Abstract
Objective:
To compare the course of change in individual PTSD symptoms during Prolonged Exposure (PE) and Cognitive Processing Therapy (CPT).
Method:
We analyzed data from a previously published randomized clinical trial comparing PE and CPT among male and female US military veterans with PTSD (Schnurr et al., 2022). Using data from the PTSD Checklist for DSM-5 administered before each therapy session, we evaluated individual symptom change from pre-treatment to final therapy session (N = 802). Then, using Network Intervention Analysis, we modeled session-by-session PTSD symptom networks that included treatment allocation (CPT vs. PE) as a node in the networks, allowing us to compare individual symptom change following each session in each treatment.
Results:
Relative to CPT, PE was associated with greater reduction in ten PTSD symptoms from first to final session of therapy. Numerous treatment-specific effects on individual symptoms emerged during the treatment period; these session-level effects occurred only in symptoms relatively specific to the diagnosis of PTSD (e.g., avoidance, hypervigilance). PE was associated with greater reduction in avoidance following the introduction and early weeks of imaginal exposure. The treatments yielded comparable effects on trauma-related blame and negative beliefs from pre-treatment to final therapy session. However, there were differences in session-level change in these symptoms that may reflect differential timing of interventions that reduce distorted cognitions within each treatment.
Conclusions:
Findings may facilitate the shared decision-making process for patients choosing between CPT and PE. Session-level results provide direction for future research on the specific intervention components of CPT and PE.
Keywords: Posttraumatic stress disorder, Cognitive processing therapy, Prolonged exposure, Randomized trial, Network analysis
Public Health Significance Statement: In this study we investigated the pattern of change in the individual symptoms of PTSD occurring during two of the most strongly recommended psychotherapies for PTSD, Cognitive Processing Therapy and Prolonged Exposure. The results help contribute to hypotheses about how these treatments work and provide information about which specific symptoms of PTSD are most responsive to these therapies.
Prolonged Exposure (PE) and Cognitive Processing Therapy (CPT) are manualized trauma-focused psychotherapies that are among the most strongly recommended treatments for posttraumatic stress disorder (PTSD; Hamblen et al., 2019). Both therapies have demonstrated efficacy in a range of patient populations including military veterans and survivors of sexual abuse and assault (e.g., Chard, 2005; Monson et al., 2006; Resick et al., 2002; Schnurr et al., 2007). Studies also support their effectiveness in real-world clinical settings (e.g., Eftekhari et al., 2013; Forbes et al., 2012).
PE and CPT are both short-term treatments aligned with general cognitive-behavioral principles. However, they are based on different theoretical models of PTSD. PE developed out of Emotional Processing Theory (Foa & Kozak, 1986), which posits that PTSD is caused by a pathological fear structure in which stimuli associated with the trauma memory become inaccurately associated with danger (Foa et al., 2019). PE aims to modify the fear structure by reactivating the trauma memory and presenting evidence that disconfirms inaccurate meanings and perceptions. The major treatment elements of PE include imaginal exposure to traumatic memories and in-vivo exposure to feared or avoided trauma-related stimuli.
CPT is based on social cognitive and information processing theories of PTSD (Resick et al., 2016). From this perspective, PTSD develops when traumatic events disrupt schemas about self, others, and the world. Individuals may develop erroneous beliefs related to the traumatic event that cause guilt or shame, as well as over-generalized beliefs about themselves and the world (Resick et al., 2014). In CPT, therapists use techniques such as Socratic dialogue, didactic instruction on identifying and challenging “stuck points” (inaccurate or unhelpful beliefs related to the trauma), and daily worksheets with the aim of helping the patient integrate the trauma into their existing schemas in a more balanced way.
To our knowledge, the efficacy of PE and CPT have been directly compared in only two randomized controlled trials. In the first trial, Resick and colleagues (2002) compared CPT, PE, and a minimal attention condition among 171 female rape victims. CPT and PE demonstrated similar efficacy in reducing PTSD and depression outcomes, and CPT was associated with greater reductions on two out of four guilt subscales. In the second trial, Schnurr and colleagues (2022) compared PE and CPT in a randomized clinical trial of over 900 US Veterans with PTSD. PE produced a statistically significant greater mean improvement in PTSD symptoms relative to CPT, but this difference was not clinically significant (defined a priori as standardized mean difference of at least .25). Patients assigned to PE had higher odds of PTSD response, remission, and loss of diagnosis than those assigned to CPT. However, PE was also associated with a significantly higher number of dropouts (56% compared with 47%) and did not differ from CPT in outcomes such as depression, functioning, and quality of life. Therefore, the authors recommended that when choosing between PE and CPT, clinicians and patients engage in shared decision making to select the treatment that best matches the patient’s needs and preferences (Schnurr et al., 2022).
Despite strong empirical support for PE and CPT, a substantial minority of patients do not respond to treatment or continue to retain a PTSD diagnosis after treatment (e.g., Steenkamp et al., 2015). Developing a stronger understanding of how symptom reduction occurs in CPT and PE is crucial for improving outcomes for these individuals. Given their differing theoretical backgrounds and treatment targets, the two therapies may arrive at similar clinical outcomes through different processes of change. One common strategy for investigating mechanisms of change in CPT and PE is to identify mediators of treatment response. For instance, numerous studies have found that in both CPT and PE, changes in post-traumatic cognitions are related to symptom reduction, and a number of others demonstrate that between-session habituation to exposure (i.e., a reduction in distress elicited during exposure) is associated with more positive treatment outcomes in PE (for review, see Alpert et al., 2023).
Another path toward clarifying how CPT and PE exert their effects is to develop stronger understanding of the sequence of change in PTSD symptoms over time. Understanding how individual symptoms of PTSD change differentially in response to CPT and PE, and at what point in treatment such change occurs, has potential to aid hypothesis development regarding the active treatment ingredients underlying the therapies. Knowledge regarding individual symptom change is especially valuable when viewing PTSD from the framework of Network Theory, which proposes that mental disorders are the product of dynamic interactions between symptoms (Borsboom & Cramer, 2013). That is, instead of being caused and maintained by a latent disease entity, PTSD can be conceptualized as a network of symptoms that causally influence one another. From this perspective, treatments for PTSD may exert their effects by directly reducing specific symptoms, changes that trigger indirect alleviation of other symptoms in the network.
