Heterogeneity in patterns of DSM-5 posttraumatic stress disorder and depression symptoms: Latent profile analyses

Journal of Affective Disorders 212 (2017) 17–24

Contents lists available at ScienceDirect

Journal of Affective Disorders journal homepage: www.elsevier.com/locate/jad

Research paper

Heterogeneity in patterns of DSM-5 posttraumatic stress disorder and depression symptoms: Latent profile analyses☆

MARK

⁎

Ateka A. Contractora, , Michelle E. Roley-Robertsb, Susan Lagdonc, Cherie Armourd a

Department of Psychology, University of North Texas, Denton, USA Department of Psychiatry and Behavioral Health, The Ohio State University Wexner Medical Center, Columbus, USA c School of Nursing & Midwifery, Queens University, Belfast, Northern Ireland, UK d Psychology Research Institute, Ulster University, Coleraine, Northern Ireland, UK b

A R T I C L E I N F O

A BS T RAC T

Keywords: DSM-5 PTSD Depression Latent profile analyses Covariates

Background: Posttraumatic stress disorder (PTSD) and depression co-occur frequently following the experience of potentially traumatizing events (PTE; Morina et al., 2013). A person-centered approach to discern heterogeneous patterns of such co-occurring symptoms is recommended (Galatzer-Levy and Bryant, 2013). We assessed heterogeneity in PTSD and depression symptomatology; and subsequently assessed relations between class membership with psychopathology constructs (alcohol use, distress tolerance, dissociative experiences). Methods: The sample consisted of 268 university students who had experienced a PTE and susequently endorsed clinical levels of PTSD or depression severity. Latent profile analyses (LPA) was used to identify the best-fitting class solution accouring to recommended fit indices (Nylund et al., 2007a); and the effects of covariates was analyzed using a 3-step approach (Vermunt, 2010). Results: Results of the LPA indicated an optimal 3-class solutions: high severity (Class 2), lower PTSD-higher depression (Class 1), and higher PTSD-lower depression (Class 3). Covariates of distress tolerance, and different kinds of dissociative experiences differentiated the latent classes. Limitations: Use of self-report measure could lead to response biases; and the specific nature of the sample limits generalizability of results. Conclusion: We found evidence for a depressive subtype of PTSD differentiated from other classes in terms of lower distress tolerance and greater dissociative experiences. Thus, transdiagnostic treatment protocols may be most beneficial for these latent class members. Further, the distinctiveness of PTSD and depression at comparatively lower levels of PTSD severity was supported (mainly in terms of distress tolerance abilities); hence supporting the current classification system placement of these disorders.

1. Introduction

individualized treatment planning. A person-centered approach to discern such subgroup categorizations is recommended (GalatzerLevy and Bryant, 2013). Thus, we assessed heterogeneity in PTSDdepression symptomatology using latent profile analyses; and subsequently assessed relations between class membership and psychopathology constructs. There are several explanations for the widespread comorbidity between PTSD and depression. According to the “causality” explanation, PTSD may be a causal risk factor for depression; or depression may be a causal risk factor for additional PTE experiences, and thereby may contribute to the risk of PTSD (Breslau, 2009; Strander et al., 2014). In contrast, the “common factors” explanation highlights shared environmental and genetic risk or buffering factors for these disorders

Posttraumatic stress disorder (PTSD) and major depressive disorder (MDD) co-occur frequently following the experience of potentially traumatic events (PTE; Bonde et al., 2016). A recent metaanalytical study indicated that 52% of individuals with a current PTSD diagnosis had co-occurring depression (Rytwinski et al., 2013). People with co-occurring PTSD and depression experience greater functional impairment (Post et al., 2012), greater number of PTEs, and greater psychological distress (Morina et al., 2013) compared to those with only PTSD or depression. Further, PTSD as a construct is highly heterogeneous (Galatzer-Levy and Bryant, 2013); hence identifying subgroups of people following the experience of PTEs is critical for

☆ ⁎

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. Correspondence to: University of North Texas, Department of Psychology, 369 Terrill Hall, Denton, TX 76203, USA. E-mail address: [email protected] (A.A. Contractor).

http://dx.doi.org/10.1016/j.jad.2017.01.029 Received 22 October 2016; Received in revised form 16 December 2016; Accepted 23 January 2017 Available online 23 January 2017 0165-0327/ © 2017 Elsevier B.V. All rights reserved.

Journal of Affective Disorders 212 (2017) 17–24

A.A. Contractor et al.

(Breslau, 2009; Strander et al., 2014). Finally, the “confounding factors” explanation indicates that the associations between PTSD and depression are coincidental (Flory and Yehuda, 2015; Strander et al., 2014) attributed to secondary influences on both disorders (e.g., effect of personal expectations on the diagnostic process) (Strander et al., 2014), or shared symptoms across the disorders (e.g., sleep difficulties) (Spitzer et al., 2007; Strander et al., 2014). The empirically supported quantitative hierarchical model (Watson, 2005, 2009), proposes a higher-order factor of emotional disorders which encompasses subclasses of 1) bipolar disorders, 2) distress disorders (e.g., MDD, PTSD, generalized anxiety disorder), and 3) fear disorders. PTSD and depression are conceptualized as sharing non-specific negative affect (Watson et al., 2011), represented by PTSD's dysphoria symptoms (Contractor et al., 2014; Elhai et al., 2015). Consequently, one debates whether PTSD and depression belong to a larger spectrum of a posttraumatic stress syndrome or whether they are distinct disorders (Strander et al., 2014). Ultimately, such questions raise concerns regarding their distinctiveness in the current classification system. There is an on-going debate on the existence of a depressive subtype of PTSD (Flory and Yehuda, 2015). People with PTSD, high negative affectivity and lower positive affectivity are more likely to have comorbid depression; whereas people with PTSD, higher negative affectivity and lower constraint are more likely to have comorbid substance-use (Miller et al., 2004). Thus, an underlying internalizing symptomatic pattern characterized by anxiety and sadness may underlie comorbid PTSD and depression. Thus, researchers question if respondents with comorbid PTSD and depression differ from those with only PTSD or depression on biological and psychological indicators (Flory and Yehuda, 2015). Person-centered approaches accounting for population heterogeneity are used to examine sub-types of diagnostic patterns. One such approach is Latent Class Analysis (LCA) or Latent Profile Analysis (LPA); the former assesses categorical symptom indicators while the latter assesses continuous indicators for latent group classification. Such approaches transcend the limitations imposed by use of diagnostic categories; they classify individuals into latent homogenous classes based on similar response patterns (McCutcheon, 1987). Thus, latent profiles are compared with reference to shape (qualitative differences) and symptom levels (quantitative differences) (Nugent et al., 2012). A tabulated summary of studies assessing latent classes of PTSD and depression symptoms is presented in Table 1. Most studies indicate that there is a comparable and parallel level of severity across the disorders mainly represented as a four- (Au et al., 2013; Hruska et al., 2014; Lai et al., 2015) or three-class solution (Armour et al., 2015; Contractor et al., 2015). Only one study found latent classes differing in type (predominantly depression versus predominantly

