- Email: [email protected]
Patterns of DSM-5 posttraumatic stress disorder and depression symptoms in an epidemiological sample of Chinese earthquake survivors: A latent profile analysis
Journal of Affective Disorders 186 (2015) 58–65
Contents lists available at ScienceDirect
Journal of Affective Disorders journal homepage: www.elsevier.com/locate/jad
Research report
Patterns of DSM-5 posttraumatic stress disorder and depression symptoms in an epidemiological sample of Chinese earthquake survivors: A latent profile analysis Xing Cao a,b, Li Wang a,n, Chengqi Cao a,b, Jianxin Zhang a, Ping Liu a,b,c, Biao Zhang c, Qi Wu c, Hong Zhang c, Zhihong Zhao d, Gaolin Fan d, Jon D. Elhai e a
Key Laboratory of Mental Health, Institute of Psychology, Chinese Academy of Sciences, Beijing, China Graduate University of Chinese Academy of Sciences, Beijing, China People's Hospital of Deyang City, Deyang, Sichuan, China d Hanwang People's Hospital, Deyang, Sichuan, China e Department of Psychology, and Department of Psychiatry, University of Toledo, Toledo, USA b c
art ic l e i nf o
a b s t r a c t
Article history: Received 11 April 2015 Accepted 2 June 2015 Available online 21 July 2015
Background: Posttraumatic stress disorder (PTSD) and depression are highly comorbid in association with serious clinical consequences. Nevertheless, to date, no study using latent class or latent profile analysis (LCA/LPA) has examined patterns of co-occurring PTSD and depression symptoms among natural disaster survivors, nor has the distinctiveness of DSM-5 PTSD and depression symptoms been clarified in the aftermath of trauma. This study was primarily aimed at filling these gaps. Methods: LPA was used to examine self-reported PTSD and depression symptoms in an epidemiological sample of 1196 Chinese earthquake survivors. Results: A 4-class solution characterized by low symptoms (53.9%), predominantly depression (18.2%), predominantly PTSD (18.9%) and combined PTSD-depression (9.0%) patterns fit the data best. Demographic characteristics and earthquake-related exposures were specifically or consistently associated with the non-parallel profiles varying in physical health impairment. Limitations: A sample exposed to specific traumatic events was assessed by self-report measures. Conclusions: The distinctiveness of DSM-5 PTSD and depression symptoms following an earthquake suggests that PTSD and depression may be independent sequelae of psychological trauma rather than a manifestation of a single form of psychopathology. The current findings support the distinction between PTSD and depression constructs, and highlight the need for identifications of natural disaster survivors at high risk for PTSD and/or depression, and interventions individually tailored to one's symptom presentations. & 2015 Elsevier B.V. All rights reserved.
Keywords: Posttraumatic stress disorder DSM-5 Depression Latent profile analysis Natural disaster
1. Introduction Comorbidity appears to be more the rule than the exception in psychopathological conditions (Clark et al., 1995). The comorbidity of posttraumatic stress disorder (PTSD) with other disorders is so prevalent that more than 80% of individuals with PTSD meet diagnostic criteria for at least one other disorder (Creamer et al., 2001; Kessler et al., 1995). Major depressive disorder (MDD) is one of the most common comorbid disorders with PTSD. According to a meta-analysis of 57 studies (N ¼ 6670) estimating the mean rate of MDD co-occurrence among PTSD samples, over half of n
Corresponding author. E-mail address: [email protected] (L. Wang).
http://dx.doi.org/10.1016/j.jad.2015.06.058 0165-0327/& 2015 Elsevier B.V. All rights reserved.
individuals with PTSD (52%) had comorbid MDD (Rytwinski et al., 2013), which is corresponding to the rate of comorbidity between PTSD and MDD (48–55%) found in national epidemiological surveys (Elhai et al., 2008; Kessler et al., 1995). Meanwhile, co-occurring PTSD and MDD is generally associated with greater distress, impairment, and health care utilization than PTSD alone (Chan et al., 2009; Ikin et al., 2010; Momartin et al., 2004). Thus, given the high prevalence and serious clinical consequences of PTSD-MDD comorbidity, a better understanding of the relationship between PTSD and depression constructs is desperately needed. One of the explanations for high rate of comorbidity between PTSD and depression in the aftermath of trauma is that they are part of a single general traumatic stress construct (O’Donnell et al., 2004). Supporting this hypothesis, many studies have revealed
X. Cao et al. / Journal of Affective Disorders 186 (2015) 58–65
that PTSD and depression share multiple overlapping risk factors and common vulnerabilities, such as childhood adversity, psychiatric history, trauma severity, lack of social support, and genetic risk (e.g. Breslau et al., 2000; Brewin et al., 2000; Kendler et al., 2002; Koenen et al., 2008; O’Donnell et al., 2004). In addition, by using factor analysis, some studies have found that PTSD and depression load onto the same higher-order factor (Anxious Misery/ Distress; Cox et al., 2002; Miller et al., 2008; Slade and Watson, 2006). And in the quantitative hierarchical model of mood and anxiety disorders proposed by Watson (2005), PTSD is grouped into the subclass of distress disorders along with MDD, suggesting that PTSD and depression share identical elements of one form of psychopathology. Moreover, the bidirectional associations between PTSD and depression in their parallel courses have been found in some longitudinal studies (Dekel et al., 2014; Norman et al., 2011), which provides further evidence for a single construct embodied by PTSD and depression following trauma. However, given the distinctiveness of PTSD and depression in theoretical conceptualizations, symptom patterns, information processing, and underlying physiological and neurobiological mechanisms (Zoellner et al., 2014), many researchers have argued that PTSD and depression are independent sequelae of psychological trauma rather than a manifestation of a single form of psychopathology (e.g. Blanchard et al., 1998; Shalev et al., 1998). Although PTSD and depression share some common risk factors in the aftermath of trauma, unique predictors also have been found (e.g. O’Donnell et al., 2004; Seal et al., 2009; Sharkansky et al., 2000). Besides, some studies have challenged the findings of grouping PTSD and depression into the same underlying dimension, with evidence that most PTSD and depression symptoms load onto two distinct but highly correlated factors (Blanchard et al., 1998; Grant et al., 2008; Gros et al., 2010). Furthermore, even in longitudinal studies revealing bidirectional associations between PTSD and depression following trauma, the impact of PTSD on subsequent depression has been greater and more consistent than that of depression on PTSD (Erickson et al., 2001; Nickerson et al., 2013), which adds to support for a distinction between PTSD and depression constructs. Putting aside the mixed findings from previous studies, it is noteworthy that the majority of them have relied on the traditional technique of factor analysis, which may overlook the differences in co-occurring symptom presentations among heterogeneous groups of individuals. Latent class or latent profile analysis (LCA/LPA, using categorical or continuous indicators, respectively), however, is a person-centered method that identifies subgroups of individuals who share common comorbidity patterns (McCutcheon, 1987), and may therefore contribute to a better understanding of the relationship between PTSD and depression following trauma. For instance, Hruska et al. (2014) used LCA to examine patterns of psychiatric comorbidity (PTSD, MDD and alcohol/other drug use disorder) among 249 motor vehicle accident (MVA) victims 6-weeks post-MVA, and found that PTSD and MDD were parallel to each other in a 4-class model consisting of resilient, mild psychopathology, moderate psychopathology, and severe psychopathology classes. Similarly, by using LPA to examine profiles of 140 parents' distress (PTSD, depression and anxiety symptoms) 2 years after war exposure, three parallel subgroups named low distress, moderate distress, and high distress respectively were identified (Lai et al., 2014), which also supports the hypothesis of a general traumatic stress response. Furthermore, focusing only on co-occurring PTSD and depression symptoms, Au et al. (2013) conducted an LPA on 119 female sexual assault survivors at 1- to 4-months post- assault. A 4-class model comprised of four parallel profiles with low, low-moderate, high-moderate, and severe levels of both PTSD and depression symptoms provided best fit to the data across time. Recently, Armour et al. (2015) used
59
LPA to identify profiles of PTSD and MDD symptoms in 283 Canadian veterans, and found three parallel patterns characterized by high, moderate, and low symptoms of PTSD and MDD, likewise suggesting that PTSD and depression are a manifestation of a single form of psychopathology rather than independent sequelae of trauma. Nevertheless, given that the presentations of PTSD and other trauma-related psychopathology may vary for different trauma types (DiMauro et al., 2014), the scarcity of studies using LCA or LPA on co-occurring PTSD and depression symptoms following trauma leaves doubt about the extension of comorbidity patterns found previously to other trauma populations, for example, natural disaster survivors. Besides, the relatively small sample size (e.g. 119; Au et al., 2013) and high level of symptom severity in the sample (e.g. a 78.0% prevalence of probable PTSD; Au et al., 2013) of previous studies partly limit the sample heterogeneity, which may preclude the emergence of possible subgroups with other symptom presentations (e.g. mild PTSD/severe depression). More importantly, as DSM-5 comes with reorganization and redefinition of PTSD symptoms (American Psychiatric Association, 2013), the Criterion D (negative alterations in cognitions and mood) cluster is expanded from the emotional numbing cluster with persistent distorted blame of oneself or others and pervasive negative emotional state, and meanwhile the Criterion E (alterations in arousal and reactivity) cluster is expanded from the hyperarousal cluster with a symptom of reckless or self-destructive behavior. Gootzeit and Markon (2011) have argued that these expanded dimensions represent general factors of negative affect rather than specific components of PTSD, thus their inclusion in DSM-5 may further decrease the distinction between PTSD and depression. However, to date, no study using LCA or LPA has examined whether DSM-5 PTSD symptoms are distinct from co-occurring depression symptoms following trauma. For the above reasons, we conducted an LPA on epidemiological data drawn from a relatively large sample of earthquake survivors to examine patterns of DSM-5 PTSD and depression symptoms. We primarily aimed to identify subgroups differentiated by symptom profiles, which would shed further light on the relationship between PTSD and depression constructs. Our second aim was to examine determinants of symptom profiles in this sample, including demographic characteristics and earthquake-related exposures. Identifying specific predictors of group membership would have potential implications for early preventions, identifications and treatments targeting individuals at high risk for PTSD and/or depression, which, however, has been overlooked by most previous studies. In addition, we also tested whether physical health impairment varies for different symptom profiles. The differences in associations between specific symptom presentations of PTSD and depression with physical health outcomes (e.g. Aversa et al., 2012), serving as external validity, would not only support the distinctiveness of PTSD and depression in theory, but also suggest the preferential interventions for individuals who suffer more clinical consequences in practice.