Given that clinical trials of CPT and PE typically focus on total PTSD symptom severity outcomes, there is relatively little understanding of how the two therapies influence individual symptom severity. The few studies available on this topic have focused on PE or other exposure therapies and have found unique effects of exposure on avoidance, numbing, and re-experiencing symptoms relative to comparison conditions such as supportive counseling and relaxation training (e.g., Bryant et al., 2003; Schnurr et al., 2015; Taylor et al., 2003). There is also a small literature on residual symptoms following PTSD treatment suggesting that arousal and hypervigilance symptoms may be the least likely to resolve following trauma-focused psychotherapies, including CPT and PE (e.g., Larsen et al., 2019; Schnurr & Lunney, 2019; Tripp et al., 2020).
Although these studies provide some valuable information regarding individual symptom presentation at the end of treatment, there is a lack of information on how individual PTSD symptoms change over the course of CPT and PE. Network Intervention Analysis (NIA), developed as a tool to characterize treatment-induced changes in a symptom network over time (Blanken et al., 2019), can help us fill this knowledge void. In NIA, network analytic techniques are used to model a symptom network, and a treatment allocation variable is included as a node in the network. When the symptom network is modeled in this way using repeated measurements over the course of treatment, it allows for detailed observation of treatment-induced changes in specific symptoms in the network over the course of the treatment period, while accounting for the interrelationships among symptoms (Blanken et al., 2019).
In this study, we applied NIA to data from the randomized clinical trial of PE and CPT by Schnurr and colleagues (2022), with a goal of characterizing the differential course of individual symptom change in the two therapies during treatment. We used data from the PTSD Checklist (PCL-5) administered before each therapy session to model session-by-session symptom networks over the course of treatment. We included treatment allocation (CPT vs. PE) as a node in the symptom networks, allowing for evaluation of the relative impact of each treatment on the PTSD symptom network at a given timepoint. Our analysis was largely exploratory in nature; however, given that CPT emphasizes modification of negative trauma-related beliefs, we hypothesized that CPT would be more strongly associated with reductions in the PTSD symptoms of negative beliefs and distorted blame for self or others. Given that PE emphasizes exposure to previously avoided situations and memories, we hypothesized that PE would be associated more strongly than CPT with reductions in the avoidance symptoms and psychological distress and physiological reactions upon exposure to trauma reminders.
Method
Participants
The parent study to this analysis was a prospective randomized clinical trial comparing PE to CPT (see Schnurr et al., 2022; ClinicalTrials.gov Identifier NCT01928732) that included 916 Veterans Health Administration (VHA) users with PTSD stemming from any type of traumatic military experience. In addition to PTSD diagnosis, participants were required to score 25 or higher on the Clinician Administered PTSD Scale for DSM-5. Exclusion criteria included lack of substance use disorder remission for at least one-month, current psychotic symptoms, manic episodes, or bipolar disorder, significant current suicidal or homicidal intent with a specific plan, and moderate to severe cognitive impairment. Individuals using psychotropic medications were required to have no changes in their prescribed drugs or their dosages for 1–2 months prior to study entry.
Participants were recruited from 17 Department of Veterans Affairs (VA) medical centers across the United States from October 2014 to February 2018. The study was approved by VA’s Central Institutional Review Board and all participants provided written informed consent prior to participation. Our analyses focus on a subsample of 802 Veterans (402 assigned to CPT and 400 assigned to PE) with at least two sequentially completed PCL-5 questionnaires with complete data (i.e., making it possible to compute a between-session change score). Sample characteristics are reported in Table 1. We report how we determined our sample size, all data exclusions (if any), all manipulations, and all measures in the study.
Table 1.
Sample characteristics by treatment assignment
| Variable | CPT group (n = 402) n (%) |
PE group (n = 400) n (%) |
|---|---|---|
| Gender | ||
| Men | 320 (79.6) | 317 (79.3) |
| Women | 82 (20.4) | 83 (20.8) |
| Age, Mean (S.D.) | 45.40 (13.60) | 46.18 (13.54) |
| Service Era | ||
| Vietnam | 74 (18.4) | 76 (19.0) |
| Gulf War | 77 (19.2) | 79 (19.8) |
| OEF/OIF/OND | 232 (57.5) | 222 (55.5) |
| Associates degree or higher | 171 (42.6) | 189 (47.4) |
| Unemployed | 125 (31.1) | 132 (33.70) |
| Married or cohabitating | 212 (52.7) | 230 (57.5) |
| Race | ||
| American Indian or Alaskan Native | 14 (3.5) | 15 (3.8) |
| Asian | 10 (2.5) | 13 (3.3) |
| Black | 108 (26.9) | 104 (26.0) |
| Native Hawaiian or Pacific Islander | 6 (1.5) | 6 (1.5) |
| White | 256 (63.7) | 267 (66.8) |
| Other | 25 (6.2) | 20 (5.0) |
| Spanish, Hispanic, or Latino ethnicity | 61 (15.2) | 63 (15.8) |
| Positive VA screen | ||
| Military sexual trauma | 118 (29.6) | 118 (29.6) |
| Traumatic brain injury | 239 (59.5) | 257 (65.1) |
| Current comorbid psychiatric disorders | ||
| Depressive disorder | 279 (69.4) | 261 (64.9) |
| Anxiety disorder* | 152 (37.8) | 116 (28.9) |
| Alcohol use disorder | 82 (20.4) | 83 (20.9) |
| Substance use disorder | 37 (9.2) | 24 (6.0) |
| Obsessive compulsive disorder | 26 (6.5) | 15 (3.8) |
Indicates a statistically significant difference between treatment conditions (p < .05)
Abbreviations: PTSD = posttraumatic stress disorder, CPT = Cognitive Processing Therapy, PE = Prolonged Exposure Therapy
Procedure
The parent study compared the effectiveness of PE and CPT for attenuating PTSD symptom severity among Veterans receiving care for PTSD within the VA healthcare system. Following screening and confirmation of eligibility, participants were randomly assigned on a 1:1 ratio to receive either PE or CPT, with 455 participants randomized to PE and 461 participants randomized to CPT. During the treatment phase, in keeping with the PE and CPT protocols, participants completed self-rated symptom severity measures immediately prior to each therapy session. Blinded assessment of study outcomes was conducted at post-treatment and 3- and 6- month follow ups. Because the focus of our analysis is on the session-level impact of PE and CPT, post-treatment and follow-up outcomes are not reported.