PTSD latent classes (Cao et al., 2016). Thus, most evidence affirms the existence of a depressive-subtype of PTSD and the idea that PTSD and depression symptoms may be a part of a common post-traumatic response (Norman et al., 2011). In light of these findings, an improved conceptual and empirical understanding of the co-occurrence of PTSD and depression symptom patterns is critical to inform a more reliable and differentiated diagnostic and treatment framework (Rytwinski et al., 2013). No study to our knowledge has examined latent subgroupings in a sample with clinical levels of PTSD or depression severity. Although depression's diagnostic criteria are unchanged from DSM-IV to DSM-5, the PTSD diagnostic criteria have undergone significant changes. In particular, the DSM-5 added several new symptoms represented under the negative alternations in cognitions and mood (NACM) sub-cluster (American Psychiatric Association, 2013), which seem to represent PTSD's non-specific distress and may explain PTSD's comorbidity with depression (Elhai et al., 2015). So far, only one study has examined latent subgroupings of PTSD and depression using DSM-5 PTSD criteria (Cao et al., 2016). However, this particular study by Cao et al. (2016) did not use DSM-based criteria for depression. Thus, analyzing latent classes based on DSM-5 indicators of both PTSD and depression is unprecedented. Moreover, to the extent known to us, no study has analyzed data from a student sample to assess PTSDdepression heterogeneity. Mental health among university students has been a constant focus of research. Not only are single and multiple experiences of PTEs highly prevalent among college students; there is also a high prevalence of clinical and sub-clinical PTSD (Elhai et al., 2012; Vrana and Lauterbach, 1994). Accounting for the aforementioned, we assessed heterogeneity in PTSD and depression symptomatology using data collected from a trauma-exposed university sample with clinical levels of PTSD or depression severity. Our study aims were two-fold: (1) to examine the best-fitting latent class solution in categorizing participants based on their responses to PTSD and depression items; and (2) to examine the construct validity of the optimal latent class-solution. Individuals with PTSD only, depression only, and co-morbid PTSD and depression differ on the extent of alcohol use (Bailey et al., 2012; Hruska et al., 2014). Individuals with PTSD only have greater alcohol use than those with depression only; and those with comorbid depression and PTSD have greater alcohol use than either those with PTSD only and depression only (Bailey et al., 2012; Hruska et al., 2014). Given that the ability to tolerate distress is inversely related to PTSD severity regardless of depression (Vinci et al., 2016), one could infer a functional coping-related role of alcohol for people with clinical PTSD severity. Additionally, dissociative symptoms of depersonalization/derealization have been found to differentiate PTSD only from comorbid PTSD and depression (Armour et al., 2014b). Hence, the

Table 1 Summary of findings of latent class solutions of PTSD and depression symptoms. Study

Sample

Best-fitting class solution

Severity/type differences

Nature of classes

PTSD and depression symptoms Armour et al. (2015)

283 Canadian Veterans.

3-class

Severity

Low, moderate, and high severity classes.

Au et al. (2013) Contractor et al. (2015)* Cao et al. (2016)

119 female sexual assault survivors. 1266 trauma-exposed soldiers.

3-class 3-class

Severity Severity

Low, low-moderate, high-moderate, and severe symptoms. Mild, moderate, and severe classes.

1196 Chinese earthquake survivors.

4-class

Severity and type

Hruska et al. (2014)*

356 motor vehicle accident victims admitted to a Level−1 trauma center. 353 children affected by Hurricane Katrina.

4-class

Severity

3-class solution

Severity

Low symptoms, predominantly depression, predominantly PTSD, and combined PTSD-depression Resilient, mild psychopathology, moderate psychopathology, severe psychopathology. No disturbance; posttraumatic stress only; mixed-internalizing (moderate PTSD severity, clinically significant anxiety and depression).

Lai et al. (2015)*

Note. * indicates studies that used other indicators besides PTSD and depression to inform latent class formations.

18

Journal of Affective Disorders 212 (2017) 17–24

A.A. Contractor et al.

via the media (Elhai et al., 2012). Participants endorsing more than one traumatic event were asked to specify which traumatic event was most distressing (Elhai et al., 2012). PTSD Checklist for DSM-5 (PCL-5; Weathers et al., 2013). The PCL-5 is a 20-item self-report measure that assesses severity of PTSD symptoms referencing the past month. Response options range from 0 ("Not at all") to 4 ("Extremely"). The PCL-5 has shown excellent internal consistency reliability, good test-retest reliability, and good convergent and discriminant validity (Blevins et al., 2015). Internal consistency in the present study was .93. Patient Health Questionnaire-9 (PHQ-9; Kroenke and Spitzer, 2002). The PHQ-9 is a 9-item self-report measure that assesses severity of depression symptoms referencing the past 2 weeks. Response options range from 0 ("Not at all") to 3 ("Nearly every day"). Further, there is an additional item assessing the degree of functional impairment with a response scale ranging from "not difficult at all" to "extremely difficult". The PHQ-9 has shown excellent internal consistency reliability and validity (Kroenke and Spitzer, 2002; Kroenke et al., 2001). Internal consistency in this study was .83. Dissociative Symptoms Scale (DSS; Carlson et al., 2016). The DSS is a 20 item self-report measure assessing facets of dissociation including depersonalization (4 items, e.g., "I felt like I wasn't myself") and derealization (2 items, e.g., "Things around me seemed strange or unreal"); cognitive-behavioral re-experiencing (i.e., intrusion) (4 items; "I got reminded of something upsetting and then spaced out for a while"); gaps in awareness or memory (5 items, e.g., "I couldn't remember things that had happened during the day even when I tried to"); and trauma-related re-experiencing of sensations, thoughts, and behaviors (5 items, e.g., "I smelled something that I know wasn't really there"). The response scale ranges from 0 ("not at all") to 4 ("more than once a day"). The DSS (overall and its subscales) has adequate to excellent internal consistency reliability, and adequate to excellent testretest reliability and validity (Carlson et al., 2016). In the current study, internal consistency of the subscales of depersonalization/derealization, gaps in memory, sensory misperceptions, and cognitive-behavioral re-experiencing were .90, .84, .85, and .77 respectively. Distress Tolerance Scale (DTS; Simons and Gaher, 2005). The DTS is a 15-item measure used to assess one's perceived ability to tolerate emotional distress. It has a response scale ranging from 5 (strongly disagree) to 1 (strongly agree), where higher scores indicate higher distress tolerance ability (Simons and Gaher, 2005). The DTS has good internal consistency and test-retest reliability, as well as good to excellent validity (Simons and Gaher, 2005). Internal consistency in the present study was .90. Alcohol Use Disorders Identification Test (AUDIT; Saunders et al., 1993). The AUDIT is a 10-item measure that assesses extent of alcohol use. Three items assess amount and frequency of alcohol use, three items assessing alcohol dependency, and four items assess problems stemming from alcohol use. The response scale ranges from 0 to 4, where higher scores indicate harmful abuse (Saunders et al., 1993). The AUDIT has adequate to excellent internal consistency and good validity (Allen et al., 1997). Cronbach's alpha in the present study is .84.