2. Methods 2.1. Participants and procedures Secondary analyses were conducted on data from a study of Chinese earthquake survivors (Liu et al., 2014). The sample was recruited from one of the largest rebuilt communities located in Hanwang Town, Mianzhu City, which was almost completely destroyed by the 2008 Wenchuan Earthquake with more than 5000 deaths. The survey was conducted from November 6 to 18, 2013, approximately 5.5 years after the earthquake. Household served as
60
X. Cao et al. / Journal of Affective Disorders 186 (2015) 58–65
the basic sample unit, and only one member within each household was randomly selected for participation. Participants met inclusion criteria if they were at least 16 years old without mental retardation or any major psychosis (e.g. schizophrenia and organic mental disorders), and experienced the disaster personally. The sample of 1196 people (response rate¼95%) consisted of 67.7% females and 32.3% males. Age ranged from 16 to 73 years (M ¼47.9 years; SD ¼ 10.0). Ethnicity was self-identified as predominantly Han (99.7%) and other ethnicities in China (0.3%). In regard of educational level, 31.6% completed high school or above, and 67.1% did not complete high school. As for marital status, 86.8% were married, and 13.0% were unmarried (single/divorced/separated/ widowed). Investigators including trained clinical psychologists, psychiatrists, psychotherapists, and psychology graduate students explained the aim and significance of the study, and obtained written informed consent before administering self-reported questionnaires to the participants. The study protocol was approved by the Institutional Review Board of the Institute of Psychology, Chinese Academy of Sciences. Additional details on sampling procedures are described elsewhere (Liu et al., 2014). 2.2. Measures Demographic characteristics assessed included age, sex, ethnicity, educational level and marital status. Earthquake-related exposures assessed included (1) being trapped under rubble; (2) being injured; (3) being disabled due to injuries; (4) participating in rescue efforts; (5) witnessing a death of someone; (6) exposure to mutilated bodies; (7) traumatic death of a family member; (8) traumatic injury of a family member; (9) traumatic death of a friend or neighbor; and (10) losing livelihood due to the disaster. Features of trauma exposure of the sample are shown elsewhere (Liu et al., 2014). The total number of exposures above ranging from 0 to 10, reflected severity of earthquake-related exposure. PTSD symptoms were assessed with the PTSD Checklist for DSM-5 (PCL-5; Weathers et al., 2013). The PCL-5 is a 20-item selfreport instrument adapted from the original PCL to map directly onto DSM-5 criteria for PTSD, and each item is rated on a 5-point Likert scale reflecting severity of PTSD symptoms from 0 (not at all) to 4 (extremely) during the past month. The original PCL has demonstrated sound psychometric properties (McDonald and Calhoun, 2010; Wilkins et al., 2011), and its Chinese version has been validated and widely used in Chinese trauma populations (e.g. Li et al., 2010). The Chinese version of the PCL-5 was obtained by a translation and back-translation process based on the Chinese version of the original PCL. In this study, the PCL-5 was completed with respect to the 2008 Wenchuan Earthquake. Total scores were calculated, along with subscores for intrusion, avoidance, negative alterations in cognitions and mood, and hyperarousal, based on the four-factor model defined by DSM-5 (American Psychiatric Association, 2013). Cronbach's α was 0.94 for the total scale, and 0.86, 0.81, 0.86, and 0.85 for the intrusion, avoidance, negative alterations in cognitions and mood, and hyperarousal subscales respectively in the current sample. Depression symptoms were assessed with the Center for Epidemiological Studies-Depression Scale (CES-D; Radloff, 1977). The CES-D is a 20-item self-report instrument adopting a 4-point Likert scale to indicate severity of depression symptoms from 0 (rare or none of the time/less than 1 day) to 3 (most or all of the time/ 5–7 days) during the past week. The CES-D has demonstrated favorable psychometric properties across various samples (e.g. Lewinsohn et al., 1997; Radloff, 1977), and its Chinese version has been validated and widely used in Chinese populations (e.g. Chi, 1995). In this study, four items describing positive affect were
reversed before calculating total scores as well as subscores for depressive affect, positive affect, somatic complaints, and interpersonal problems, based on the four-factor model identified by Radloff (1977). Cronbach's α was 0.87 for the total scale, and 0.85, 0.67, 0.77, and 0.67 for the depressive affect, positive affect, somatic complaints, and interpersonal problems subscales respectively in the current sample. Physical health impairment was assessed with the Physical Health Questionnaire (PHQ; Schat et al., 2005). The PHQ is a 14item self-report instrument for four dimensions of somatic health, including sleep disturbances, headaches, gastrointestinal problems, and respiratory infections. Items are rated on a 7-point Likert scale ranging from 1 (not at all) to 7 (all of the time), reflecting the frequency with which the respondents experienced the symptoms during the past year. The PHQ has demonstrated good reliability and validity (Schat et al., 2005), and has been used in studies of various samples (e.g. Tinlin et al., 2013). The Chinese version of the PHQ was adapted by a two-stage process of translation and back translation. In this study, total scores were calculated (reverse-scoring Item 4, the endorsement of which indicated the absence of symptoms), and higher scores reflected more severe physical health impairment. Cronbach's α was 0.83 for the total scale in the current sample. 2.3. Statistical analysis In order to minimize the number of indicators and facilitate model convergence, while maximizing the interpretability of class solution, as suggested by Au et al. (2013), we conducted LPA on eight sum scores for four PTSD factors (i.e. intrusion, avoidance, negative alterations in cognitions and mood, and hyperarousal) and four depression factors (i.e. depressive affect, positive affect, somatic complaints, and interpersonal problems) using Mplus 7.0 maximum likelihood estimation with robust standard errors (Muthén and Muthén, 1998–2012). Given the difference in scale range between the PCL-5 and CES-D, the eight sum scores were converted to T scores for the LPA. 1- to 5-class solutions of LPA models were evaluated and compared based on fit indices, parsimony, and interpretability. A lower Bayesian Information Criterion (BIC) value, a lower Akaike Information Criterion (AIC) value, a higher entropy value, a significant Lo–Mendell–Rubin likelihood ratio test (LMR LRT), a significant Lo–Mendell–Rubin adjusted likelihood ratio test (ALMR LRT) and a significant bootstrap likelihood ratio test (BLRT) are indicative that the class solutions have better goodness of fit and more accurate classification (Akaike, 1987; Lo et al., 2001; Nylund et al., 2007; Schwarz, 1978). After the optimal class solution was determined, we conducted subsequent analyses with the information of class assignments in SPSS 19.0. We used ANOVAs to examine the between-class differences in PTSD symptoms (PCL-5 total score, and the four PCL-5 subscores), depression symptoms (CES-D total score, and the four CES-D subscores), and physical health impairment (PHQ total score). The Games–Howell tests, which do not assume equal variances or sample sizes, were used for post-hoc pairwise comparisons (Games and Howell, 1976). To examine determinants of symptom profiles, we conducted multinomial logistic regression analyses. The class assignments served as the dependent variables, and the predictors included age, sex, educational level, marital status, and earthquake-related exposures (both the exposure severity and specific exposures were analyzed respectively).
3. Results The mean scores on the PCL-5 and CES-D were 18.8 (SD ¼13.5, range: 1–77) and 27.8 (SD ¼ 6.5, range: 15–51) respectively in this
X. Cao et al. / Journal of Affective Disorders 186 (2015) 58–65
Table 1 Fit indices for latent profile analyses.
1-class 2-class 3-class 4-class 5-class
BIC
AIC
Entropy
LMR LRT
ALMR LRT
BLRT
71320.40 68515.55 67598.83 67152.89 66826.92
71239.01 68388.39 67425.88 66934.16 66562.41
– 0.92 0.88 0.88 0.84
– o 0.001 o 0.001 0.002 0.336
– o 0.001 o 0.001 0.002 0.340
– o 0.001 o 0.001 o 0.001 o 0.001
BIC, Bayesian Information Criterion; AIC, Akaike Information Criterion; LMR LRT, Lo–Mendell–Rubin likelihood ratio test; ALMR LRT, Lo–Mendell–Rubin adjusted likelihood ratio test; BLRT, Bootstrap likelihood ratio test.
sample. Based on the DSM-5 diagnostic algorithm of at least one intrusion symptom, one avoidance symptom, two negative alterations in cognitions and mood symptoms, and two arousal symptoms of at least moderate (2 or higher) severity, adapted from the algorithm of the original PCL (Cook et al., 2003), the prevalence of probable PTSD was 13.8% (n¼ 165). Based on a cutoff score of 28 (Radloff, 1991), the caseness rate for depression was 42.2% (n ¼505). Fit indices for 1- to 5-class solutions of LPA models are presented in Table 1. The 4-class solution was found to have the best fit for the data, on account of lower BIC and AIC values, and significant LMR, ALMR, and BLRT tests. In addition, given that the non-significant LMR test suggested stopping increasing the number of classes (Nylund et al., 2007), the 4-class solution should not be rejected in favor of the 5-class solution, although the 5-class solution was found to have lower BIC and AIC values. Moreover, the 4-class solution was superior to the 5-class solution given the higher entropy value reflecting higher classification accuracy, despite the significant BLRT test of the 5-class solution. Consequently, we selected the 4-class solution as the optimal class solution. This solution was characterized by low symptoms (53.9%), predominantly depression (18.2%), predominantly PTSD (18.9%), and combined PTSD-depression (9.0%) patterns. Fig. 1 displays the PCL-5 and CES-D standardized subscores for the 4-class model of symptom profiles. PCL-5 and CES-D unstandardized total scores, and the four PCL-5 and four CES-D unstandardized subscores, as well as the results of ANOVAs and post-hoc pairwise comparisons are shown in Table 2. The between-class differences in PTSD symptoms (PCL-5 total score, and the four PCL-5 subscores) and depression symptoms (CES-D total score, and the four CES-D subscores) were found to be significant. Class 1 (low symptoms) individuals presented the lowest symptom severity of PTSD and
61
depression. Class 2 (predominantly depression) individuals presented relatively low PTSD symptom severity, while relatively high depression symptom severity with CES-D total scores above the cutoff. Class 3 (predominantly PTSD) individuals, on the contrary, presented relatively high PTSD symptom severity while relatively low depression symptom severity falling below the cutoff. Class 4 (combined PTSD-depression) individuals presented the highest symptom severity of PTSD and depression, with extremely severe PTSD symptoms and CES-D total scores above the cutoff. All four classes were significantly differentiated from each other in overall symptom severity of PTSD and depression, and all four PTSD factors and two depression factors (i.e. depressive affect and somatic complaints). The other two depression factors (i.e. positive affect and interpersonal problems), however, differentiated less well among the four classes, with comparable symptom severity between Class 1 and 3, and Class 2 and 4. Table 3 displays the results of multinomial logistic regression analyses examining determinants of symptom profiles. Compared to those in the low symptoms class, individuals in the predominantly depression, predominantly PTSD, and combined PTSDdepression classes were more likely to be older, female, and to report greater severity of earthquake-related exposure. Additionally, higher educational level was protective against the inclusion in the predominantly PTSD class. Table 4 displays specific earthquake-related exposures associated with symptom profiles after demographic characteristics shown in Table 3 were controlled. Being trapped under rubble, witnessing a death of someone, traumatic death of a family member, and losing livelihood due to the disaster were associated with symptom profiles of predominantly depression, predominantly PTSD, and combined PTSD-depression. Being disabled due to injuries and exposure to mutilated bodies were additionally associated with symptom profile of combined PTSD-depression. Significant between-class differences in physical health impairment were found (F (3, 1192) ¼179.97, p o0.001). The mean scores on the PHQ were 38.2 (SD ¼10.1), 51.6 (SD ¼ 12.0), 47.3 (SD ¼11.2), and 60.7 (SD ¼ 14.1) for Class 1, Class 2, Class 3, and Class 4 respectively. Post-hoc pairwise comparisons indicated that all four classes were significantly differentiated from each other; and based on the degree of severity, they ranked in the following order: Class 4, Class 2, Class 3, and Class 1.
Fig. 1. PCL-5 and CES-D standardized subscores for the 4-class model of symptom profiles. PTSD, post-traumatic stress disorder; In, intrusion; Av, avoidance; NACM, negative alterations in cognitions and mood; Hy, hyperarousal; DEP, depressive affect; POS, positive affect; SOM, somatic complaints; INT, interpersonal problems; PCL-5, PTSD Checklist for DSM-5; CES-D, Center for Epidemiological Studies-Depression Scale.
62
X. Cao et al. / Journal of Affective Disorders 186 (2015) 58–65
Table 2 Class means for 4-class model. Class 1 (low symptoms) M (SD) In 3.89 (2.12) Av 1.20 (1.09) NACM 1.53 (1.80) Hy 2.99 (2.18) DEP 7.34 (1.77) POS 6.39 (2.09) SOM 8.57 (1.83) INT 1.83 (0.58) PCL-5 9.60 (5.09) CES-D 24.13 (4.00)
Class 2 (predominantly depression) M (SD)
Class 3 (predominantly PTSD) M (SD)
Class 4 (combined PTSDdepression) M (SD)
Group comparisons Post-hoc pairwise comF parisons (p o0.05)
6.86 (3.01) 2.21 (1.33) 4.64 (2.88) 7.20 (3.01) 12.29 (2.26) 7.25 (1.66) 12.60 (1.90) 2.97 (0.76) 20.92 (6.71) 35.10 (3.98)
10.15 (3.05) 3.76 (1.71) 6.55 (3.75) 8.10 (3.08) 8.46 (1.81) 6.29 (1.86) 10.17 (1.84) 1.78 (0.51) 28.56 (6.45) 26.70 (3.69)
14.62 5.31 13.86 14.83 13.64 6.90 14.08 2.91 48.62 37.53
704.00*** 430.49*** 684.77*** 707.25*** 555.56*** 13.12*** 413.64*** 221.28*** 1543.32*** 603.46***
(3.39) (1.59) (4.56) (3.83) (2.73) (1.86) (2.53) (1.19) (9.04) (5.69)
1 o2 o3 o4 1 o2 o3 o4 1 o2 o3 o4 1 o2 o3 o4 1 o3 o2 o4 (1 ¼ 3) o(2 ¼ 4) 1 o3 o2 o4 (1 ¼ 3) o(2 ¼ 4) 1 o2 o3 o4 1 o3 o2 o4
PTSD, post-traumatic stress disorder; In, intrusion; Av, avoidance; NACM, negative alterations in cognitions and mood; Hy, hyperarousal; DEP, depressive affect; POS, positive affect; SOM, somatic complaints; INT, interpersonal problems; PCL-5, PTSD Checklist for DSM-5; CES-D, Center for Epidemiological Studies-Depression Scale. ***
p o 0.001.
Table 3 Determinants of symptom profiles.
Age Sex Male (reference) Female Education Less than high school (reference) High school or above Marital status Unmarried (reference) Married Exposure severity
Class 2 (predominantly depression) RRR (95% CI)
Class 3 (predominantly PTSD) RRR (95% CI)
Class 4 (combined PTSD-depression) RRR (95% CI)
1.02 (1.00-1.04)*
1.05 (1.03-1.07)***
1.09 (1.06-1.12)***
– 1.84 (1.29–2.64)**
– 2.02 (1.41–2.88)***
2.25 (1.38–3.68)**
– 1.09 (0.75–1.57)
– 0.71 (0.48–1.05)a
0.92 (0.54–1.55)
– 0.77 (0.49–1.23) 1.29 (1.18–1.42)***
1.03 (0.63–1.69) 1.29 (1.18–1.41)***
0.68 (0.37–1.25) 1.77 (1.56–2.01)***
PTSD, post-traumatic stress disorder; RRR, relative risk ratio; CI, confidence interval. Reference class ¼ Class 1 (low symptoms). Bolded RRRs (95% CI) indicate a significant association with the class. a
p ¼ 0.084. p o 0.05. po 0.01. *** p o 0.001. *
**
Table 4 Specific earthquake-related exposures associated with symptom profiles.