Treatments
Both PE and CPT were delivered in outpatient VHA clinics according to the standardized manuals used in the VHA’s national training rollouts of the two therapies (Karlin et al., 2010). PE consisted primarily of in vivo and imaginal exposure and processing of exposure experiences (Foa et al., 2019). CPT involved cognitive therapy focused on challenging trauma-related “stuck points” as well as writing two accounts of the index trauma (this is now considered an optional component of the newer version of CPT; Resick et al., 2016). There were 12 weekly sessions for each treatment, with 90-minute standard PE sessions and 60-minute standard CPT sessions. However, participants could complete treatment in 10 or 11 sessions if they reported a PCL-5 severity score of 18 or less in 2 consecutive sessions starting at Session 8. Conversely, participants with PCL-5 scores of 38 or higher at Session 12 could opt for 2 additional treatment sessions. Participants could also request up to two “Stressor sessions” outside of the treatment protocols to address significant distress that could potentially impede participation in the study. Please contact the treatment developers for the treatment manuals.
Measures
PTSD.
PTSD symptom severity was assessed throughout the treatment period using the PTSD Checklist for DSM-5 (PCL-5; Weathers et al., 2013). The PCL-5 assesses the 20 symptoms of PTSD by asking participants to rate how much each symptom bothered them during the past month on a 5-point Likert-type scale ranging from 0 (“Not at all”) to 4 (“Extremely”), with higher scores indicating higher severity of symptoms (total range is from 0 to 80). We used the past-week version of the PCL-5, administered immediately before each therapy session, to evaluate between-session change in severity scores. The PCL-5 has strong psychometric properties (e.g., Bovin et al., 2016). Cronbach’s alpha exceeded .90 at all timepoints.
Demographic characteristics.
We collected demographic characteristics including gender, age, race, ethnicity, education level, and marital status via self-report questionnaires.
Statistical Analysis
To aid in interpretation of NIA results, we evaluated individual symptom outcomes from baseline to each patient’s final therapy session using PCL-5 data. First, we plotted the session-by-session means of each individual PCL-5 item. We then conducted within sample t-tests to evaluate symptom change from pre-treatment to the last attended therapy session for each individual PTSD symptom in each treatment group. Finally, we compared the two treatment groups’ mean residualized change scores from baseline to last available therapy session using t-tests. Residualized changes scores were calculated for each participant by regressing each PCL-5 item value at pre-treatment (i.e., the Session 1 PCL-5 administration) onto its value at the final available therapy session and saving the residuals of these regression equations. In these analyses, missing data were handled using pairwise deletion; missing observations were infrequent (i.e., less than 1% of all possible observations were missing from baseline and final therapy session PCL-5 administrations).
We applied NIA (Blanken et al., 2019) to investigate differential symptom change between PE and CPT groups following each treatment session. Specifically, we estimated a Mixed Graphical Model (MGM; Haslbeck & Waldorp, 2020) at each session of therapy through Session 9. The baseline network consisted of nodes representing each of the PCL-5 item scores collected immediately prior to treatment Session 1 and a treatment allocation variable (CPT = 0, PE = 1); this network allowed for evaluation of potential pre-treatment differences in symptom presentation. Each of the remaining MGMs were modeled using residualized change scores for each PCL-5 item and also included the dichotomous treatment allocation variable. The residualized change scores were calculated using data from the PCL-5 assessment administered immediately prior to the therapy session of interest (Pre) and the PCL-5 administered at the beginning of the next therapy session (Post; this took place approximately one week later but varied by participant). We regressed each variable’s value at Pre onto its value at Post, and entered the saved residuals of these regression equations as nodes in the MGM. Therefore, each symptom node in the network represented the degree of change in that PCL-5 item following that therapy session, while accounting for pre-session severity level. The “post” PCL-5 of each model was from the PCL-5 administered at the next protocol session (e.g., if modeling Session 2, it was the PCL-5 administered at Session 3). If a Stressor session occurred prior to this, the PCL-5 administered at this visit was used instead.
We did not include data collected at Sessions 11 or 12 in the NIA because at this point in the study, participants began to discontinue therapy early on the basis of obtaining two consecutive PCL-5 severity scores of 18 or less, introducing collider bias into the sample. Collider bias (also referred to as conditioning on a collider) can cause spurious correlations and occurs when selecting a sample based on a variable that is influenced by both exposure and outcome (Holmberg & Andersen, 2022). In this case, the sample was selected on PCL-5 total scores, which were influenced by treatment condition (i.e., more participants in the PE condition were eligible to end therapy early than CPT patients; Χ2 (1, N = 802) = 4.10, p = 0.04), and were also influenced by the individual PCL-5 items that were the outcome of the NIA.
R package mgm was used to estimate the MGMs. To prevent the inclusion of spurious edges due to variation in sampling, we applied LASSO regularization; cross-validation was used to select the LASSO tuning parameter (Haslbeck & Waldrop, 2020). We report on the edge weights (EW) between the treatment allocation node and all other PTSD symptom change scores in the session-level networks. EWs can be interpreted as the conditional dependencies between two nodes, accounting for all other edges in the network. A positive EW between the treatment node and a symptom node was indicative of a greater reduction in that symptom in the CPT condition; a negative EW between the treatment node and a symptom node was indicative of a greater reduction in that symptom in the PE condition.
Using the resample feature of the mgm package, we evaluated edge weight stability of each network by repeating the model in 100 bootstrap samples. This provided a stability estimate reflecting the proportion of runs in which an edge weight was non-zero. We followed the approach of Fishbein and colleagues (2022) and focus our reporting and interpretation only on the edges between the treatment node and symptom nodes that yielded a stability estimate of .60 or greater (see Supplement for all results, including those with lower stability estimates). This approach was selected to avoid over-interpretation of results that were highly dependent on sample composition; however, it should be noted that there is no empirical research investigating optimal cut points for determining which network edges are interpreted. Each MGM was also visualized using R package qgraph.
Because NIA procedures require complete data, there was variability in sample size at each time point due to missing data from individuals who skipped specific PCL-5 items or dropped out or completed treatment early. Table 2 reports on attendance and provision of complete PCL-5 data at each therapy session. Of all PCL-5 administrations throughout the treatment period, incomplete data occurred on 3.2% and 3.3% of occasions in the CPT and PE groups, respectively. Dropout and early completion of treatment were more frequent causes of missing data. In the analytic sample, 50% (n = 201) of PE patients completed treatment, compared with 61% (n = 246) of CPT patients, Χ2 (1, N = 802) = 9.73, p = 0.002. A significantly greater proportion of the sample dropped out of PE relative to CPT following Sessions 2 and 4 (X2 (1, N = 745) = 8.44, p = .003, and X2 (1, N = 624) = 6.26, p = .012, respectively). To address the possibility that differential patterns of dropout may influence the NIA results, we conducted sensitivity analyses in which we re-ran all network models in the treatment completer sample. Analyses were not pre-registered. Data from the study are not available; because the study was approved and initiated before VA implemented its current data sharing policies, the consent form did not include consent language for open data sharing.