constructs of alcohol use, distress tolerance, and dissociative experiences were used to establish construct validity of the optimal class solution. We hypothesized finding an optimal three- (Armour et al., 2015; Contractor et al., 2015) or four-class solution (Au et al., 2013; Hruska et al., 2014; Lai et al., 2015) differing in severity. Additionally, we expected the latent classes to differ in type based on recent research with DSM-5 PTSD symptoms (Cao et al., 2016) and the clinical nature of our sample, which may increase the likelihood of finding a PTSD only and depression only latent class. Regarding the covariate analyses, we predicted that dissociative experiences (Blevins et al., 2014) would differentiate latent classes with predominant/higher PTSD symptom severity from other latent classes. Additionally, given that individuals with comorbid PTSD and depression have greater distress (Morina et al., 2013), we speculate that distress tolerance will distinguish a comorbid PTSD-depression class from other latent classes. Regarding alcohol use, we speculated that members in the latent class of comorbid PTSD - depression symptoms would be more likely to utilize alcohol than members of the other latent classes (Bailey et al., 2012; Hruska et al., 2014). 2. Method 2.1. Procedure and participants Data was collected from university students who were sent a survey invitation email distributed by the university central Information and Communication Technology (ICT) services. The survey invitation was only sent to registered students (full-time and part-time) who were over 18 years of age at the time of the email distribution. The email invitation was sent on multiple instances over a six-month period, and targeted 25,000 students. All project procedures were approved by the University Ethics Committee prior to data collection. The email invitation consisted of a brief description of the overall project, including the purpose, eligibility criteria, an estimation of the amount of time required to complete the survey (approximately 20– 30 min), information about a lottery incentive (six chances to win £50 [roughly $61] Amazon vouchers), and a link to the survey which was hosted on Qualtrics online software site (Qualtrics, 2015). Clicking the link to the survey led to a participant information sheet including a description of the study and details of any potential risks (e.g., questions may elicit upset feelings), potential benefits, confidentiality, and information on voluntary participation and on-going use of data. If the participant selected to continue, this led to a consent form. The consent form contained questions designed to verify that participants understood the nature of the study and what they would be asked to do, including an example of the type of questions they may be asked during the survey. No identifying information was requested. When participants had completed the survey they were assigned a random ID code by the online survey software. Participants were then asked to email this random ID code to an email account which was set up for the lottery incentive. 2.2. Measures

2.3. Exclusions and missing data Demographic Information. Information on age, gender, employment (i.e., full-time, part-time, retired, unemployed, unemployed student), and ethnicity was obtained. Stressful Life Events Screening Questionnaire (SLESQ; Goodman et al., 1998). The SLESQ is a 13-item self-report measure that assesses exposure to PTEs (e.g., childhood sexual abuse). Three additional items were added to address changes in DSM-5 criteria for a qualifying traumatic event. These items were: clarifying whether the participant directly witnessed rather than observed the traumatic events via the media, whether there was repeated exposure to details of the traumatic event, and whether those details were obtained via the person's job or

We restricted our sample of 1416 participants to those endorsing at least one PTE on the SLESQ (n =940) and not missing 30% or more items on either the PCL-5 ( > 6 items) or the PHQ-9 ( > 3 items; n =118) resulting in a sample size of 822 participants. Finally, we restricted the sample of 822 participants to those endorsing clinical level severity on the PCL-5 ( > /=31; Bovin et al., 2015) or the PHQ-9 ( > /=10; Kroenke et al., 2001). Missing data for the effective sample of 268 participants was minimal and was estimated with Maximum Likelihood (ML) in Mplus version 7 for the primary analyses. The sample of 268 participants averaged 24.17 years of age (SD =7.50) and 19

Journal of Affective Disorders 212 (2017) 17–24

A.A. Contractor et al.

Table 2 Descriptive information on demographics and psychopathology constructs for the entire sample and each latent class.

Mean (SD) Age PTSD severity Depression severity Alcohol use Distress tolerance Depersonalization and derealization Gaps in memory and cognition Sensory misperception Cognitive-behavioral re-experiencing n (%)* Female Employed Ethnic status White Asian African American Mixed ethnicity Probable PTSD ( > /= 31) Probable depression ( > /= 10)

Full sample (n=268)

Class 1 (n=63)

Class 2 (n=70)

Class 3 (n=135)

24.17 (7.50) 37.50 (18.91) 13.11 (4.67) 9.46 (6.72) 38.41 (12.80) 3.88 (5.24) 6,24 (5.06) 2.65 (3.98) 4.40 (3.57)

22.42 (4.36) 12.05 (8.55) 12.98 (3.19) 10.31 (8.11) 43.57 (13.58) 1.65 (2.23) 4 (3.96) 1.14 (2.16) 2.69 (2.42)

25.31 (7.91) 59.04 (11.42) 17.67 (3.23) 10.13 (6.97) 32.28 (11.92) 8.40 (6.88) 10.72 (5.48) 5.66 (5.64) 7.37 (3.86)