Trapped under rubble Injured Disabled due to injuries Participated in rescue efforts Witnessed a death of someone Exposed to mutilated bodies Traumatic death of family member Traumatic injury of family member Traumatic death of friend or neighbor Lost livelihood due to the disaster
Class 2 (predominantly depression) RRR (95% CI)
Class 3 (predominantly PTSD) RRR (95% CI)
Class 4 (combined PTSD-depression) RRR (95% CI)
2.04 (1.10–3.78)* 0.93 (0.54–1.59) 1.02 (0.33–3.19) 0.99 (0.68–1.42) 1.82 (1.21–2.74)** 1.19 (0.83–1.71) 1.54 (1.05–2.25)* 0.89 (0.62–1.28) 1.33 (0.85–2.06)
2.36 (1.27–4.38)** 0.60 (0.33–1.07) 1.68 (0.59–4.73) 0.97 (0.67–1.40) 1.86 (1.25–2.76)** 1.19 (0.83–1.71) 1.92 (1.32–2.79)** 0.83 (0.58–1.20) 1.20 (0.79–1.82)
2.21 (1.06–4.62)* 1.22 (0.63–2.36) 3.12 (1.03–9.43)* 1.16 (0.71–1.90) 2.02 (1.11–3.67)* 1.58 (0.97–2.58)a 2.69 (1.67–4.35)*** 1.38 (0.86–2.22) 1.82 (0.92–3.63)
1.66 (1.17–2.35)**
1.63 (1.15–2.32)**
2.09 (1.31–3.34)**
PTSD, post-traumatic stress disorder; RRR, relative risk ratio; CI, confidence interval. Reference class ¼ Class 1 (low symptoms). Relative risk ratios are adjusted for demographic characteristics shown in Table 3. Bolded RRRs (95% CI) indicate a significant association with the class. a
p ¼ 0.067. p o 0.05. po 0.01. *** p o 0.001. *
**
X. Cao et al. / Journal of Affective Disorders 186 (2015) 58–65
4. Discussion In this study, we used LPA to examine patterns of DSM-5 PTSD and depression symptoms in an epidemiological sample of Chinese earthquake survivors. Four non-parallel profiles of low symptoms, predominantly depression, predominantly PTSD, and combined PTSD-depression were identified, and found to be determined by specific and common factors and vary in physical health impairment. These findings fill the gaps in the existing literature using LCA or LPA to examine patterns of co-occurring PTSD and depression symptoms among natural disaster survivors, and to clarify the distinctiveness of DSM-5 PTSD and depression symptoms following trauma. Contrary to previous findings of parallel patterns of PTSD and depression symptoms in the aftermath of trauma (Armour et al., 2015; Au et al., 2013; Hruska et al., 2014; Lai et al., 2014), we found that the severity of PTSD and depression symptoms was not consistently reflective of one another, and there were subgroups containing individuals with predominantly PTSD or predominantly depression symptoms. On the one hand, these discrepancies may be attributable to the differences in trauma types (e.g. natural disasters vs. interpersonal traumas) and assessment time points (e.g. in the chronic vs. acute aftermath of trauma); while on the other hand, considering that the relatively large sample size and low-moderate level of symptom severity in this heterogeneous sample might have given rise to the emergency of subgroups with less common but reliable symptom presentations (e.g. mild PTSD/severe depression, eliminated due to its relatively small proportion of the sample; Au et al., 2013), the seemingly novel finding of a 4-class model comprised of four non-parallel profiles (i.e. low symptoms, predominantly depression, predominantly PTSD, and combined PTSD-depression) in this study may provide a more authentic depiction of co-occurring PTSD and depression symptoms following trauma. Moreover, the existence of subgroups with predominantly PTSD or predominantly depression symptoms in the aftermath of trauma is in line with previous findings of traumatized individuals with either PTSD or depression in the absence of the other across various samples (e.g. Lai et al., 2013; Morina et al., 2013), and is also consistent with previous studies suggesting that PTSD and depression are independent sequelae of trauma rather than a manifestation of a single form of psychopathology (e.g. Blanchard et al., 1998; Shalev et al., 1998). Some researchers have argued that the inclusion of expanded negative cognitions/mood and hyperarousal dimensions overlapping with depression and other mood and anxiety disorders in DSM-5 may decrease the distinction between PTSD and depression or other disorders (Gootzeit and Markon, 2011). However, the current finding of predominantly PTSD and predominantly depression symptom presentations demonstrates that DSM-5 PTSD symptoms can differentiate from co-occurring depression symptoms following trauma, which adds to evidence for the distinctiveness of PTSD and depression from the perspective of comorbidity patterns, and also generally supports the strong diagnostic construct validity of PTSD in DSM-5. In addition, some previous studies have found that removing symptoms theorized to overlap with mood and anxiety disorders from PTSD's criteria has little impact on comorbidity rates (Elhai et al., 2008; Grubaugh et al., 2010), which suggests that the association between PTSD and other disorders may be a natural aspect of the disorder's phenomenology rather than an artifact of poor differential diagnosis (Elhai et al., 2008), and the overlapping symptoms may not be responsible for the comorbidity between PTSD and other disorders. Thus, given that PTSD and depression represent two distinct constructs rather than a single general traumatic stress construct despite sharing some common symptoms, other
63
etiological hypotheses for PTSD-depression comorbidity, for example, preexisting PTSD causes depression (Stander et al., 2014) are supposed to be considered and evaluated in future research. Another noteworthy finding in this study is that all four PTSD factors could differentiate well among the four subgroups characterized by parallel varying severity of PTSD symptoms, which is in line with findings of previous studies using LCA or LPA to examine the heterogeneity of PTSD symptoms (e.g. Rosellini et al., 2014). Nevertheless, although overall depression symptom severity was able to differentiate between subgroups, only depressive affect and somatic complaints, but not positive affect and interpersonal problems, could distinguish one subgroup from another, indicating that the patterns of depression symptoms were not completely parallel across four factors. This finding is partly consistent with previous evidence that there are classes of individuals with relatively high negative affect and somatic symptoms but relatively low anhedonia, or relatively high anhedonia but relatively low negative affect (e.g. Hybels et al., 2013; Rodgers et al., 2014), which provides further support for the distinction between negative and positive affect suggested by previous theoretical and empirical studies (Watson, 2005, 2009), as well as the Research Domain Criteria (RDoC) project (Cuthbert and Kozak, 2013). In this study, we identified some factors consistently associated with symptom profiles of predominantly depression, predominantly PTSD, and combined PTSD-depression among earthquake survivors. These findings roughly accord with meta-analyses of risk factors for PTSD (Brewin et al., 2000) and depression following natural disasters (Tang et al., 2014), and highlight the importance of early preventions, identifications and treatments targeting high-risk individuals, especially those who are older, female, at high levels of exposure severity, and experienced some specific earthquake-related exposures, including being trapped under rubble, witnessing a death of someone, traumatic death of a family member, and losing livelihood due to the disaster. Notably, some factors were specifically related to certain symptom profiles, which provides evidence for the differences in predictions of specific symptom presentations of PTSD and depression, and suggests the specific identifications of individuals at high risk for PTSD and/or depression. In addition, we found that physical health impairment varied for different symptom profiles: individuals with combined PTSDdepression symptoms suffered the most, and individuals with predominantly depression and predominantly PTSD symptoms were the second and third most severe subgroups, respectively, while individuals with low symptoms suffered the least. These differences in associations between specific symptom presentations of PTSD and depression with physical health outcomes are consistent with previous findings that PTSD and depression symptoms differentially affect the various physical health status domains (Aversa et al., 2012), which further validates the distinction between PTSD and depression with evidence of clinical meaningfulness, and indicates the preferential interventions for individuals who may suffer more negative consequences, especially those with combined PTSD-depression symptoms. Moreover, given that LCA or LPA is a person-centered method which can identify subgroups of individuals with specific symptom presentations, some studies using LCA or LPA to examine PTSD symptom profiles have supported the need for interventions individually tailored to one's severity of symptoms and the extent of functional impairment (Hellmuth et al., 2014). In the same way, the findings of symptomatic profiles of predominantly depression, predominantly PTSD, and combined PTSD-depression varying in physical health impairment in this study suggest that individually tailored interventions according to one's symptom presentations of PTSD and depression as well as the severity of symptoms and the extent of physical health impairment may contribute to
64
X. Cao et al. / Journal of Affective Disorders 186 (2015) 58–65
improving the efficiency and effectiveness of treatments. There are several noteworthy limitations to this study. First, the current findings might be limited by the specific trauma type, thus further replications across various trauma types are warranted. Second, this study only relied on self-report measures, and therefore the findings of this study also need to be further tested using data from clinical interviews. Third, the cross-sectional design at a single time point of this study precluded the examination of changes in symptom profiles and class membership over time. Future studies should conduct latent transition analysis on longitudinal data to demonstrate the more reliable symptom presentations of PTSD and depression over time. Fourth, psychiatric history prior to the earthquake, as well as stressful life events before and after the earthquake were not assessed in this study, thus future studies examining these determinants of symptom profiles of PTSD and depression are needed. Fifth, only PTSD but not depression symptoms were assessed based on DSM-5 criteria in this study. Therefore, further replications based on DSM-5 criteria for PTSD and MDD are also needed. Finally, we only adopted physical health impairment as the clinical variable to test external validity of the 4-class model of symptom profiles. Future studies should further examine the differentiable associations between specific symptom presentations of PTSD and depression with a range of external psychological and biological variables. Despite these limitations, to our knowledge this is the first study using LPA to examine patterns of co-occurring PTSD and depression symptoms among natural disaster survivors, and to clarify the distinctiveness of DSM-5 PTSD and depression symptoms following trauma. The current findings of non-parallel profiles determined by specific and common factors and varying in physical health impairment demonstrate that DSM-5 PTSD symptoms can differentiate from co-occurring depression symptoms, which supports the distinction between PTSD and depression constructs, and highlights the need for identifications of natural disaster survivors at high risk for PTSD and/or depression, and interventions individually tailored to one's symptom presentations. Conflict of interest None.
Contributors Xing Cao conducted the literature review and the analysis, and wrote the first draft of the manuscript. Li Wang conceptualized the study design, conducted the study, assisted with the analysis, and commented on drafts. Chengqi Cao and Jianxin Zhang assisted with the study and the analysis, and commented on drafts. Ping Liu, Biao Zhang, Qi Wu, Hong Zhang, Zhihong Zhao, and Gaolin Fan collected the data and commented on drafts. Jon D. Elhai revised the first draft of the manuscript and commented on drafts. All authors have approved the final paper.
Role of the funding source The conduct of this study was financially supported in part by the National Key Technology Research and Development Program of China (No. 2013BAI08B02), the Key Research Program of the Chinese Academy of Sciences (No. KJZD-EW-L04), the National Natural Science Foundation of China (No. 31271099 and No. 31471004), and the Outstanding Young Investigator Foundation of
the Institute of Psychology, Chinese Academy of Sciences. The funding sources had no role in study design; in the collection, analysis and interpretation of data; in the writing of the report; or in the decision to submit the paper for publication.
Acknowledgments This study was partially supported by the National Key Technology Research and Development Program of China (No. 2013BAI08B02), the Key Research Program of the Chinese Academy of Sciences (No. KJZD-EW-L04), the National Natural Science Foundation of China (No. 31271099 and No. 31471004), the Outstanding Young Investigator Foundation of the Institute of Psychology, Chinese Academy of Sciences, and the Key Laboratory of Mental Health, Institute of Psychology, Chinese Academy of Sciences.