Table 2.
Participant attendance and patterns of data completion at each therapy session by treatment condition
| Session Number | Attended Session | Provided Complete PCL-5 Data | Final session attended | |||
|---|---|---|---|---|---|---|
| CPT (n = 402) |
PE (n = 400) |
CPT | PE | CPT | PE | |
| 1 | 401 | 399 | 395 | 384 | 0 | 0 |
| 2 | 399 | 399 | 388 | 377 | 18 | 39* |
| 3 | 380 | 357 | 371 | 339 | 30 | 33 |
| 4 | 354 | 325 | 343 | 311 | 21 | 37* |
| 5 | 332 | 290 | 320 | 279 | 26 | 15 |
| 6 | 307 | 272 | 292 | 267 | 12 | 13 |
| 7 | 292 | 260 | 282 | 255 | 16 | 14 |
| 8 | 276 | 247 | 266 | 243 | 13 | 16 |
| 9 | 266 | 232 | 254 | 229 | 6 | 8 |
| 10 | 260 | 225 | 249 | 221 | 21 | 54* |
| 11 | 239 | 171 | 230 | 167 | 13 | 24* |
| 12 | 225 | 146 | 221 | 142 | 184 | 115* |
| 13 | 41 | 32 | 38 | 32 | 17 | 15 |
| 14 | 25 | 17 | 25 | 17 | 25 | 17 |
Abbreviations: PCL-5 = PTSD Checklist for DSM-5, CPT = Cognitive Processing Therapy, PE = Prolonged Exposure Therapy
Significantly higher rate of attrition or early treatment completion in PE relative to CPT after this session, p < .05
Results
Table 1 presents the characteristics of the analytic sample by treatment group. The two groups did not differ by gender, age, service era, employment status, education level, marital status, race or ethnic identity, trauma exposure, or PTSD disability claim status. There was a significantly higher percentage of participants with current comorbid anxiety disorders in the CPT group relative to the PE group, Χ2 (1, N = 802) = 7.12, p = 0.008. There were no differences between groups in the percentage of participants with current depressive, substance use, alcohol use, or obsessive-compulsive disorders.
Change in Individual PTSD Symptoms from Pre-Treatment to Final Attended Therapy Session
Figure 1 depicts the session-by-session change in the mean of each PCL-5 item by treatment condition. Paired t-tests demonstrated that in both treatment conditions, there was a significant reduction in each individual PCL-5 item from baseline to the participants’ final attended therapy session (see Table 3, Cohen’s d ranging from 0.27 – 0.98). Relative to CPT, PE was associated with significantly greater reduction in ten symptoms: flashbacks, psychological distress in response to trauma reminders, physiological reactions to trauma reminders, avoidance of thoughts and feelings, avoidance of external reminders, persistent negative emotions, detachment from others, irritable or aggressive behavior, hypervigilance, and sleep disturbance (see Table 3, Cohen’s d ranging from 0.08 to 0.32). Nine of these symptoms (all but sleep disturbance) remained significantly favoring PE in a sensitivity analysis of the treatment completer sample (see Supplement).
Figure 1.
Mean PCL-5 item rating by treatment condition over the course of treatment
Note. PCL-5 = PTSD Checklist for DSM-5
Table 3.
Change in individual PTSD symptoms within and between treatment condition according to PCL-5 data collected at pre-treatment and at each participant’s final therapy session.
| PCL-5 Item | Pre-treatment to final therapy session effect size a | Between treatment effect size | ||||
|---|---|---|---|---|---|---|
|
| ||||||
| CPT | PE | |||||
|
| ||||||
| d | [95% CI] | d | [95% CI] | d | [95% CI] | |
| 1. Intrusive memories | 0.61 | [0.51, 0.72] | 0.69 | [0.58, 0.80] | .11 | [−0.03, 0.25] |
| 2. Nightmares | 0.53 | [0.43, 0.64] | 0.59 | [0.48, 0.70] | .08 | [−0.06, 0.22] |
| 3. Flashbacks | 0.33 | [0.23, 0.43] | 0.44 | [0.34, 0.55] | .18* | [0.04, 0.32] |
| 4. Psychological distress in response to reminders | 0.56 | [0.46, 0.67] | 0.84 | [0.72, 0.95] | .24* | [0.10, 0.37] |
| 5. Physiological reactions to reminders | 0.57 | [0.47, 0.68] | 0.70 | [0.59, 0.81] | .15* | [0.01, 0.29] |
| 6. Avoidance of thoughts and feelings | 0.64 | [0.53, 0.74] | 0.93 | [0.81, 1.04] | .24* | [0.10, 0.38] |
| 7. Avoidance of external reminders | 0.68 | [0.57, 0.79] | 0.98 | [0.86, 1.10] | .32* | [0.18, 0.46] |
| 8. Amnesia | 0.27 | [0.17, 0.37] | 0.38 | [0.28, 0.48] | .11 | [−0.03, 0.24] |
| 9. Distorted negative beliefs | 0.52 | [0.42, 0.63] | 0.63 | [0.53, 0.74] | .13 | [−0.01, −0.27] |
| 10. Distorted blame | 0.50 | [0.40, 0.60] | 0.60 | [0.49, 0.70] | .07 | [−0.07, 0.21] |
| 11. Persistent negative emotions | 0.56 | [0.46, 0.67] | 0.81 | [0.70, 0.93] | .20* | [0.06, 0.34] |
| 12. Loss of interest | 0.67 | [0.56, 0.78] | 0.78 | [0.66, 0.89] | .13 | [−0.01, 0.27] |
| 13. Detachment | 0.67 | [0.56, 0.78] | 0.85 | [0.73, 0.96] | .18* | [0.04, 0.32] |
| 14. Inability to experience positive emotions | 0.52 | [0.42, 0.63] | 0.65 | [0.54, 0.76] | .13 | [−0.01, 0.27] |
| 15. Irritable or aggressive behavior | 0.57 | [0.46, 0.67] | 0.73 | [0.61, 0.84] | .15* | [0.02, 0.29] |
| 16. Reckless behavior | 0.38 | [0.28, 0.48] | 0.53 | [0.43, 0.64] | .12 | [−0.02, 0.26] |
| 17. Hypervigilance | 0.61 | [0.50, 0.72] | 0.82 | [0.71, 0.94] | .22* | [0.08, 0.36] |
| 18. Exaggerated Startle | 0.63 | [0.53, 0.74] | 0.69 | [0.58, 0.80] | .11 | [−0.03, 0.25] |
| 19. Concentration problems | 0.54 | [0.43, 0.64] | 0.65 | [0.54, 0.75] | .09 | [−0.05, −.23] |
| 20. Sleep disturbance | 0.54 | [0.44, 0.65] | 0.62 | [0.51, 0.72] | .15* | [0.01, 0.29] |
All pre-treatment to - final treatment session comparisons were statistically significant (p < .05)
Indicates statistical significance in mean residualized change score between groups (p < .05)
Abbreviations: PCL-5 = PTSD Checklist for DSM-5, CPT = Cognitive Processing Therapy, PE = Prolonged Exposure Therapy
Network Intervention Analysis
In the baseline network model, the treatment condition node was not associated with any other nodes in the network. There were differential treatment effects after 7 of the 9 therapy sessions modeled, with effects in favor of CPT after two sessions (Sessions 3 and 6), and effects in favor of PE after 9 sessions (Sessions 1, 3, 4, 5, 6, 8, and 9). As hypothesized, change in avoidance was greater in the PE condition; this occurred following Sessions 3 and 5. As hypothesized, CPT was associated with greater reduction in blame for self/others; this occurred following Session 6. However, PE was associated with greater reduction in blame following Session 3, and negative beliefs after Sessions 6 and 8. Table 4 reports all symptom nodes that were significantly associated with treatment condition at each timepoint. Sensitivity analyses conducted in the treatment completer sample were broadly consistent with the results in the entire sample and suggested that it was unlikely that the treatment associations described above were due to uneven rates of dropout/missing data (see Supplement).