24.62 (8.31) 38.21 (8.01) 10.81 (4.14) 8.78 (5.85) 39.39 (11.68) 2.47 (3.49) 4.88 (3.65) 1.83 (2.62) 3.63 (2.89)

213 (81.30%) 159 (59.30%)

51 (81.0%) 35 (55.60%)

61 (88.40%) 41 (58.60%)

101 (77.70%) 83 (61.50%)

263 (98.10%) 4 (1.50%) 0 (0%) 1 (.40%) 186 (69.4%) 215 (80.20%)

63 (100%) 0 (0%) 0 (0%) 0 (0%) 0 (0%) 63 (100%)

68 (97.10%) 1 (1.40%) 0 (0%) 1 (1.40%) 70 (100%) 67 (95.70%)

132 (97.8%) 3 (2.20%) 0 (0%) 0 (0%) 116 (85.90%) 85 (63%)

Note. *All reported percentages are valid percentages to account for missing data; Class 1 is lower PTSD-higher depression; Class 2 is high overall; Class 3 is higher PTSD-lower depression.

as the optimal class solution (DiStefano and Kamphaus, 2006; Nylund et al., 2007a). According to LRT value guidelines, the 2-class solution would be the best-fitting model. However, we chose the 3-class model as the best-fitting class solution based on subsequently declining BIC values, a relatively smaller difference in BIC values between the 3- and 4-class models, and conceptual and interpretative meaning. Further, entropy values were quite similar for all the models tested. See Fig. 1 for a graphical depiction of the 3-class solution. Results of the one-way ANOVA indicated that all latent classes significantly differed in total PTSD severity, F (2265) =439.40, p < .001, partial eta2 =.77; and total depression severity, F (2265) =78.90, p < .001, partial eta2 =,37. Tukey's post-hoc comparisons indicated that Class 2 members reported significantly greater PTSD severity (M =59.04, SD =11.42) compared to Class 1 members (M =12.05, SD =8.55), p < .001; and Class 3 members (M =38.21, SD =8.01), p < .001. Further, Class 3 members reported significantly greater PTSD severity compared to Class 1 members, p < .001. Further, Tukey's post-hoc comparisons indicated that Class 2 members reported significantly greater depression severity (M =17.67, SD =3.23) compared to Class 1 members (M =12.98, SD =3.19), p < .001; and Class 3 members (M =10.81, SD =4.14), p < .001. Further, Class 1 members reported significantly greater depression severity compared to Class 3 members, p < .001. Thus, Class 1 is characterized by relatively lower PTSD item-level severity; however, higher depression item-level severity (particularly PHQ-9 items of sleep changes, appetite changes, and fatigue). In comparison, Class 3 is characterized by relatively higher PTSD itemlevel severity and lower depression item-level severity. Thus Class 1 is labelled as a lower PTSD-higher depression class, and Class 3 is labelled as a higher PTSD-lower depression class. Class 2 is characterized by relatively higher PTSD and depression level severity compared to the other classes and hence is termed as a high severity class. This is representative of the depressive subtype of PTSD (Flory and Yehuda, 2015). See Table 4 for detailed results of the multinomial logistic regression analyses (n =201). Results indicated that distress tolerance scores (B =.04, p =.03, OR =1.04) were significant in predicting the lower PTSD-higher depression versus the higher PTSD-lower depression class membership. Higher distress tolerance scores by every single unit increment increased the chances of being in the lower PTSDhigher depression class compared to the higher PTSD-lower depression class by 4%.

the majority were females (n =213, 81.30%). Detailed information on demographics is provided in Table 2. 2.4. Data analysis We conducted a latent profile analyses (LPA) to categorize participants into latent sub-groups based on their endorsed item-level responses on the PCL-5 and the PHQ-9 (McLachlan and Peel, 2000; Muthén, 2004). LPA was conducted using Mplus 7 and the estimator was Maximum Likelihood estimation with robust standard errors (ML). One- through four-class models were analyzed based on prior research findings (e.g., Armour et al., 2015; Contractor et al., 2015). In terms of recommended fit indices, the optimal class solution had lower Akaike Information Criterion (AIC) values, Bayesian Information Criterion (BIC) values and sample-size adjusted BIC values (SSABIC); a significant Lo–Mendell–Rubin likelihood (LRT) test value; a significant Bootstrapped Likelihood Ratio Test (BLRT) p value; relatively higher entropy values; and conceptual meaning (DiStefano and Kamphaus, 2006; Nylund et al., 2007a, 2007b). A model with a 10point lower BIC value has a 150:1 likelihood to be the better fitting model (Raftery, 1995). When comparing a K-class model with a K-1 class model, a significant LRT test indicates that the model with K classes is optimal (Nylund et al., 2007a). After identifying the best-fitting class solution, we first tested if the latent classes significantly differed in total PTSD and depression severity using a one-way ANOVA in SPSS. Further, we tested the effect of covariates (alcohol use, distress tolerance, depersonalization and derealization, gaps in memory and awareness, sensory misperception, and cognitive-behavioral re-experiencing) on latent class membership. We used the three-step approach (multinomial logistic regression) by estimating class membership in relation to auxiliary variables of interest while accounting for misspecification bias (Asparouhov and Muthén, 2014; Vermunt, 2010). For this part of the analyses, the default in dealing with missing data is list-wise deletion in Mplus, thus reducing our sample to 201 participants. 3. Results Descriptive information on demographics and psychopathology constructs is provided in Table 2. Table 3 indicates results of the LPA. Based on established guidelines, we selected the 3-class solution 20

Journal of Affective Disorders 212 (2017) 17–24

A.A. Contractor et al.

