Early Predictors of Chronic PTSD Symptom Trajectories in US Army Soldiers Deployed to the Iraq War

doi:10.21203/rs.3.rs-2244497/v1

Download PDF

Research Article

Early Predictors of Chronic PTSD Symptom Trajectories in US Army Soldiers Deployed to the Iraq War

https://doi.org/10.21203/rs.3.rs-2244497/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Purpose

The course of posttraumatic stress disorder (PTSD) symptoms varies among veterans of warzones, but sources of variation in long-term course remain poorly understood. Modeling of symptom growth trajectories facilitates understanding predictors of individual outcomes over time. Although growth mixture modeling (GMM) has been applied to military populations, few studies have incorporated both pre-deployment and follow-up measurements over an extended time.

Methods

In this prospective study, 1087 US Army soldiers with varying military occupational specialties and geographic locations were assessed before and after deployment to the Iraq War, with long-term follow-up assessment occurring at least five years after return from deployment. The primary outcome variable was the PTSD Checklist, civilian version, summary score.

Results

GMM yielded four latent profiles: (1) primarily asymptomatic (n = 194; 17·8%), (2) post-deployment worsening symptoms (n = 84; 7·7%), (3) mild symptoms (n = 320; 29·4%), and (4) pre-existing, with chronic post-deployment elevation of symptoms (n = 489; 45·0%). Regression models comparing the primarily asymptomatic class to symptomatic classes revealed that chronic symptom classes were associated with greater stress exposure, less pre-deployment social support, identifying as a racial/ethnic minority, military reservist or veteran status at most recent assessment, and poorer pre-deployment visual memory.

Conclusions

PTSD symptom course varies considerably over time after military deployment and is associated with potentially modifiable biopsychosocial factors occurring early in its course, in addition to exposures and military status.

Trial Registration

: This study reported in this article was formally preregistered September 9, 2008 with ClinicalTrials.gov (https://clinicaltrials.gov/ct2/show/NCT00748995).

PTSD

warzone

biopsychosocial risk

trajectories

prospective

Warzone service often involves exposure to psychologically traumatic events, leading to elevated risk of posttraumatic stress disorder (PTSD) symptoms among deployed military personnel. Yet not all veterans of warzones develop PTSD symptoms and, among those who do, the course of symptoms varies considerably [1–3]. Prevention of chronic PTSD symptoms among warzone veterans requires better prediction of the varying trajectories of such symptoms.

Statistical analyses that aggregate longitudinal data across participants can fail to detect meaningful variation in PTSD symptom course. Growth mixture modeling (GMM) has been used to identify individual PTSD symptom trajectories in both civilian and military populations. Within military populations, however, few studies using GMM have incorporated prospective measurement starting prior to deployment, and among those that have, fewer have followed warzone veterans over sufficient time to capture longer-term outcomes. A systematic review [4] found that studies without prospective data consistently underestimated resilience and that among military samples, veterans were less likely than military personnel continuing to serve to be categorized as “resilient.” These findings highlight the value of both prospectively assessing service members prior to deployment and following them long enough to understand their outcomes after separation from service.

Results have varied among prospective military studies following study populations over longer periods, although most have identified primarily “resilient” trajectories among their samples. For example, 85% of a Dutch military sample assessed for PTSD symptoms prospectively over a 10-year period were classified as resilient, with early factors such as younger age, lower rank, more deployment stressors, and less social support predicting more symptomatic trajectories [2]. Similarly, most (90.2%) of a sample of deployed British military personnel followed over a 14-year period were classified as resilient [3]. In a sample of US service members enrolled in the Millennium Cohort Study, 90% of non-combat exposed vs. 80% of combat exposed US service members remained resilient over a 10-year period, highlighting the increased risk among those with combat exposure [1].

In a sample of primarily deployment naïve US Army Soldiers, we applied latent GMM to prospective data collected beginning prior to deployment to the Iraq War and extending to a long-term follow-up assessment conducted over five years following return from deployment to identify: (1) common PTSD symptom courses, and (2) predictors of symptom course, including stress exposure, social resources, demographic and military factors, and neurocognitive resources – a domain not typically included in similar studies but of high theoretical relevance as a biologically-relevant predictor [5, 6]. We hypothesized that in regression models adjusted for rank and ethnicity, as compared with trajectories characterized by a) pre-existing and chronically elevated post-deployment symptoms or b) with post-deployment worsening symptoms, greater social and neurocognitive resources at pre- and/or post-deployment assessments, higher perceived preparedness, as well as lower levels of pre-military and deployment stress exposures would be associated with a course characterized by mild but sustained systems. We further hypothesized that continued military service at long-term follow-up would be associated with a primarily asymptomatic course as compared with symptomatic courses.