Table 4.
Summary of significant edges between treatment condition node and PCL-5 item residualized change score nodes following each therapy session in Network Intervention Analysis (NIA), with accompanying overview of CPT and PE session.
| NIA indicates greater relative reduction in: | CPT Agenda | PE Agenda | |
|---|---|---|---|
| Session 1 | PE: Hypervigilance (EW = −.14) | • Psychoeducation and treatment rationale • Review index trauma • Assign written impact statement |
• Psychoeducation and treatment rationale • Trauma interview • Breathing retraining |
| Session 2 | None | • Read impact statement, identify stuck points • Discuss event-thought-feeling connection and introduce ABC monitoring worksheet |
• Psychoeducation • Rationale for in vivo exposure • Build in vivo exposure hierarchy • Plan exposure homework |
| Session 3 | CPT: Intrusive Memoriesa (EW = .08) and Nightmaresa (EW = .04) PE: Flashbacksb (EW = −.04), Avoidance of external reminders (EW = −.09), and Blame for self/others (EW = −.08) |
• Review monitoring practice • Begin examining and challenging assimilated stuck points • Introduce written trauma account |
• Review Homework • Rationale for imaginal exposures • Conduct and process imaginal exposure |
| Session 4 | PE: Intrusive memories (EW = −.12) and Exaggerated startle (EW = −.05) | • Read trauma account, identify stuck points and examine using Socratic questioning • Discuss difference between responsibility and blame • Assign second written trauma account |
• Review Homework • Conduct and process imaginal exposure • Discuss in vivo exposures |
| Session 5 | PE: Avoidance of Thoughts and Memories (EW = −.13) | • Read second trauma account • Continue to challenge stuck points, focus on self-blame • Introduce Challenging Questions Worksheet |
• Review Homework • Conduct and process imaginal exposure (hot spot work) • Discuss in vivo exposures |
| Session 6 | CPT: Blame for self and others (EW = .08) PE: Negative beliefs (−.05) and Hypervigilance (−.08) |
• Review Challenging Questions worksheet • Continue cognitive therapy for stuck points • Introduce Patterns of problematic thinking worksheet |
• Review Homework • Conduct and process imaginal exposure (hot spot work) • Discuss in vivo exposures |
| Session 7 | None | • Review Patterns of Problematic Thinking worksheets • Introduce Challenging Beliefs Worksheet, practice with a stuck point • Introduce Safety theme |
• Review Homework • Conduct and process imaginal exposure (hot spot work) • Discuss in vivo exposures |
| Session 8 | PE: Negative Beliefs (EW = −.06) | • Review Safety theme worksheets/stuck points, help patient confront and challenge problematic cognitions • Introduce Trust theme |
• Review Homework • Conduct and process imaginal exposure (hot spot work) • Discuss in vivo exposures |
| Session 9 | PE: Physiological reactions to reminders (EW = −.11) | • Review Trust theme worksheets/stuck points, help patient confront and challenge problematic cognitions • Introduce Power/Control theme |
• Review Homework • Conduct and process imaginal exposure (hot spot work) • Discuss in vivo exposures |
| Session 10 | Network not modeled | • Review Trust and Power and Control worksheets/stuck points, help patient confront and challenge • Review Giving and Taking Power handout • Introduce Esteem theme |
• Continue as needed. • If final session, recount trauma memory only once and discuss progress over the course of treatment, relapse prevention, and termination. |
| Session 11 | Network not modeled | • Review Esteem theme and stuck points • Discuss termination • Introduce Intimacy theme |
• Continue as needed. • If final session, recount trauma memory only once and discuss progress over the course of treatment, relapse prevention, and termination. |
| Session 12 | Network not modeled | • Review Intimacy theme stuck points • Read new and original impact statements • Review and discuss treatment progress and future goals |
• Continue as needed. • If final session, recount trauma memory only once and discuss progress over the course of treatment, relapse prevention, and termination. |
Abbreviations: CPT = Cognitive Processing Therapy, PE = Prolonged Exposure Therapy
The mean scores on these items after Session 2 and 3 suggest that these symptoms were exacerbated in the PE group rather than improved in the CPT group.
The mean scores on this item after Session 2 and 3 suggest that these symptoms were exacerbated in the CPT group rather than improved in the PE group
Note. Networks for Sessions 10, 11, and 12 were not modeled due to introduction of collider bias into sample (see Method)
Discussion
In this study, we examined the influence of CPT and PE on individual PTSD symptoms throughout the treatment period. Evaluation of symptom change from baseline to the final therapy session revealed several noteworthy findings. First, there were moderate to large within-subject reductions in nearly all DSM-5 symptoms in both PE and CPT groups. Assignment to PE was associated with significantly greater reduction in ten of the twenty DSM-5 symptoms relative to CPT at one’s final therapy session. These ten symptoms were a mix of specific (i.e., psychological distress and physiological reactions to reminders, avoidance, and hypervigilance) and non-specific PTSD symptoms (e.g., sleep disturbance, persistent negative emotion, aggressive behavior). These results contrast with the one prior study which compared PE and CPT directly on individual residual symptom outcomes (Larsen et al., 2019) and found only minimal, non-significant differences. It is likely that the larger size of the current trial allowed for enhanced power to detect the small effect sizes that we observed between groups. Further replication of these results is warranted, particularly using clinician-rated, follow-up data in addition to the self-rated, last-available treatment session data examined here. However, these findings suggest that some symptoms of PTSD are more likely to be more adequately addressed using PE, and this information may be useful for shared decision-making purposes.