References Akaike, H., 1987. Factor analysis and AIC. Psychometrika 52, 317–332. Armour, C., Contractor, A., Elhai, J.D., Stringer, M., Lyle, G., Forbes, D., Richardson, J. D., 2015. Identifying latent profiles of posttraumatic stress and major depression symptomssymptoms in Canadian veterans; exploring differences across profiles in health related functioning. Psychiatry Res. 228, 1–7. American Psychiatric Association, 2013. Diagnostic and Statistical Manual of Mental Disorders, 5th ed. American Psychiatric Association, Washington, DC. Au, T.M., Dickstein, B.D., Comer, J.S., Salters-Pedneault, K., Litz, B.T., 2013. Co-occurring posttraumatic stress and depression symptoms after sexual assault: a latent profile analysis. J. Affect. Disord. 149, 209–216. Aversa, L.H., Stoddard, J.A., Doran, N.M., Au, S., Chow, B., McFall, M., Saxon, A., Baker, D.G., 2012. PTSD and depression as predictors of physical health-related quality of life in tobacco-dependent veterans. J. Psychosomat. Res. 73, 185–190. Blanchard, E.B., Buckley, T.C., Hickling, E.J., Taylor, A.E., 1998. Posttraumatic stress disorder and comorbid major depression: Is the correlation an illusion? J. Anxiety Disord. 12, 21–37. Breslau, N., Davis, G.C., Peterson, E.L., Schultz, L.R., 2000. A second look at comorbidity in victims of trauma: the posttraumatic stress disorder-major depression connection. Biol. Psychiatry 48, 902–909. Brewin, C.R., Andrews, B., Valentine, J.D., 2000. Meta-analysis of risk factors for posttraumatic stress disorder in trauma-exposed adults. J. Consult. Clin. Psychol. 68, 748–766. Chan, D., Cheadle, A.D., Reiber, G., Unutzer, J., Chaney, E.F., 2009. Health care utilization and its costs for depressed veterans with and without comorbid PTSD symptoms. Psychiatr. Serv. 60, 1612–1617. Chi, I., 1995. Mental health of the old–old in Hong Kong. Clin. Gerontol 15, 31–44. Clark, L.A., Watson, D., Reynolds, S., 1995. Diagnosis and classification of psychopathology: challenges to the current system and future directions. Annu. Rev. Psychol. 46, 121–153. Cook, J.M., Thompson, R., Coyne, J.C., Sheikh, J.I., 2003. Algorithm versus cut-point derived PTSD in ex-prisoners of war. J. Psychopathol. Behav. Assess. 25, 267–271. Cox, B.J., Clara, I.P., Enns, M.W., 2002. Posttraumatic stress disorder and the structure of common mental disorders. Depress. Anxiety 15, 168–171. Creamer, M., Burgess, P., McFarlane, A.C., 2001. Post-traumatic stress disorder: findings from the Australian National Survey of Mental Health and Well-being. Psychol. Med. 31, 1237–1247. Cuthbert, B.N., Kozak, M.J., 2013. Constructing constructs for psychopathology: the NIMH research domain criteria. J. Abnorm. Psychol. 122, 928–937. Dekel, S., Solomon, Z., Horesh, D., Ein-Dor, T., 2014. Posttraumatic stress disorder and depressive symptoms: joined or independent sequelae of trauma? J. Psychiatr. Res. 54, 64–69. DiMauro, J., Carter, S., Folk, J.B., Kashdan, T.B., 2014. A historical review of traumarelated diagnoses to reconsider the heterogeneity of PTSD. J. Anxiety Disord. 28, 774–786. Elhai, J.D., Grubaugh, A.L., Kashdan, T.B., Frueh, B.C., 2008. Empirical examination of a proposed refinement to DSM-IV posttraumatic stress disorder symptom criteria using the National Comorbidity Survey Replication data. J. Clin. Psychiatry 69, 597–602. Erickson, D.J., Wolfe, J., King, D.W., King, L.A., Sharkansky, E.J., 2001. Posttraumatic stress disorder and depression symptomatology in a sample of Gulf War veterans: a prospective analysis. J. Consult. Clin. Psychol. 69, 41–49. Games, P.A., Howell, J.F., 1976. Pairwise multiple comparison procedures with unequal n's and/or variances: a Monte Carlo study. J. Educ. Behav. Stat. 1, 113–125. Gootzeit, J., Markon, K., 2011. Factors of PTSD: differential specificity and external correlates. Clin. Psychol. Rev. 31, 993–1003. Grant, D.M., Beck, J.G., Marques, L., Palyo, S.A., Clapp, J.D., 2008. The structure of distress following trauma: posttraumatic stress disorder, major depressive disorder, and generalized anxiety disorder. J. Abnorm. Psychol. 117, 662–672. Gros, D.F., Simms, L.J., Acierno, R., 2010. Specificity of posttraumatic stress disorder symptoms: an investigation of comorbidity between posttraumatic stress disorder symptoms and depression in treatment-seeking veterans. J. Nerv. Ment. Dis. 198, 885–890. Grubaugh, A.L., Long, M.E., Elhai, J.D., Frueh, B.C., Magruder, K.M., 2010. An
X. Cao et al. / Journal of Affective Disorders 186 (2015) 58–65
examination of the construct validity of posttraumatic stress disorder with veterans using a revised criterion set. Behav. Res. Ther. 48, 909–914. Hellmuth, J.C., Jaquier, V., Swan, S.C., Sullivan, T.P., 2014. Elucidating posttraumatic stress symptom profiles and their correlates among women experiencing bidirectional intimate partner violence. J. Clin. Psychol. 70, 1008–1021. Hruska, B., Irish, L.A., Pacella, M.L., Sledjeski, E.M., Delahanty, D.L., 2014. PTSD symptom severity and psychiatric comorbidity in recent motor vehicle accident victims: a latent class analysis. J. Anxiety Disord. 28, 644–649. Hybels, C.F., Landerman, L.R., Blazer, D.G., 2013. Latent subtypes of depression in a community sample of older adults: can depression clusters predict future depression trajectories? J. Psychiatr. Res. 47, 1288–1297. Ikin, J.F., Creamer, M.C., Sim, M.R., McKenzie, D.P., 2010. Comorbidity of PTSD and depression in Korean War veterans: prevalence, predictors, and impairment. J. Affect. Disord. 125, 279–286. Kendler, K.S., Gardner, C.O., Prescott, C.A., 2002. Toward a comprehensive developmental model for major depression in women. Am. J. Psychiatry 159, 1133–1145. Kessler, R.C., Sonnega, A., Bromet, E., Hughes, M., Nelson, C.B., 1995. Posttraumatic stress disorder in the national comorbidity survey. Arch. Gen. Psychiatry 52, 1048–1060. Koenen, K.C., Fu, Q.J., Ertel, K., Lyons, M.J., Eisen, S.A., True, W.R., Goldberg, J., Tsuang, M.T., 2008. Common genetic liability to major depression and posttraumatic stress disorder in men. J. Affect. Disord. 105, 109–115. Lai, B.S., Hadi, F., Llabre, M.M., 2014. Parent and child distress after war exposure. Br. J. Clin. Psychol. 53, 333–347. Lai, B.S., La Greca, A.M., Auslander, B.A., Short, M.B., 2013. Children's symptoms of posttraumatic stress and depression after a natural disaster: comorbidity and risk factors. J. Affect. Disord. 146, 71–78. Lewinsohn, P.M., Seeley, J.R., Roberts, R.E., Allen, N.B., 1997. Center for Epidemiologic Studies Depression Scale (CES-D) as a screening instrument for depression among community-residing older adults. Psychol. Aging 12, 277–287. Li, H.H., Wang, L., Shi, Z.B., Zhang, Y.Q., Wu, K.K., Liu, P., 2010. Diagnostic utility of the PTSD checklist in detecting PTSD in Chinese earthquake victims. Psychol. Rep. 107, 733–739. Liu, P., Wang, L., Cao, C., Wang, R., Zhang, J., Zhang, B., Wu, Q., Zhang, H., Zhao, Z., Fan, G., Elhai, J.D., 2014. The underlying dimensions of DSM-5 posttraumatic stress disorder symptoms in an epidemiological sample of Chinese earthquake survivors. J. Anxiety Disord. 28, 345–351. Lo, Y.T., Mendell, N.R., Rubin, D.B., 2001. Testing the number of components in a normal mixture. Biometrika 88, 767–778. McCutcheon, A.L., 1987. Latent Class Analysis, Quantitative Applications in the Social Sciences Series, No. 64. Sage Publications, Newbury Park, CA. McDonald, S.D., Calhoun, P.S., 2010. The diagnostic accuracy of the PTSD Checklist: a critical review. Clin. Psychol. Rev. 30, 976–987. Miller, M.W., Fogler, J.M., Wolf, E.J., Kaloupek, D.G., Keane, T.M., 2008. The internalizing and externalizing structure of psychiatric comorbidity in combat veterans. J. Trauma Stress 21, 58–65. Momartin, S., Silove, D., Manicavasagar, V., Steel, Z., 2004. Comorbidity of PTSD and depression: associations with trauma exposure, symptom severity and functional impairment in Bosnian refugees resettled in Australia. J. Affect. Disord. 80, 231–238. Morina, N., Ajdukovic, D., Bogic, M., Franciskovic, T., Kucukalic, A., Lecic-Tosevski, D., Morina, L., Popovski, M., Priebe, S., 2013. Co-occurrence of major depressive episode and posttraumatic stress disorder among survivors of war: How is it different from either condition alone? J. Clin. Psychiatry 74, e212–e218. Muthén, L.K., Muthén, B.O., 1998-2012. Mplus User's Guide, 7th ed. Muthén & Muthén, Los Angeles, CA. Nickerson, A., Steenkamp, M., Aerka, I.M., Salters-Pedneault, K., Carper, T.L., Barnes, J.B., Litz, B.T., 2013. Prospective investigation of mental health following sexual assault. Depress. Anxiety 30, 444–450. Norman, S.B., Trim, R.S., Goldsmith, A.A., Dimsdale, J.E., Hoyt, D.B., Norman, G.J.,
65
Stein, M.B., 2011. Role of risk factors proximate to time of trauma in the course of PTSD and MDD symptoms following traumatic injury. J. Trauma. Stress 24, 390–398. Nylund, K.L., Asparoutiov, T., Muthen, B.O., 2007. Deciding on the number of classes in latent class analysis and growth mixture modeling: a Monte Carlo simulation study. Struct. Equ. Model. 14, 535–569. O’Donnell, M.L., Creamer, M., Pattison, P., 2004. Posttraurnatic stress disorder and depression following traurna: understanding comorbidity. Am. J. Psychiatry 161, 1390–1396. Radloff, L.S., 1977. The CES-D scale: a self-report depression scale for research in the general population. Appl. Psychol. Meas. 1, 385–401. Radloff, L.S., 1991. The use of the Center for Epidemiologic Studies Depression Scale in adolescents and young adults. J. Youth Adolesc. 20, 149–166. Rodgers, S., Grosse Holtforth, M., Muller, M., Hengartner, M.P., Rossler, W., AjdacicGross, V., 2014. Symptom-based subtypes of depression and their psychosocial correlates: a person-centered approach focusing on the influence of sex. J. Affect. Disord. 156, 92–103. Rosellini, A.J., Coffey, S.F., Tracy, M., Galea, S., 2014. A person-centered analysis of posttraumatic stress disorder symptoms following a natural disaster: predictors of latent class membership. J. Anxiety. Disord. 28, 16–24. Rytwinski, N.K., Scur, M.D., Feeny, N.C., Youngstrom, E.A., 2013. The co-occurrence of major depressive disorder among individuals with posttraumatic stress disorder: a meta-analysis. J. Trauma. Stress 26, 299–309. Schat, A.C., Kelloway, E.K., Desmarais, S., 2005. The Physical Health Questionnaire (PHQ): construct validation of a self-report scale of somatic symptoms. J. Occup. Health Psychol. 10, 363–381. Schwarz, G., 1978. Estimating the dimension of a model. Ann. Stat. 6, 461–464. Seal, K.H., Metzler, T.J., Gima, K.S., Bertenthal, D., Maguen, S., Marmar, C.R., 2009. Trends and risk factors for mental health diagnoses among iraq and afghanistan veterans using department of veterans affairs health care, 2002–2008. Am. J. Public Health 99, 1651–1658. Shalev, A.Y., Freedman, S., Peri, T., Brandes, D., Sahar, T., Orr, S.P., Pitman, R.K., 1998. Prospective study of posttraumatic stress disorder and depression following trauma. Am. J. Psychiatry 155, 630–637. Sharkansky, E.J., King, D.W., King, L.A., Wolfe, J., Erickson, D.J., Stokes, L.R., 2000. Coping with Gulf War combat stress: mediating and moderating effects. J. Abnorm. Psychol. 109, 188–197. Slade, T., Watson, D., 2006. The structure of common DSM-IV and ICD-10 mental disorders in the Australian general population. Psychol. Med. 36, 1593–1600. Stander, V.A., Thomsen, C.J., Highfill-McRoy, R.M., 2014. Etiology of depression comorbidity in combat-related PTSD: a review of the literature. Clin. Psychol. Rev. 34, 87–98. Tang, B., Liu, X., Liu, Y., Xue, C., Zhang, L., 2014. A meta-analysis of risk factors for depression in adults and children after natural disasters. BMC Public Health 14, 623. Tinlin, R.M., Watkins, C.D., Welling, L.L.M., DeBruine, L.M., Al-Dujaili, E.A.S., Jones, B. C., 2013. Perceived facial adiposity conveys information about women's health. Br. J. Psychol. 104, 235–248. Watson, D., 2005. Rethinking the mood and anxiety disorders: a quantitative hierarchical model for DSM-V. J. Abnorm. Psychol. 114, 522–536. Watson, D., 2009. Differentiating the mood and anxiety disorders: a quadripartite model. Annu. Rev. Clin. Psychol. 5, 221–247. Weathers, F.W., Litz, B.T., Keane, T.M., Palmieri, P.A., Marx, B.P., Schnurr, P.P., 2013. The PTSD checklist for DSM-5 (PCL-5). Scale available from the National Center for PTSD. Wilkins, K.C., Lang, A.J., Norman, S.B., 2011. Synthesis of the psychometric properties of the PTSD checklist (PCL) military, civilian, and specific versions. Depress. Anxiety 28, 596–606. Zoellner, L.A., Pruitt, L.D., Farach, F.J., Jun, J.J., 2014. Understanding heterogeneity in PTSD: fear, dysphoria, and distress. Depress. Anxiety 31, 97–106.
