensive screening of trauma and PTSD symptoms among public mental health clients in a statewide community mental health system. In 851 individuals with SMI and probable PTSD, childhood sexual abuse was the most commonly endorsed “index” trauma, followed closely by the sudden death of a loved one. Participants had typically experienced an average of 7 types of traumatic events in their lifetime. The number of types of traumatic events experienced and Hispanic ethnicity were significantly associated with PTSD symptom severity. Clients reported experiencing PTSD in relation to events which occurred on average 20 years earlier, suggesting the clinical need to address trauma and loss throughout the lifespan, including their prolonged after-effects.
Over the past two decades, a growing body of research has shown that individuals with severe mental illness (SMI) are at greatly increased risk for trauma exposure (see Grubaugh, Zinzow, Paul, Egede, & Frueh, 2011 for a review). While national surveys indicate that more than half of people in the general population report exposure to at least one event that meets DSM-IV criteria for trauma (Sledjeski, Speisman, & Dierker, 2008; Kessler, Onnega, Bromet, Hughes & Nelson, 1995), studies of people with a SMI (such as schizophrenia, bipolar disorder, or major depression) suggest that trauma exposure is nearly universal, with multiple traumas being the norm (Goodman et al., 1997; Mueser et. al, 1998; Mueser et. al., 2004a). Violent victimization, a particularly toxic class of trauma, is unusually common in people with SMI, with 34–53% reporting child abuse, and 43–81% reporting lifetime victimization (Coverdale & Turbott, 2000; Mueser et al, 1998).
The high rates of trauma exposure among people with SMI, combined with possibly increased vulnerability to the effects of trauma (Bebbington et al, 2004; Garno et. al, 2005; Read et. al, 2001), are associated with an increased prevalence of PTSD in this population (Grubaugh, Elhai, Cusack, Wells & Frueh, 2007). Specifically, in most studies, the current prevalence of PTSD among persons with SMI has been found to range from 28 to 43% (Calhoun et al., 2007; Cascardi et al., 1996; Craine et al., 1988; Cusack et al., 2006; Ford et al., 2007; Goldberg & Garno, 2005; Howgego et al., 2005; McFarlane, 2001; Mueser et al., 1998, 2001, 2004; Picken & Tarrier, 2011; Strauss et al., 2006), although a few studies have reported lower, but nevertheless increased rates ranging from 16–18% (Fan et al., 2008; Lommen et al., 2009; Neria et al., 2002). This contrasts with an estimated current rate of 3.5% for PTSD in the general population (Kessler, Chiu, Demler, & Walters, 2005). Despite evidence that PTSD is a significant clinical problem among people with SMI, many questions remain regarding the nature of PTSD in this population (Grubaugh et al., 2011; Rosenberg et al., 2002). Although the types of traumatic exposure commonly experienced by people with SMI have been previously reported (e.g., Mueser et al., 1998; McFarlane et al., 2001; Goldberg & Garno, 2005; Goodman et al., 2001), limited data are available on which events are most distressing and most likely to lead to PTSD. In a survey of trauma exposure and associated distress and PTSD symptoms in people with SMI, O’Hare and Sherrer (2011) reported that the most distressing event was sexual assault (either in childhood or adulthood), followed by physical assault and the sudden unexpected death of a loved one; sexual assault was the strongest predictor of PTSD symptoms, followed by unexpected death. Another study of individuals with SMI reported that exposure to childhood sexual abuse was more uniquely predictive of PTSD than any other types of trauma (Mueser et al., 1998), whereas Goldberg and Garno (2005) found that history of adult sexual assault or a history of suicide or homicide in a close friend or relative were more strongly related to PTSD.
No studies that we know of have evaluated the relationship between exposure to different types of traumatic events and PTSD symptom severity among people with SMI and probable PTSD. A better understanding of which traumatic events clients with SMI and PTSD find most distressing, and which events are most strongly related to PTSD symptom severity, could inform specific trauma interventions for this population. The experience of traumatic events and their relationship to PTSD symptom severity tends to differ by gender (Kessler, Sonnega, Bromet, Hughes, & Nelson, 1995; Breslau, Davis, Andreski, & Peterson, 1991; Norris, Foster, & Weishaar, 2002), so the differential impact of traumatic events on PTSD among people with SMI also needs to be examined. Consistent with research in the general population, studies suggest that women with SMI are significantly more likely to experience sexual violence than men, both in childhood and adulthood (see Grubaugh et al., 2011 for a review).