Study design and participants

This prospective, observational cohort study of 1087 men and women who deployed to Iraq at least once between 2003 and 2005 incorporated longitudinal data collection starting before an index Operation Iraqi Freedom (OIF) deployment as part of the Neurocognition Deployment Health Study (NDHS) [7] and extending through its most recent assessment phase, Department of Veterans Affairs (VA) Cooperative Studies Program Study #566 (CSP#566) [8]. Archived pre-deployment, post-deployment, and short-term follow-up (1–3 years post-deployment) data were enriched with CSP#566 long-term (≥ 5 years post-deployment) follow-up data.

Initial NDHS sampling entailed cluster-based sampling of regular active duty and activated reservist battalion-level US Army units [9]. Reflecting occupational and geographic dispersion of NDHS cohort members to other military unit assignments and non-military careers, at long-term follow-up (CSP#566) we sampled participants individually using stratified random sampling based on the initial battalion-level strata to create seven recruitment waves chronologically spaced over five years [8]. Only those NDHS and CSP#566 participants who a) completed a pre-deployment assessment within 180 days of the initiation of their index OIF deployment, b) deployed > 30 days in support of OIF at least once after their pre-deployment assessment, c) completed at least one assessment following their index deployment, and d) provided valid self-report (indicated by the absence of uniform endorsement of unidirectional extreme responses on measures in which the pathological response varied in direction) were included in the analytic sample. (See Fig. 1 for sample derivation).

Data Sources

Pre- and post-deployment and short-term follow-up assessments were conducted at military installations. Long-term follow-up data were derived from written mail surveys. Deployment information and duty status (active duty vs. military veteran or reservist) were verified with Defense Manpower Data Center (DMDC) administrative records. Participants self-reported demographic information.

Measures

PTSD symptom severity. The PTSD Checklist-Civilian version (PCL-C) [10], yielding a summary score of 17–85, was administered at all time-points as a self-report measure of Diagnostic and Statistical Manual, 4th Edition text revision (DSM-IV-TR) [11] PTSD symptom severity. A 5–10 point change is considered reliable; a 10–20 point change is considered clinically significant [12]. The PCL-C, as a written survey instrument, was suitable to field research conducted at Army installations, reflected the current DSM version at the outset of the study, has strong psychometric properties [13], and at long-term follow-up was highly correlated (r = 0.82, P < .0001) with the summary score of the gold standard Clinician Administered PTSD Scale [14].

Risk and resilience factors. We considered variables with conceptual relevance to, or prior evidence of association with, PTSD symptom severity for potential inclusion in regression models. Domains included demographic factors, military factors, neurocognitive resources, social resources, and stress exposures. (See the Supplementary Information for details.) Measures were derived from self-report survey, DMDC records, the Deployment Risk and Resilience Inventory (DRRI) [15], and objective Wechsler Memory Scale (WMS) [16] performance.

Statistical analyses

Data preparation. Item-level missing values for PCL-C were substituted with the corresponding symptom cluster mean score (< 4% of participants) for PCL-C assessments missing: (1) < 6 items overall, and (2) < 50% of items in each symptom cluster. A total of 537 participants (49.4%) completed two PCL-C assessments, 423 (38.9%) completed three PCL-C assessments and 127 (11.7%) completed four PCL-C assessments. To establish trajectory classes, missing PCL-C assessments among participants with at least one pre- and post-deployment PCL-C assessment were handled using full information maximum likelihood robust estimation, which accommodates non-normal data distribution and assumes data are missing at random. To examine change over time in PTSD symptom severity, PCL-C scores were centered around the pre-deployment PCL-C sample mean value; interpretations for change were equivalent to deviations from this pre-deployment mean.