Contents lists available at ScienceDirect
Journal of Affective Disorders journal homepage: www.elsevier.com/locate/jad
Research report
Patterns of DSM-5 posttraumatic stress disorder and depression symptoms in an epidemiological sample of Chinese earthquake survivors: A latent profile analysis Xing Cao a,b, Li Wang a,n, Chengqi Cao a,b, Jianxin Zhang a, Ping Liu a,b,c, Biao Zhang c, Qi Wu c, Hong Zhang c, Zhihong Zhao d, Gaolin Fan d, Jon D. Elhai e a
Key Laboratory of Mental Health, Institute of Psychology, Chinese Academy of Sciences, Beijing, China Graduate University of Chinese Academy of Sciences, Beijing, China People's Hospital of Deyang City, Deyang, Sichuan, China d Hanwang People's Hospital, Deyang, Sichuan, China e Department of Psychology, and Department of Psychiatry, University of Toledo, Toledo, USA b c
art ic l e i nf o
a b s t r a c t
Article history: Received 11 April 2015 Accepted 2 June 2015 Available online 21 July 2015
Background: Posttraumatic stress disorder (PTSD) and depression are highly comorbid in association with serious clinical consequences. Nevertheless, to date, no study using latent class or latent profile analysis (LCA/LPA) has examined patterns of co-occurring PTSD and depression symptoms among natural disaster survivors, nor has the distinctiveness of DSM-5 PTSD and depression symptoms been clarified in the aftermath of trauma. This study was primarily aimed at filling these gaps. Methods: LPA was used to examine self-reported PTSD and depression symptoms in an epidemiological sample of 1196 Chinese earthquake survivors. Results: A 4-class solution characterized by low symptoms (53.9%), predominantly depression (18.2%), predominantly PTSD (18.9%) and combined PTSD-depression (9.0%) patterns fit the data best. Demographic characteristics and earthquake-related exposures were specifically or consistently associated with the non-parallel profiles varying in physical health impairment. Limitations: A sample exposed to specific traumatic events was assessed by self-report measures. Conclusions: The distinctiveness of DSM-5 PTSD and depression symptoms following an earthquake suggests that PTSD and depression may be independent sequelae of psychological trauma rather than a manifestation of a single form of psychopathology. The current findings support the distinction between PTSD and depression constructs, and highlight the need for identifications of natural disaster survivors at high risk for PTSD and/or depression, and interventions individually tailored to one's symptom presentations. & 2015 Elsevier B.V. All rights reserved.
Keywords: Posttraumatic stress disorder DSM-5 Depression Latent profile analysis Natural disaster
1. Introduction Comorbidity appears to be more the rule than the exception in psychopathological conditions (Clark et al., 1995). The comorbidity of posttraumatic stress disorder (PTSD) with other disorders is so prevalent that more than 80% of individuals with PTSD meet diagnostic criteria for at least one other disorder (Creamer et al., 2001; Kessler et al., 1995). Major depressive disorder (MDD) is one of the most common comorbid disorders with PTSD. According to a meta-analysis of 57 studies (N ¼ 6670) estimating the mean rate of MDD co-occurrence among PTSD samples, over half of n
Corresponding author. E-mail address: [email protected] (L. Wang).
http://dx.doi.org/10.1016/j.jad.2015.06.058 0165-0327/& 2015 Elsevier B.V. All rights reserved.
individuals with PTSD (52%) had comorbid MDD (Rytwinski et al., 2013), which is corresponding to the rate of comorbidity between PTSD and MDD (48–55%) found in national epidemiological surveys (Elhai et al., 2008; Kessler et al., 1995). Meanwhile, co-occurring PTSD and MDD is generally associated with greater distress, impairment, and health care utilization than PTSD alone (Chan et al., 2009; Ikin et al., 2010; Momartin et al., 2004). Thus, given the high prevalence and serious clinical consequences of PTSD-MDD comorbidity, a better understanding of the relationship between PTSD and depression constructs is desperately needed. One of the explanations for high rate of comorbidity between PTSD and depression in the aftermath of trauma is that they are part of a single general traumatic stress construct (O’Donnell et al., 2004). Supporting this hypothesis, many studies have revealed
X. Cao et al. / Journal of Affective Disorders 186 (2015) 58–65
that PTSD and depression share multiple overlapping risk factors and common vulnerabilities, such as childhood adversity, psychiatric history, trauma severity, lack of social support, and genetic risk (e.g. Breslau et al., 2000; Brewin et al., 2000; Kendler et al., 2002; Koenen et al., 2008; O’Donnell et al., 2004). In addition, by using factor analysis, some studies have found that PTSD and depression load onto the same higher-order factor (Anxious Misery/ Distress; Cox et al., 2002; Miller et al., 2008; Slade and Watson, 2006). And in the quantitative hierarchical model of mood and anxiety disorders proposed by Watson (2005), PTSD is grouped into the subclass of distress disorders along with MDD, suggesting that PTSD and depression share identical elements of one form of psychopathology. Moreover, the bidirectional associations between PTSD and depression in their parallel courses have been found in some longitudinal studies (Dekel et al., 2014; Norman et al., 2011), which provides further evidence for a single construct embodied by PTSD and depression following trauma. However, given the distinctiveness of PTSD and depression in theoretical conceptualizations, symptom patterns, information processing, and underlying physiological and neurobiological mechanisms (Zoellner et al., 2014), many researchers have argued that PTSD and depression are independent sequelae of psychological trauma rather than a manifestation of a single form of psychopathology (e.g. Blanchard et al., 1998; Shalev et al., 1998). Although PTSD and depression share some common risk factors in the aftermath of trauma, unique predictors also have been found (e.g. O’Donnell et al., 2004; Seal et al., 2009; Sharkansky et al., 2000). Besides, some studies have challenged the findings of grouping PTSD and depression into the same underlying dimension, with evidence that most PTSD and depression symptoms load onto two distinct but highly correlated factors (Blanchard et al., 1998; Grant et al., 2008; Gros et al., 2010). Furthermore, even in longitudinal studies revealing bidirectional associations between PTSD and depression following trauma, the impact of PTSD on subsequent depression has been greater and more consistent than that of depression on PTSD (Erickson et al., 2001; Nickerson et al., 2013), which adds to support for a distinction between PTSD and depression constructs. Putting aside the mixed findings from previous studies, it is noteworthy that the majority of them have relied on the traditional technique of factor analysis, which may overlook the differences in co-occurring symptom presentations among heterogeneous groups of individuals. Latent class or latent profile analysis (LCA/LPA, using categorical or continuous indicators, respectively), however, is a person-centered method that identifies subgroups of individuals who share common comorbidity patterns (McCutcheon, 1987), and may therefore contribute to a better understanding of the relationship between PTSD and depression following trauma. For instance, Hruska et al. (2014) used LCA to examine patterns of psychiatric comorbidity (PTSD, MDD and alcohol/other drug use disorder) among 249 motor vehicle accident (MVA) victims 6-weeks post-MVA, and found that PTSD and MDD were parallel to each other in a 4-class model consisting of resilient, mild psychopathology, moderate psychopathology, and severe psychopathology classes. Similarly, by using LPA to examine profiles of 140 parents' distress (PTSD, depression and anxiety symptoms) 2 years after war exposure, three parallel subgroups named low distress, moderate distress, and high distress respectively were identified (Lai et al., 2014), which also supports the hypothesis of a general traumatic stress response. Furthermore, focusing only on co-occurring PTSD and depression symptoms, Au et al. (2013) conducted an LPA on 119 female sexual assault survivors at 1- to 4-months post- assault. A 4-class model comprised of four parallel profiles with low, low-moderate, high-moderate, and severe levels of both PTSD and depression symptoms provided best fit to the data across time. Recently, Armour et al. (2015) used
59
LPA to identify profiles of PTSD and MDD symptoms in 283 Canadian veterans, and found three parallel patterns characterized by high, moderate, and low symptoms of PTSD and MDD, likewise suggesting that PTSD and depression are a manifestation of a single form of psychopathology rather than independent sequelae of trauma. Nevertheless, given that the presentations of PTSD and other trauma-related psychopathology may vary for different trauma types (DiMauro et al., 2014), the scarcity of studies using LCA or LPA on co-occurring PTSD and depression symptoms following trauma leaves doubt about the extension of comorbidity patterns found previously to other trauma populations, for example, natural disaster survivors. Besides, the relatively small sample size (e.g. 119; Au et al., 2013) and high level of symptom severity in the sample (e.g. a 78.0% prevalence of probable PTSD; Au et al., 2013) of previous studies partly limit the sample heterogeneity, which may preclude the emergence of possible subgroups with other symptom presentations (e.g. mild PTSD/severe depression). More importantly, as DSM-5 comes with reorganization and redefinition of PTSD symptoms (American Psychiatric Association, 2013), the Criterion D (negative alterations in cognitions and mood) cluster is expanded from the emotional numbing cluster with persistent distorted blame of oneself or others and pervasive negative emotional state, and meanwhile the Criterion E (alterations in arousal and reactivity) cluster is expanded from the hyperarousal cluster with a symptom of reckless or self-destructive behavior. Gootzeit and Markon (2011) have argued that these expanded dimensions represent general factors of negative affect rather than specific components of PTSD, thus their inclusion in DSM-5 may further decrease the distinction between PTSD and depression. However, to date, no study using LCA or LPA has examined whether DSM-5 PTSD symptoms are distinct from co-occurring depression symptoms following trauma. For the above reasons, we conducted an LPA on epidemiological data drawn from a relatively large sample of earthquake survivors to examine patterns of DSM-5 PTSD and depression symptoms. We primarily aimed to identify subgroups differentiated by symptom profiles, which would shed further light on the relationship between PTSD and depression constructs. Our second aim was to examine determinants of symptom profiles in this sample, including demographic characteristics and earthquake-related exposures. Identifying specific predictors of group membership would have potential implications for early preventions, identifications and treatments targeting individuals at high risk for PTSD and/or depression, which, however, has been overlooked by most previous studies. In addition, we also tested whether physical health impairment varies for different symptom profiles. The differences in associations between specific symptom presentations of PTSD and depression with physical health outcomes (e.g. Aversa et al., 2012), serving as external validity, would not only support the distinctiveness of PTSD and depression in theory, but also suggest the preferential interventions for individuals who suffer more clinical consequences in practice.
