The Comparative Effectiveness of Cognitive Processing Therapy for MaleVeterans Treated in a VHA Posttraumatic Stress Disorder Residential

Rehabilitation Program

Jennifer Alvarez, Caitlin McLean,and Alex H. S. Harris

Veterans Affairs Palo Alto Health Care System

Craig S. Rosen and Josef I. RuzekVeterans Affairs Palo Alto Health Care System and Stanford

University

Rachel KimerlingVeterans Affairs Palo Alto Health Care System

Objective: To examine the effectiveness of group cognitive processing therapy (CPT) relative totrauma-focused group treatment as usual (TAU) in the context of a Veterans Health Administration(VHA) posttraumatic stress disorder (PTSD) residential rehabilitation program. Method: Participantswere 2 cohorts of male patients in the same program treated with either CPT (n � 104) or TAU (n �93; prior to the implementation of CPT). Cohorts were compared on changes from pre- to posttreatmentusing the PTSD Checklist (PCL; Weathers, Litz, Herman, Huska, & Keane, 1993) and other measuresof symptoms and functioning. Minorities represented 41% of the sample, and the mean age was 52 years(SD � 9.22). The CPT group was significantly younger and less likely to receive disability benefits forPTSD; however, these variables were not related to outcome. Results: Analyses of covariance controllingfor intake symptom levels and cohort differences revealed that CPT participants evidenced moresymptom improvement at discharge than TAU participants on the PCL, F(3, 193) � 15.32, p � .001, b �6.25, 95% CI [3.06, 9.44], and other measures. In addition, significantly more patients treated with CPTwere classified as “recovered” or “improved” at discharge, �2(1, N � 197) � 4.93, p � .032.Conclusions: There is still room for improvement, as substantial numbers of veterans continue toexperience significant symptoms even after treatment with CPT in a residential program. However, CPTappears to produce significantly more symptom improvement than treatment conducted before theimplementation of CPT. The implementation of this empirically supported treatment in VHA settings isboth feasible and sustainable and is likely to improve care for male veterans with military-related PTSD.

Keywords: PTSD, veterans, cognitive processing therapy, group/residential therapy, dissemination

Military service places individuals at high risk for exposure totrauma and for posttraumatic stress disorder (PTSD). Studies varybased on measurement and population sampled, but even the mostconservative estimates indicate that almost 18.7% of Vietnamveterans have reported symptoms consistent with PTSD at somepoint since their service (Dohrenwend et al., 2006). The NationalVietnam Readjustment Study reported lifetime PTSD prevalencerates of over 30% (Kulka et al., 1990). As soldiers return from Iraqand Afghanistan, it appears that a new generation of veterans is

struggling with this disorder and its consequences. A study ofArmy and Marine combat infantry units deployed to Iraq reporteda 18%–20% prevalence using a broad definition of PTSD (meetingDiagnostic and Statistical Manual of Mental Disorders [4th ed.;DSM–IV; American Psychiatric Association, 1994] symptom cri-teria according to the PTSD Checklist [PCL; Weathers, Litz,Herman, Huska, & Keane, 1993]) and 12%–13% using a narrowdefinition (meeting DSM–IV criteria and a severity score of at least50 on the PCL) 3–4 months after returning to the United States

This article was published Online First July 11, 2011.Jennifer Alvarez, Veterans Affairs Palo Alto Health Care System; Cait-

lin McLean, National Center for PTSD, Veterans Affairs Palo Alto HealthCare System; Alex H. S. Harris, Center for Health Care Evaluation,Veterans Affairs Palo Alto Health Care System; Craig S. Rosen, NationalCenter for PTSD and Center for Health Care Evaluation, Veterans AffairsPalo Alto Health Care System, and Department of Psychiatry and Behav-ioral Sciences, Stanford University; Josef I. Ruzek, National Center forPTSD, Veterans Affairs Palo Alto Health Care System, and Department ofPsychiatry and Behavioral Sciences, Stanford University; Rachel Ki-merling, National Center for PTSD and Center for Health Care Evaluation,Veterans Affairs Palo Alto Health Care System.

The views expressed in this article are those of the authors and do notnecessarily reflect the position or policy of the Department of VeteransAffairs, the U.S. government, or their respective institutions. This workwas supported in part by Department of Defense Grant W81XWH-08-2-0659 and by the Department of Veterans Affairs, Palo Alto Health CareSystem. We thank Kent Drescher, Patricia Resick, and the staff of theVeterans Affairs Palo Alto Heath Care System Trauma Recovery Programsfor their contributions to this article.