Latent trajectories. GMM was conducted using Mplus version 8 [17] following established recommendations [18, 19]. The time axis represented the span of time from assessments to the index deployment. Time = 0 reflected the midpoint of each participant’s deployment; negative and positive values represented the number of months an assessment was conducted prior to, or following, deployment, respectively.

We used latent growth curve analysis to estimate candidate models of average PTSD symptoms trajectory from pre-deployment to long-term follow-up: no change, linear change, quadratic change, and spline models comprising two linear slopes joined at a knot point defined as the participant’s index deployment midpoint. All models included fixed and random effects of intercept and slopes and were compared with likelihood ratio tests [20].

Starting with the best-fitting one-class trajectory model, we increased the number of classes incrementally by one until models were no longer well-identified. Trajectory classes differed in mean estimates of PCL-C intercepts and slopes. For each class, we also tested four configurations of the variance-covariance matrix for the intercepts and slopes [21]. In doing so, we began by testing the least restrictive variance-covariance configuration, allowing parameters to covary. When a configuration produced uninterpretable estimates for one or more classes within a model, those parameters were subsequently constrained to zero for testing in subsequent model configurations. We did not consider models with zero variance-covariance configurations, because this assumes no differences in intercept or slope variances across the classes, which can lead to inaccurate representations of the data [22]. Models were evaluated based on numerous fit indices and classification diagnostics.

Prediction of trajectory class. As described in the Supplementary Information, to minimize chance findings, one to two variables from each conceptual domain were selected for model inclusion based on Wald tests, conceptual salience, and completeness of data in the absence of significant associations. The final set of covariates included self-reported race/ethnicity, military duty status at most recent PCL-C administration, pre-deployment visual learning (WMS Visual Reproduction test), pre-deployment unit support (DRRI Unit Support scale), cumulative stressful life events prior to deployment (DRRI Life Events scale), and index deployment warzone events (DRRI combined Combat Experiences and Post-Battle scales). We used multinomial logistic regression to examine associations between the final set of covariates and trajectory class membership, comparing the primarily asymptomatic class with all other classes.

Sample characteristics

At pre-deployment, most participants were male (93.0%), White/Caucasian (61.9%), regular active duty (87.5%), junior enlisted rank (69.8%), and had no prior deployments (74.5%). See Table 1 for all sample characteristics.

Table 1. Sample characteristics (N=1087)

Characteristic	Mean ± SD or No. (%)
Demographics (pre-deployment)
Age, y	25.8 (5.9)
Gender
Male	1011 (93.0%)
Female	76 (7.0%)
Race/ethnicity
White, non-Hispanic	673 (61.9%)
Black, non-Hispanic	156 (14.4%)
Asian, non-Hispanic	35 (3.2%)
Hispanic	127 (11.7%)
Other	95 (8.7%)
(missing)	1 (0.1%)
Military variables
Duty status, pre-deployment
Regular Active Duty	951 (87.5%)
Activated Reservist	136 (12.5%)
Rank, pre-deployment
Non-commissioned officers	286 (26.3%)
Junior enlisted	759 (69.8%)
Officers (commissioned or warrant)	31 (2.9%)
(missing)	11 (1.0%)
Perceived preparedness, pre-deployment
Yes, prepared	729 (67.1%)
No, not prepared	356 (32.7%)
(missing)	2 (0.2%)
Military occupation type, pre-deployment
Combat arms	529 (48.6%)
Combat support	173 (15.9%)
Service support	379 (34(9%)
(missing)	6 (0.6%)
Prior operational deployment, pre-deployment
Yes	32 (2.9%)
No	810 (74.5%)
(missing)	245 (22.6%)
Length of military service, pre-deployment, y	4.6 (4.8)
Duty status at last PCL-C assessment
Regular Active Duty	486 (44.7%)
Reservist	271 (24.9%)
Military Veteran	310 (28.5%)
(missing)	20 (1.9%)
No. deployments subsequent to index deployment (at last PCL-C assessment)	1.6 (0.8)
Neurocognitive resources
Native intellectual potential (pre-deployment NES3 Vocabulary, summary score)	17.0 (4.2)
Pre-deployment visual learning (WMS Visual Reproductions, immediate recall, summary score)	7.6 (2.4)
Pre-deployment visual memory (WMS Visual Reproductions, delayed recall, summary score)	7.7 (2.4)
Change in visual learning after deployment (post-deployment – pre-deployment WMS Visual Reproductions immediate recall summary scores)	0.1 (2.1)
Change in visual memory after deployment (post-deployment – pre-deployment WMS Visual Reproductions, delayed recall, summary scores)	0.0 (2.2)
Social resources
Pre-deployment unit support, DRRI Unit Support scale, summary score	37.4 (10.9)
Deployment unit support, DRRI Unit Support scale, summary score	38.4 (11.3)
Post-deployment general social support, DRRI Social Support scale, summary score	56.7 (9.9)
Change in marital status (at any assessment)
Married-to-Single	55 (5.1%)
Single-to-Married	86 (7.9%)
No-Change	808 (74.3%)
(missing)	138 (12.7%)
Stress Exposures
Pre-military stressful life events, DRRI Life Events scale, summary score	4.5 (3.4)
Perceived war-zone threat, DRRI Deployment Concerns scale, summary score	45.8 (10.6)
Stressful war-zone events, DRRI combined Combat Experiences and Post-battle scales, summary score	17.2 (7.7)
Deployment home front concerns, DRRI Life and Family Concerns summary score	25.0 (7.6)