2. Methods 2.1. Participants and procedures Secondary analyses were conducted on data from a study of Chinese earthquake survivors (Liu et al., 2014). The sample was recruited from one of the largest rebuilt communities located in Hanwang Town, Mianzhu City, which was almost completely destroyed by the 2008 Wenchuan Earthquake with more than 5000 deaths. The survey was conducted from November 6 to 18, 2013, approximately 5.5 years after the earthquake. Household served as
60
X. Cao et al. / Journal of Affective Disorders 186 (2015) 58–65
the basic sample unit, and only one member within each household was randomly selected for participation. Participants met inclusion criteria if they were at least 16 years old without mental retardation or any major psychosis (e.g. schizophrenia and organic mental disorders), and experienced the disaster personally. The sample of 1196 people (response rate¼95%) consisted of 67.7% females and 32.3% males. Age ranged from 16 to 73 years (M ¼47.9 years; SD ¼ 10.0). Ethnicity was self-identified as predominantly Han (99.7%) and other ethnicities in China (0.3%). In regard of educational level, 31.6% completed high school or above, and 67.1% did not complete high school. As for marital status, 86.8% were married, and 13.0% were unmarried (single/divorced/separated/ widowed). Investigators including trained clinical psychologists, psychiatrists, psychotherapists, and psychology graduate students explained the aim and significance of the study, and obtained written informed consent before administering self-reported questionnaires to the participants. The study protocol was approved by the Institutional Review Board of the Institute of Psychology, Chinese Academy of Sciences. Additional details on sampling procedures are described elsewhere (Liu et al., 2014). 2.2. Measures Demographic characteristics assessed included age, sex, ethnicity, educational level and marital status. Earthquake-related exposures assessed included (1) being trapped under rubble; (2) being injured; (3) being disabled due to injuries; (4) participating in rescue efforts; (5) witnessing a death of someone; (6) exposure to mutilated bodies; (7) traumatic death of a family member; (8) traumatic injury of a family member; (9) traumatic death of a friend or neighbor; and (10) losing livelihood due to the disaster. Features of trauma exposure of the sample are shown elsewhere (Liu et al., 2014). The total number of exposures above ranging from 0 to 10, reflected severity of earthquake-related exposure. PTSD symptoms were assessed with the PTSD Checklist for DSM-5 (PCL-5; Weathers et al., 2013). The PCL-5 is a 20-item selfreport instrument adapted from the original PCL to map directly onto DSM-5 criteria for PTSD, and each item is rated on a 5-point Likert scale reflecting severity of PTSD symptoms from 0 (not at all) to 4 (extremely) during the past month. The original PCL has demonstrated sound psychometric properties (McDonald and Calhoun, 2010; Wilkins et al., 2011), and its Chinese version has been validated and widely used in Chinese trauma populations (e.g. Li et al., 2010). The Chinese version of the PCL-5 was obtained by a translation and back-translation process based on the Chinese version of the original PCL. In this study, the PCL-5 was completed with respect to the 2008 Wenchuan Earthquake. Total scores were calculated, along with subscores for intrusion, avoidance, negative alterations in cognitions and mood, and hyperarousal, based on the four-factor model defined by DSM-5 (American Psychiatric Association, 2013). Cronbach's α was 0.94 for the total scale, and 0.86, 0.81, 0.86, and 0.85 for the intrusion, avoidance, negative alterations in cognitions and mood, and hyperarousal subscales respectively in the current sample. Depression symptoms were assessed with the Center for Epidemiological Studies-Depression Scale (CES-D; Radloff, 1977). The CES-D is a 20-item self-report instrument adopting a 4-point Likert scale to indicate severity of depression symptoms from 0 (rare or none of the time/less than 1 day) to 3 (most or all of the time/ 5–7 days) during the past week. The CES-D has demonstrated favorable psychometric properties across various samples (e.g. Lewinsohn et al., 1997; Radloff, 1977), and its Chinese version has been validated and widely used in Chinese populations (e.g. Chi, 1995). In this study, four items describing positive affect were
reversed before calculating total scores as well as subscores for depressive affect, positive affect, somatic complaints, and interpersonal problems, based on the four-factor model identified by Radloff (1977). Cronbach's α was 0.87 for the total scale, and 0.85, 0.67, 0.77, and 0.67 for the depressive affect, positive affect, somatic complaints, and interpersonal problems subscales respectively in the current sample. Physical health impairment was assessed with the Physical Health Questionnaire (PHQ; Schat et al., 2005). The PHQ is a 14item self-report instrument for four dimensions of somatic health, including sleep disturbances, headaches, gastrointestinal problems, and respiratory infections. Items are rated on a 7-point Likert scale ranging from 1 (not at all) to 7 (all of the time), reflecting the frequency with which the respondents experienced the symptoms during the past year. The PHQ has demonstrated good reliability and validity (Schat et al., 2005), and has been used in studies of various samples (e.g. Tinlin et al., 2013). The Chinese version of the PHQ was adapted by a two-stage process of translation and back translation. In this study, total scores were calculated (reverse-scoring Item 4, the endorsement of which indicated the absence of symptoms), and higher scores reflected more severe physical health impairment. Cronbach's α was 0.83 for the total scale in the current sample. 2.3. Statistical analysis In order to minimize the number of indicators and facilitate model convergence, while maximizing the interpretability of class solution, as suggested by Au et al. (2013), we conducted LPA on eight sum scores for four PTSD factors (i.e. intrusion, avoidance, negative alterations in cognitions and mood, and hyperarousal) and four depression factors (i.e. depressive affect, positive affect, somatic complaints, and interpersonal problems) using Mplus 7.0 maximum likelihood estimation with robust standard errors (Muthén and Muthén, 1998–2012). Given the difference in scale range between the PCL-5 and CES-D, the eight sum scores were converted to T scores for the LPA. 1- to 5-class solutions of LPA models were evaluated and compared based on fit indices, parsimony, and interpretability. A lower Bayesian Information Criterion (BIC) value, a lower Akaike Information Criterion (AIC) value, a higher entropy value, a significant Lo–Mendell–Rubin likelihood ratio test (LMR LRT), a significant Lo–Mendell–Rubin adjusted likelihood ratio test (ALMR LRT) and a significant bootstrap likelihood ratio test (BLRT) are indicative that the class solutions have better goodness of fit and more accurate classification (Akaike, 1987; Lo et al., 2001; Nylund et al., 2007; Schwarz, 1978). After the optimal class solution was determined, we conducted subsequent analyses with the information of class assignments in SPSS 19.0. We used ANOVAs to examine the between-class differences in PTSD symptoms (PCL-5 total score, and the four PCL-5 subscores), depression symptoms (CES-D total score, and the four CES-D subscores), and physical health impairment (PHQ total score). The Games–Howell tests, which do not assume equal variances or sample sizes, were used for post-hoc pairwise comparisons (Games and Howell, 1976). To examine determinants of symptom profiles, we conducted multinomial logistic regression analyses. The class assignments served as the dependent variables, and the predictors included age, sex, educational level, marital status, and earthquake-related exposures (both the exposure severity and specific exposures were analyzed respectively).
3. Results The mean scores on the PCL-5 and CES-D were 18.8 (SD ¼13.5, range: 1–77) and 27.8 (SD ¼ 6.5, range: 15–51) respectively in this
X. Cao et al. / Journal of Affective Disorders 186 (2015) 58–65
Table 1 Fit indices for latent profile analyses.
1-class 2-class 3-class 4-class 5-class
BIC
AIC
Entropy
LMR LRT
ALMR LRT
BLRT
71320.40 68515.55 67598.83 67152.89 66826.92
71239.01 68388.39 67425.88 66934.16 66562.41
– 0.92 0.88 0.88 0.84
– o 0.001 o 0.001 0.002 0.336
– o 0.001 o 0.001 0.002 0.340
– o 0.001 o 0.001 o 0.001 o 0.001
BIC, Bayesian Information Criterion; AIC, Akaike Information Criterion; LMR LRT, Lo–Mendell–Rubin likelihood ratio test; ALMR LRT, Lo–Mendell–Rubin adjusted likelihood ratio test; BLRT, Bootstrap likelihood ratio test.
sample. Based on the DSM-5 diagnostic algorithm of at least one intrusion symptom, one avoidance symptom, two negative alterations in cognitions and mood symptoms, and two arousal symptoms of at least moderate (2 or higher) severity, adapted from the algorithm of the original PCL (Cook et al., 2003), the prevalence of probable PTSD was 13.8% (n¼ 165). Based on a cutoff score of 28 (Radloff, 1991), the caseness rate for depression was 42.2% (n ¼505). Fit indices for 1- to 5-class solutions of LPA models are presented in Table 1. The 4-class solution was found to have the best fit for the data, on account of lower BIC and AIC values, and significant LMR, ALMR, and BLRT tests. In addition, given that the non-significant LMR test suggested stopping increasing the number of classes (Nylund et al., 2007), the 4-class solution should not be rejected in favor of the 5-class solution, although the 5-class solution was found to have lower BIC and AIC values. Moreover, the 4-class solution was superior to the 5-class solution given the higher entropy value reflecting higher classification accuracy, despite the significant BLRT test of the 5-class solution. Consequently, we selected the 4-class solution as the optimal class solution. This solution was characterized by low symptoms (53.9%), predominantly depression (18.2%), predominantly PTSD (18.9%), and combined PTSD-depression (9.0%) patterns. Fig. 1 displays the PCL-5 and CES-D standardized subscores for the 4-class model of symptom profiles. PCL-5 and CES-D unstandardized total scores, and the four PCL-5 and four CES-D unstandardized subscores, as well as the results of ANOVAs and post-hoc pairwise comparisons are shown in Table 2. The between-class differences in PTSD symptoms (PCL-5 total score, and the four PCL-5 subscores) and depression symptoms (CES-D total score, and the four CES-D subscores) were found to be significant. Class 1 (low symptoms) individuals presented the lowest symptom severity of PTSD and
61
depression. Class 2 (predominantly depression) individuals presented relatively low PTSD symptom severity, while relatively high depression symptom severity with CES-D total scores above the cutoff. Class 3 (predominantly PTSD) individuals, on the contrary, presented relatively high PTSD symptom severity while relatively low depression symptom severity falling below the cutoff. Class 4 (combined PTSD-depression) individuals presented the highest symptom severity of PTSD and depression, with extremely severe PTSD symptoms and CES-D total scores above the cutoff. All four classes were significantly differentiated from each other in overall symptom severity of PTSD and depression, and all four PTSD factors and two depression factors (i.e. depressive affect and somatic complaints). The other two depression factors (i.e. positive affect and interpersonal problems), however, differentiated less well among the four classes, with comparable symptom severity between Class 1 and 3, and Class 2 and 4. Table 3 displays the results of multinomial logistic regression analyses examining determinants of symptom profiles. Compared to those in the low symptoms class, individuals in the predominantly depression, predominantly PTSD, and combined PTSDdepression classes were more likely to be older, female, and to report greater severity of earthquake-related exposure. Additionally, higher educational level was protective against the inclusion in the predominantly PTSD class. Table 4 displays specific earthquake-related exposures associated with symptom profiles after demographic characteristics shown in Table 3 were controlled. Being trapped under rubble, witnessing a death of someone, traumatic death of a family member, and losing livelihood due to the disaster were associated with symptom profiles of predominantly depression, predominantly PTSD, and combined PTSD-depression. Being disabled due to injuries and exposure to mutilated bodies were additionally associated with symptom profile of combined PTSD-depression. Significant between-class differences in physical health impairment were found (F (3, 1192) ¼179.97, p o0.001). The mean scores on the PHQ were 38.2 (SD ¼10.1), 51.6 (SD ¼ 12.0), 47.3 (SD ¼11.2), and 60.7 (SD ¼ 14.1) for Class 1, Class 2, Class 3, and Class 4 respectively. Post-hoc pairwise comparisons indicated that all four classes were significantly differentiated from each other; and based on the degree of severity, they ranked in the following order: Class 4, Class 2, Class 3, and Class 1.
Fig. 1. PCL-5 and CES-D standardized subscores for the 4-class model of symptom profiles. PTSD, post-traumatic stress disorder; In, intrusion; Av, avoidance; NACM, negative alterations in cognitions and mood; Hy, hyperarousal; DEP, depressive affect; POS, positive affect; SOM, somatic complaints; INT, interpersonal problems; PCL-5, PTSD Checklist for DSM-5; CES-D, Center for Epidemiological Studies-Depression Scale.
62
X. Cao et al. / Journal of Affective Disorders 186 (2015) 58–65
Table 2 Class means for 4-class model. Class 1 (low symptoms) M (SD) In 3.89 (2.12) Av 1.20 (1.09) NACM 1.53 (1.80) Hy 2.99 (2.18) DEP 7.34 (1.77) POS 6.39 (2.09) SOM 8.57 (1.83) INT 1.83 (0.58) PCL-5 9.60 (5.09) CES-D 24.13 (4.00)
Class 2 (predominantly depression) M (SD)
Class 3 (predominantly PTSD) M (SD)
Class 4 (combined PTSDdepression) M (SD)
Group comparisons Post-hoc pairwise comF parisons (p o0.05)
6.86 (3.01) 2.21 (1.33) 4.64 (2.88) 7.20 (3.01) 12.29 (2.26) 7.25 (1.66) 12.60 (1.90) 2.97 (0.76) 20.92 (6.71) 35.10 (3.98)
10.15 (3.05) 3.76 (1.71) 6.55 (3.75) 8.10 (3.08) 8.46 (1.81) 6.29 (1.86) 10.17 (1.84) 1.78 (0.51) 28.56 (6.45) 26.70 (3.69)
14.62 5.31 13.86 14.83 13.64 6.90 14.08 2.91 48.62 37.53
704.00*** 430.49*** 684.77*** 707.25*** 555.56*** 13.12*** 413.64*** 221.28*** 1543.32*** 603.46***
(3.39) (1.59) (4.56) (3.83) (2.73) (1.86) (2.53) (1.19) (9.04) (5.69)
1 o2 o3 o4 1 o2 o3 o4 1 o2 o3 o4 1 o2 o3 o4 1 o3 o2 o4 (1 ¼ 3) o(2 ¼ 4) 1 o3 o2 o4 (1 ¼ 3) o(2 ¼ 4) 1 o2 o3 o4 1 o3 o2 o4
PTSD, post-traumatic stress disorder; In, intrusion; Av, avoidance; NACM, negative alterations in cognitions and mood; Hy, hyperarousal; DEP, depressive affect; POS, positive affect; SOM, somatic complaints; INT, interpersonal problems; PCL-5, PTSD Checklist for DSM-5; CES-D, Center for Epidemiological Studies-Depression Scale. ***
p o 0.001.
Table 3 Determinants of symptom profiles.
Age Sex Male (reference) Female Education Less than high school (reference) High school or above Marital status Unmarried (reference) Married Exposure severity
Class 2 (predominantly depression) RRR (95% CI)
Class 3 (predominantly PTSD) RRR (95% CI)
Class 4 (combined PTSD-depression) RRR (95% CI)
1.02 (1.00-1.04)*
1.05 (1.03-1.07)***
1.09 (1.06-1.12)***
– 1.84 (1.29–2.64)**
– 2.02 (1.41–2.88)***
2.25 (1.38–3.68)**
– 1.09 (0.75–1.57)
– 0.71 (0.48–1.05)a
0.92 (0.54–1.55)
– 0.77 (0.49–1.23) 1.29 (1.18–1.42)***
1.03 (0.63–1.69) 1.29 (1.18–1.41)***
0.68 (0.37–1.25) 1.77 (1.56–2.01)***
PTSD, post-traumatic stress disorder; RRR, relative risk ratio; CI, confidence interval. Reference class ¼ Class 1 (low symptoms). Bolded RRRs (95% CI) indicate a significant association with the class. a
p ¼ 0.084. p o 0.05. po 0.01. *** p o 0.001. *
**
Table 4 Specific earthquake-related exposures associated with symptom profiles.