In the NIA comparing session-by-session change in individual PTSD symptoms, treatment assignment was associated with ten DSM-5 PTSD symptoms following at least one session. Most of the symptoms that demonstrated greater change in the PE group at the end of therapy were directly associated with the treatment allocation node in the NIA at some point during treatment. However, some were not; these include persistent negative emotions, detachment from others, aggressive/irritable behavior, and sleep disturbance. Therefore, differences in these symptoms between PE and CPT may occur indirectly, as a result of change in other symptoms.
Interestingly, all ten PTSD symptoms associated with treatment condition in the NIA can be counted among those that are relatively specific to PTSD. This may reflect that both CPT and PE are designed to target the unique symptoms of PTSD (i.e., avoidance, intrusive memories), but vary in the specific techniques, and timing of techniques used to do so. Therefore, differences may emerge as the treatments exert effects on these targeted symptoms at different points in their protocols. However, because the two therapies do not contain specific intervention elements designed to target transdiagnostic symptoms (e.g., concentration difficulties, negative emotions), this could lead to a lack of session-level differences in these symptoms despite overall reductions in these symptoms across treatments.
Results supported our hypothesis that reductions in avoidance would be more pronounced in PE relative to CPT. Avoidance of external reminders demonstrated the largest between-group difference of any of PTSD symptom, with larger reductions in the PE group between baseline and last available therapy session. NIA results showed that assignment to PE was associated with greater change in avoidance of external trauma reminders relative to CPT following Session 3. This timing aligns with the introduction of imaginal exposures in PE sessions. Then, following Session 5, the PE group also experienced significantly greater reduction in avoidance of trauma-related thoughts and memories. As treatment progressed, although imaginal exposure continued in PE, there were no further associations between treatment condition and change in avoidance symptoms, indicating that the impact of PE relative to CPT on avoidance symptoms occurs in the early stages of treatment. Although significant reductions in avoidance symptoms occurred in both conditions, these results suggest that the emphasis placed on facing feared situations and trauma memories in PE makes it particularly well-suited for patients seeking to address avoidance quickly.
Results did not support our hypothesis that CPT would be associated with more robust reductions in inaccurate cognitions. Studies show that clinicians sometimes choose to use CPT with patients who present with strong negative cognitions (e.g., Simiola et al., 2019). However, change in distorted blame or negative beliefs about oneself, others, and the world from baseline to participants’ final therapy session did not differ between treatment conditions. Despite this, there were some between-group session-level effects. The results may reflect differential timing of interventions that reduce distorted blame within each treatment. PE patients showed greater reduction in distorted blame after Session 3, when imaginal exposures begin, suggesting that exposure may directly and quickly target this symptom. In CPT, distorted blame is most explicitly addressed in Sessions 4 and 5. That blame was reduced relative to PE following Session 6 instead of these sessions raises the possibility that the impact of blame-focused cognitive interventions in CPT is not immediate. The session-by-session data for this symptom is consistent with a pattern of more gradual change; in the CPT group, most of the reduction in distorted blame occurred in the latter half of treatment, particularly following Sessions 6 through 9. This could be explained in part by the fact that blame-related cognitions are associated with decreased cognitive flexibility (Keith et al., 2015) and increased feelings of shame (Beck et al., 2015), which may complicate the response to cognitive intervention strategies.
Results also showed that PE was associated with greater reduction in distorted negative beliefs following Sessions 6 and 8. This timing is noteworthy; because these CPT sessions are focused on challenging overgeneralized negative beliefs, it would have been reasonable to expect the results to favor CPT at these timepoints. There are several possible interpretations of this pattern of results. It may again signal that change in cognitions occurs gradually in response to CPT. It could also reflect that the impact of PE on negative beliefs is particularly strong at these points of therapy, when patients have engaged in several weeks of repeated exposure exercises and have begun more difficult exposures (i.e., in-vivo exposures toward the top of their hierarchy and imaginal exposure to “hot spots” of the trauma). Although these findings may help to generate hypotheses about differences between CPT and PE, continued research is needed to explore this unexpected pattern of results. It would be particularly valuable to collect qualitative or process-related data after these sessions of treatment.
The NIA suggested differential effects of PE and CPT on the individual intrusion symptoms of PTSD at several points during treatment. Differences first appeared following Session 3, when it appears that PE led to mild exacerbation of intrusive memories and nightmares relative to CPT, while CPT was associated with mild exacerbation of flashbacks relative to PE. It is not surprising that intrusion symptoms increased at this time; in PE, patients had just completed their first imaginal exposures, and in CPT, patients had completed a written account of their index traumatic event as homework. However, it is important to note that despite the small increases in these symptoms at this timepoint, consistent with prior research (e.g., Larsen et al., 2016), there was clear evidence of improvement over the duration of treatment. Notably, PE was associated with a greater reduction in intrusive memories relative to CPT following Session 4, which suggests that the uptick in intrusive memories after the introduction to imaginal exposure was short-lived.
Some of the differential effects found between PE and CPT during the treatment period are not easily explained by existing theories or research findings. This highlights the potential value of NIA; such findings may aid hypothesis development about how change occurs in response to treatment. For instance, at Session 1, both therapies include psychoeducation and treatment rationale components, but PE also involves breathing retraining. Therefore, it may be that the reduced hypervigilance demonstrated in the PE group following Session 1 reflects the effects of breathing retraining. At Session 4, both treatment groups have begun to engage in recounting of the trauma memory (PE through repeated imaginal exposure, and CPT through two written trauma accounts). That the PE group saw a greater reduction in exaggerated startle at this time may indicate the importance of repeated exposure for change in this symptom. However, it is also possible that such results reflect fluctuations in symptoms that are statistically significant but not clinically meaningful. Although NIA may be a valuable tool for hypothesis generation, application of other research methods for exploring session-level responses to CPT and PE are needed to better understand to what extent these findings reflect meaningful differences in the change processes of the two treatments.
