Review

Effectiveness of Cognitive Processing Therapy and Prolonged Exposure in the Department of Veterans Affairs

Psychological Reports 0(0) 1–21 ! The Author(s) 2017 Reprints and permissions: sagepub.com/journalsPermissions.nav DOI: 10.1177/0033294117727746 journals.sagepub.com/home/prx

Benjamin T. Rutt Dwight D. Eisenhower VAMC, Leavenworth, KS, USA; University of Kansas, Leavenworth, KS, USA

Mary E. Oehlert Dwight D. Eisenhower VAMC, Leavenworth, KS, USA

Thomas S. Krieshok and James W. Lichtenberg University of Kansas, KS, USA

Abstract Objective: This study evaluated the effectiveness of cognitive processing therapy and prolonged exposure in conditions reflective of current clinical practice within the Veterans Health Administration. Method: This study involved a retrospective review of 2030 charts. A total of 750 veterans from 10 U.S. states who received cognitive processing therapy or prolonged exposure in individual psychotherapy were included in the study (participants in cognitive processing therapy, N ¼ 376; participants in prolonged exposure, N ¼ 374). The main dependent variable was self-reported posttraumatic stress disorder symptoms as measured by total scores on the Posttraumatic Stress Disorder Checklist. The study used multilevel modeling to evaluate the absolute and relative effectiveness of both treatments and determine the relationship between patientlevel variables and total Posttraumatic Stress Disorder Checklist scores during treatment. Results: Cognitive processing therapy and prolonged exposure were equally effective at reducing total Posttraumatic Stress Disorder Checklist scores. Veterans who completed therapy reported significantly larger reductions in the Posttraumatic Stress Disorder Checklist than patients who did not complete therapy. There

Corresponding Author: Benjamin T. Rutt, Dwight D. Eisenhower VAMC, 4101 South 4th Street, Leavenworth, KS 66048, USA. Email: [email protected]

2

Psychological Reports 0(0)

were no significant differences in the improvement of posttraumatic stress disorder symptoms with respect to age and three racial/ethnic groups (Caucasian, African American, and Hispanic). Conclusions: Cognitive processing therapy and prolonged exposure were shown to be effective in conditions highly reflective of clinical practice and with a highly diverse sample of veterans. Challenges related to dropout from trauma focused therapy should continue to be researched. Keywords posttraumatic stress disorder, psychotherapy, cognitive processing therapy, prolonged exposure therapy, veterans

Introduction The Department of Veterans Affairs (VA) currently serves over 9.5 million veterans in hundreds of facilities across the United States (National Center for Veterans Analysis and Statistics, 2016). During the last decade, VA has responded to a large influx of veterans from Operation Enduring Freedom/ Operation Iraqi Freedom/Operation New Dawn (OEF-OIF-OND). Many veterans have returned from OEF-OIF-OND with deployment-related mental health problems. The most common mental health diagnosis among veterans served by VA is posttraumatic stress disorder (PTSD; Department of Veterans Affairs, 2015). Since 2002, VA has treated nearly 380,000 OEFOIF-OND veterans with PTSD. There also continues to be large cohorts of veterans from the Vietnam War and Operation Desert Storm who are in need of treatment for PTSD. PTSD is associated with a number of harmful outcomes including poor physical health (Norman et al., 2006); comorbid mental health diagnoses (Kessler, Sonnega, Bromet, Hughes, & Nelson, 1995); substance abuse (Zatzick et al., 1997); as well as an increased risk of unemployment, homelessness, and divorce (Schnurr, Lunney, Bovin, & Marx, 2009). To help prevent these negative outcomes and meet the needs of a large influx of OEF-OIF-OND veterans in recent years, VA implemented a nationwide initiative to research and disseminate evidence-based treatments for PTSD. Two treatments highly promoted by VA are cognitive processing therapy (CPT) and prolonged exposure (PE). Much past research has demonstrated the efficacy of both CPT and PE (e.g., Alvarez et al., 2011; Chard, Schumm, Owens, & Cottingham, 2010; Chard, Ricksecker, Healy, Karlin, & Resick, 2012; Eftekhari et al., 2013; Goodson, Leftkowitz, Helstrom, & Gawrysiak, 2013; Jeffreys et al., 2014; Powers, Halpern, Ferenschak, Gillihan, & Foa, 2010; Resick, Nishith, Weaver, Astin, & Feuer, 2002; Tuerk et al., 2011; Yoder et al., 2012). However, some argue that the effectiveness of CPT and PE is exaggerated (e.g., Steenkamp, Litz, Hoge, & Marmer, 2015).

Rutt et al.

3

Steenkamp et al.’s (2015) meta-analysis of evidence-based treatments for PTSD contributed to an ongoing debate in the field. Steenkamp and coworkers argue that the way CPT and PE have been implemented across VA is not reflective of true evidence-based practice. They argue the national rollout efforts rarely take into account individual preferences or clinical judgment. Others (e.g., Norman et al., 2016) have defended the high promotion of CPT and PE within VA, arguing that Steenkamp’s meta-analysis omitted several important studies and unfairly disregarded significant available data. The present study sought to inform this discussion with an analysis of the effectiveness of CPT and PE and other relevant clinical outcomes in a large sample of veterans receiving treatment within VA. One limitation of the past literature involves generalizing findings from randomized controlled trials (RCTs) to the larger population of veterans who are currently receiving these treatments. Experimental controls have limited the external validity of past research. For example, many of the thousands of veterans participating in CPT or PE in VA may not have met the inclusion criteria for past studies. Furthermore, the supports provided to therapists providing the treatments in past studies such as weekly consultation or supervision are not reflective of the conditions under which therapists actually provide these treatments within VA. This study addressed these limitations by examining the absolute and relative effectiveness of CPT and PE in a large sample of veterans who received treatment in 13 VA facilities across 10 states. An additional limitation of past research on CPT and PE is that individuals of racial/ethnic minority groups have largely been underrepresented or even ignored. This is a common limitation of RCTs (American Psychological Association, 2003). Of the studies that investigated the efficacy or effectiveness of CPT and PE, three studies have directly tested the influence of race/ethnicity on treatment outcomes. Tuerk et al. (2011) found that race/ethnicity was related to significantly higher initial self-reported PTSD symptoms. Tuerk et al. (2011) did not find that race/ethnicity predicted any clinical or statistical differences after the patients completed PE. Consistent with the prior literature, Resick et al. (2008) found that minority status was significantly associated with a lower rate of treatment completion. Alvarez et al. (2011) did not find any significant treatment effects of CPT due to race/ethnicity. Many have argued that we must make our research of empirically supported treatments more inclusive of all patients, especially those who are from racial/ethnic minority groups (see e.g., Sue & Zane, 2006). The current study addressed this limitation of past research by evaluating the effectiveness of CPT and PE in a highly diverse sample of Caucasian, African American, and Hispanic veterans. The literature on CPT and PE has also suggested that certain groups of veterans respond differently to treatment. For example, there is evidence that veterans from the Vietnam War are less responsive to trauma-focused therapy than OEF-OIF veterans (Chard et al., 2010). Other studies have suggested that