Correspondence concerning this article should be addressed to JenniferAlvarez, Department of Veterans Affairs Palo Alto Health Care System,795 Willow Road (352-117), Menlo Park, CA 94025. E-mail:jennifer.alvarez@va.gov

Journal of Consulting and Clinical Psychology In the public domain2011, Vol. 79, No. 5, 590–599 DOI: 10.1037/a0024466

590

(Hoge et al., 2004). A more recent investigation revealed 11.8% ofactive duty men and women reported PTSD immediately postde-ployment in Iraq, and 16.7% reported PTSD 3–6 months later(Milliken, Auchterlonie, & Hoge, 2007).

Not surprisingly, PTSD is prevalent among veterans served bythe Veterans Health Administration (VHA) health care system.Nearly one in four (24.5%) Iraq and Afghanistan veterans treatedby VHA have received a PTSD diagnosis (VHA Office of PublicHealth and Environmental Hazards, 2009), and among all veteransserved by VHA, the proportion diagnosed with PTSD increased by60% between 2001 (4.8%) and 2007 (7.6%). This increase is likelydue to multiple factors, including the following: (a) implementa-tion of mandated screening for PTSD and improved detection; (b)an influx of newly returning veterans with high rates of PTSD; (c)increased prevalence of PTSD diagnoses among Vietnam-era vet-erans, potentially due to retirement and/or the current wars; and (d)more Vietnam veterans enrolled after a policy change to beginproviding free diabetes care for those exposed to Agent Orange.

As a result, many VHA resources have been and will continueto be devoted to PTSD treatment. For example, large-scale dis-semination efforts have been executed throughout the VHA healthcare system to train clinicians in evidence-based treatment inter-ventions for PTSD (Karlin et al., 2010), such as cognitive process-ing therapy (CPT; Resick, Monson, & Chard, 2007; Resick,Nishith, Weaver, Astin, & Feuer, 2002; Resick & Schnicke, 1992)and prolonged exposure (PE; Foa et al., 1999; Foa, Rothbaum,Riggs, & Murdock, 1991), and to implement these interventions inexisting VHA treatment settings. A randomized, controlled trialhas provided promising evidence regarding the efficacy of CPTwith veteran populations. Monson et al. (2006) reported that CPTadministered in an individual, outpatient therapy setting was su-perior to a wait-list control in treating military-related PTSD inVHA patients.

It is extremely important to evaluate the comparative effective-ness and sustainability of new treatments for PTSD implementedin VHA settings. The randomized, controlled trial discussed aboveprovided valuable efficacy data but does not address whetherevidence-based treatments for PTSD can be successfully inte-grated into existing program structures. Two recent open trialsdemonstrate the feasibility of delivering CPT in VHA settings.Chard, Schumm, Owens, and Cottingham (2010) examined CPTdelivered in individual, outpatient VHA clinic settings and sug-gested that the intervention was effective overall and that Iraq andAfghanistan veterans may report fewer symptoms of PTSD post-treatment compared to Vietnam veterans. In a small-scale exami-nation of CPT delivered in a group format in the context of aresidential PTSD program for female veterans, Zappert andWestrup (2008) reported that veterans significantly improved on ameasure of PTSD symptoms after treatment. However, neither ofthese studies included a comparison group, so they could notassess the comparative effectiveness of CPT relative to treatmentas usual (TAU).

The implementation literature has emphasized that in addition toinitial staff training, elements such as ongoing technical support,feedback/refinement of the treatment package, and evaluation maybe important to implement effective health care interventions innew settings (Kilbourne, Neumann, Pincus, Bauer, & Stall, 2007).For example, one study indicated that on-going feedback, coach-ing, and/or supervision in addition to one-time training are an

essential feature in achieving and retaining new psychotherapeuticskills (Miller, Yahne, Moyers, Martinez, & Pirritano, 2004). Astrength of the current implementation study is that it incorporateselements identified as important by the literature, such as a de-scription of refinements to the treatment package and expert con-sultation.