Table 3 Fit indices of the latent-class models. Model

AIC

BIC

SSABIC

Entropy

Lo-Mendell –Rubin adjusted LRT (p)

Bootstrapped LRT (p)

1 class 2 class 3 class 4 class

25643.475 24216.325 23427.335 23040.495

25851.752 24532.331 23851.072 23571.961

25667.856 24253.317 23476.939 23102.710

.95 .94 .95

1478.337 (p=.002) 843.958 (p=.12) 444.192 (p= .17)

p < .001 p < .001 p < .001

(Armour et al., 2015; Contractor et al., 2015) which offers support for our study results. However, most prior study results differ from the current study in terms of the nature of classes. Specifically, classes have been found to be different mainly in terms of severity rather than type (Armour et al., 2015; Au et al., 2013; Contractor et al., 2015; Hruska et al., 2014). One exception is the study by Cao et al. (2016) which found classes differing in severity and type; these results are consistent with our study findings. Methodological differences can explain most of the discrepant findings, mainly that our sample was restricted to those reporting clinically significant levels of PTSD and/or depression symptoms. Such sample characteristics would eliminate a low PTSD and depression severity group as found in most studies (e.g., Armour et al., 2015; Contractor et al., 2015). Further, differing data collection procedures, and differing measures may also explain some of the discrepant results. Extensive changes from DSM-IV to DSM-5 in PTSD's diagnostic criteria (American Psychiatric Association, 2013) may influence the cooccurrence patterns of PTSD and depression symptoms. At relatively higher levels of PTSD severity, we find comparable levels of higher PTSD and depression severity (depressive subtype of PTSD). Such results may support the idea of a common latent factor of negative affect underlying PTSD and depression symptoms (Watson et al., 2011) especially when participants endorse a higher level of PTSD severity. The negative affect may be captured by the greater number of dysphoria-laden DSM-5 NACM symptoms (Elhai et al., 2015). We could additionally infer that at relatively lower levels of PTSD severity, a class with prominent depression symptoms may be expected. Our current study findings of a predominant PTSD severity class and a predominant depression severity class extends support for the distinctiveness of PTSD and depression as independent sequelae of post-traumatic experiences (Strander et al., 2014). Although there is a probable PTSD subtype characterized by prominent depression and negative-affect based symptoms; it is posible that following the experience of PTE, participants could have prominent PTSD or depression symptoms only. Such PTSD and depression symptom heterogeneity well demonstrated by the current study findings is important to acknowledge in clinical work. We see a parallel trend in severity for most of the individual PTSD

Comparing the high severity class to the higher PTSD-lower depression class, results indicated significant covariates of distress tolerance (B =−.07, p=.005, OR =.93); and dissociative experiences of gaps in memory (B =.22, p=.009, OR =1.25). Greater dissociative experiences of gaps in memory by every single unit increment increased the chances of being in the high severity class compared to the higher PTSD-lower depression class by 24%. Higher distress tolerance scores by every single unit increment decreased the chances of being in the high severity class compared to the higher PTSD-lower depression class by 7%. Comparing the higher severity class to the lower PTSD-higher depression class, results indicated significant covariates of distress tolerance skills (B =−.11, p < .001, OR =.90); dissociative experiences of derealization and depersonalization (B =.22, p=.03, OR =1.24); and dissociative experiences of gaps in memory (B =.23, p=.05, OR =1.26). Greater disssociative experiences of derealization/depersonalization, and greater dissociative experiences of gaps in memory by every single unit increment increased the chances of being in the higher severity class compared to the lower PTSD-higher depression class by 24%, and 26% respectively. Further, higher distress tolerance scores by every single unit increment decreased the chances of being in the higher severity class compared to the lower PTSD-higher depression class by 10%. 4. Discussion The current study examined latent subgroupings of participants based on the endorsed responses to items assessing PTSD and depression severity. The sample consisted of 268 university students who had experienced a PTE and subsequently endorsed clinical levels of PTSD or depression severity. Results of the LPA indicated an optimal 3-class solution: high severity (Class 2), lower PTSD-higher depression (Class 1), and higher PTSD-lower depression (Class 3). Thus, we found evidence for a depressive subtype of PTSD (high severity class). Additionally, several covariates such as distress tolerance abilities, and different kinds of dissociative experiences differentiated the latent classes. Prior studies have indicted an optimal fitting 3-class solution

Fig. 1. Latent profiles of participants based on responses to PTSD and depression indicators.

21

Journal of Affective Disorders 212 (2017) 17–24

A.A. Contractor et al.

Table 4 Results of multinomial logistic regression analyses.

OR (95% CI) Alcohol use Distress tolerance Depersonalization and derealization Gaps in memory and cognition Sensory misperception Cognitive-behavioral re-experiencing

Class 1 vs. 3#

Class 2 vs. 3#

Class 2 vs. 1#

1.06 (1.00 − 1.13) 1.04 (1.00 − 1.07)** .89 (.74 − 1.08) .99 (.84 − 1.17) .75 (.53 − 1.07) .98 (.80 – 1.20)

1.01 (.93 – 1.09) .93 (.89 – 98)** 1.11 (.97 – 1.27) 1.24 (1.06 – 1.47)** 1.07 (.87 – 1.31) .95 (.79 – 1.15)

.95 (.87 − 1.04) .90 (.85 − .95)* 1.24 (1.02 – 1.52)** 1.26 (1.00 – 1.59) p 1.42 (.96 – 2.08) .98 (.76 – 1.24)

= .05

Note. Class 1 is lower PTSD-higher depression; Class 2 is high overall; Class 3 is higher PTSD-lower depression; * p < .001; ** p < .05; # indicates the reference class.

the latent classes. Alcohol use is highly comorbid with both PTSD and depression (Erbes et al., 2007; Morina et al., 2013), both of which share underlying negative affectivity (Watson, 2009). It could be speculated that alcohol use may serve a similar functional coping role in dealing with the commonly shared negative affect associated with PTSD and depressive symptoms (Martens et al., 2008). Thus, the construct of alcohol use would not distinguish the three latent classes. Indeed, in an experimental study examining alcohol demand, individuals with both elevated PTSD and depressive symptoms had increased alcohol demand intensity (Murphy et al., 2013).