4

Psychological Reports 0(0)

veterans with more severe PTSD symptoms may experience a significantly larger reduction in PTSD symptoms (Jeffreys et al., 2014; Schumm, Walter, & Chard, 2013). The present study included age and initial PTSD symptom severity in the analysis to add to this body of literature. We are unaware of past literature regarding comparative effectiveness of either treatment with respect to gender, so we also compared clinical outcomes between men and women. The primary analysis of this study involved using multilevel modeling (MLM) of outcome data from CPT and PE in a sample of 750 veterans. We created a two-level hierarchical model: repeated measures over time (Level 1) nested within individual patients (Level 2). We sought to evaluate the absolute and relative effectiveness of each treatment, as well as which patient variables (i.e., completion status, initial PTSD symptom severity, age, sex, and racial/ethnic status) were related to the reduction in PTSD symptoms. Some (e.g., Imel, Laska, Jakupcak, & Simpson, 2013) have raised concerns that CPT and PE lead to higher dropout rates than present-focused therapy for PTSD. In light of the current concern in the field regarding the level of engagement in CPT and PE, we conducted secondary analyses on engagement in treatment. Although we were unable to compare the dropout rate between traumafocused and present-centered therapies, we were able to estimate the level of engagement of CPT and PE within VA, measured by number of sessions attended. We also tested whether number of sessions attended significantly varied according to treatment, sex, and minority status. We included the following hypotheses in the present study: 1. Both CPT and PE will result in significant reductions in PTSD symptoms over time. 2. Both CPT and PE will be equally effective at reducing PTSD symptoms. 3. Lower session attendance will be related to significantly smaller declines in PTSD symptoms. 4. Higher initial PTSD symptom severity will be related to significantly greater declines in PTSD symptoms. 5. Younger veterans will experience significantly higher reductions in PTSD symptoms. 6. The reduction in PTSD symptoms over time will not be significantly related to sex. 7. The rate of change in PTSD symptoms is predicted to be the same across Caucasian, African American, and Hispanic veterans. 8. Number of sessions attended will not significantly vary across treatment type, sex, or race/ethnicity. We discuss the significance of these findings in light of prior research on trauma-focused therapy, future directions for research, and implications for practice.

Rutt et al.

5

Method Sampling procedures and patient characteristics The study involved a retrospective chart review. In keeping with the principles of effectiveness research, exclusionary criteria for this study were minimal. The goal was to reflect as accurately as possible the overall population of veterans in VA settings that have received CPT and PE in individual therapy. All data were obtained from the VA Informatics and Computing Infrastructure following approval from the VA Eastern Kansas IRB and Human Subjects Committee of Lawrence. Chart reviews were conducted using the VA’s Compensation and Pension Records Interchange. To obtain a sample high in racial/ethnic diversity, the data selection procedure targeted specific Veterans Integrated Service Networks known to be high in diversity according to the 2010 U.S. Census. Veterans who received treatments in VA Medical Centers and Community-Based Outpatient Clinics in Arizona (N ¼ 306), California (N ¼ 178), Georgia (N ¼ 67), Missouri (N ¼ 54), Wisconsin (N ¼ 42), Illinois (N ¼ 39), Texas (N ¼ 29), and Alabama (N ¼ 12) were included in this study. Decisions regarding the assignment of participants to treatment groups were based on the clinical judgment of mental health providers, treatment teams, as well as the consent of the individual patients. The main inclusion criterion was whether patients had received CPT or PE in individual therapy. Patients who completed the Posttraumatic Stress Disorder Checklist (PCL; Weathers, Litz, Herman, Juska, & Keane, 1993) three or more times within a six-month period from 1 January 2006 to 31 July 2013 were randomly selected for a chart review. This is due to the fact that veterans who completed the PCL three or more times within a six-month period were more likely to be engaging in one of the evidence-based treatments for PTSD. The date range was chosen due to the adoption of new DSM 5 criteria. At the time of this study, the DSM 5 version of the PCL was not yet developed. Two thousand thirty patients’ charts were reviewed. One thousand thirty patients were identified as having received either CPT or PE. Eight hundred forty patients received one of those treatments in individual therapy. If patients completed both therapies, data from the second therapy they received were excluded from the analysis. Ninety patients were excluded from statistical analyses for the following reasons: (1) the patient’s initial measurement did not take place until after the third session (n ¼ 20), (2) the treatment consisted of a combination of individual and group psychotherapy (n ¼ 23), (3) there was a gap of six or more sessions between measurements (n ¼ 29), or (4) patients who received more than 30 weeks of individual therapy (n ¼ 18). PCL scores for 38 patients who received more than 14 sessions of psychotherapy were included. However, scores from session 15 and beyond were not included in the statistical analysis. These scores were removed because the trajectory of symptoms reported by these patients were not reflective of the rest of the group and held the potential to bias

6

Psychological Reports 0(0)

the results. Twenty-nine of these patients received PE while eight had received CPT. Demographic information and descriptive statistics for the final sample of 750 veterans are located in Table 1.