In addition to evaluating the importance of exposure to different types of traumatic events in people with SMI, there is a need to further examine the role of ethnicity in the experience of these events and their effects on PTSD symptoms. Some have suggested that culture may have an influence on the impact of traumatic events (Carlson, 2005; Fontes, 1995), e.g., by moderating the relationship between trauma exposure and development of psychopathology (Garcia-Coll & Garrido, 2000). Studies in the general population have found that Hispanic individuals are more vulnerable to developing PTSD when exposed to sexual, assaultive, or combat-related traumatic events (Ortega & Rosenheck, 2000; Torres & Han, 2000), and, among those with PTSD, Hispanics experience more severe symptoms than persons from other ethnic backgrounds (Marshall, Schell & Miles, 2009; Pole et al., 2005). While one study found higher rates of PTSD among Hispanic individuals with SMI (Mueser et al., 2004), we know of no other studies that have examined the relationship between ethnicity and PTSD symptom severity in this population.
To address these gaps in the literature, this study reports findings from a comprehensive screening of trauma and PTSD symptoms in public mental health clients in a statewide community mental health system. Among a large group of individuals with SMI and probable PTSD, we examined: the types of trauma experienced; which traumatic events were most distressing to participants; and the association between traumatic events, demographic and clinical characteristics, and PTSD symptom severity.
The study protocol, informed consent, and all study-related materials were reviewed and approved by the Institutional Review Boards at Dartmouth Medical School and the University of Medicine and Dentistry of New Jersey (UMDNJ) – Robert Wood Johnson Medical School.
Study participants were clients with SMI (defined by the State of New Jersey) receiving services at the University of Medicine and Dentistry of New Jersey—University Behavioral HealthCare (UMDNJ-UBHC). UBHC serves approximately 15,000 clients annually, and is one of the largest mental health specialty providers in the United States. In addition to outpatient clinics and partial hospitalization clinics (five of which participated in the study), UBHC is also equipped with programs such as intensive case management services, residential programs, an emergency room, and an inpatient unit. UBHC serves clients on Medicaid/Medicare (56%) as well as uninsured/self-pay clients (20%).
Acceptance into services at UMDNJ-UBHC requires meeting New Jersey criteria for SMI, which include diagnosis: DSM-IV diagnosis; disability: within the past 3–6 months, the mental disorder has resulted in functional limitations in major life activities that would be appropriate for the client’s developmental stage; and duration: during the past 2 years, and due to a mental disorder, a) there were two or more treatment episodes of greater intensity than outpatient services, such as inpatient, emergency or partial hospitalization care, or a single episode lasting 3 months or more, or b) the normal living situation was disrupted to the point that supportive services were required to maintain the client in that home or residence or housing, or law enforcement officials intervened. While these criteria are similar to “broad” criteria for SMI which have been discussed in the literature (e.g., Ruggieri, Leese, Thornicroft, Bisoffi, & Tansella, 2000), we removed participants with no Axis I diagnosis other than substance use, as this is a further criterion for SMI in other jurisdictions.
Study sites included 5 outpatient and partial hospitalization programs located in three different cities in central and northern New Jersey. A comprehensive screening of trauma exposure and PTSD symptoms was implemented at these sites as part of a research study aimed at evaluating two different treatments for PTSD in people with SMI. Clients were not paid for their participation in the screening. This screening sought to identify clients with SMI and probable PTSD, who were then approached for participation in a treatment study.
An abbreviated 16-item version of the Traumatic Life Events Questionnaire (TLEQ) (Kubany et al., 2000) was used to screen lifetime trauma history for all clients at one of the five sites. For each event on the scale, the client indicated whether he or she had ever experienced it over their lifetime in a binary (yes/no) format (e.g., “Has anyone threatened to kill you or seriously hurt you?”). The TLEQ asks about the experience of traumatic events using wording that corresponds with the DSM-IV criterion A for PTSD. This version of the TLEQ was used to screen for trauma exposure in previous studies with persons with SMI (Mueser et al., 2008).