and depression symptoms. Noteworthy is that severity of the PCL-5 item assessing traumatic amnesia is lower relative to other PCL-5 items for the high severity and the higher PTSD-lower depression latent classes. These results are consistent with prior research indicating low base rates for the traumatic amnesia endorsement (Baschnagel et al., 2005; Boal et al., 2016; Palmieri et al., 2007) and may be linked to studies demonstrating that this item often produces the lowest factor loading in PTSD confirmatory factor analytic studies (Armour et al., 2016; Contractor et al., 2014). However, this item has a reverse trend comparative to other PCL-5 items in the lower PTSD-higher depression latent class. Thus, difficulties remembering parts of PTEs increase with higher levels of depression. Prior research has shown that people with higher depression severity have lower memory specificity for stressful experiences (Kleim and Ehlers, 2008; Williams et al., 2007). Results confirmed the construct validity of the latent classes. The depressive subtype of PTSD comprising of co-occuring PTSD and depression symptoms at an equivalent and high level of severity was distinct from other latent classes on certain aspects. Specifically, individuals with comorbid depression and PTSD were more likely to endorse greater dissociative experiences of gaps in memory and lower distress tolerance skills compared to individuals with prominent PTSD symptoms. Further, individuals with the depressive subtype of PTSD were more likely to endorse greater derealization/depersonalization, greater dissociative experiences of gaps in memory, and lower distress tolerance compared to individuals with prominent depression symptoms. Thus, dissociative experiences of gaps in memory and depersonalization/derealization are more associated with co-occurring PTSD and depression severity. There is significant correlation between dissociation and depression severity (Lipsanen et al., 2004); and between dissociation and PTSD severity (Briere et al., 2005). Thus experiences of derealization, depersonalization, and difficulties in memory could be conceptualized as common vulnerability or covariate to co-occurring PTSD and depression. Further, ineffective distress tolerance skills may explain the detrimental functional impairment seen among people with comorbid PTSD and depression severity (Post et al., 2012). The latent class with prominent depression severity was distinct in distress tolerance skills from the latent class with prominent PTSD symptom severity. Specifically, individuals with prominent depression symptoms were more likely to have better distress tolerance abilities compared to individuals with prominent PTSD severity. There is considerable research indicating that people with increased PTSD severity struggle with effective emotion regulation and distress tolerance skills; which in turn, could lead to engagement in risky and impulsive behaviors as coping strategies (Weiss et al., 2013). Dissociative experiences was not distinguishing of latent classes with prominent PTSD and prominent depression symptoms. Dissociative experiences may be an indicator of increased severity of both PTSD and depression symptoms; rather than just one symptomatology. The construct of alcohol use was not significant in distinguishing

4.1. Implications Our study results have several conceptual and clinical implications. The existence of a distinct depression subtype of PTSD indicates the clinical importance of assessing for co-occuring depression following the experience of a PTE. Theoretically, one could speculate that nonspecific negative affect underlies co-occuring PTSD and depression symptoms (Contractor et al., 2015; Watson, 2009) especially when people endorse higher levels of PTSD severity. Clinicians may benefit from using a transdiagnostic treatment protocol targetting negative affect for participants with the co-occuring symptom patterns. Such treatment protocols would additionally benefit from targeting disssociative experiences of gaps in memory, and distress tolerance skills. Further, considering the level of PTSD severity may be critical to identifying latent subgroupings. At lower levels of PTSD severity, classes with predominant PTSD severity and predominant depression severity were distinct. Thus, current study results suggest differing underlying mechanisms for both disorders and confirm the continued separation of the disorders in diagnostic manuals (i.e., DSM/ICD). It may be beneficial to use specific treatment protocols to match the distinct symptom patterns. The psychopathology construct of dissociation is most distinguishing of the latent classes. It is thus important for clinicians to assess for dissociative symptoms, and in particular, the type of dissociation a given client is experiencing (Armour et al., 2014a). With co-occuring depression and PTSD, experience of derealization, depersonalization, and memory difficulties may be critical to assess and target in treatment. Additionally, results support literature on the importance of assessing distress tolerance skills among people with higher PTSD severity irrespective of the level of co-occuring depression severity. Targeting effective distress tolerance and emotion regulation skills as emphasized in certain trauma treatment protocols (Cloitre et al., 2002) may be particularly effective. 4.2. Limitations and future research Results of the current study need to be interpreted in the light of several limitations. The specific nature of our sample restricts its generalizability to other types of samples. Further, we did not assess 22

Journal of Affective Disorders 212 (2017) 17–24

A.A. Contractor et al.