Due to missing data, the sample size varies across measures.

Abbreviations: PCL-C=PTSD Checklist, civilian version. NES3=Neurobehavioral Evaluation System, 3^rd revision. WMS=Wechsler Memory Scale. DRRI=Deployment Risk and Resilience Inventory.

PTSD symptom trajectories

The spline model that included random intercept, and random pre- and post-deployment slope parameters provided the best fit as the unconditional one-class model. In latent class growth mixture models, overall improvement of model fit indices with higher number of classes suggests the presence of trajectory subgroups (Table 2). A four-class model with class-specific variance-covariance matrices for two subgroups (Classes 2 and 4) and zero variance and covariance for two subgroups (Classes 1 and 3) consistently performed well across fit indices. We selected this model as the final unconditional latent profile model because it was theoretically grounded and produced classes with relatively balanced size memberships to provide stable estimates.

Table 2. Fit indices across best-fitting latent growth mixture models of PCL-C severity score

Model^a	npar	n-adj BIC	CAIC	AWE	Bayes Factor	Entropy	Smallest Class Size
1 class	10	22544.50	22586.27	22676.18	2.95e-179	-	-
2 class	21	21687.40	21775.10	21963.91	4.48e-53	0.76	39%
3 class	23	21439.97	21536.03^b	21742.82^b	0.24^b	0.73	17%
4 class	31	21411.72^b	21541.18^b	21819.90^b	1297.90^b	0.72	8%
5 class	36	21410.17^b	21560.51	21884.20	-	0.75	2%

^aFor the 3, 4, and 5-class models, the least restrictive variance-covariance configuration produced uninterpretable estimates. Consequently, for the 3-class model, the covariances of all classes were constrained to zero; for the 4-class model, the variances and covariances of classes 1 and 3 were constrained to zero; for the 5-class model, the variances and covariances of classes 2, 4, and 5 were constrained to zero.

^bIndicates the two top-performing classes for each fit index.

Abbreviations: PCL-C=PTSD Checklist, civilian version. n-adj BIC=sample-adjusted Bayesian Information Criteria. CAIC=Consistent Akaike's Information Criterion. AWE=Approximate Weight of Evidence Criterion. BF=Bayes Factor.

The four-class latent profiles are presented in Figure 2: (1) primarily asymptomatic (“asymptomatic”) (n=194; 17.8%), (2) post-deployment worsening symptoms (“post-deployment worsening”) (n=84; 7.7%), (3) mild symptoms (“mild”) (n=320; 29.4%), and (4) pre-existing, chronically elevated post-deployment symptoms (“high, chronic”) (n=489; 45.0%). In sum, over half of the sample (52.7%) belonged to classes indicating poor long-term outcomes.

Associations between covariates and trajectory class

As shown in Table 3, higher odds of belonging to the high, chronic symptom class, relative to the asymptomatic class, were associated with reservist or veteran status (vs. active duty) at most recent PCL-C assessment, lower pre-deployment Visual Reproduction performance, lower pre-deployment unit support, more cumulative pre-deployment stressful life events, and more index deployment warzone events. Higher odds of belonging to the mild symptom class, relative to the asymptomatic class, were associated with identifying as non-Hispanic White vs. racial/ethnic minority, lower pre-deployment Visual Reproduction performance, lower pre-deployment unit support, more higher pre-deployment stressful life events, and more index deployment warzone events. No covariates significantly distinguished post-deployment worsening vs. asymptomatic classes.