This study evaluates implementation of CPT in a PTSD Resi-dential Rehabilitation Program in the VHA health care system. Theprimary aim was to determine whether CPT was more effectivethan the treatment that was being delivered prior to the implemen-tation effort. We employed a cohort design, comparing clinicaloutcomes for veterans treated with CPT groups to outcomes for aprior cohort of veterans treated in the same residential programwith trauma-focused groups conducted before the implementationof CPT. We hypothesized that the cohort treated with CPT wouldshow more improvement on clinical measures than the cohorttreated prior to the CPT implementation. The rationale for thishypothesis was derived from a review of previous research sug-gesting larger treatment effect sizes for CPT (Monson et al., 2006)compared to a similar type of trauma-focused group therapy as theTAU examined here (Schnurr et al., 2003). A secondary aim of thestudy was to determine how patient variables such as demograph-ics might be associated with variation in outcomes within the CPTcohort. To help assess whether CPT could be sustained effectively,another aim was to compare outcomes of patients treated in thefirst year of adoption, when clinicians were receiving regularexpert case consultation, and in subsequent years after intensiveconsultation ended.

Method

Design

This study employed a retrospective, quasi-experimental, cohortdesign within a VHA residential treatment program for PTSD.Data were obtained from the clinical database of the residentialtreatment program as described below in the Procedure section.Clinical outcomes for a cohort of patients treated followingprogram-wide implementation of CPT (CPT cohort) were com-pared to a historical control group of patients treated in the sameprogram prior to the implementation of CPT (TAU cohort). Datawere compared at two time periods: intake and discharge from theprogram. The primary outcome measure was the self-reportedseverity of PTSD symptoms. Secondary outcome measures wereself-reported depression, psychological distress, quality of life, andcoping. To identify potential sources of bias, we compared cohortson demographics and baseline clinical characteristics. Any signif-icant differences in demographic or baseline characteristics werecontrolled for in all analyses of between-group differences.

To determine the relative effectiveness of treatment during theCPT cohort, compared to during the TAU cohort, we comparedprimary and secondary outcome measures for each cohort at dis-charge, controlling for symptoms at intake and cohort differences.As a secondary analysis of relative treatment effectiveness, wecompared the proportions of patients with clinically significantimprovement on the primary outcome measure for both cohorts.To determine patient characteristics associated with treatment out-come in the CPT cohort, we examined the relationship betweendemographic variables and primary and secondary outcome mea-

591COMPARATIVE EFFECTIVENESS OF CPT FOR VETERANS

sures at discharge, controlling for intake symptoms. Implementa-tion factors were examined within the CPT cohort by assessing therelationship of regular therapist case consultation to primary andsecondary outcome measures.

Setting

This study examined implementation of CPT in a PTSD Resi-dential Rehabilitation Program in the VHA health care system,replacing TAU trauma-focused groups. This 60–90 day, residen-tial treatment program has existed since 1978 and provides treat-ment for men of all ages and eras of service with military-relatedPTSD and related problems. The program has a national catchmentarea, receiving referrals from VHA hospitals/clinics, VeteransCenters, and private practitioners around the country. Veterans arereferred to the residential program when a more intensive, resi-dential treatment environment is indicated. Often, this means thatPTSD symptoms have been treatment-refractory in outpatient set-tings, but referrals are also made directly from acute psychiatricinpatient or residential substance abuse settings or occasionallywhen time for treatment is limited by life circumstances (e.g.,individual has to return to work). Residents live in a therapeuticmilieu setting and participate in group interventions throughout theday and evening, including cognitive therapy, communicationskills, psychoeducation, process groups, parenting skills, and rec-reation therapy. A type of trauma-focused group therapy consistingof a life-span developmental model and incorporating some ele-ments of CBT had been provided in this program for approxi-mately 15 years but had never been empirically validated. In thefall of 2004, three clinical psychologists and one licensed clinicalsocial worker were trained by Patricia Resick, the developer ofCPT. These providers had all worked in this VHA PTSD residen-tial treatment program for at least 5 years (three providers hadworked in this program for more than 11 years) and were trainedor familiar with cognitive-behavioral therapy techniques but not

CPT specifically. These same four providers then began conduct-ing CPT in place of the previous type of trauma-focused grouptherapy. Over the course of the next year, providers in the programparticipated in weekly telephone case consultation with CPT ex-perts. Treatment providers were consistent across the entire studyperiod. Other than changes related to the implementation andintegration of CPT (described below), other changes to the resi-dential treatment program (e.g., other groups, staffing) were min-imal over the study period. In addition, the overall theoreticalorientation of the program was primarily cognitive-behavioral andremained constant over the entire study period. CPT/TAU groupswere the only form of trauma-focused treatment that patients inthis program received; none received individual therapy.