Bonde, J.P., Utzon-Frank, N., Bertelsen, M., Borritz, M., Eller, N.H., Nordentoft, M., 2016. Risk of depressive disorder following disasters and military deployment: systematic review with meta-analysis. Br. J. Psychiatry 208, 330–336. Bovin, M.J., Marx, B.P., Weathers, F.W., Gallagher, M.W., Rodriguez, P., Schnurr, P.P., Keane, T.M., 2015. Psychometric properties of the PTSD Checklist for Diagnostic and Statistical Manual of Mental Disorders–Fifth Edition (PCL-5) in veterans. Psychological Assessment Advance Online Publication. Breslau, N., 2009. The epidemiology of trauma, PTSD, and other posttrauma disorders. Trauma Violence Abus. 10, 198–210. Briere, J., Scott, C., Weathers, F., 2005. Peritraumatic and persistent dissociation in the presumed etiology of PTSD. Am. J. Psychiatry 162, 2295–2301. Cao, X., Wang, L., Cao, C., Zhang, J., Liu, P., Zhang, B., Wu, Q., Zhang, H., Zhao, Z., Fan, G., Elhai, J.D., 2016. Patterns of DSM-5 posttraumatic stress disorder and depression symptoms in an epidemiological sample of Chinese earthquake survivors: a latent profile analysis. J. Affect. Disord. 186, 58–65. Carlson, E.B., Waelde, L.C., Palmieri, P.A., Macia, K.S., Smith, S.R., McDade-Montez, E., 2016. Development and validation of the dissociative symptoms scale. Assessment. Cloitre, M., Koenen, K.C., Cohen, L.R., Han, H., 2002. Skills training in affective and interpersonal regulation followed by exposure: a phase-based treatment for PTSD related to childhood abuse. J. Consult. Clin. Psychol. 70, 1067–1074. Contractor, A.A., Durham, T.A., Brennan, J.A., Armour, C., Wutrick, H.R., Frueh, C.B., Elhai, J.D., 2014. DSM-5 PTSD's symptom dimensions and relations with major depression's symptom dimensions in a primary care sample. Psychiatry Res. 215, 146–153. Contractor, A.A., Elhai, J.D., Fine, T.H., Tamburrino, M.B., Cohen, G.H., Shirley, E., Chan, P.K., Liberzon, I., Galea, S., Calabrese, J.R., 2015. Latent Profile Analyses of Posttraumatic Stress Disorder, Depression and Generalized Anxiety Disorder symptoms in Trauma-Exposed Veterans. J. Psychiatr. Res. 68, 19–26. Dalenberg, C.J., Glaser, D., Alhassoon, O.M., 2012. Statistical support for subtypes in posttraumatic stress disorder: the how and why of subtype analysis. Depress. Anxiety 29, 671–678. Davies, L., Ford-Gilboe, M., Willson, A., Varcoe, C., Wuest, J., Campbell, J., Scott-Storey, K., 2015. Patterns of cumulative abuse among female survivors of intimate partner violence links to women's health and socioeconomic status. Violence Women 21, 30–48. DiStefano, C., Kamphaus, R.W., 2006. Investigating subtypes of child development. A comparison of cluster analysis and latent class cluster analysis in typology creation. Educ. Psychol. Meas. 66, 778–794. Elhai, J.D., Contractor, A.A., Tamburrino, M., Fine, T.H., Cohen, G., Shirley, E., Chan, P.K., Liberzon, I., Calabrese, J.R., Galea, S., 2015. Structural relations between DSM5 PTSD and Major Depression symptoms in military soldiers. J. Affect. Disord. 175, 373–378. Elhai, J.D., Miller, M.E., Ford, J.D., Biehn, T.L., Palmieri, P.A., Frueh, B.C., 2012. Posttraumatic stress disorder in DSM-5: estimates of prevalence and symptom structure in a nonclinical sample of college students. J. Anxiety Disord. 26, 58–64. Erbes, C., Westermeyer, J., Engdahl, B., Johnsen, E., 2007. Post-Traumatic Stress Disorder and Service Utilization in a Sample of Service Members from Iraq and Afghanistan. Mil. Med. 172, 359–363. Flory, J.D., Yehuda, R., 2015. Comorbidity between post-traumatic stress disorder and major depressive disorder: alternative explanations and treatment considerations. Dialog-. Clin. Neurosci. 17, 141–150. Galatzer-Levy, I.R., Bryant, R.A., 2013. 636,120 Ways to Have Posttraumatic Stress Disorder. Perspect. Psychol. Sci. 8, 651–662. Golder, S., Connell, C.M., Sullivan, T.P., 2012. Psychological distress and substance use among community-recruited women currently victimized by intimate partners A latent class analysis and examination of between-class differences. Violence Women 18, 934–957. Goodman, L.A., Corcoran, C., Turner, K., Yuan, N., Green, B.L., 1998. Assessing traumatic event exposure: general issues and preliminary findings for the Stressful Life Events Screening Questionnaire. J. Trauma. Stress 11, 521–542. 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. Kleim, B., Ehlers, A., 2008. Reduced autobiographical memory specificity predicts depression and posttraumatic stress disorder after recent trauma. J. Consult. Clin. Psychol. 76, 231–242. Kroenke, K., Spitzer, R.L., 2002. The PHQ-9: a new depression diagnostic and severity measure. Psychiatr. Ann. 32, 509–515. Kroenke, K., Spitzer, R.L., Williams, J.B.W., 2001. The PHQ 9 validity of a brief depression severity measure. J. Gen. Intern. Med. 16, 606–613. Lai, B.S., Kelley, M.L., Harrison, K.M., Thompson, J.E., Self-Brown, S., 2015. Posttraumatic stress, anxiety, and depression symptoms among children after Hurricane Katrina: a latent profile analysis. J. Child Fam. Stud. 24, 1262–1270. Lipsanen, T., Saarijärvi, S., Lauerma, H., 2004. Exploring the relations between depression, somatization, dissociation and alexithymia–overlapping or independent constructs? Psychopathology 37, 200–206. Martens, M.P., Neighbors, C., Lewis, M.A., Lee, C.M., Oster-Aaland, L., Larimer, M.E., 2008. The roles of negative affect and coping motives int he relationship between alcohol use and alcohol-related problems among college students. J. Stud. Alcohol Drugs 69, 412–419. McCutcheon, A.L., 1987. Latent Class Analysis. McLachlan, G.J., Peel, D., 2000. Finite mixture models. John Wiley & Sons, Inc, Hoboken, NJ. Miller, M.W., Kaloupek, D.G., Dillon, A.L., Keane, T.M., 2004. Externalizing and internalizing subtypes of combat-related PTSD: a replication and extension using the PSY–5 scales. J. Abnorm. Psychol. 113, 636–645.

for other possible factors that may be critical in distinguishing the latent classes such as rumination; this is an avenue for future research. Additionally, the use of self-report measures entails the possibility of response bias. Hence, a multimethod assessment is recommended in subsequent studies. Lack of a longitudinal study precludes any causal interpretations between PTSD and depression severity. Latent transition analyses could capitalize on the temporal nature of the data to answer questions on causality as well as trends in class membership across time; this is an avenue for future research. Lastly, research indicates that specific types of traumatic experiences (e.g., interpersonal traumas) and the cumulative impact of multiple trauma types are associated with a greater risk of PTSD and depression (Davies et al., 2015; Golder et al., 2012). Thus, it would be important to address the interacting patterns of the PTSD-depression latent classes with the number, type, frequency, and age of onset for endorsed traumatic events. In conclusion, we see that PTSD and depression represent disctinct constructs; however seem to co-occur at higher levels of PTSD severity. Thus, identifying and addressing the distinctive nature of the depressive subtype of PTSD is critical. There is extensive discussion on the dissociative and childhood subtypes of PTSD (Dalenberg et al., 2012); however, less on the depressive (internalizing) versus externalizing subtypes of PTSD (Miller et al., 2004). Sufficient research on PTSD subtypes may inform the forumlation of a check-box approach which will allow clinicians to identify the subtypes of PTSD. This would be an alternative to more additions to the diagnostic criteria for PTSD, which risks adding more negative-affect based non-specific symptoms and creating fuzzy diagnostic boundaries across distress-based disorders. Additional research on the nature and prevalence of the depressive subtype of PTSD is needed given the extensive comorbidity across these two disorders. References Allen, J.P., Litten, R.Z., Fertig, J.B., Babor, T.F., 1997. A review of research on the Alcohol Use Disorders Identification Test (AUDIT). Alcohol. Clin. Exp. Res. 21, 613–619. American Psychiatric Association, 2013. Diagnostic and Statistical Manual of Mental Disorders 5th ed.. American Psychiatric Association, Washington, DC. Armour, C., Contractor, A.A., Elhai, J.D., Stringer, M., Lyle, G., Forbes, D., Richardon, J.D., 2015. Identifying latent profiles of Posttraumatic Stress and Major Depression symptoms in Canadian veterans; exploring differences across profiles in health related functioning. Psychiatry Res. 228, 1–7. Armour, C., Contractor, A.A., Palmieri, P.A., Elhai, J.D., 2014a. Assessing latent level associations between PTSD and dissociative factors: is depersonalization and derealization related to PTSD factors more so than alternative dissociative factors? Psychol. Inj. Law 7, 131–142. Armour, C., Contractor, A.A., Shea, M.T., Elhai, J.D., Pietrzak, R.H., 2016. Factor structure of the PCL-5: relationships among symptom clusters, anger, and impulsivity. J. Nerv. Ment. Dis. 204, 108–115. Armour, C., Elklit, A., Lauterbach, D., Elhai, J.D., 2014b. The DSM-5 dissociative-PTSD subtype: can levels of depression, anxiety, hostility, and sleeping difficulties differentiate between dissociative-PTSD and PTSD in rape and sexual assault victims? J. Anxiety Disord. 28, 418–426. Asparouhov, T., Muthén, B., 2014. Auxiliary variables in mixture modeling: three-step approaches using Mplus. Struct. Equ. Model.: A Multidiscip. J. 21, 1–13. Au, T.M., Dickstein, B.D., Comer, J.S., Salters-Pedneault, K., Litz, B.T., 2013. Cooccurring posttraumatic stress and depression symptoms after sexual assault: a latent profile analysis. J. Affect. Disord. 149, 209–216. Bailey, K., Webster, R., Baker, A.L., Kavanagh, D.J., 2012. Exposure to dysfunctional parenting and trauma events and posttraumatic stress profiles among a treatment sample with coexisting depression and alcohol use problems. Drug Alcohol Rev. 31, 529–537. Baschnagel, J.S., O'Connor, R.M., Colder, C.R., Hawk, L.W., 2005. Factor structure of posttraumatic stress among Western New York undergraduates following the September 11th terrorist attack on the World Trade Center. J. Trauma. Stress 18, 677–684. Blevins, C.A., Weathers, F.W., Davis, M.T., Witte, T.K., Domino, J.L., 2015. The Posttraumatic Stress Disorder Checklist for DSM-5 (PCL-5): development and initial psychometric evaluation. J. Trauma. Stress 28, 489–498. Blevins, C.A., Weathers, F.W., Witte, T.K., 2014. Dissociation and Posttraumatic Stress Disorder: a Latent Profile Analysis. J. Trauma. Stress 27, 388–396. Boal, A.L., Vaughan, C.A., Sims, C.S., Miles, J.N.V., 2016. Measurement invariance across administration mode: examining the Posttraumatic Stress Disorder (PTSD) Checklist. Psychol. Assess. Adv. Online Publ..