Table 3. Multinomial logistic regression predicting latent trajectory class membership, in asymptomatic class vs. other classes

	Asymptomatic (n=190) vs.
	High, chronic (n=481)		Post-deployment worsening (n=78)		Mild (n=318)
	OR (95% CI)	P-value	OR (95% CI)	P-value	OR (95% CI)	P-value
Race/ethnicity
Racial/ethnic minority	(ref)	--	(ref)	--	(ref)	--
White/non-Hispanic	1.08 (0.69-1.69)	0.75	1.67 (0.80-3.49)	0.17	2.74 (1.60-4.68)	< 0.0001
Duty status, last PCL-C
Active duty	(ref)	--	(ref)	--	(ref)	--
Reservist	1.73 (1.04-2.87)	0.04	1.05 (0.43-2.55)	0.91	1.22 (0.70-2.12)	0.48
Veteran	2.90 (1.67-5.07)	< 0.0001	1.85 (0.84-4.11)	0.13	1.13 (0.60-2.10)	0.71
Pre-deployment VR learning	0.87 (0.81-0.94)	< 0.0001	0.90 (0.81-1.01)	0.07	0.90 (0.83-0.98)	0.02
Pre-deployment unit support	0.95 (0.93-0.96)	< 0.0001	1.00 (0.98-1.02)	0.94	0.98 (0.96-1.00)	0.01
Pre-deployment DRRI Life Events	1.22 (1.14-1.31)	< 0.0001	1.00 (0.88, 1.14)	0.99	1.15 (1.06-1.23)	< 0.0001
DRRI war-zone events	1.12 (1.09-1.15)	< 0.0001	1.02 (0.98-1.07)	0.30	1.04 (1.01-1.07)	0.01

^aCombined score from the DRRI Combat Experiences and Post-Battle Events scales.OR=odds ratio. CI=confidence interval. PCL-C=PTSD Checklist, civilian version. VR=Visual Reproductions. DRRI=Deployment Risk and Resilience Inventory.

In this prospective study of a large sample of US Army Soldiers and military veterans who deployed to the Iraq War and were followed longitudinally for five or more years after their return from deployment, we identified four PTSD symptom trajectories: primarily asymptomatic, post-deployment worsening symptoms, mild (persisting) symptoms, and pre-existing chronically elevated post-deployment symptoms. Although characteristics of the four classes parallel those identified by other studies of military personnel, our results differ from prior literature in that a smaller proportion of our sample (17.8%) was categorized as primarily asymptomatic (or “resilient” as labeled in some other studies).

Previous research has indicated that veterans are more likely to be in relatively symptomatic classes than military personnel continuing service [4]. Thus, the smaller number of primarily asymptomatic participants in our sample may reflect that more of our participants were military veterans at their last PTSD symptom assessment, as compared with other studies. This finding also reflects our previous work showing that symptom increases over time aggregated across individuals were more likely among military veterans relative to current service members [23]. The relatively large proportion of the sample expressing elevated pre-deployment symptoms (45.0%) likely reflects correspondingly high endorsement of pre-military stressful life events and is consistent with recent literature highlighting the potential role of a broad range of life stressors on PTSD symptom development [24].

Relatedly, in examining predictors of class membership, we found in the current analyses that being a reservist or military veteran was associated with higher odds of belonging to the high, chronic symptom class, relative to the asymptomatic class. Other factors differentiating asymptomatic from mild and/or chronically elevated symptomatic trajectories included, as expected from previous literature, stress exposures (both prior to and during deployment), access to social support prior to deployment, and identifying as a racial/ethnic minority. These findings collectively support prior research suggesting a dose-relationship between stress exposure and PTSD symptom severity [2,25], the benefits of social support in managing stress exposures and their consequences [2,26], and that burdens may be placed on members of racial-ethnic minorities that serve to increase stress responses [24,27,28].