The PTSD Checklist (PCL) (Blanchard, Jones-Alexander, Buckley, & Forneris, 1996) was used to screen and identify cases with probable PTSD, as well as to assess PTSD symptom severity. The PCL includes one question for each DSM-IV PTSD symptom, requiring the respondent to rate the severity of each symptom over the past month on a 5-point Likert scale (range: 17–85). The PCL has good test-retest reliability and convergent validity in people with SMI (Mueser et al., 2001; Grubaugh et al., 2007). A total score of 45 or greater on the PCL was used to identify cases of probable PTSD during the initial screening (Blanchard et al., 1996). A mean imputation procedure whereby the scale mean for the participant was substituted for missing items was used in cases where individual items were missing for participants. A total of 89 items (or less than .01%) across the 14467 responses were replaced according to this procedure.
From 1/12/2007 to 8/2/2010 clinicians conducted routine screening (using the aforementioned measures) of trauma history and PTSD with their clients, either at the second intake session for new clients or at regular sessions for clients who were already in treatment (note: treatment participants were receiving was routine mental health treatment and not the specific treatment targeting PTSD for which they were being screened). When clients were grossly psychotic or suicidal, the screening was deferred until a later time when the person was more clinically stable.
Clients first completed the TLEQ. If they indicated ‘yes’ to any of the 16-items, they then completed the PCL, based on the most upsetting event identified on the TLEQ (only two clients in the current study were unable to identify one most distressing event, and data were noted as “missing” for these participants). Clients with probable PTSD (PCL ≥ 45) were then asked if they were willing to have their screening data, and other pertinent clinical information, provided to the research team for possible participation in a treatment study. Clients who agreed then completed a consent form, and the results of the screening and other clinical information were then provided to the research team.
Data on primary psychiatric diagnoses, ethnicity, education level, and age were drawn from participants’ medical records after they had provided consent.
Within a 31-month period, a total of 851 clients endorsed at least one traumatic event on the TLEQ, had a total PCL score of at least 45 indicating probable PTSD, had a chart diagnosis of an Axis I diagnosis other than substance use disorder, and expressed an interest in the study. Data are not available on screened individuals who did not show evidence of likely PTSD, as clinicians were instructed to seek consent for release of the information to research team only if the participant showed evidence of likely PTSD and would be a candidate for targeted PTSD treatment. However, we note that the ethnic and diagnostic characteristics of the study sample were similar to the characteristics of clients at the participating sites of UBHC (demographic information available for four out of five participating sites was as follows: demographic information for participants from these four sites was the following: 30% European Americans, 47% African Americans, 14% Hispanic, 8% other; 17% with schizophrenia-spectrum disorders).
Table 1 summarizes the demographic and clinical characteristics of the study sample. The participants were predominantly female (64%), in their early 40’s, and had completed high school. Participants were ethnically diverse, with most self-identifying as African-American (44%) or European-American (33%); 14 % identified themselves as Hispanic. The most common principal Axis I diagnoses in clients’ charts were major depressive disorder and other depressive disorders (46%), schizophrenia and other psychotic disorders (17%), and bipolar disorders I and II (22%). Of note, only 5% of the sample had PTSD listed as a primary diagnosis in their medical record.
Table 2 lists the total number of types of traumatic events reported by the participants on the TLEQ by gender. On average, both male and female participants reported experiencing 7 or more types of traumatic events listed on the TLEQ. The most common traumatic event was the sudden death of a loved one, which did not differ in frequency by gender. Car accidents and childhood physical abuse were also reported at similar frequencies by both men and women. However, men more frequently experienced robbery, stranger assault, and being threatened, and were significantly more likely than women to report experiencing combat and other accidents. Women more often experienced domestic violence, childhood sexual abuse, adult sexual assault, and stalking than men.