Several limitations to this study should be acknowledged. First, the NIA analysis presented here provides information about the response to PE and CPT only in comparison to one another, and only for the first nine sessions of therapy. It would be valuable to conduct NIA in studies of PE or CPT in which a no-treatment control group was available for comparison so that more conclusions could be drawn about the sequence of change within, rather than between, treatments. Second, our analysis modeled group averages in symptom severity over the course of treatment, and therefore cannot capture within-individual variations in the dynamics of symptom change in response to PE and CPT. Additionally, because the NIA modeled and compared change from one session to the next, the analyses may not capture treatment differences that may develop over the course of multiple sessions.
Like other PTSD treatment studies with military and veteran samples (e.g., Steenkamp et al., 2015), attrition was high in our study. Participants who dropped out stopped providing weekly PCL-5 data, even if they completed post-treatment assessments at later timepoints. As a result, the network models reported here represent only those who chose to continue with treatment after a given session. This prevents us from identifying changes in symptom dynamics that may have contributed to attrition. It also raises the possibility that differences in patterns of dropout between groups might account for some of the treatment associations found in the session-by-session networks, although this concern is somewhat assuaged by the consistency in results between the analytic sample and the treatment completer sample.
Despite such limitations, our study helps to elucidate the differential effects of CPT and PE on individual symptoms of PTSD in a large, randomized trial of veterans with PTSD. PE was associated with greater change in several PTSD symptoms at participants’ final therapy session but performed comparably with CPT for others; these findings may facilitate the shared decision-making process as patients consider their unique treatment goals based on their specific symptom profiles. Network analyses showed emergence of differential treatment effects over the course of treatment on the individual symptoms of PTSD, primarily those that are specific to the diagnosis of PTSD as opposed to transdiagnostic symptoms reflecting general psychological distress. Continued investigation of symptom-specific changes in response to treatment is an important path for optimizing response to PE and CPT.
Supplementary Material
Acknowledgments
Study funding was provided by the Veterans Affairs (VA) Cooperative Studies Program. The VA Cooperative Studies program was involved in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, and approval of the manuscript; and decision to submit the manuscript for publication. The views expressed in this article are those of the authors and do not necessarily represent the views of the Department of Veterans Affairs or any US government agency.
Footnotes
This manuscript has not been published previously and is not under consideration for publication elsewhere. Data from this trial have been previously published in a manuscript focusing on the primary and secondary study outcomes (Schnurr et al., 2022). This is the first paper that has reported the weekly PCL-5 data from this trial. There are no published or submitted papers from this dataset that address related questions.
Data sharing statement:
Data are available to VA investigators who obtain a signed data use agreement from the VA Cooperative Studies Program. Because the study was approved and initiated before VA implemented its current data sharing policies, the consent form did not include consent language for open data sharing. Treatment manuals and materials are available from the treatment developers.
References
- Alpert E, Tabke CS, Cole TA, Lee DJ, & Sloan DM (2023). A systematic review of literature examining mediators and mechanisms of change in empirically supported treatments for posttraumatic stress disorder. Clinical Psychology Review, 102300. 10.1016/j.cpr.2023.102300 [DOI] [PubMed] [Google Scholar]
- Beck JG, Reich CM, Woodward MJ, Olsen SA, Jones JM, & Patton SC (2015). How do negative emotions relate to dysfunctional posttrauma cognitions? An examination of interpersonal trauma survivors. Psychological Trauma: Theory, Research, Practice, and Policy, 7(1), 3–10. 10.1037/a0032716 [DOI] [PubMed] [Google Scholar]
- Blanken TF, Van Der Zweerde T, Van Straten A, Van Someren EJ, Borsboom D, & Lancee J (2019). Introducing network intervention analysis to investigate sequential, symptom-specific treatment effects: a demonstration in co-occurring insomnia and depression. Psychotherapy and Psychosomatics, 88(1), 52–54. 10.1159/000495045 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Borsboom D, & Cramer AO (2013). Network analysis: an integrative approach to the structure of psychopathology. Annual Review of Clinical Psychology, 9(1), 91–121. 10.1146/annurev-clinpsy-050212-185608 [DOI] [PubMed] [Google Scholar]
- Bovin MJ, Marx BP, Weathers FW, Gallagher MW, Rodriguez P, Schnurr PP, & Keane TM (2016). Psychometric properties of the PTSD checklist for diagnostic and statistical manual of mental disorders–fifth edition (PCL-5) in veterans. Psychological Assessment, 28(11), 1379–1391. 10.1037/pas0000254 [DOI] [PubMed] [Google Scholar]
- Bryant RA, Moulds ML, Guthrie RM, Dang ST, & Nixon RD (2003). Imaginal exposure alone and imaginal exposure with cognitive restructuring in treatment of posttraumatic stress disorder. Journal of Consulting and Clinical Psychology, 71(4), 706–712. 10.1037/0022-006X.71.4.706 [DOI] [PubMed] [Google Scholar]
- Chard KM (2005). An evaluation of cognitive processing therapy for the treatment of posttraumatic stress disorder related to childhood sexual abuse. Journal of Consulting and Clinical Psychology, 73(5), 965–971. 10.1037/0022-006x.73.5.965 [DOI] [PubMed] [Google Scholar]
- Eftekhari A, Ruzek JI, Crowley JJ, Rosen CS, Greenbaum MA, & Karlin BE (2013). Effectiveness of national implementation of prolonged exposure therapy in Veterans Affairs care. JAMA Psychiatry, 70(9), 949–955. 10.1001/jamapsychiatry.2013.36 [DOI] [PubMed] [Google Scholar]