Consistent with our previous findings [29,30] and past research highlighting the benefit of visual imagery in reducing anxiety [31], we also found that less proficient visual memory at pre-deployment was associated with greater odds of belonging to a symptomatic trajectory. This finding represents a novel area of study, given that cognitive performance has been rarely examined as a potential predictor of PTSD symptom growth trajectories. Results from a prospective study of 473 US active duty Army personnel stationed at Fort Campbell indicated that pre-deployment inhibitory control [5], sustained attention [5,6], and cognitive flexibility [6] were risk factors for a trajectory characterized by post-deployment development of PTSD. Our findings complement the Fort Campbell study by identifying pre-deployment memory as a predictor of longitudinal expression of PTSD symptoms in a larger sample and across two distinct symptom trajectories (pre-existing, chronically elevated and persistent mild symptoms). Taken together, these findings highlight the importance of cognitive health across multiple domains of cognition.

Strengths and Limitations

We did not find significant associations distinguishing the post-deployment worsening from the primarily asymptomatic class. Our failure to detect significant associations between predictors and the post-deployment worsening trajectory may reflect the relatively small membership of the post-deployment worsening class (n=78). It may also be that our predictors more effectively distinguished chronic symptom courses as compared to the more dynamic symptom course of the post-deployment worsening class. Specifically, we had less predictor data on variables chronologically distal from the deployment (e.g., post-deployment stress exposures, social support) that could be hypothesized to have had greater impact on changes over time.

Other limitations include a predominantly male, White, US Army sample that may not fully generalize to the broader international military population but that was nonetheless reasonably representative of the US Army at the time of study enrollment. Because of the timing of the study and the necessity of reducing respondent burden, we also relied on a validated, DSM-IV-based, self-report measure of PTSD symptom severity. However, the PCL-C has strong psychometric properties [13] and is highly correlated with both CAPS symptom severity [21] and the PCL for DSM-5 [32].

Limitations notwithstanding, this study extends knowledge of PTSD symptom courses and their predictors in warzone deployed military samples. Specifically, we address gaps in previous literature by applying GMM as a method of examining individual symptom trajectories in the context of both prospective measurement and a long-term follow-up that incorporated assessment of a nationally dispersed sample of both participants still serving in the military and those separated from military service. We additionally introduced objective neurocognitive performances as a rarely studied, biologically relevant predictor of PTSD symptom growth trajectories.

Our findings emphasize the potential value of bolstering cognitive and social resources among military personnel who have deployed and attending to their stressful early life or pre-military experiences in addition to potentially traumatic stressors encountered in the warzone. Furthermore, associations between racial/ethnic minority status and military veteran status with poorer PTSD symptom outcomes following deployment suggest that social determinants may play a central role in the long-term course of PTSD symptoms and highlight the need to develop mechanisms to optimize both instrumental and emotional support in addressing hardships associated with social factors.

Acknowledgments: We thank Ms. Patty Crutchfield for her administrative contributions, and Drs. Grant Huang and Terri Gleason for their support of the study.

Author Contributions: All authors contributed to, read, and approved the final report. JK and MA had direct access to and verified the underlying data. JJV, MF, LOL, APK, SPP, BPM, PPS, JC, and MA contributed to conceptualization. JJV, SPP, and JC were responsible for funding acquisition. JJV, SPP, and BM were directly involved in investigation. JJV, SPP, JC, and MA had roles in project administration. MF, LOL, JK, and MA conducted formal analyses. SPP, APK, and JK were responsible for data curation, JJV, LOL, SPP contributed directly to methodology.

Funding: Funding was provided by the Department of Veterans Affairs (VA) Cooperative Studies Program (CSP#566). Support for previous collection of archived data used in this study’s longitudinal analyses was provided by the US Army Medical Research and Development Command (DAMD 17-03-0020) and VA Clinical Sciences Research and Development. The funding agencies provided general guidance and policies for conducting human subjects research. The manuscript underwent administrative review at the US Army Research Institute of Environmental Medicine. The views expressed in this article are those of the authors and do not reflect official policy or position of any aspect of the US government.

Conflicts of Interest: The authors declare that they have no conflicts of interest to disclose.

Ethics Approval: Human subjects protection approvals were obtained from the VA Central Institutional Review Board for CSP#566 and from human subjects review boards of the US Army, Tulane University Health Sciences Center, and affiliated VA facilities for earlier NDHS phases. VA CSP#566 participants provided phone consent; written consent, including permission to be contacted for future research, was obtained for earlier NDHS assessments. The procedures used in this study adhere to the tenets of the Declaration of Helsinki.