Measures Time. Time was operationalized as number of weeks and was included as a Level 1 predictor. Time was centered so that time ¼ 0 was session 1 of the treatment received by each patient.

Table 1. Demographics of sample and descriptive statistics (N ¼ 750). Variable Sex Men Women Age (years) Race/ethnicity White African American Hispanic Other Initial total PCL score Number of sessions Cognitive processing therapy Prolonged exposure

N

%

625 125

83.3 16.7

440 122 126 62

Race/ethnicity White African American Hispanic Other Completed full therapy protocol? Yes No

SD

46.01

15.55

61.73

12.15

8.54 7.98

3.78 3.52

58.7 16.3 16.8 8.3

CPT Variable

M

PE

N

%

N

%

225 47 71 33

30.0 6.3 9.5 8.8

215 75 55 29

28.7 10.0 7.3 7.8

148 228

39.4 60.6

209 165

55.9 44.1

CPT: cognitive processing therapy; PCL: Posttraumatic Stress Disorder Checklist; PE: PE: prolonged exposure.

Rutt et al.

7

Sessions attendance. Session attendance was operationalized as a binary variable for the MLM. This variable was included in analyses as a Level 2 predictor. Patients who completed a full protocol of CPT (n of sessions  12) and PE (n of sessions  8) were considered treatment completers (completer ¼ 1). Patients who received fewer than the above number of sessions in each treatment were considered treatment noncompleters (noncompleter ¼ 0). In the secondary analysis of engagement in treatment, engagement was operationalized as number of sessions attended. Initial PTSD symptom severity. Initial PTSD symptom severity was defined as the first PCL measurement available for each patient. It was included in the analyses as a Level 2 predictor. Session 1 PCL scores were available for 83.5% (n ¼ 626) of the participants. Session 2 PCL scores (13.2% of sample; n ¼ 99) and session 3 PCL scores (3.3% of sample; n ¼ 25) were used when session 1 PCL scores were unavailable. Initial PCL scores were included in all models as a mean-centered variable. Sex. Veterans were identified as either male or female based on the information in their charts. Sex was operationalized as a binary variable (Male ¼ 1, Female ¼ 0). Age. Age was operationalized as the number of years old the patient was at the beginning of treatment. It was included as a mean-centered variable. Treatment. The treatment modality for all patients was individual psychotherapy. Treatment was operationalized as participation in CPT (Treatment ¼ 1) or PE (Treatment ¼ 0) and was included as a Level 2 predictor. For patients who received both CPT and PE, only the data for the first psychotherapy they participated in were included in the analysis Race/ethnicity. The race/ethnicity of each client was included as an additional independent variable. Due to their prevalence within the VA system, three racial/ethnic groups were included in the study: Caucasians, African Americans, and Hispanics. These variables were reported by veterans upon their entry into the VA system. African American status and Hispanic status were included as two separate binary variables in the MLM. Both variables were included as Level 2 predictors. In the secondary analysis of treatment engagement, race/ethnicity was entered as a binary variable indicating majority and minority racial/ethnic status. Posttraumatic Stress Disorder Checklist. The primary dependent variable for this study was the PCL (Weathers et al., 1993). The PCL is a 17-item self-report measure based on the diagnostic criteria for PTSD. At the time of this study, the

8

Psychological Reports 0(0)

PCL was being updated for revised DSM 5 criteria. Results in this study are based on the old DSM-IV-TR criteria for PTSD. Scores range from 17 to 85, with higher scores indicating more severe PTSD symptoms. The PCL’s reliability and validity are well established (Weathers et al., 1993). Weathers et al. (1993) recommend a cut score of 50 for the diagnosis of PTSD. The National Center for PTSD (2014) recommends a decline of five points as a minimum threshold for determining whether a person has responded to treatment. A 10-point decline is the minimum threshold for clinically meaningful change.

Treatments CPT. CPT is a 12-session manualized treatment for PTSD adapted for veterans by Resick, Monson, and Chard (2007). The main components of the intervention involve psychoeducation regarding PTSD symptoms, Socratic questioning of assimilated and overaccommodated cognitive distortions (i.e., stuck points), and teaching clients cognitive therapy techniques to challenge distorted and extreme thinking. The therapy emphasizes five common themes that affect the following trauma: safety, trust, power and control, esteem, and intimacy. There are currently two versions of the treatment. The original version (CPT) involves completing two written trauma accounts while a cognitive-only version (CPT-C) does not include any written trauma accounts. This study only includes veterans who participated in the original version of CPT. PE. PE is typically conducted in 8 to 15 sessions that are between 60 and 90 minutes in length (Foa, 2011). The core components of PE include breathing retraining, psychoeducation about PTSD, imaginal exposures of the most traumatic event, and in vivo exposures of people, places, and situations that were previously avoided (Hembree, Rauch, & Foa, 2003). Treatment provided. During the chart review, the investigators verified via standard note templates used in each VA facility that either CPT or PE was provided to all participants. Providers. Veterans received treatment from psychologists (N ¼ 72), unlicensed psychology trainees (N ¼ 46), social workers (N ¼ 31), and other allied healthcare professionals (N ¼ 13).

Data analysis All statistical tests were conducted using version 20 of SPSS. Prior to conducting MLM, we conducted an analysis of pre–post PCL scores with the intent-to-treat sample as well as treatment completers. We then determined whether there was a significant difference in treatment engagement with respect to treatment type,

Rutt et al.

