Upload
others
View
0
Download
0
Embed Size (px)
Citation preview
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 [email protected].
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