Consent to Participate: All participants gave their informed consent to participate prior to inclusion in the study and for additional procedures included in the long-term follow-up assessment.

Consent for Publication: Not applicable.

Donoho CJ, Bonanno GA, Porter B, Kearney L, Powell TM. A decade of war: prospective trajectories of posttraumatic stress disorder symptoms among deployed US military personnel and the influence of combat exposure. Am J Epidemiol. 2017;186(12): 1310-1318. doi: 10.1093/aje/kwx318
van der Wal SJ, Vermetten E, Elbert G. Long-term development of post-traumatic stress symptoms and associated risk factors in military service members deployed to Afghanistan: Results from the PRISMO 10-year follow-up. Eur Psychiatry. 2020;64(1): e10. Published 2020 Dec 21. doi:10.1192/j.eurpsy.2020.113
Palmer L, Thandi G, Norton S, et al. Fourteen-year trajectories of posttraumatic stress disorder (PTSD) symptoms in UK military personnel, and associated risk factors. J Psychiatr Res. 2019;109:156-163. doi:10.1016/j.jpsychires.2018.11.023
Galatzer-Levy IR, Huang SH, Bonanno GA. Trajectories of resilience and dysfunction following potential trauma: A review and statistical evaluation. Clin Psychol Rev. 2018;63:41-55. doi: 10.1016/j.cpr.2018.05.008
Samuelson KW, Newman J, Abu Amara D, et al. Predeployment neurocognitive functioning predicts postdeployment posttraumatic stress in Army personnel. Neuropsychology. 2020;34(3):276-287. doi: 10.1037/neu0000603
Schultebraucks K, Qian M, Abu-Amara D, et al. Pre-deployment risk factors for PTSD in active-duty personnel deployed to Afghanistan: a machine-learning approach for analyzing multivariate predictors. Mol Psychiatry. 2021;26(9):5011-5022. doi:10.1038/s41380-020-0789-2
Vasterling JJ, Proctor SP, Amoroso P, Kane R, Heeren T, White RF. Neuropsychological outcomes of army personnel following deployment to the Iraq war. JAMA. 2006; 296(5):519-529. doi: 10.1001/jama.296.5.519
Aslan M, Concato J, Peduzzi PN, et al. Design of "neuropsychological and mental health outcomes of operation Iraqi freedom: a longitudinal cohort study”. J Investig Med. 2013; 61(3):569-577. doi:10.2310/JIM.0b013e31828407ff
Vasterling JJ, Proctor SP, Amoroso P, et al. The Neurocognition Deployment Health Study: a prospective cohort study of Army Soldiers. Mil Med. 2006;171(3):253-260. doi:10.72015/milmed.171.3.253
Ruggiero KJ, Del Ben K, Scotti JR, Rabalais AE. Psychometric properties of the PTSD Checklist-Civilian Version. J Trauma Stress. 2003;16(5):495-502. doi:10.1023/A:1025714729117
American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders, 4^th ed, text revision. Washington, DC: American Psychiatric Association; 2000.
Lang AJ, Wilkins K, Roy-Byrne PP, et al. Abbreviated PTSD Checklist (PCL) as a guide to clinical response. Gen Hosp Psychiatry. 2012;34(4):332-338. doi: 10.1016/j.genhospsych.2012.02.003
Wilkins KC, Lang AJ, Norman SB. Synthesis of the psychometric properties of the PTSD checklist (PCL) military, civilian, and specific versions. Depress Anxiety. 2011;28(7):596-606. doi:10.1002/da.20837
Blake DD, Weathers FW, Nagy LM, et al. The development of a Clinician-Administered PTSD Scale. J Trauma Stress. 1995;8(1):75-90. doi: 10.1007/BF02105408
Vogt DS, Proctor SP, King DW, King LA, Vasterling JJ. Validation of scales from the Deployment Risk and Resilience Inventory in a sample of Operation Iraqi Freedom veterans. Assessment. 2008;15(4):391-403. doi: 10.1177/1073191108316030
Wechsler, D. A standardized memory scale for clinical use. J Psychol. 1945;19:87-95.
Muthén LK, & Muthén BO. Mplus user’s guide, 8^th ed. Los Angeles, CA: Muthén & Muthén; 1988-2017.