9

sex, and minority status. All statistical tests were two-tailed. Familywise Type 1 error was .05. MLM analyses. The MLM procedure followed methods described by Singer and Willett (2003) and Peugh (2010). Peugh (2010) outlines a number of steps for conducting a multilevel model including picking a parameter estimator, determining whether multilevel modeling is needed, building each level of the model, reporting multilevel effect sizes, and testing competing models using the likelihoods ratio test. We tested an unconditional model (i.e., a one-way random effects analysis of variance) to calculate the intraclass correlation coefficient (ICC) to determine whether MLM was an appropriate statistical procedure for the data. The ICC for a level is the percentage of total variance accounted for by that level in the model. The total variance in a two-level hierarchical model is given by 2 2  2 ¼ time þ patient

2 time is the amount of variance in PCL scores due to time. Variance in PCL scores due to time is typically a reflection of a patient’s participation in psycho2 therapy. patient is the amount of variance in PCL scores due to patients. Variance at this level may be due to individual differences between patients such as trauma history, level of PTSD symptom severity, or level of engagement in the treatment. In a two-level hierarchical model, the total variance in PCL scores due to patients can be defined as

2 ICCpatient ¼ patient



2 2 þ time patient



We ran a one-way random effects analysis of variance with two levels (Level 1 ¼ time; Level 2 ¼ patients) with no predictors at either level. The purpose of this approach is to determine what percentage of change in PCL scores was due to each level. Level 2 variance must be high enough in order to use MLM. Unfortunately, there is no agreed-upon standard that tells us how large an ICC needs to be in order to justify using multilevel modeling. However, an ICC as small as 0.05 can indicate that if nestedness of the data is ignored, hypothesis tests may be invalidated if multilevel modeling is not used (Kreft & de Leeuw, 1998). An additional preliminary step involved determining which variables would be fixed across participants (i.e., fixed effects) and which variables would be allowed to vary across participants (i.e., random effects). Preliminary models failed to converge when the patient-level variables (i.e., treatment type, race/ethnicity, age, sex) were included as both fixed and random effects. Due to this fact, we only included time as a random variable. This allowed for each

10

Psychological Reports 0(0)

individual to have a different rate of reduction in PTSD symptoms over time. We used the full information maximum likelihood method when estimating parameters in the multilevel model. Residuals were assumed to be uncorrelated in all analyses. Summary of models tested with MLM. We used a linear slope for all models. The initial model tested the relationship between PTSD symptoms and time, treatment type (CPT vs. PE), treatment completion, initial PTSD symptom severity, age, sex, African American status, and Hispanic status. We included a main effect and interaction effect parameters to test the significance of each variable. For all parameter estimates, a p value of .05 or below was considered statistically significant. Based on the results of the initial model, we removed all nonsignificant fixed effects to create a more parsimonious model. The final model included time, treatment type, completion status, and PTSD symptom severity as fixed effects and time as a random effect.

Results Analysis of intent-to-treat sample and treatment completers. Pre- and post-PCL scores for the intent-to-treat sample and for treatment completers are shown in Table 2. The within-group Hedge’s g for CPT in the intent-to-treat sample was 0.63. The within-group Hedge’s g for PE in the intent-to-treat sample was 0.77. Overall effect sizes were higher for treatment completers. The within-group Hedge’s g for CPT for treatment completers was 1.03. The within-group Hedge’s g for PE for treatment completers was 0.99. Between-group Hedge’s g was small (g ¼ 0.11 for ITT; g ¼ 0.14 for completers). Based on Weathers et al.’s (1993) cut score of

Table 2. CPT and PE mean scores over time and unbiased Hedge’s g: Intent-to-treat sample and completers. Pretreatment n Intent-to-treat sample CPT 374 PE 367 Completer sample CPT 148 PE 209

Posttreatment

M

SD

M

SD

Within-group Hedge’s g

61.57 62.01

12.29 12.02

51.44 49.54

17.80 17.41

0.63 0.77

60.97 62.74

12.05 12.09

43.72 46.06

16.39 16.88

1.03 0.99

CPT: cognitive processing therapy; PE: prolonged exposure.

Rutt et al.

11

50 for the PCL, 34.5% of veterans who completed CPT still met diagnostic criteria for PTSD and 39.2% of veterans who completed PE still met diagnostic criteria for PTSD. According to Weathers et al. (1993), a decline of 10 or more points on the PCL is a clinically meaningful decrease. Using this definition, 33.1% of veterans who completed CPT and 34.0% of veterans who completed PE did not have a meaningful decrease in the PCL scores. Preliminary MLM analyses. The ICC for patients was 67.3%(166.75/ [166.75 + 80.80] ¼ .673). This indicates that 67.3% of the variance in PTSD symptoms was due to patient characteristics. The ICC for patients indicated that MLM was an appropriate statistical method to analyze the data from this study (Peugh, 2010). Absolute and relative effectiveness of CPT and PE. Table 3 shows the results from the initial and final MLM. Both treatments resulted in significant decreases in PTSD symptoms over time. There was no significant difference in baseline PCL scores between the two treatments, (B ¼ 0.32, SE ¼ 0.40, p > .05). Figure 1 shows how the final model predicts the PCL scores for veterans in CPT and PE. There was a statistically significant interaction between time and treatment (B ¼ 0.22, SE ¼ 0.09, p < .05). According to the final MLM model, veterans in PE improved significantly faster than veterans enrolled in CPT. However, the model predicts the PCL score after 12 sessions is 54.8 for CPT and 52.7 for PE. This difference is unlikely to be clinically meaningful. Due to this finding, we concluded that both treatments were equally effective. Hypotheses 1 and 2 were supported. Session attendance. Figure 1 shows the predictions of the final model for treatment completers and noncompleters. Veterans attended a mean (SD) of 8.54 (3.78) sessions of CPT and 7.98 (3.52) sessions of PE. There was no significant difference in initial PTSD symptom severity between veterans who completed treatment and veterans who did not complete treatment (B ¼  0.24, SE ¼ 0.40, p > .05). There was a significant Completion Status  Time interaction (B ¼  0.37, SE ¼ 0.02, p < .001). This indicates that the rate of PTSD symptom reduction was significantly higher for patients who completed treatment. This interaction is shown in Figure 1. Hypothesis 3 was supported. Initial PTSD symptom severity. There was a significant main effect for initial PTSD symptom severity (B ¼ 0.97, SE ¼ 0.02, p < .001). There was also a significant Initial PTSD Symptom Severity  Time interaction (B ¼  0.01, SE ¼ 0.004, p < .01). Patients with higher initial self-reported PTSD symptoms tended to experience a larger decrease in symptoms over the course of the therapy. However, due to the small size of this finding, we did not consider this finding clinically meaningful. Hypothesis 4 was not supported.

