1

Running Head. Latent Classes of PTSD and Depression

Heterogeneity in Patterns of DSM-5 Posttraumatic Stress Disorder and Depression

Symptoms: Latent Profile Analyses.

Ateka A. Contractor

Department of Psychology, University of North Texas, Denton, USA.

Michelle E. Roley-Roberts

Department of Psychiatry, The Ohio State University Wexner Medical Center,

Columbus, USA.

Susan Lagdon

School of Nursing & Midwifery, Queens University, Belfast, Northern Ireland.

Cherie Armour

Psychology Research Institute, Ulster University, Coleraine, Northern Ireland.

Author Notes: Correspondence about this paper can be addressed to Ateka A. Contractor,

Ph.D. University of North Texas, Department of Psychology, 369 Terrill Hall, Denton,

TX 76203; 940-369-8228. E-mail correspondence may be sent to ateka.c@gmail.com.

This research did not receive any specific grant from funding agencies in the public,

commercial, or not-for-profit sectors.

Word count: 4,999 (excluding abstract, references and tables)

2

Abstract

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 anayzed using a 3-step appoach

(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 treatement

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.

Keywords. DSM-5; PTSD; depression; latent profile analyses; covariates.

3

Introduction

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 meta-analytical 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

individualized treatment planning. A person-centered approach to discern such subgroup

categorizations is recommended (Galatzer-Levy and Bryant, 2013). Thus, we assessed

heterogeneity in PTSD-depression 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 (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;

4

Strander et al., 2014). An empirically supported latent-level explanation, the quantitative

hierarchical model (Watson, 2005, 2009), proposes that a higher-order factor of emotional

disorders encompasses subclasses of 1) bipolar disorders 2) distress disorders (e.g., MDD,

PTSD, generalized anxiety disorder), and 3) fear disorders. Thus, PTSD and depression are

conceptualized as sharing non-specific negative affect (Watson et al., 2011), mainly represented

by PTSD’s dysphoria symptoms (Contractor et al., 2014; Elhai et al., 2015). Consequently, there

is a debate regarding 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

5

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

6

extent known to us, no study has analyzed data from a student sample to assess PTSD-depression

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 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.

7

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).

Method

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 minutes), 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

8

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.

Measures

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 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

9

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 test-retest 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.

10

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 & 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.

Exclusions and Missing Data

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 the majority were females (n = 213, 81.30%). Detailed information on

demographics is provided in Table 2.

Data Analysis

11

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;

Nylund et al., 2007b). A model with a 10-point 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.

Results

12

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 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 Figure 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 (2,265) =

439.40, p < .001, partial eta2 = .77; and total depression severity, F (2,265) = 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

item-level severity and lower depression item-level severity. Thus Class 1 is labelled as a lower

13

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 PTSD-higher depression class compared to the higher PTSD-lower

depression class by 4%.

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

14

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%.

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 (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. Methological 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).

15

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 co-occurence 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 sequalae 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 expereince of PTE, participants could have prominent

PTSD or depression symptoms only. Such PTSD and depression symptom heterogenity 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 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)

16

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

expereinces 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).

17

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

distiguinshing of latent classes with promiment 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 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

commnly 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).

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

non-specific negative affect underlies co-occuring PTSD and depression symptoms (Contractor

18

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 seperation of the disorders

in diagnostic manuals (i.e., DSM-5/ICD). It may be beneficial to use specific treatment protocols

to match the distinct symptom patterns.

The psychopathology construct of dissociation is most distingushing 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.

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

19

did not assess 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

reommended 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 dissocative 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. Addtional

reseach on the nature and prevalence of the depressive subtype of PTSD is needed given the

extensive comorbidty across these two disorders.

20

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). Alcoholism: Clinical and Experimental Research. 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

Research 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? Psychological Injury and

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. Journal of Nervous &

Mental Disease 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? Journal of Anxiety

Disorders 28, 418-426.

Asparouhov, T., Muthén, B., 2014. Auxiliary variables in mixture modeling: Three-step

approaches using Mplus. Structural Equation Modeling: A Multidisciplinary Journal 21, 1-13.

21

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.

Journal of Affective Disorders 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 and Alcohol Review 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. Journal of Traumatic 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.

Journal of Traumatic Stress 28, 489-498.