Trapped under rubble Injured Disabled due to injuries Participated in rescue efforts Witnessed a death of someone Exposed to mutilated bodies Traumatic death of family member Traumatic injury of family member Traumatic death of friend or neighbor Lost livelihood due to the disaster
Class 2 (predominantly depression) RRR (95% CI)
Class 3 (predominantly PTSD) RRR (95% CI)
Class 4 (combined PTSD-depression) RRR (95% CI)
2.04 (1.10–3.78)* 0.93 (0.54–1.59) 1.02 (0.33–3.19) 0.99 (0.68–1.42) 1.82 (1.21–2.74)** 1.19 (0.83–1.71) 1.54 (1.05–2.25)* 0.89 (0.62–1.28) 1.33 (0.85–2.06)
2.36 (1.27–4.38)** 0.60 (0.33–1.07) 1.68 (0.59–4.73) 0.97 (0.67–1.40) 1.86 (1.25–2.76)** 1.19 (0.83–1.71) 1.92 (1.32–2.79)** 0.83 (0.58–1.20) 1.20 (0.79–1.82)
2.21 (1.06–4.62)* 1.22 (0.63–2.36) 3.12 (1.03–9.43)* 1.16 (0.71–1.90) 2.02 (1.11–3.67)* 1.58 (0.97–2.58)a 2.69 (1.67–4.35)*** 1.38 (0.86–2.22) 1.82 (0.92–3.63)
1.66 (1.17–2.35)**
1.63 (1.15–2.32)**
2.09 (1.31–3.34)**
PTSD, post-traumatic stress disorder; RRR, relative risk ratio; CI, confidence interval. Reference class ¼ Class 1 (low symptoms). Relative risk ratios are adjusted for demographic characteristics shown in Table 3. Bolded RRRs (95% CI) indicate a significant association with the class. a
p ¼ 0.067. p o 0.05. po 0.01. *** p o 0.001. *
**
X. Cao et al. / Journal of Affective Disorders 186 (2015) 58–65
4. Discussion In this study, we used LPA to examine patterns of DSM-5 PTSD and depression symptoms in an epidemiological sample of Chinese earthquake survivors. Four non-parallel profiles of low symptoms, predominantly depression, predominantly PTSD, and combined PTSD-depression were identified, and found to be determined by specific and common factors and vary in physical health impairment. These findings fill the gaps in the existing literature using LCA or LPA to examine patterns of co-occurring PTSD and depression symptoms among natural disaster survivors, and to clarify the distinctiveness of DSM-5 PTSD and depression symptoms following trauma. Contrary to previous findings of parallel patterns of PTSD and depression symptoms in the aftermath of trauma (Armour et al., 2015; Au et al., 2013; Hruska et al., 2014; Lai et al., 2014), we found that the severity of PTSD and depression symptoms was not consistently reflective of one another, and there were subgroups containing individuals with predominantly PTSD or predominantly depression symptoms. On the one hand, these discrepancies may be attributable to the differences in trauma types (e.g. natural disasters vs. interpersonal traumas) and assessment time points (e.g. in the chronic vs. acute aftermath of trauma); while on the other hand, considering that the relatively large sample size and low-moderate level of symptom severity in this heterogeneous sample might have given rise to the emergency of subgroups with less common but reliable symptom presentations (e.g. mild PTSD/severe depression, eliminated due to its relatively small proportion of the sample; Au et al., 2013), the seemingly novel finding of a 4-class model comprised of four non-parallel profiles (i.e. low symptoms, predominantly depression, predominantly PTSD, and combined PTSD-depression) in this study may provide a more authentic depiction of co-occurring PTSD and depression symptoms following trauma. Moreover, the existence of subgroups with predominantly PTSD or predominantly depression symptoms in the aftermath of trauma is in line with previous findings of traumatized individuals with either PTSD or depression in the absence of the other across various samples (e.g. Lai et al., 2013; Morina et al., 2013), and is also consistent with previous studies suggesting that PTSD and depression are independent sequelae of trauma rather than a manifestation of a single form of psychopathology (e.g. Blanchard et al., 1998; Shalev et al., 1998). Some researchers have argued that the inclusion of expanded negative cognitions/mood and hyperarousal dimensions overlapping with depression and other mood and anxiety disorders in DSM-5 may decrease the distinction between PTSD and depression or other disorders (Gootzeit and Markon, 2011). However, the current finding of predominantly PTSD and predominantly depression symptom presentations demonstrates that DSM-5 PTSD symptoms can differentiate from co-occurring depression symptoms following trauma, which adds to evidence for the distinctiveness of PTSD and depression from the perspective of comorbidity patterns, and also generally supports the strong diagnostic construct validity of PTSD in DSM-5. In addition, some previous studies have found that removing symptoms theorized to overlap with mood and anxiety disorders from PTSD's criteria has little impact on comorbidity rates (Elhai et al., 2008; Grubaugh et al., 2010), which suggests that the association between PTSD and other disorders may be a natural aspect of the disorder's phenomenology rather than an artifact of poor differential diagnosis (Elhai et al., 2008), and the overlapping symptoms may not be responsible for the comorbidity between PTSD and other disorders. Thus, given that PTSD and depression represent two distinct constructs rather than a single general traumatic stress construct despite sharing some common symptoms, other
63
etiological hypotheses for PTSD-depression comorbidity, for example, preexisting PTSD causes depression (Stander et al., 2014) are supposed to be considered and evaluated in future research. Another noteworthy finding in this study is that all four PTSD factors could differentiate well among the four subgroups characterized by parallel varying severity of PTSD symptoms, which is in line with findings of previous studies using LCA or LPA to examine the heterogeneity of PTSD symptoms (e.g. Rosellini et al., 2014). Nevertheless, although overall depression symptom severity was able to differentiate between subgroups, only depressive affect and somatic complaints, but not positive affect and interpersonal problems, could distinguish one subgroup from another, indicating that the patterns of depression symptoms were not completely parallel across four factors. This finding is partly consistent with previous evidence that there are classes of individuals with relatively high negative affect and somatic symptoms but relatively low anhedonia, or relatively high anhedonia but relatively low negative affect (e.g. Hybels et al., 2013; Rodgers et al., 2014), which provides further support for the distinction between negative and positive affect suggested by previous theoretical and empirical studies (Watson, 2005, 2009), as well as the Research Domain Criteria (RDoC) project (Cuthbert and Kozak, 2013). In this study, we identified some factors consistently associated with symptom profiles of predominantly depression, predominantly PTSD, and combined PTSD-depression among earthquake survivors. These findings roughly accord with meta-analyses of risk factors for PTSD (Brewin et al., 2000) and depression following natural disasters (Tang et al., 2014), and highlight the importance of early preventions, identifications and treatments targeting high-risk individuals, especially those who are older, female, at high levels of exposure severity, and experienced some specific earthquake-related exposures, including being trapped under rubble, witnessing a death of someone, traumatic death of a family member, and losing livelihood due to the disaster. Notably, some factors were specifically related to certain symptom profiles, which provides evidence for the differences in predictions of specific symptom presentations of PTSD and depression, and suggests the specific identifications of individuals at high risk for PTSD and/or depression. In addition, we found that physical health impairment varied for different symptom profiles: individuals with combined PTSDdepression symptoms suffered the most, and individuals with predominantly depression and predominantly PTSD symptoms were the second and third most severe subgroups, respectively, while individuals with low symptoms suffered the least. These differences in associations between specific symptom presentations of PTSD and depression with physical health outcomes are consistent with previous findings that PTSD and depression symptoms differentially affect the various physical health status domains (Aversa et al., 2012), which further validates the distinction between PTSD and depression with evidence of clinical meaningfulness, and indicates the preferential interventions for individuals who may suffer more negative consequences, especially those with combined PTSD-depression symptoms. Moreover, given that LCA or LPA is a person-centered method which can identify subgroups of individuals with specific symptom presentations, some studies using LCA or LPA to examine PTSD symptom profiles have supported the need for interventions individually tailored to one's severity of symptoms and the extent of functional impairment (Hellmuth et al., 2014). In the same way, the findings of symptomatic profiles of predominantly depression, predominantly PTSD, and combined PTSD-depression varying in physical health impairment in this study suggest that individually tailored interventions according to one's symptom presentations of PTSD and depression as well as the severity of symptoms and the extent of physical health impairment may contribute to
64
X. Cao et al. / Journal of Affective Disorders 186 (2015) 58–65
improving the efficiency and effectiveness of treatments. There are several noteworthy limitations to this study. First, the current findings might be limited by the specific trauma type, thus further replications across various trauma types are warranted. Second, this study only relied on self-report measures, and therefore the findings of this study also need to be further tested using data from clinical interviews. Third, the cross-sectional design at a single time point of this study precluded the examination of changes in symptom profiles and class membership over time. Future studies should conduct latent transition analysis on longitudinal data to demonstrate the more reliable symptom presentations of PTSD and depression over time. Fourth, psychiatric history prior to the earthquake, as well as stressful life events before and after the earthquake were not assessed in this study, thus future studies examining these determinants of symptom profiles of PTSD and depression are needed. Fifth, only PTSD but not depression symptoms were assessed based on DSM-5 criteria in this study. Therefore, further replications based on DSM-5 criteria for PTSD and MDD are also needed. Finally, we only adopted physical health impairment as the clinical variable to test external validity of the 4-class model of symptom profiles. Future studies should further examine the differentiable associations between specific symptom presentations of PTSD and depression with a range of external psychological and biological variables. Despite these limitations, to our knowledge this is the first study using LPA to examine patterns of co-occurring PTSD and depression symptoms among natural disaster survivors, and to clarify the distinctiveness of DSM-5 PTSD and depression symptoms following trauma. The current findings of non-parallel profiles determined by specific and common factors and varying in physical health impairment demonstrate that DSM-5 PTSD symptoms can differentiate from co-occurring depression symptoms, which supports the distinction between PTSD and depression constructs, and highlights the need for identifications of natural disaster survivors at high risk for PTSD and/or depression, and interventions individually tailored to one's symptom presentations. Conflict of interest None.
Contributors Xing Cao conducted the literature review and the analysis, and wrote the first draft of the manuscript. Li Wang conceptualized the study design, conducted the study, assisted with the analysis, and commented on drafts. Chengqi Cao and Jianxin Zhang assisted with the study and the analysis, and commented on drafts. Ping Liu, Biao Zhang, Qi Wu, Hong Zhang, Zhihong Zhao, and Gaolin Fan collected the data and commented on drafts. Jon D. Elhai revised the first draft of the manuscript and commented on drafts. All authors have approved the final paper.
Role of the funding source The conduct of this study was financially supported in part by the National Key Technology Research and Development Program of China (No. 2013BAI08B02), the Key Research Program of the Chinese Academy of Sciences (No. KJZD-EW-L04), the National Natural Science Foundation of China (No. 31271099 and No. 31471004), and the Outstanding Young Investigator Foundation of
the Institute of Psychology, Chinese Academy of Sciences. The funding sources had no role in study design; in the collection, analysis and interpretation of data; in the writing of the report; or in the decision to submit the paper for publication.
Acknowledgments This study was partially supported by the National Key Technology Research and Development Program of China (No. 2013BAI08B02), the Key Research Program of the Chinese Academy of Sciences (No. KJZD-EW-L04), the National Natural Science Foundation of China (No. 31271099 and No. 31471004), the Outstanding Young Investigator Foundation of the Institute of Psychology, Chinese Academy of Sciences, and the Key Laboratory of Mental Health, Institute of Psychology, Chinese Academy of Sciences.