12

Psychological Reports 0(0)

Table 3. Results from multilevel models of CPT and PE. Parameters Regression coefficients (fixed effects) Intercept Week CPT treatment main effect CPT treatment interaction Completer main effect Completer interaction Initial PTSD score main effect Initial PTSD score interaction African American main effect African American interaction Hispanic main effect Hispanic interaction Male main effect Male interaction Age main effect Age interaction Variance components (random effects) Level 1 residual (r2) Level 2 linear intercept ( 00) Level 2 linear slope ( 11) Level 2 linear covariance ( 01) Model summary Deviance statistic AIC BIC Number of estimated parameters Method of estimation

Initial model

62.06 0.94 0.28 0.23 0.28 0.38 0.97 0.01 0.55 0.11 0.30 0.15 0.42 0.10 0.01 .05). These findings suggest there was no significant difference between men and women with respect to initial PTSD symptom severity or changes in PTSD symptoms over time during treatment. Hypothesis 6 was supported. Race/ethnicity. The MLM model did not detect significant baseline differences in PCL scores for African American veterans (B ¼ 0.55, SE ¼ 0.56, p > .05) or Hispanic veterans (B ¼ 0.30, SE ¼ 0.54, p > .05). The African American  Time interaction was not statistically significant (B ¼ 0.11, SE ¼ 0.13, p > .05). The Hispanic  Time interaction was also not statistically significant (B ¼ 0.15, SE ¼ 0.13, p > .05). Thus, there were no significant differences with respect to the initial PTSD symptom severity or the rate of reduction in PTSD symptom reduction between Caucasian, African American, and Hispanic veterans. Hypothesis 7 was supported. Analysis of treatment engagement. An independent samples t test indicated that the number of sessions attended for CPT (M ¼ 8.47, SD ¼ 4.41) and PE (M ¼ 8.77, SD ¼ 4.12) was not significantly different, t(748) ¼ 0.949, p > .05. We subjected number of sessions attended to a two-way analysis of variance having two levels

14

Psychological Reports 0(0)

of treatment (CPT or PE) and two levels of sex (Male and Female). There were no significant main effects or interactions, F (3, 746) ¼ 0.985, p > .05. There were no significant differences in number of sessions attended for men in CPT (M ¼ 8.75, SD ¼ 4.22), women in CPT (M ¼ 8.83, SD ¼ 3.74), men in PE (M ¼ 8.59, SD ¼ 4.45), or women in PE (M ¼ 7.63, SD ¼ 4.06). We also subjected number of sessions to a two-way analysis of variance with two levels of treatment (CPT and PE) and minority status (Caucasian veterans and minority veterans). There were no significant main effects or interactions, F (3, 746) ¼ 0.642, p > .05. There were no significant differences in number of sessions attended for Caucasian Veterans in CPT (M ¼ 8.87, SD ¼ 4.14), minority veterans in CPT (M ¼ 8.60, SD ¼ 4.09), Caucasian veterans in PE (M ¼ 8.32, SD ¼ 4.17), or minority veterans in PE (M ¼ 8.68, SD ¼ 4.73). Hypothesis 8 was supported.

Discussion The current study adds to the literature of the effectiveness of CPT and PE for the treatment of military-related PTSD. The absolute effectiveness of both treatments in the present study was consistent with prior research. All past RCTs of CPT (e.g., Monson et al., 2006) and PE (e.g., Foa, Hembree, Cahill, Rauch, & Riggs, 2005) have found that patients who receive either treatment tend to experience a reduction in self-reported PTSD symptoms over time. The current study’s results also align with prior research that included veterans receiving care in conditions more reflective of clinical practice (e.g., Chard et al., 2012; Jeffreys et al., 2014, Tuerk et al., 2011). After controlling for the number of sessions attended and initial PTSD symptom severity, veterans who completed therapy in the present study experienced an average reduction in PTSD symptoms of 20% in CPT and 22% in PE. Percent reductions in the present study for CPT align with the results of past effectiveness studies (e.g., Alvarez et al., 2011; Jeffreys et al., 2014). The 23% reduction in PTSD symptoms reported by veterans who completed PE was close to Eftekari et al.’s (2011) finding of 24%. The present study found smaller reductions in PTSD symptoms than past RCTs on CPT (e.g., Forbes et al., 2012; Monson et al., 2006) and PE (e.g., Goodson et al., 2013; Tuerk et al., 2011). Treatment effects in psychotherapy outcome research tend to be weaker in conditions where there are fewer experimental controls. The present study’s sample was highly heterogeneous, which tends to reduce observed treatment effects. Patients in the present study were treated by providers of varying levels of training ranging from licensed providers to trainees. Therapists in past studies were either restricted to doctoral level clinicians (e.g., Monson et al., 2006), experienced members of clinical teams (e.g. Tuerk et al., 2011), or receiving clinical training in CPT or PE concurrently to their clinical work

Rutt et al.