Blevins, C.A., Weathers, F.W., Witte, T.K., 2014. Dissociation and Posttraumatic Stress

Disorder: A Latent Profile Analysis. Journal of Traumatic 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.

Psychological Assessment Advance online publication.

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. The British Journal of 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

22

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 Abuse 10, 198-210.

Briere, J., Scott, C., Weathers, F., 2005. Peritraumatic and persistent dissociation in the

presumed etiology of PTSD. American Journal of 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. Journal of

Affective Disorders 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. Journal of Consulting and Clinical Psychology 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 Research 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. Journal of Psychiatric Research 68, 19-26.

23

Dalenberg, C.J., Glaser, D., Alhassoon, O.M., 2012. Statistical support for subtypes in

posttraumatic stress disorder: The how and why of subtype analysis. Depression and 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 against 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. Educational

and Psychological Measurement. 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 DSM-5 PTSD and

Major Depression symptoms in military soldiers. Journal of Affective Disorders 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. Journal of Anxiety Disorders 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. . Military

Medicine 172, 359-363.

Flory, J.D., Yehuda, R., 2015. Comorbidity between post-traumatic stress disorder and major

depressive disorder: alternative explanations and treatment considerations. Dialogues in Clinical

Neuroscience 17, 141-150.

Galatzer-Levy, I.R., Bryant, R.A., 2013. 636,120 Ways to Have Posttraumatic Stress Disorder.

Perspectives on Psychological Science 8, 651-662.

24

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 against 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. Journal of Traumatic 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. Journal of Anxiety Disorders 28, 644-649.

Kleim, B., Ehlers, A., 2008. Reduced autobiographical memory specificity predicts depression

and posttraumatic stress disorder after recent trauma. Journal of Consulting and Clinical

Psychology 76, 231-242.

Kroenke, K., Spitzer, R.L., 2002. The PHQ-9: a new depression diagnostic and severity measure.

Psychiatric Annals 32, 509-515.

Kroenke, K., Spitzer, R.L., Williams, J.B.W., 2001. The PHQ 9 validity of a brief depression

severity measure. Journal of General Internal Medicine 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. Journal of Child and Family Studies 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.

25

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. Journal of Studies on Alcohol and 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. Journal

of Abnormal Psychology 113, 636-645.

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? The Journal of Clinical 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 and Alcohol Dependence 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., pp. 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. Journal of Traumatic Stress 24, 390-398.

26

Nugent, N.R., Koenen, K.C., Bradley, B., 2012. Heterogeneity of posttraumatic stress symptoms

in a highly traumatized low income, urban, African American sample. Journal of Psychiatric

Research 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. Structural

Equation Modeling. 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 Development 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. Journal of abnormal psychology 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. Journal of Anxiety

Disorders 25, 1123-1130.

Qualtrics, 2015. Qualtrics, Provo, Utah.

Raftery, A.E., 1995. Bayesian model selection in social research. Sociological Methodology 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 meta‐analysis. .

Journal of Traumatic Stress 26, 299-309.

27

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. Journal of

Anxiety Disorders 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. Clinical Psychology Review 34, 87-98.

Vermunt, J.K., 2010. Latent class modeling with covariates: Two improved three-step

approaches. Political Analysis 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.

Military Psychology 28, 104-114.

Watson, D., 2005. Rethinking the mood and anxiety disorders: A quantitative hierarchical model

for DSM-V. Journal of Abnormal Psychology 114, 522-536.

Watson, D., 2009. Differentiating the mood and anxiety disorders: A quadripartite model. The

Annual Review of Clinical Psychology 5, 221-247.

Watson, D., Clark, L.A., Stasik, S.M., 2011. Emotions and the emotional disorders: A

quantitative hierarchical perspective. International Journal of Clinical and Health Psychology 11,

429-442.

28

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 and Alcohol Dependence 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. Psychological Bulletin 133,

122-148.

29

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) 119 female sexual

assault survivors.

3-class Severity Low, low-moderate, high-moderate, and severe

symptoms.

Contractor et al.

(2015)*

1266 trauma-exposed

soldiers.

3-class Severity Mild, moderate, and severe classes.

Cao et al.

(2016)

1196 Chinese

earthquake survivors.

4-class Severity and

type

Low symptoms, predominantly depression,

predominantly PTSD, and combined PTSD-

depression

Hruska et al.