Table 3 lists the traumatic events identified by participants as most distressing on the TLEQ, upon which the PCL was based. The most frequently endorsed distressing event, across gender, was childhood sexual abuse (22%), followed by the sudden death of a loved one/friend (20%). Inspection of the specific nature of the sudden death of a loved one/friend found the following causes: murder (12%), suicide (9%), witnessing death/finding someone dead (7%), car accident/fire (4%), drug overdose (1%), or unspecified (68%). Among women, the next most common most distressing events were childhood sexual abuse (27%), sudden death of a loved one (19%), and being a victim of domestic violence (9%). Among men, the next most common most distressing events were the sudden death of a loved one (21%), childhood sexual abuse (13%), and robbery/stranger assault (10%). On average and across gender, the reported index trauma had occurred almost 19 years prior to the screening (mean = 18.71, SD = 14.30, range from 0–54.71).
Table 4 reports correlations between endorsement of specific traumatic events on the TLEQ and PCL total score. The overall number of types of trauma exposed to as reported in the TLEQ was moderately and significantly correlated with PCL total score (r = .27, p < .001), and specific traumatic experiences (with the exception of car accidents and warfare) were also significantly correlated with PTSD symptom severity. To evaluate diagnostic and demographic correlates of overall trauma exposure and PTSD symptom severity we also performed t-tests or one-way analyses of variance (for categorical variables), or computed Pearson correlations (for continuous variables). Hispanic participants had significantly higher PCL total scores than participants of other ethnic groups (F(3, 838) = 5.19, p = .001). Education level was modestly negatively correlated with PCL total score (r = −.09, p = .008). Age, gender, diagnoses (psychotic vs. other), and index trauma identified as most distressing were not significantly related to PTSD symptom severity.
Finally, we used linear regression to examine which combination of demographic and trauma variables best predicted PTSD symptom severity. For this analysis, we used a step-wise regression approach and included all variables which showed a significant bivariate relationship with PCL total score. Thus, variables included TLEQ total score, all TLEQ items with the exception of warfare and car accidents, ethnicity (dummy coded so that Hispanic participants were contrasted to all other race/ethnic groups) and years of education. We confirmed that PCL scores were normally distributed in the sample, despite our sample being restricted to persons with high scores. The overall equation was significant (R2 = .09, F(4, 824) = 20.99, p < .001). The final equation included four variables: TLEQ total (β = .17, t = 3.80, p < .001), Hispanic Ethnicity (β = .09, t = 2.66, p = .008), “being threatened to be killed” (β = .10, t = 2.46, p = .014) and “being stalked” (β = .09, t = 2.42, p = .016).
The present study represents the first effort (to our knowledge) to examine the predictors of PTSD symptom severity among a large sample of persons with SMI and probable PTSD. Findings provide important implications for the targeting of trauma-based interventions for this population, in that they identify factors associated with risk for particularly severe PTSD symptoms. One important finding was that exposure to more types of traumatic events was a strong predictor of PTSD symptom severity in this sample. These findings extend previous research showing that exposure to a number of types of trauma is predictive of PTSD diagnosis and symptom severity in clients with SMI (Goldberg & Garno 2005; McFarlane et al., 2001; Mueser et al.,1998; O’Hare & Sherrer, 2009), by demonstrating a similar association within the subgroup of clients with probable PTSD.
Among the range of traumatic events experienced, childhood sexual abuse was the most commonly endorsed “index trauma” leading to PTSD symptoms. This is consistent with previous research showing that childhood sexual abuse is uniquely predictive of PTSD in people with SMI (Mueser et al., 1998). The findings support the growing evidence documenting the significant public health burden of childhood sexual abuse in the public sector (Gilbert et al., 2009; Talbot et al., 2009), and, considering the increased rate of such abuse in SMI, underscore the importance of treating the most common sequelae of abuse, PTSD. However, although the most frequently endorsed “index trauma” was childhood sexual abuse, it is interesting to note that the traumatic events that were associated with the most severe PTSD symptoms in the regression equation were being stalked and being threatened. This indicates that, although these events were not most frequently identified by clients as most distressing, the experience of these events was associated with significantly elevated PTSD symptoms above and beyond the effect of experiencing a number of different types of traumatic events.