15

(e.g., Chard et al., 2012). It is likely that many providers in the present study were not enrolled in CPT or PE consultation, which may also influence the effectiveness of both treatments. Despite losing some of their potency, both CPT and PE remained effective treatments for PTSD when practiced under conditions highly reflective of clinical practice. With respect to the relative effectiveness of CPT and PE, there have been some conflicting findings in the literature. In the only RCT that compared both treatments, Resick et al. (2002) found there was insufficient evidence to suggest that CPT and PE led to significantly different outcomes. However, Jeffreys et al. (2014) concluded that PE was significantly more effective than CPT in a large sample of over 500 veterans after controlling for age, service era, and race/ethnicity. The current study failed to replicate Jeffreys et al.’s (2014) finding. There was a small but nonclinically significant difference in effectiveness between CPT and PE in the present study. However, the size of the effect was small and unlikely to be clinically meaningful. We conclude that, on average, both treatments are equally effective in conditions reflective of clinical practice. The finding of equivalent effectiveness across both treatments is consistent with past studies on PTSD treatments (e.g., Wampold et al., 2010) and the larger body of effectiveness research (e.g., Stiles, Barkham, Mellor-Clark, & Connell, 2008). When the effectiveness of two or more bona fide psychotherapies is compared in naturalistic settings, they tend to be equally effective. By verifying the delivery of the intended treatment via confirmation of CPT and PE note templates, the present study addresses one important limitations of the past effectiveness literature. All providers in the present study were either full-time staff who received training and consultation in the CPT or PE rollouts or, in the case of trainees, were practicing under clinical supervision by licensed mental health practitioners. The most clinically meaningful interaction in the present study was related to whether patients completed treatment. Patients who completed treatment tended to have clinically significant declines in PTSD symptoms, whereas patients who did not complete treatment experienced little to no change in their symptoms. However, dropout rates in the present study appear to be higher than in past studies. When dropout is defined as failure to complete a full therapy protocol (8 sessions for PE and 12 sessions for CPT), 44% of patients dropped out from PE while 61% of patients dropped out from CPT. It should be noted that there was no significant difference in engagement in each treatment, which is defined as number of sessions attended. Estimates of dropout in past studies have been much smaller. For example, a recent meta-analysis on dropout in PTSD treatments estimated that 18% of patients dropout prematurely (Imel et al., 2013). However, the prevalence of dropout from PE in the present study was similar to Jeffreys et al.’s (2014) finding. The higher dropout rate in the present study could be due to a number of factors. Veterans with co-occurring substance use and

16

Psychological Reports 0(0)

complex presenting concerns including multiple psychiatric diagnoses were not screened out of the present study. Barriers related to low socioeconomic status and a high rate of comorbid physical health problems may have also led to a higher dropout rate. Due to the relative paucity of research on the relationship between initial PTSD symptom severity and the trajectory of PTSD symptom reduction during treatment, the finding that there is a significant interaction between the initial PCL score and time has relatively few precedents in the literature. However, it should be noted that the role of initial PTSD symptoms severity was small and unlikely to be clinically meaningful. Veterans beginning treatment with more severe PTSD symptoms may be more likely to experience declines in their symptoms simply due to regression to the mean. Regarding the association between race/ethnicity and treatment outcomes, the results from the present study suggest that Caucasian veterans, African American veterans, and Hispanic veterans all experienced similar rates of improvement in their PTSD symptoms. The present study had one of the largest samples of African American and Hispanic veterans in the literature to date, so this finding is encouraging. This finding aligns with past research on the association between race and response to trauma focused therapy. Tuerk et al. (2011) found that race was not a significant predictor of response to PE, while Jeffreys et al. (2014) found no significant difference in outcomes between Caucasian and Hispanic veterans who participated in CPT or PE. Due to the fact that this is an effectiveness study, a main limitation of this study is the lack of an experimental design. Patients were not randomly assigned into treatment groups. There was no control group of untreated patients, a typical limitation of most effectiveness research. Although CPT and PE have been implemented across the VA system in a standardized format, presentfocused therapy has not received systematic implementation across VA. Thus, we were not able to identify a single control group with uniform procedures to include in this analysis. This is the primary limitation of the present study. One cannot conclude from this study alone that clinically significant declines in PCL scores are due to the trauma-focused nature of the treatments. That conclusion would rest on the body of literature that has come before which has demonstrated the efficacy of CPT and PE in experimental conditions. There were also multiple unaccounted for variables in the present study that may have influenced the progress including lack of control for therapist effects, alternative explanations for recovery (e.g., pharmacotherapy, participation in additional psychotherapies), trauma type, trauma severity, homework completion, patient comprehension of topics covered in therapy, and motivational factors. Furthermore, all data on patient progress in therapy were obtained from selfreport measures. Limitations of self-report measures include biases due to social desirability and different response styles while completing the assessments, which increase measurement error. Finally, while the presence of CPT or PE template

Rutt et al.

17

notes was used to identify which treatment was delivered, there were no additional measures of adherence to the therapy protocols. There are a number of potential future issues regarding trauma-focused therapy that additional research could address in light of the findings from the present study. The finding that completion status was the most clinically relevant factor that was related to outcomes in therapy raises an important question: What factors lead to favorable versus nonfavorable responses to treatment? Dropout from both treatments is a significant issue. About half of the participants in the present study did not complete a full therapy protocol. Furthermore, many patients who complete CPT or PE may still have clinically significant PTSD symptoms. Additional research should seek to help this sizeable group of veterans. The results from this analysis have many implications for therapists who provide trauma-focused therapy to veterans. Perhaps the most encouraging finding for mental health practitioners is that the effectiveness of both CPT and PE was confirmed in conditions that are reflective of clinical practice. When patients are assigned to one of these treatments based on their preferences and based on the ability of their therapist to provide the treatment, patients tend to have beneficial outcomes. However, although some veterans may experience a full remission of symptoms, many will not. The findings from this study indicate that clinicians can communicate that if veterans participate in the sessions and do work in between sessions, they can expect a modest improvement in their symptoms. The reduced effectiveness of CPT and PE in the present study when compared to past RCTs speaks to an ongoing debate in the field regarding the overall effectiveness of trauma-focused therapy. Proponents of evidence-based, traumafocused therapies consider CPT and PE to be the most effective psychotherapies for PTSD. Others such as Steenkamp et al. (2015) argue that the definition of evidence-based treatments is too limited and the effectiveness of CPT and PE is exaggerated. The results of the present study, in particular the high dropout rate from treatment, speak to the many challenges that veterans have completing these treatments. Whether reduced treatment effects are due to aspects of the treatments themselves or due to external factors such as SES or comorbid conditions will remain a subject of debate. However, the results of the present study do suggest providers should communicate realistic expectations to veterans regarding how much they can expect to improve. While CPT and PE are not panaceas for PTSD, this study shows that most veterans who complete one of these treatments can expect a noticeable improvement in symptoms. Acknowledgments This research is the result of work supported with resources and the use of facilities at the VA Eastern Kansas Healthcare System. The authors would like to thank Vicky Peyton, David Hansen, Bruce Wampold, and Tracela White for their assistance with this study.

