Typologies of PTSD clusters and reckless/self-destructive behaviors: A latent profile analysis
Introduction
Although most individuals experience traumatizing events (TEs; 89.70% of U.S. adults), few develop lifetime posttraumatic stress disorder (PTSD; 8.30%; Kilpatrick et al., 2013). PTSD is characterized by a constellation of intrusion, avoidance of trauma reminders, negative alternations in cognitions and mood (NACM), and alterations in arousal and reactivity (AAR) symptoms (American Psychiatric Association, 2013). PTSD typologies (i.e., latent distinct subgroups of individuals based on endorsed response patterns) have been identified with person-centered approaches of latent class (LCA; categorical indicators) and latent profile (LPA; continuous indicators) analyses (Dalenberg et al., 2012, Contractor et al., 2017b). Rarely explored are typologies characterized by PTSD and reckless and self-destructive behaviors (RSDBs) despite their established theoretical and empirical relations. Thus, the current study examined the nature and construct validity of the best-fitting class solution in categorizing participants based on PTSD symptom and RSDB endorsement.
Theoretically, the disinhibition viewpoint suggests that individuals with PTSD symptoms may engage in RSDBs when perceiving rewarding situations (Casada and Roache, 2005); the negative/positive reinforcement viewpoint highlights the functional role of RSDBs in reducing/escaping negative affect and/or eliciting/maintaining/increasing positive affect (Baker et al., 2004, Simpson et al., 2014); and the cognitive explanation indicates that a narrow attention span and information processing capacity post-TE increases RSDBs (Ben-Zur and Zeidner, 2009). Unsurprisingly, empirical evidence indicates a strong link between PTSD and diverse RSDBs such as substance use (Strom et al., 2012, Weiss et al., 2018), gambling (Scherrer et al., 2007), problematic technology use (Contractor et al., 2017d, Schimmenti et al., 2017), disordered eating (Forman-Hoffman et al., 2012), risky sexual behaviors (Weiss et al., 2013), aggression (Lusk et al., 2017), illegal behaviors (Pat-Horenczyk et al., 2007), reckless driving (Stecklov and Goldstein, 2004), and suicidal behaviors (Pat-Horenczyk et al., 2007, Weiss et al., 2015a).
To account for the role and salience of RSDBs, a new E2 symptom was added to the DSM-5 criteria for PTSD (American Psychiatric Association, 2013). Existing research on PTSD's externalizing and impulsive typologies provides preliminary evidence for a reckless behaviors subtype of PTSD. For instance, subgroups with greater PTSD severity and impulsivity (personality trait related to RSDBs; Stanford et al., 1996) report greater RSDBs such as substance use and aggression, distinct mental health outcomes, differing coping styles, and unique genetic risk factors compared to subgroups reporting greater/lower PTSD severity and lower impulsivity (Miller, 2003, Miller et al., 2004, Flood et al., 2010, Wolf et al., 2010, Castillo et al., 2014, Contractor et al., 2016a, Contractor et al., 2018c). Additionally, research indicates that individuals with comorbid PTSD and RSDBs report worse treatment outcomes (Ouimette et al., 1998) and more severe mental health correlates than individuals with only PTSD or RSDBs (Saladin et al., 1995). Overall, comorbid PTSD and RSDBs is associated with worse psychosocial functional impairment and poorer treatment outcomes (Tull et al., 2015).
However, the existence and nature of a reckless behaviors subtype of PTSD (individuals endorsing higher PTSD severity and more RSDBs compared to other subgroups) has been explored in a limited manner. First, person-centered research has focused on singular constructs of either PTSD (e.g., Breslau et al., 2005, Dickstein et al., 2010) or RSDBs (e.g., Mueller et al., 2010, Carragher and McWilliams, 2011, Connor et al., 2014, Elhai and Contractor, 2018). Second, PTSD's externalizing and impulsive subtypes (Miller et al., 2004, Flood et al., 2010, Wolf et al., 2010, Contractor et al., 2016a, Contractor et al., 2018c) are determined based on endorsed PTSD symptoms and personality traits (related to RSBDs), rather than based on engagement in RSDBs. Hence, these personality-based typologies have limited conceptual overlap with a reckless behaviors subtype of PTSD; however support its investigation. Third, limited research has examined PTSD in combination with RSDBs, and these studies have focused on specific and singular RSDBs vs. considering the impact of several RSDBs simultaneously (e.g., Anderson et al., 2017). However, RSDBs often co-occur; engagement in one RSDB increases the likelihood of other RSDBs (White and Hansell, 1998, Cooper, 2002). Fourth, limited research has examined relations between PTSD clusters and different RSDBs, which may influence the categorization and heterogeneity (and thereby utility and meaning) of PTSD-RSDB typologies. For example, research indicates that PTSD's intrusion symptoms relate to substances such as depressants and cannabis (Khoury et al., 2010, Avant et al., 2011, Contractor et al., 2016b), aggressive behaviors (Hellmuth et al., 2012), and non-suicidal self-injury (NSSI; Weierich and Nock, 2008); PTSD's numbing/avoidance symptoms relate to substances such as depressants, heroine, and sedatives (Tull et al., 2010, Avant et al., 2011, Dworkin et al., 2018) and NSSI (Weierich and Nock, 2008); PTSD's NACM symptoms associate with suicidal ideation/attempts (Pietrzak et al., 2015, Guina et al., 2017), problematic smartphone use (Contractor et al., 2017a), and NSSI (Weierich and Nock, 2008); and PTSD's arousal symptoms associate with aggression (Hellmuth et al., 2012), suicidal ideation/attempts (Pietrzak et al., 2015), and risky sexual behaviors (Weiss et al., 2013). Thus, heterogeneity in PTSD clusters needs to be considered in relation to different RSDBs.