Participants

Outcome data for 104 male veterans treated with CPT groupsand 93 male veterans treated with TAU prior to the implementa-tion of CPT were examined for this study. Sample characteristicsare described in Table 1. Participants were two retrospectivecohorts of male veterans treated in the same residential treatmentprogram for PTSD in a VHA Medical Center 2 years before and 2years after CPT was disseminated and implemented in the program(September 2004). Inclusion criteria were as follows: (a) partici-pation in a trauma-focused group in this program during the studyperiod and (b) have previously provided informed consent thattheir data could be used for research purposes and, thus, partici-pated in routine data collection during their stay in the program.All participants who initiated treatment in a trauma-focused group(for either the CPT or TAU cohorts) were included in analyses,including those who terminated treatment early or were irregularlydischarged before completing a trauma-focused group. There wereno significant cohort differences in trauma group completion rate(CPT � 89.4%, TAU � 94.6%), �2(1, N � 197) � 1.779, p �.182. Similarly, there were no significant demographic or baseline

Table 1Participant Characteristics

Variable

CPT (n � 104) TAU (n � 93) Total (N � 197)

n (%) n (%) n (%)

Age in years, M � SD� 50.20 � 11.55 54.51 � 4.66 52.23 � 9.22Non-White race 39 (37.5) 41 (44.1) 80 (40.6)Married/partnered 49 (47.6) 46 (53.5) 95 (50.3)Education in years, M � SD 12.98 � 2.51 12.68 � 2.52 12.84 � 2.51Income �$50,000 71 (87.7) 68 (78.2) 139 (82.7)Period of service�

Vietnam 63 (64.3) 82 (94.3) 145 (78.4)Iraq/Afghanistan 15 (15.3) 0 (0.0) 15 (8.1)Other 20 (20.4) 5 (5.7) 25 (13.5)

Branch of serviceArmy 55 (55.6) 55 (60.4) 110 (57.9)Navy 17 (17.2) 5 (5.5) 22 (11.6)Air Force 6 (6.1) 7 (7.7) 13 (6.8)Marines 21 (21.2) 24 (26.4) 45 (23.7)

Disability compensation for PTSD� 54 (51.9) 63 (67.7) 117 (59.4)Seeking compensation 41 (53.9) 58 (63.7) 99 (59.3)

Note. Data are given as a number (valid percentage) of participants, except where indicated otherwise. CPT �cognitive processing therapy; TAU � treatment as usual; PTSD � posttraumatic stress disorder.� Indicates a significant difference between groups at the p � .05 level.

592 ALVAREZ ET AL.

symptom differences between veterans who completed CPT orTAU groups and those who discharged before completing a traumagroup. The average length of stay was 86 days for both cohorts(CPT � 85, TAU � 88), with no significant difference by cohort.Due to rapid/unexpected discharge or procedural error in theclinical program, only 84.3% of patients who provided intake datacompleted a discharge assessment. Veterans who completed theirdischarge assessment were more likely to report more symptomson the Beck Depression Inventory (BDI; Beck, Steer, & Garbin,1988; Beck, Ward, Mendelsohn, Mock, & Erbaugh, 1961) atbaseline, t(197) � 2.10, p � .037, than veterans who did notcomplete their discharge assessment. There were no other signif-icant demographic or baseline symptom differences between vet-erans who completed a discharge assessment and those who didnot. A multiple imputation strategy was used to handle missingdata (described below).

Measures

Demographics. Demographic variables were measured by aself-report demographic questionnaire that collected informationabout age, race/ethnicity, marital status, education, income, periodof military service, branch of military service, disability compen-sation, and whether the individual was seeking disability compen-sation.