(2014)*

356 motor vehicle

accident victims

admitted to a Level-1

trauma center.

4-class Severity Resilient, mild psychopathology, moderate

psychopathology, severe psychopathology.

30

Lai et al. (2015)* 353 children affected

by Hurricane Katrina.

3-class

solution

Severity No disturbance; posttraumatic stress only;

mixed-internalizing (moderate PTSD severity,

clinically significant anxiety and depression).

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

31

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

Full Sample

(n = 268)

Class 1

(n = 63)

Class 2

(n = 70)

Class 3

(n = 135)

Mean (SD)

Age 24.17 (7.50) 22.42 (4.36) 25.31 (7.91) 24.62 (8.31)

PTSD severity 37.50 (18.91) 12.05 (8.55) 59.04 (11.42) 38.21 (8.01)

Depression severity 13.11 (4.67) 12.98 (3.19) 17.67 (3.23) 10.81 (4.14)

Alcohol use 9.46 (6.72) 10.31 (8.11) 10.13 (6.97) 8.78 (5.85)

Distress tolerance 38.41 (12.80) 43.57 (13.58) 32.28 (11.92) 39.39 (11.68)

Depersonalization and derealization 3.88 (5.24) 1.65 (2.23) 8.40 (6.88) 2.47 (3.49)

Gaps in memory and cognition 6,24 (5.06) 4 (3.96) 10.72 (5.48) 4.88 (3.65)

Sensory misperception 2.65 (3.98) 1.14 (2.16) 5.66 (5.64) 1.83 (2.62)

Cognitive-behavioral re-experiencing 4.40 (3.57) 2.69 (2.42) 7.37 (3.86) 3.63 (2.89)

n (%)*

Female 213 (81.30%) 51 (81.0%) 61 (88.40%) 101 (77.70%)

Employed 159 (59.30%) 35 (55.60%) 41 (58.60%) 83 (61.50%)

Ethnic status

White 263 (98.10%) 63 (100%) 68 (97.10%) 132 (97.8%)

32

Asian 4 (1.50%) 0 (0%) 1 (1.40%) 3 (2.20%)

African American 0 (0%) 0 (0%) 0 (0%) 0 (0%)

Mixed ethnicity 1 (.40%) 0 (0%) 1 (1.40%) 0 (0%)

Probable PTSD (>/= 31) 186 (69.4%) 0 (0%) 70 (100%) 116 (85.90%)

Probable depression (>/= 10) 215 (80.20%) 63 (100%) 67 (95.70%) 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.

33

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 25643.475 25851.752 25667.856

2 class 24216.325 24532.331 24253.317 .95 1478.337 (p = .002) p < .001

3 class 23427.335 23851.072 23476.939 .94 843.958 (p = .12) p < .001

4 class 23040.495 23571.961 23102.710 .95 444.192 (p= .17) p < .001

34

Table 4. Results of multinomial logistic regression analyses.

Class 1 vs. 3# Class 2 vs. 3# Class 2 vs. 1#

OR (95% CI)

Alcohol use 1.06 (1.00 - 1.13) 1.01 (.93 – 1.09) .95 (.87 - 1.04)

Distress tolerance 1.04 (1.00 - 1.07)** .93 (.89 – 98)** .90 (.85 - .95)*

Depersonalization and derealization .89 (.74 - 1.08) 1.11 (.97 – 1.27) 1.24 (1.02 – 1.52)**

Gaps in memory and cognition .99 (.84 - 1.17) 1.24 (1.06 – 1.47)** 1.26 (1.00 – 1.59) p

= .05

Sensory misperception .75 (.53 - 1.07) 1.07 (.87 – 1.31) 1.42 (.96 – 2.08)

Cognitive-behavioral re-experiencing .98 (.80 – 1.20) .95 (.79 – 1.15) .98 (.76 – 1.24)

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.

35

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

PCL_1

PCL_3

PCL_5

PCL_7

PCL_9

PCL_11

PCL_13

PCL_15

PCL_17

PCL_19PHQ_1PHQ_3PHQ_5PHQ_7PHQ_9

0

0.5

1

1.5

2

2.5

3

3.5

4Class 1 (Low PTSD-High Depression)Class 2 (High Overall)Class 3 (High PTSD-Low Depression)

Indicators

Mea

n Sc

ores