References Akaike, H., 1987. Factor analysis and AIC. Psychometrika 52, 317–332. Armour, C., Contractor, A., Elhai, J.D., Stringer, M., Lyle, G., Forbes, D., Richardson, J. D., 2015. Identifying latent profiles of posttraumatic stress and major depression symptomssymptoms in Canadian veterans; exploring differences across profiles in health related functioning. Psychiatry Res. 228, 1–7. American Psychiatric Association, 2013. Diagnostic and Statistical Manual of Mental Disorders, 5th ed. American Psychiatric Association, Washington, DC. Au, T.M., Dickstein, B.D., Comer, J.S., Salters-Pedneault, K., Litz, B.T., 2013. Co-occurring posttraumatic stress and depression symptoms after sexual assault: a latent profile analysis. J. Affect. Disord. 149, 209–216. Aversa, L.H., Stoddard, J.A., Doran, N.M., Au, S., Chow, B., McFall, M., Saxon, A., Baker, D.G., 2012. PTSD and depression as predictors of physical health-related quality of life in tobacco-dependent veterans. J. Psychosomat. Res. 73, 185–190. Blanchard, E.B., Buckley, T.C., Hickling, E.J., Taylor, A.E., 1998. Posttraumatic stress disorder and comorbid major depression: Is the correlation an illusion? J. Anxiety Disord. 12, 21–37. Breslau, N., Davis, G.C., Peterson, E.L., Schultz, L.R., 2000. A second look at comorbidity in victims of trauma: the posttraumatic stress disorder-major depression connection. Biol. Psychiatry 48, 902–909. Brewin, C.R., Andrews, B., Valentine, J.D., 2000. Meta-analysis of risk factors for posttraumatic stress disorder in trauma-exposed adults. J. Consult. Clin. Psychol. 68, 748–766. Chan, D., Cheadle, A.D., Reiber, G., Unutzer, J., Chaney, E.F., 2009. Health care utilization and its costs for depressed veterans with and without comorbid PTSD symptoms. Psychiatr. Serv. 60, 1612–1617. Chi, I., 1995. Mental health of the old–old in Hong Kong. Clin. Gerontol 15, 31–44. Clark, L.A., Watson, D., Reynolds, S., 1995. Diagnosis and classification of psychopathology: challenges to the current system and future directions. Annu. Rev. Psychol. 46, 121–153. Cook, J.M., Thompson, R., Coyne, J.C., Sheikh, J.I., 2003. Algorithm versus cut-point derived PTSD in ex-prisoners of war. J. Psychopathol. Behav. Assess. 25, 267–271. Cox, B.J., Clara, I.P., Enns, M.W., 2002. Posttraumatic stress disorder and the structure of common mental disorders. Depress. Anxiety 15, 168–171. Creamer, M., Burgess, P., McFarlane, A.C., 2001. Post-traumatic stress disorder: findings from the Australian National Survey of Mental Health and Well-being. Psychol. Med. 31, 1237–1247. Cuthbert, B.N., Kozak, M.J., 2013. Constructing constructs for psychopathology: the NIMH research domain criteria. J. Abnorm. Psychol. 122, 928–937. Dekel, S., Solomon, Z., Horesh, D., Ein-Dor, T., 2014. Posttraumatic stress disorder and depressive symptoms: joined or independent sequelae of trauma? J. Psychiatr. Res. 54, 64–69. DiMauro, J., Carter, S., Folk, J.B., Kashdan, T.B., 2014. A historical review of traumarelated diagnoses to reconsider the heterogeneity of PTSD. J. Anxiety Disord. 28, 774–786. Elhai, J.D., Grubaugh, A.L., Kashdan, T.B., Frueh, B.C., 2008. Empirical examination of a proposed refinement to DSM-IV posttraumatic stress disorder symptom criteria using the National Comorbidity Survey Replication data. J. Clin. Psychiatry 69, 597–602. Erickson, D.J., Wolfe, J., King, D.W., King, L.A., Sharkansky, E.J., 2001. Posttraumatic stress disorder and depression symptomatology in a sample of Gulf War veterans: a prospective analysis. J. Consult. Clin. Psychol. 69, 41–49. Games, P.A., Howell, J.F., 1976. Pairwise multiple comparison procedures with unequal n's and/or variances: a Monte Carlo study. J. Educ. Behav. Stat. 1, 113–125. Gootzeit, J., Markon, K., 2011. Factors of PTSD: differential specificity and external correlates. Clin. Psychol. Rev. 31, 993–1003. Grant, D.M., Beck, J.G., Marques, L., Palyo, S.A., Clapp, J.D., 2008. The structure of distress following trauma: posttraumatic stress disorder, major depressive disorder, and generalized anxiety disorder. J. Abnorm. Psychol. 117, 662–672. Gros, D.F., Simms, L.J., Acierno, R., 2010. Specificity of posttraumatic stress disorder symptoms: an investigation of comorbidity between posttraumatic stress disorder symptoms and depression in treatment-seeking veterans. J. Nerv. Ment. Dis. 198, 885–890. Grubaugh, A.L., Long, M.E., Elhai, J.D., Frueh, B.C., Magruder, K.M., 2010. An
X. Cao et al. / Journal of Affective Disorders 186 (2015) 58–65
examination of the construct validity of posttraumatic stress disorder with veterans using a revised criterion set. Behav. Res. Ther. 48, 909–914. Hellmuth, J.C., Jaquier, V., Swan, S.C., Sullivan, T.P., 2014. Elucidating posttraumatic stress symptom profiles and their correlates among women experiencing bidirectional intimate partner violence. J. Clin. Psychol. 70, 1008–1021. Hruska, B., Irish, L.A., Pacella, M.L., Sledjeski, E.M., Delahanty, D.L., 2014. PTSD symptom severity and psychiatric comorbidity in recent motor vehicle accident victims: a latent class analysis. J. Anxiety Disord. 28, 644–649. Hybels, C.F., Landerman, L.R., Blazer, D.G., 2013. Latent subtypes of depression in a community sample of older adults: can depression clusters predict future depression trajectories? J. Psychiatr. Res. 47, 1288–1297. Ikin, J.F., Creamer, M.C., Sim, M.R., McKenzie, D.P., 2010. Comorbidity of PTSD and depression in Korean War veterans: prevalence, predictors, and impairment. J. Affect. Disord. 125, 279–286. Kendler, K.S., Gardner, C.O., Prescott, C.A., 2002. Toward a comprehensive developmental model for major depression in women. Am. J. Psychiatry 159, 1133–1145. Kessler, R.C., Sonnega, A., Bromet, E., Hughes, M., Nelson, C.B., 1995. Posttraumatic stress disorder in the national comorbidity survey. Arch. Gen. Psychiatry 52, 1048–1060. Koenen, K.C., Fu, Q.J., Ertel, K., Lyons, M.J., Eisen, S.A., True, W.R., Goldberg, J., Tsuang, M.T., 2008. Common genetic liability to major depression and posttraumatic stress disorder in men. J. Affect. Disord. 105, 109–115. Lai, B.S., Hadi, F., Llabre, M.M., 2014. Parent and child distress after war exposure. Br. J. Clin. Psychol. 53, 333–347. Lai, B.S., La Greca, A.M., Auslander, B.A., Short, M.B., 2013. Children's symptoms of posttraumatic stress and depression after a natural disaster: comorbidity and risk factors. J. Affect. Disord. 146, 71–78. Lewinsohn, P.M., Seeley, J.R., Roberts, R.E., Allen, N.B., 1997. Center for Epidemiologic Studies Depression Scale (CES-D) as a screening instrument for depression among community-residing older adults. Psychol. Aging 12, 277–287. Li, H.H., Wang, L., Shi, Z.B., Zhang, Y.Q., Wu, K.K., Liu, P., 2010. Diagnostic utility of the PTSD checklist in detecting PTSD in Chinese earthquake victims. Psychol. Rep. 107, 733–739. Liu, P., Wang, L., Cao, C., Wang, R., Zhang, J., Zhang, B., Wu, Q., Zhang, H., Zhao, Z., Fan, G., Elhai, J.D., 2014. The underlying dimensions of DSM-5 posttraumatic stress disorder symptoms in an epidemiological sample of Chinese earthquake survivors. J. Anxiety Disord. 28, 345–351. Lo, Y.T., Mendell, N.R., Rubin, D.B., 2001. Testing the number of components in a normal mixture. Biometrika 88, 767–778. McCutcheon, A.L., 1987. Latent Class Analysis, Quantitative Applications in the Social Sciences Series, No. 64. Sage Publications, Newbury Park, CA. McDonald, S.D., Calhoun, P.S., 2010. The diagnostic accuracy of the PTSD Checklist: a critical review. Clin. Psychol. Rev. 30, 976–987. Miller, M.W., Fogler, J.M., Wolf, E.J., Kaloupek, D.G., Keane, T.M., 2008. The internalizing and externalizing structure of psychiatric comorbidity in combat veterans. J. Trauma Stress 21, 58–65. Momartin, S., Silove, D., Manicavasagar, V., Steel, Z., 2004. Comorbidity of PTSD and depression: associations with trauma exposure, symptom severity and functional impairment in Bosnian refugees resettled in Australia. J. Affect. Disord. 80, 231–238. Morina, N., Ajdukovic, D., Bogic, M., Franciskovic, T., Kucukalic, A., Lecic-Tosevski, D., Morina, L., Popovski, M., Priebe, S., 2013. Co-occurrence of major depressive episode and posttraumatic stress disorder among survivors of war: How is it different from either condition alone? J. Clin. Psychiatry 74, e212–e218. Muthén, L.K., Muthén, B.O., 1998-2012. Mplus User's Guide, 7th ed. Muthén & Muthén, Los Angeles, CA. Nickerson, A., Steenkamp, M., Aerka, I.M., Salters-Pedneault, K., Carper, T.L., Barnes, J.B., Litz, B.T., 2013. Prospective investigation of mental health following sexual assault. Depress. Anxiety 30, 444–450. Norman, S.B., Trim, R.S., Goldsmith, A.A., Dimsdale, J.E., Hoyt, D.B., Norman, G.J.,
65
Stein, M.B., 2011. Role of risk factors proximate to time of trauma in the course of PTSD and MDD symptoms following traumatic injury. J. Trauma. Stress 24, 390–398. Nylund, K.L., Asparoutiov, T., Muthen, B.O., 2007. Deciding on the number of classes in latent class analysis and growth mixture modeling: a Monte Carlo simulation study. Struct. Equ. Model. 14, 535–569. O’Donnell, M.L., Creamer, M., Pattison, P., 2004. Posttraurnatic stress disorder and depression following traurna: understanding comorbidity. Am. J. Psychiatry 161, 1390–1396. Radloff, L.S., 1977. The CES-D scale: a self-report depression scale for research in the general population. Appl. Psychol. Meas. 1, 385–401. Radloff, L.S., 1991. The use of the Center for Epidemiologic Studies Depression Scale in adolescents and young adults. J. Youth Adolesc. 20, 149–166. Rodgers, S., Grosse Holtforth, M., Muller, M., Hengartner, M.P., Rossler, W., AjdacicGross, V., 2014. Symptom-based subtypes of depression and their psychosocial correlates: a person-centered approach focusing on the influence of sex. J. Affect. Disord. 156, 92–103. Rosellini, A.J., Coffey, S.F., Tracy, M., Galea, S., 2014. A person-centered analysis of posttraumatic stress disorder symptoms following a natural disaster: predictors of latent class membership. J. Anxiety. Disord. 28, 16–24. Rytwinski, N.K., Scur, M.D., Feeny, N.C., Youngstrom, E.A., 2013. The co-occurrence of major depressive disorder among individuals with posttraumatic stress disorder: a meta-analysis. J. Trauma. Stress 26, 299–309. Schat, A.C., Kelloway, E.K., Desmarais, S., 2005. The Physical Health Questionnaire (PHQ): construct validation of a self-report scale of somatic symptoms. J. Occup. Health Psychol. 10, 363–381. Schwarz, G., 1978. Estimating the dimension of a model. Ann. Stat. 6, 461–464. Seal, K.H., Metzler, T.J., Gima, K.S., Bertenthal, D., Maguen, S., Marmar, C.R., 2009. Trends and risk factors for mental health diagnoses among iraq and afghanistan veterans using department of veterans affairs health care, 2002–2008. Am. J. Public Health 99, 1651–1658. Shalev, A.Y., Freedman, S., Peri, T., Brandes, D., Sahar, T., Orr, S.P., Pitman, R.K., 1998. Prospective study of posttraumatic stress disorder and depression following trauma. Am. J. Psychiatry 155, 630–637. Sharkansky, E.J., King, D.W., King, L.A., Wolfe, J., Erickson, D.J., Stokes, L.R., 2000. Coping with Gulf War combat stress: mediating and moderating effects. J. Abnorm. Psychol. 109, 188–197. Slade, T., Watson, D., 2006. The structure of common DSM-IV and ICD-10 mental disorders in the Australian general population. Psychol. Med. 36, 1593–1600. Stander, V.A., Thomsen, C.J., Highfill-McRoy, R.M., 2014. Etiology of depression comorbidity in combat-related PTSD: a review of the literature. Clin. Psychol. Rev. 34, 87–98. Tang, B., Liu, X., Liu, Y., Xue, C., Zhang, L., 2014. A meta-analysis of risk factors for depression in adults and children after natural disasters. BMC Public Health 14, 623. Tinlin, R.M., Watkins, C.D., Welling, L.L.M., DeBruine, L.M., Al-Dujaili, E.A.S., Jones, B. C., 2013. Perceived facial adiposity conveys information about women's health. Br. J. Psychol. 104, 235–248. Watson, D., 2005. Rethinking the mood and anxiety disorders: a quantitative hierarchical model for DSM-V. J. Abnorm. Psychol. 114, 522–536. Watson, D., 2009. Differentiating the mood and anxiety disorders: a quadripartite model. Annu. Rev. Clin. Psychol. 5, 221–247. Weathers, F.W., Litz, B.T., Keane, T.M., Palmieri, P.A., Marx, B.P., Schnurr, P.P., 2013. The PTSD checklist for DSM-5 (PCL-5). Scale available from the National Center for PTSD. Wilkins, K.C., Lang, A.J., Norman, S.B., 2011. Synthesis of the psychometric properties of the PTSD checklist (PCL) military, civilian, and specific versions. Depress. Anxiety 28, 596–606. Zoellner, L.A., Pruitt, L.D., Farach, F.J., Jun, J.J., 2014. Understanding heterogeneity in PTSD: fear, dysphoria, and distress. Depress. Anxiety 31, 97–106.