- Fishbein JN, Haslbeck J, & Arch JJ (2023). Network intervention analysis of anxiety-related outcomes and processes of acceptance and commitment therapy (ACT) for anxious cancer survivors. Behaviour Research and Therapy, 162, 104266. 10.1016/j.brat.2023.104266 [DOI] [PubMed] [Google Scholar]
- Foa EB, Hembree EA, Rothbaum BO, & Rauch S (2019). Prolonged Exposure Therapy for PTSD: Emotional processing of traumatic experiences—Therapist guide. Oxford University Press. 10.1093/med-psych/9780190926939.001.0001 [DOI] [Google Scholar]
- Foa EB, & Kozak MJ (1986). Emotional processing of fear: exposure to corrective information. Psychological Bulletin, 99(1), 20–35. 10.1037/0033-2909.99.1.20 [DOI] [PubMed] [Google Scholar]
- Forbes D, Lloyd D, Nixon RDV, Elliott P, Varker T, Perry D, Bryant RA, & Creamer M (2012). A multisite randomized controlled effectiveness trial of cognitive processing therapy for military-related posttraumatic stress disorder. Journal of Anxiety Disorders, 26(3), 442–452. 10.1016/j.janxdis.2012.01.006 [DOI] [PubMed] [Google Scholar]
- Hamblen JL, Norman SB, Sonis JH, Phelps AJ, Bisson JI, Nunes VD, ... & Schnurr PP (2019). A guide to guidelines for the treatment of posttraumatic stress disorder in adults: An update. Psychotherapy, 56(3), 359–373. 10.1037/pst0000231 [DOI] [PubMed] [Google Scholar]
- Haslbeck J, & Waldorp LJ (2020). mgm: Estimating time-varying mixed graphical models in high-dimensional data. Journal of Statistical Software, 93(8), 1–46. 10.18637/jss.v093.i08 [DOI] [Google Scholar]
- Holmberg MJ, & Andersen LW (2022) Collider bias. JAMA, 327 (13), 1282–1283. 10.1001/jama.2022.1820 [DOI] [PubMed] [Google Scholar]
- Karlin BE, Ruzek JI, Chard KM, Eftekhari A, Monson CM, Hembree EA, Resick PA, & Foa EB (2010). Dissemination of evidence-based psychological treatments for posttraumatic stress disorder in the Veterans Health Administration. Journal of Traumatic Stress, 23(6), 663–673. 10.1002/jts.20588 [DOI] [PubMed] [Google Scholar]
- Keith J, Velezmoro R, & O’Brien C (2015). Correlates of cognitive flexibility in veterans seeking treatment for posttraumatic stress disorder. The Journal of Nervous and Mental Disease, 203(4), 287–293. 10.1097/NMD.0000000000000280 [DOI] [PubMed] [Google Scholar]
- Larsen SE, Stirman SW, Smith BN, & Resick PA (2016). Symptom exacerbations in trauma-focused treatments: Associations with treatment outcome and non-completion. Behaviour Research and Therapy, 77, 68–77. 10.1016/j.brat.2015.12.009 [DOI] [PubMed] [Google Scholar]
- Larsen SE, Fleming CJ, & Resick PA (2019). Residual symptoms following empirically supported treatment for PTSD. Psychological Trauma: Theory, Research, Practice, and Policy, 11(2), 207–215. 10.1037/tra0000384 [DOI] [PubMed] [Google Scholar]
- Monson CM, Schnurr PP, Resick PA, Friedman MJ, Young-Xu Y, & Stevens SP (2006). Cognitive processing therapy for veterans with military-related posttraumatic stress disorder. Journal of Consulting and Clinical Psychology, 74(5), 898–907. 10.1037/0022-006x.74.5.898 [DOI] [PubMed] [Google Scholar]
- Resick PA, Nishith P, Weaver TL, Astin MC, & Feuer CA (2002). A comparison of cognitive-processing therapy with prolonged exposure and a waiting condition for the treatment of chronic posttraumatic stress disorder in female rape victims. Journal of Consulting and Clinical Psychology, 70(4), 867–879. 10.1037//0022-006x.70.4.867 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Resick PA, Monson CM, Chard KM (2016). Cognitive Processing Therapy for PTSD: A comprehensive manual. Guilford Press. [Google Scholar]
- Schnurr PP, Friedman MJ, Engel CC, Foa EB, Shea MT, Chow BK, Resick PA, Thurston V, Orsillo SM, Haug R, Turner C, & Bernardy N (2007). Cognitive behavioral therapy for posttraumatic stress disorder in women: A randomized controlled trial. JAMA, 297(8), 820–830. 10.1001/jama.297.8.820 [DOI] [PubMed] [Google Scholar]
- Schnurr PP, Chard KM, Ruzek JI, Chow BK, Resick PA, Foa EB, ... & Shih MC (2022). Comparison of prolonged exposure vs cognitive processing therapy for treatment of posttraumatic stress disorder among US veterans: A randomized clinical trial. JAMA Network Open, 5(1), e2136921–e2136921. 10.1001/jamanetworkopen.2021.36921 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Schnurr PP, & Lunney CA (2015). Differential effects of prolonged exposure on posttraumatic stress disorder symptoms in female veterans. Journal of Consulting and Clinical Psychology, 83(6), 1154–1160. 10.1037/ccp0000031 [DOI] [PubMed] [Google Scholar]
- Schnurr PP, & Lunney CA (2019). Residual symptoms following prolonged exposure and present-centered therapy for PTSD in female veterans and soldiers. Depression and Anxiety, 36(2), 162–169. 10.1002/da.22871 [DOI] [PubMed] [Google Scholar]
- Simiola V, Ellis AE, Thompson R, Schnurr PP, & Cook JM (2019). Provider perspectives on choosing prolonged exposure of cognitive processing therapy for PTSD: A national investigation of VA residential treatment providers. Practice Innovations, 4(3), 194–203. 10.1037/pri0000091 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Steenkamp MM, Litz BT, Hoge CW, & Marmar CR (2015). Psychotherapy for military-related PTSD: A review of randomized clinical trials. JAMA, 314(5), 489–500. 10.1001/jama.2015.8370 [DOI] [PubMed] [Google Scholar]
- Taylor S, Thordarson DS, Maxfield L, Fedoroff IC, Lovell K, & Ogrodniczuk J (2003). Comparative efficacy, speed, and adverse effects of three PTSD treatments: exposure therapy, EMDR, and relaxation training. Journal of Consulting and Clinical Psychology, 71(2), 330–338. 10.1037/0022-006X.71.2.330 [DOI] [PubMed] [Google Scholar]
- Tripp JC, Norman SB, Kim HM, Venners MR, Martis B, Simon NM, ... & PROGrESS Study Team. (2020). Residual symptoms of PTSD following Sertraline plus enhanced medication management, Sertraline plus PE, and PE plus placebo. Psychiatry Research, 291, 113279. 10.1016/j.psychres.2020.113279 [DOI] [PubMed] [Google Scholar]
- Weathers FW, Litz BT, Keane TM, Palmieri PA, Marx BP, & Schnurr PP (2013). The PTSD Checklist for DSM-5 (PCL-5). Scale available from the National Center for PTSD at www.ptsd.va.gov [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
Data are available to VA investigators who obtain a signed data use agreement from the VA Cooperative Studies Program. Because the study was approved and initiated before VA implemented its current data sharing policies, the consent form did not include consent language for open data sharing. Treatment manuals and materials are available from the treatment developers.