18

Psychological Reports 0(0)

Declaration of Conflicting Interests The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding The author(s) received no financial support for the research, authorship, and/or publication of this article.

References Alvarez, J., McClean, C., Harris, A., Rosen, C., Ruzek, J., & Kimerling, R. (2011). The comparitive effectiveness of cognitive processing therapy for male veterans treated in a VHA posttraumatic stress disorder residential rehabilitation program. Journal of Consulting and Clinical Psychology, 79(5), 590–599. doi:http://dx.doi.org/10.1037/ a0024466 American Psychological Association (2003). Guidelines on multicultural education, training, research, practice, and organizational change for psychologists. American Psychologist, 58(5), 377–402. doi:http://dx.doi.org/10.1037/0003-066X.58.5.377 Chard, K. M., Schumm, J. A., Owens, G. P., & Cottingham, S. M. (2010). A comparison of OEF and OIF veterans and Vietnam veterans receiving cognitive processing therapy. Journal of Traumatic Stress, 23(1), 25–32. doi: 10.1002/jts.20500 Chard, K. M., Ricksecker, E. G., Healy, E. T., Karlin, B. E., & Resick, P. A. (2012). Dissemination and experience with cognitive processing therapy. Journal of Rehabilitation Research & Development, 9(5), 667–678. doi:http://dx.doi.org/10.1682/ JRRD.2011.10.0198 Department of Veterans Affairs. (2015). Analysis of VA health care utilization among OEF-OIF-OND veterans. Retrieved from http://www.publichealth.va.gov/epidemiology/reports/oefoifond/health-care-utilization/ Eftekhari, A., Ruzek, J. I., Crowley, J. J., Rosen, C. S., Greenbaum, M. A., & Karlin, B. E. (2013). Effectiveness of national implementation of prolonged exposure therapy in veterans affairs care. JAMA Psychiatry, 70(9), 949–955. doi:http://dx.doi.org/10.1001/ jamapsychiatry.2013.36 Foa, E. B., Hembree, E. A., Cahill, S. P., Rauch, S. A., & Riggs, D. S. (2005). Randomized trial of prolonged exposure for posttraumatic stress disorder with and without cognitive restructuring: Outcome at academic and community clinics. Journal of Consulting and Clinical Psychology, 73(5), 953–964. doi:http://dx.doi.org/10.1037/ 0022-006X.73.5.953 Foa, E. B. (2011). Prolonged exposure therapy: Past, present, and future. Depression and Anxiety, 28(12), 1043–1047. doi: 10.1002/da.20907 Forbes, D., Lloyd, D., Nixon, R. D., Elliott, P., Varker, T., Perry, D., & Creamer, M. (2012). A multisite randomized controlled effectiveness trial of cognitive processing therapy for military-related posttraumatic stress disorder. Jounal of Anxiety Disorders, 26(3), 442–452. doi: 10.1016/j.janxdis.2012.01.006 Goodson, J. T., Leftkowitz, C. M., Helstrom, A. W., & Gawrysiak, M. J. (2013). Outcomes of prolonged exposure therapy for veterans with posttraumatic stress

Rutt et al.

19

disorder. Journal of Traumatic Stress, 26(4), 419–425. doi:http://dx.doi.org/10.1002/ jts.21830 Hembree, E. A., Rauch, S. A. M., & Foa, E. B. (2003). Beyond the manual: The insider’s guide to prolonged exposure therapy for PTSD. Cognitive and Behavioral Practice, 10, 22–30. Imel, Z. E., Laska, K., Jakupcak, M., & Simpson, T. L. (2013). Meta-analysis of dropout in treatments for posttraumatic stress disorder. Journal of Consulting and Clinical Psychology, 81(3), 394–404. doi:10.1037/a0031474 Jeffreys, M. D., Reinfeld, C., Nair, P. V., Garcia, H. A., Mata-Galan, E., & Rentz, T. O. (2014). Evaluating treatment of posttraumatic stress disorder with cognitive processing therapy and prolonged exposure in a VHA specialty clinic. Journal of Anxiety Disorders, 28, 108–114. doi:10.1016/j.janxdis.2013.04.010 Kessler, R. C., Sonnega, A., Bromet, E. J., Hughes, M., & Nelson, C. B. (1995). Posttraumatic stress disorder in the National Comorbidity Survey. Archives of General Psychiatry, 52(12), 1048–1060. Kreft, I. G. G., & de Leeuw, J. (1998). Introducing multilevel modeling. Thousand Oaks, CA: Sage. Monson, C. M., Schnurr, P. P., Resick, P. A., Friedman, M. J., Young-Xu, Y., & Stevens, S. P. (2006). Cognitive processing therapy for veterans with military-related posttraumatic stress disorder. Journal of Consulting and Clinical Psychology, 74(5), 898–907. doi:10.1037/0022-006X.74.5.898 National Center for Veterans Analysis and Statistics. (2016). Department of Veterans Affairs Statistics at a Glance. Retrieved from http://www.va.gov/vetdata/ Norman, S. B., Means-Christensen, A. J., Craske, M. G., Sherbourne, C. D., Roy-Byrne, P. P., & Stein, M. B. (2006). Associations between psychological trauma and physical illness in primary care. Journal of Traumatic Stress, 19(4), 461–470. Norman, S., Chard, K., Rauch, S., Foa, E., Monson, C., & Resick, P. (2016). True evidence-based care for posttraumatic stress disorder in military personnel and veterans—A reply. Retrieved from http://archpsyc.jamanetwork.com/article.aspx? articleid=2491947 Peugh, J. L. (2010). A practical guide to multilevel modeling. Journal of School Psychology, 48(1), 85–112. doi:http://dx.doi.org/10.1016/j.jsp.2009.09.002 Powers, M. B., Halpern, J. M., Ferenschak, M. P., Gillihan, S. J., & Foa, E. B. (2010). A meta-analytic review of prolonged exposure for posttraumatic stress disorder. Clinical Psychology Review, 30(6), 635–641. doi:http://dx.doi.org/10.1016/j. cpr.2010.04.007 Resick, P. A., Nishith, P., Weaver, T. L., Astin, M. C., & Feuer, C. A. (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. doi:http://dx. doi.org/10.1037//0022-006X.70.4.867 Resick, P. A., Monson, C. M., & Chard, K. M. (2007). Cognitive processing therapy: Veteran/military version. Washington, DC: Department of Veterans Affairs. Resick, P. A., Galovski, T. E., Uhlmansiek, M. O. B., Scher, C. D., Clum, G. A., & Young-Xu, Y. (2008). A randomized clinical trial to dismantle components of