23

Journal of Affective Disorders 212 (2017) 17–24

A.A. Contractor et al.

meta‐analysis. J. Trauma. Stress 26, 299–309. Saunders, J.B., Aasland, O.G., Babor, T.F., De la Fuente, J.R., Grant, M., 1993. Development of the alcohol use disorders identification test (AUDIT): WHO collaborative project on early detection of persons with harmful alcohol consumption‐II. Addiction 88, 791–804. Simons, J.S., Gaher, R.M., 2005. The Distress Tolerance Scale: development and validation of a self-report measure. Motiv Emot. 29, 83–102. Spitzer, R.L., First, M.B., Wakefield, J.C., 2007. Saving PTSD from itself in DSM-V. J. Anxiety Disord. 21, 233–241. Strander, 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. Vermunt, J.K., 2010. Latent class modeling with covariates: two improved three-step approaches. Political Anal. 18, 450–469. Vinci, C., Mota, N., Berenz, E., Connolly, K., 2016. Examination of the relationship between PTSD and distress tolerance in a sample of male veterans with comorbid substance use disorders. Mil. Psychol. 28, 104–114. Vrana, S., Lauterbach, D., 1994. Prevalence of traumatic events and post‐traumatic psychological symptoms in a nonclinical sample of college students. J. Trauma. Stress 7, 289–302. 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. Watson, D., Clark, L.A., Stasik, S.M., 2011. Emotions and the emotional disorders: a quantitative hierarchical perspective. Int. J. Clin. Health Psychol. 11, 429–442. 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 at 〈www.ptsd.va.gov〉. Weiss, N.H., Tull, M.T., Anestis, M.D., Gratz, K.L., 2013. The relative and unique contributions of emotion dysregulation and impulsivity to posttraumatic stress disorder among substance dependent inpatients. Drug Alcohol Depend. 128, 45–51. Williams, J.M.G., Barnhofer, T., Crane, C., Hermans, D., Raes, F., Watkins, E., Dalgleish, T., 2007. Autobiographical memory specificity and emotional disorder. Psychol. Bull. 133, 122–148.

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, 212–218. Murphy, J.G., Yurasek, A.M., Dennhardt, A.A., Skidmore, J.R., McDevitt-Murphy, M.E., MacKillop, J., Martens, M.P., 2013. Symptoms of depression and PTSD are associated with elevated alcohol demand. Drug Alcohol Depend. 127, 129–136. Muthén, B.O., 2004. Latent variable analysis: growth mixture modeling and related techniques for longitudinal data. In: Kaplan, D. (Ed.), Handbook of quantitative methodology for the social sciences.. Saga, Newbury Park, CA, 345–368. Norman, S.B., Trim, R.S., Goldsmith, A.A., Dimsdale, J.E., Hoyt, D.B., Norman, G.J., 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. Nugent, N.R., Koenen, K.C., Bradley, B., 2012. Heterogeneity of posttraumatic stress symptoms in a highly traumatized low income, urban, African American sample. J. Psychiatr. Res. 46, 1576–1583. Nylund, K., Asparouhov, T., Muthén, B.O., 2007a. 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. Nylund, K., Bellmore, A., Nishina, A., Graham, S., 2007b. Subtypes, severity, and structural stability of peer victimization: what does latent class analysis say? Child Dev. 78, 1706–1722. Palmieri, P.A., Weathers, F.W., Difede, J., King, D.W., 2007. Confirmatory factor analysis of the PTSD Checklist and the Clinician-Administered PTSD Scale in disaster workers exposed to the World Trade Center Ground Zero. J. Abnorm. Psychol. 116, 329. Post, L.M., Zoellner, L.A., Youngstrom, E., Feeny, N.C., 2012. Understanding the relationship between co-occurring PTSD and MDD: symptom severity and affect. J. Anxiety Disord. 25, 1123–1130. Qualtrics, 2015. Qualtrics. Utah, Provo. Raftery, A.E., 1995. Bayesian model selection in social research. Sociol. Methodol. 25, 111–163. 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

24