Addressing the aforementioned limitations, we examined the heterogeneity in PTSD and RSDBs. We identified the nature and construct validity of the best-fitting latent class solution in categorizing participants based on endorsed PTSD symptoms and diverse RSDBs. Based on past research (e.g., Breslau et al., 2005, Dickstein et al., 2010, Contractor et al., 2017b, Contractor et al., 2018c), we hypothesized finding an optimal three- or four-class solution (lower PTSD severity and RSDBs; higher PTSD severity and RSDBs; primarily PTSD severity and/or RSDBs). Further, we examined the covariates of age, gender, depression severity, number of trauma types, and RSDB-related functional impairment to establish the construct validity of the optimal class solution. Consistent with extant research, we hypothesized that greater PTSD severity classes would have more females (Tolin and Breslau, 2007), and greater RSDB classes would be younger (Jonah, 1990, Chen and Kandel, 1995, Kessler et al., 2005) and have more males (Brady and Randall, 1999, Nolen-Hoeksema, 2004, Pat-Horenczyk et al., 2007). Additionally, we hypothesized that classes characterized by greater PTSD severity (e.g., Contractor et al., 2018a, Contractor et al., 2018b) and more RSDBs (Davis et al., 2002, Briere et al., 2010) would report more trauma types. Further, we expected that classes characterized by greater PTSD severity (Rytwinski et al., 2013, Bonde et al., 2016) and more RSDBs (Becona et al., 1996, Swendsen and Merikangas, 2000) would report greater depression. Finally, we hypothesized that classes with greater PTSD severity and more RSDBs would report greater impairment (Read et al., 2004, Olino et al., 2012).
Section snippets
Procedure/participants
We recruited participants through Amazon's Mechanical Turk (MTurk) platform. We described the study as a 45–60 min survey to develop a novel measure assessing posttrauma risky behaviors. Inclusionary criteria included being 18 years or older, residing in North America, having a working knowledge of English, and experiencing a traumatic event (Primary Care PTSD Screen; Prins et al., 2015). Eligible participants who provided informed consent and completed the survey hosted on Qualtrics were
Results
PCL-5 intrusions, avoidance, NACM, and AAR cluster scores averaged 6.67 (range of 20), 2.99 (range of 8), 8.48 (range of 28), and 2.13 (range of 16) respectively. Table 3 provides the LPA results. There were model convergence problems with the four-class solution attributed to the small sample size of one of the classes (only 13 observations). Overall, LPA fit indices indicated the three-class solution as the optimal model based on the available results. The LMR p value being 0.03 for the
Discussion
The current study examined the heterogeneity in patterns of PTSD symptoms and RSDBs. Study results provided support for three classes characterized by relatively (1) lower PTSD severity and RSDB frequency (Class 1), (2) higher PTSD severity and lower RSDB frequency (Class 2), and (3) higher PTSD and RSDB frequency (Class 3). Our findings thus offer preliminary evidence for a reckless behaviors subtype of PTSD (Class 3) consistent with some existing research (Guina et al., 2016). Our results add
Conflicts of interest
Ateka Contractor and Nicole Weiss declare that they have no conflict of interest.
Acknowledgements
We thank Ms. Jackeline Marquez and Ms. Sara Koh for reviewing the literature on trauma, PTSD, and risky behaviors to aid in the development of the Posttrauma Risky Behaviors Questionnaire. We thank Drs. Jon D. Elhai, Tami P. Sullivan, Matthew T. Tull, Lily A. Brown, and Melanie S. Harned for their expert feedback at several stages of the development of the Posttrauma Risky Behaviors Questionnaire.
Funding
The research described here was supported, in part, by grants from the National Institute on Drug Abuse (K23DA039327 and L30DA038349) awarded to the second author.
References (133)
- et al.
Conditional risk for PTSD among Latinos: a systematic review of racial/ethnic differences and sociocultural explanations
Clin. Psychol. Rev.
(2013) - et al.
Dimensional structure of DSM-5 posttraumatic stress symptoms: Support for a hybrid Anhedonia and externalizing behaviors model
J. Psychiatr. Res.
(2015) - et al.
Gender differences in substance use disorders
Psychiatr. Clin. North Am.
(1999) - et al.
A latent class analysis of DSM-IV criteria for pathological gambling: results from the national epidemiologic survey on alcohol and related conditions
Psychiatry Res.
(2011) - et al.
Latent profiles of DSM-5 PTSD symptoms and the "Big Five" personality traits
J. Anxiety Disord.
(2016) - et al.
Relation between lifespan polytrauma typologies and post-trauma mental health
Compr. Psychiatry
(2018) - et al.
Examination of the heterogeneity in PTSD and impulsivity facets: a latent profile analysis
Personal. Individ. Differ.
(2018) - et al.
Latent profile analyses of posttraumatic stress disorder, depression and generalized anxiety disorder symptoms in trauma-exposed soldiers
J. Psychiatr. Res.
(2015) - et al.
Latent-level relations between DSM-5 PTSD symptom clusters and problematic smartphone use
Comput. Hum. Behav.
(2017) - et al.
The moderating role of dysphoria in the relationship between intrusions and alcohol use
Addict. Behav.
(2016)