20

Psychological Reports 0(0)

cognitive processing therapy for posttraumatic stress disorder in female victims of interpersonal violence. Journal of Consulting and Clinical Psychology, 76(2), 243–258. doi:http://dx.doi.org/10.1037/0022-006X.76.2.243 Schnurr, P., Lunney, C., Bovin, M., & Marx, B. (2009). Posttraumatic stress disorder and quality of life: Extension of findings to veterans of the wars in Iraq and Afghanistan. Clinical Psychology Review, 29, 727–735. doi:10.1016/j.cpr.2009.08.006 Schumm, J. A., Walter, K. H., & Chard, K. M. (2013). Latent class differences explain variability in PTSD symptom changes during cognitive processing therapy for veterans. Psychological Trauma: Theory, Research, Practice, and Policy, 5(6), 536–544. doi:http://dx.doi.org/10.1037/a0030359 Singer, J. D., & Willett, J. B. (2003). Applied longitudinal data analysis: Modeling change and event occurrence. New York, NY: Oxford University Press. Steenkamp, M. M., Litz, B. T., Hoge, C. W., & Marmar, C. R. (2015). Psychotherapy for military-related PTSD, a review of randomized clinical trials. Journal of the American Medical Association, 314(5), 489–500. doi:10.1001/jama.2015.8370 Stiles, W. B., Barkham, M., Mellor-Clark, J., & Connell, J. (2008). Effectiveness of cognitive-behavioural, person-centred, and psychodynamic therapies in UK primary-care routine practice: Replication in a larger sample. Psychological Medicine, 38, 677–688. doi:10.1017/S0033291707001511 Sue, S., & Zane, N. (2006). Ethnic minority populations have been neglected by evidencebased practices. In J. Norcross, L. Beutler & R. Levant (Eds.), Evidence-based practices in mental health (pp. 329–337). Washington, DC: American Psychological Association. Tuerk, P. W., Yoder, M., Grubaugh, A. L., Myrick, H., Hamner, M. B., & Acierno, R. E. (2011). Prolonged exposure therapy for combat-related posttraumatic stress disorder: An examination of treatment effectiveness for veterans of the wars in Afghanistan and Iraq. Journal of Anxiety Disorders, 25(3), 397–403. doi:http://dx.doi.org/10.1016/j. janxdis.2010.11.002 Wampold, B., Imel, Z., Laska, K., Benish, S., Miller, S., Flu¨ckiger, C., . . ., Budge, S. (2010). Determining what works in the treatment of PTSD. Clinical Psychology Review, 30(8), 923–933. doi:10.1016/j.cpr.2010.06.005 Weathers, F. W., Litz, B., Herman, D., Juska, J., & Keane, T. (1993). PTSD Checklist—Civilian version. Boston, MA: National Center for PTSD. Yoder, M., Tuerk, P. W., Price, M., Grubaugh, A. L., Strachan, M., Myrick, H., & Acierno, R. (2012). Prolonged exposure therapy for combat-related posttraumatic stress disorder: Comparing outcomes for veterans of different wars. Psychological Services, 9(1), 16–25. doi:http://dx.doi.org/10.1037/a0026279 Zatzick, D. F., Marmar, C. R., Weiss, D. S., Browner, W. S., Metzler, T. J., Golding, J. M., & Wells, K. B. (1997). Posttraumatic stress disorder and functioning and quality of life outcomes in a nationally representative sample of male Vietnam veterans. American Journal of Psychiatry, 154(12), 1690–1695.

Author Biographies Benjamin T. Rutt obtained a doctorate in Counseling Psychology from the University of Kansas in 2015. He completed a pre-doctoral internship and

Rutt et al.

21

postdoctoral fellowship with the Washington DC, VA Medical Center. His postdoctoral fellowship included specialized training in the diagnosis and treatment of posttraumatic stress disorder (PTSD). Both his research and clinical work have focused on evidence-based therapies for PTSD. Currently, he works in private practice in Baltimore, MD. Mary E. Oehlert has served the past 30 years at the VA Eastern Kansas Health Care System where her career has focused on the provision of psychological services to veterans, research, and training the next generation of psychologists. Regarding the latter, she served 15 years as the Psychology Director of Training and APA Site Visitor. Currently, she serves as the associate chief of Staff for Research incorporating research as a training opportunity for interns and residents in various disciplines. Thomas S. Krieshok is a Kemper Teaching Fellow and a Budig Teaching Professor of Education in the Department of Educational Psychology at the University of Kansas, a fellow of the American Psychological Association and of the American Educational Research Association, and a founding member of the Society for Vocational Psychology. He has consulted extensively with Veterans Administration medical centers and with state agencies on vocational rehabilitation and counseling. His teaching and research interests focus on career decision-making and training issues in professional psychology. James W. Lichtenberg is an emeritus professor of Counseling Psychology and former associate dean for Graduate Programs and Research at the University of Kansas, where he also served for 34 years as his program’s director of training. He is a former vice president for Science and a past president of APA’s Division 17, the Society of Counseling Psychology. He is the 2017 recipient of APA’s Award for Distinguished Contributions of Applications of Psychology to Education and Training.

Effectiveness of Cognitive Processing Therapy and Prolonged Exposure in the Department of Veterans Affairs.

Objective This study evaluated the effectiveness of cognitive processing therapy and prolonged exposure in conditions reflective of current clinical p...
201KB Sizes 0 Downloads 6 Views