The traumatic event most frequently experienced in this sample was the sudden unexpected death of a loved one, reported by 78% of the sample. This finding is consistent with other research on the SMI population showing that unexpected death is the most commonly reported traumatic event (Mueser et al., 1998; O’Hare & Sherrer, 2009), and supports Goldberg and Garno’s (2005) assertion that childhood sexual abuse and severe interpersonal loss may sensitize individuals with SMI to the development of PTSD.
The high frequency of unexpected death in this sample of clients with SMI, and its association with PTSD symptom severity, raises the question of the related disorder of complicated grief (Horowitz et al., 1997). There is significant overlap between PTSD, complicated grief, and depression (Bonanno et al., 2007; Burke, Neimeyer, & McDevitt-Murphy, 2010; Craig, Sossou, Schnak, & Exxex, 2008; Pivar & Field, 2004), and recent trials report the success of cognitive-behavioral therapy in treating grief reactions in the general population (Shear et al., 2005; de Goot et al., 2007). The findings from this study support O’Hare and Sherrer’s (2011) suggestion that clients with SMI may benefit from counseling that targets complicated grief reactions.
Despite a more ethnically diverse population in this study, consistent with previous findings among clients with SMI (Kilcommons & Morrison, 2005; Mueser et al., 1998), men were more likely to have experienced warfare, robbery, stranger assault, witnessing stranger assault, being threatened. Women were more likely to have experienced from sexual abuse in childhood and adulthood, domestic violence, being stalked and witnessing domestic violence. Men and women did not differ on exposure to childhood physical abuse, and sudden death of loved ones, and had comparable rates of traumatic event exposure, consistent with previous findings with SMI clients (Kilcommons & Morrison, 2005; Mueser et al., 1998). Hispanic ethnicity was also found to be significantly associated with PTSD symptom severity, a finding that remained significant even when controlling for the effects of education level and TLEQ score. The finding of more severe PTSD symptoms among Hispanic individuals is consistent with Marshall et al.’s (2009) and Torres and Han (2000)’s findings that Hispanic individuals with PTSD tend to report more severe symptoms. This result is also consistent with one previous report that Hispanics with SMI were more likely to have PTSD than non-Hispanic clients (Mueser et al., 2004b). While the mechanisms underlying this relationship remain unclear (Marshal et al., 2009), the finding suggests that clinicians should be aware of the risk for increased PTSD symptoms in Hispanic clients.
Although the type of trauma identified as most distressing was not found to be associated with PTSD symptom severity, the number of types traumatic events experienced was, in line with previous research showing that cumulative trauma exposure is related to PTSD diagnosis in people with SMI (Mueser et al., 1998). Of particular note, clients with SMI reported experiencing PTSD symptoms related to events that had occurred on average almost 20 years earlier. These findings highlight the need to routinely assess trauma exposure, and to address the prolonged effects of trauma and loss throughout the lifespan of individuals with SMI.
Some limitations of the present study should be noted. Diagnoses were based on clinical charts and therefore may be less reliable than research based diagnoses conducted using interview schedules such as the Structured Clinical Interview for the DSM-IV. Furthermore, while the sample was drawn from a large community mental health center as part of a comprehensive screening effort, data may not be generalizable to other groups of individuals with SMI living in less urban settings, with larger numbers of individuals with psychotic disorders, and with fewer individuals from African-American and Hispanic backgrounds. Finally, while data from the TLEQ give us a sense of the range of exposure to different types of trauma which participants experienced, we are unable to determine the number of traumatic experiences which participants had (e.g., repeated experience of multiple traumatic events of the same type). Future research needs to be conducted to more accurately assess the relationship between the number and severity of types of traumatic events experienced by people with SMI and PTSD symptom severity.
This research was supported by National Institute of Mental Health grant R01 MH064662. We wish to thank the following individuals for their assistance with this project: Edward Kim, Lee Hyer, Rachael Fite, Kenneth Gill, Rosemarie Rosati, Christopher Kosseff, Karen Somers, John Swanson, Avis Scott, Rena Gitlitz, John Markey, Zygmond Gray, Marilyn Green, Alex Shay, Leila Hosseini, and Yetunde Adetona.
Disclosure of Conflict of Interest
Disclosures: None for any authors.