Ram N, Grimm KJ. Growth mixture modeling: a method for identifying differences in longitudinal change among unobserved groups. Int J Behav Dev. 2009;33(6):565-576. doi: 10.1177/0165025409343765
van de Schoot R, Sijbrandij M, Winter SD, Depaoli S, Vermunt, JK. The GRoLTS-checklist: guidelines for reporting on latent trajectory studies. Struct Equ Modeling 2017;24:451-467.
Satorra A, Bentler PM. Ensuring positiveness of the scaled difference chi-square test statistic. Psychometrika. 2010;75(2):243-248. doi: 10.1007/s11336-009-9135-y
Masyn KE. Latent class analysis and finite mixture modeling. In TD Little, ed., The Oxford Handbook of Quantitative Methods in Psychology. New York: Oxford University Press; 2013.
Infurna FJ, Luthar SS. Re-evaluating the notion that resilience is commonplace: a review and distillation of directions for future research, practice, and policy. Clin Psychol Rev. 2018;65:43-56. doi: 10.1016/j.cpr.2018.07.003
Vasterling JJ, Aslan M, Proctor SP, et al. Longitudinal examination of posttraumatic stress disorder as a long-term outcome of Iraq War deployment. Am J Epidemiol. 2016;184(11):796-805. doi: 10.1093/aje/kww151
Gradus JL, Galea S. Reconsidering the definition of trauma. Lancet Psychiatry. 2022;9(8):608-609. doi: 10.1016/S2215-0366(22)00196-1
Mota NP, Cook JM, Smith NB, et al. Posttraumatic stress symptom courses in US military veterans: a seven-year, nationally representative, prospective cohort study. J Psychiatr Res. 2019;119:23-31. doi: 10.1016/j.jpsychires.2019.09.005
Lee DJ, Lee LO, Bovin MJ, et al. The 20-year course of posttraumatic stress disorder symptoms among veterans. J Abnorm Psychol. 2020;129(6):658-669. doi: 10.1037/abn000571
Fani N, Carter SE, Harnett NG, Ressler KJ, Bradley B. Association of racial discrimination with neural response to threat in Black women in the US exposed to trauma. JAMA Psychiatry. 2021;78(9):1005-1012. doi: 10.1001/jamapsychiatry.2021.1480
Sripada RK, Pfeiffer PN, Rampton J, et al. Predictors of PTSD symptom change among outpatients in the US Department of Veterans Affairs Health Care System. J Trauma Stress. 2017;30(1):45-53. doi: 10.1002/jts.22156
Marx BP, Doron-Lamarca S, Proctor SP, Vasterling JJ. The influence of pre-deployment neurocognitive functioning on post-deployment PTSD symptom outcomes among Iraq-deployed Army soldiers. J Int Neuropsychol Soc. 2009;15(6):840-852. doi: 10.107/S1355617709990488
Vasterling JJ, Aslan M, Lee LO, et al. Longitudinal associations among posttraumatic stress disorder symptoms, traumatic brain injury, and neurocognitive functioning in Army Soldiers deployed to the Iraq War. J Int Neuropsychol Soc. 2018;24(4):311-323. doi: 10.1017/S1355617717001059
Holmes EA & Mathews A. Mental imagery and emotion: a special relationship? Emotion. 2005;5(4):489-497. doi: 10.1037/1528-3542.5.4.489
Moshier SJ, Lee DJ, Bovin MJ, et al. An empirical crosswalk for the PTSD Checklist: translating DSM-IV to DSM-5 using a veteran sample. J Trauma Stress. 2019;32(5):799-805. doi: 10.1002/jts.22438

No competing interests reported.

SupplementaryInformationSPPE.docx

Download PDF

Version 1

posted

You are reading this latest preprint version

Early Predictors of Chronic PTSD Symptom Trajectories in US Army Soldiers Deployed to the Iraq War

Status:

Version 1

Abstract

Purpose

Methods

Results

Conclusions

Trial Registration

Figures

Introduction

Methods

Study design and participants

Measures

Statistical analyses

Results

DISCUSSION

Conclusions

Declarations

References

Additional Declarations

Supplementary Files

Status:

Version 1