Primary outcome measure. PTSD was assessed with thePCL (Weathers et al., 1993), a 17-item self-report inventory thatassesses DSM–IV Criteria B–D symptoms of PTSD. Items aboutproblems in response to “stressful military experiences” were ratedfrom 1 (not at all) to 5 (extremely) and summed to produce a totalscore and three subscales: reexperiencing (PCL-B), avoidance/numbing (PCL-C), and hyperarousal (PCL-D). Higher scores in-dicated more PTSD symptoms (range � 17–85 for the total score).A cutoff score of 50 for likely PTSD diagnosis was recommendedfor Vietnam veterans (Weathers et al., 1993) and for veterans withcombat trauma related PTSD (Forbes, Creamer, & Biddle, 2001).The PCL has very high internal consistency, with coefficients of.97 for the total scale and .92–.93 for each subscale, and test–retestreliability over 2–3 days of .96 for Vietnam veterans (Weathers etal., 1993). Cronbach’s alpha in the current sample was .90 for thetotal scale and .77–.89 for each subscale at intake, and .92 for thetotal scale and .82–.86 for each subscale at discharge.

Secondary outcome measures. Depression was assessedwith the 21-item self-report BDI (Beck et al., 1988, 1961). Itemswere scored on a scale from 0 to 3 and summed to produce a totalscore. Higher scores indicated more depressive symptoms(range � 0–63). This measure has high internal consistency inpsychiatric and nonpsychiatric samples of .87 and good test–retestreliability of greater than .60 (Beck et al., 1988). Cronbach’s alphain the current sample was .88 at intake and .90 at discharge.

Quality of life was measured with the World Health Organiza-tion Quality of Life–BREF (WHOQOL-BREF; Skevington, Lotfy,& O’Connell, 2004). It consists of 26 items; two items that providea baseline measure of quality of life and quality of health, as wellas four specific quality of life domains: physical health, psycho-logical, social relationship, and environment. For the purpose ofthis study, we examined the Physical, Psychological, and Socialsubscales. Items were rated on four types of 5-point Likert scales.Higher scores indicated a better quality of life (range � 4–20 for

transformed scores). The WHOQOL-BREF has good internal con-sistency of greater than .70 (Skevington et al., 2004). At intake,Cronbach’s alpha in the current sample was .79 for the WHOQOL-Physical subscale, .73 for the WHOQOL-Psychological subscale,and .66 for the WHOQOL-Social subscale. At discharge, Cron-bach’s alpha was .76 for the WHOQOL-Physical subscale, .75 forthe WHOQOL-Psychological subscale, and .61 for the WHOQOL-Social subscale.

Coping style was measured using the 28-item Brief COPE(Carver, 1997). Items were rated from 0 (I haven’t been doing thisat all) to 3 (I’ve been doing this a lot), querying a variety ofdifferent coping methods. Two subscales were identified thoughfactor analysis: Direct coping (COPE-Positive) and Indirect coping(COPE-Avoidant) (Lee & Liu, 2001). Higher scores on the COPE-Positive subscale indicated more positive coping (range � 10–40), and higher scores on the COPE-Avoidant subscale indicatedmore avoidant coping (range � 6–24). At intake, Cronbach’salpha in the current sample was .77 for the COPE-Avoidantsubscale and .82 for the COPE-Positive subscale. At discharge,Cronbach’s alpha was .78 for the COPE-Avoidant subscale and .75for the COPE-Positive subscale.

Psychological distress was measured with the six-item index ofthe Symptom Checklist (SCL-6; Rosen et al., 2000). This indexwas based on the SCL-90 (Derogatis, Lipman, & Covi, 1973),comprised of two questions each from the Depression, Anxiety,and Psychoticism subscales. Items were rated from 0 (not at all) to4 (extremely), with higher scores indicating more psychologicaldistress (range � 6–30). The SCL-6 has good internal consistencyof .83 and had convergent validity with the SCL-90, correlating .87(Rosen et al., 2000). Cronbach’s alpha in the current sample was.87 at intake and .86 at discharge.

Implementation Factors

To assess whether CPT could be sustained effectively, a variablewas created to identify which patients in the CPT cohort weretreated during the first year of adoption, when program clinicianswere receiving ongoing expert case consultation. Outcomes onprimary and secondary outcome measures for these patients werethen compared to patients in the CPT cohort treated in subsequentyears when intensive case consultation ended. CPT providersremained the same in Years 1 and 2.

Procedure

This study conformed to ethical guidelines set forth by theAmerican Psychological Association and was approved by theStanford University Institutional Review Board. As part of normalclinical practice, a variety of self-report assessment measures ofdemographics, symptoms, and functioning were completed by allpatients upon intake and discharge from the residential program.These assessments were distributed and collected by an assessmentcoordinator, who was part of the clinical team but who was not atrauma-focused group therapist. All patients in the program alsoreviewed an informed consent document granting permission fortheir clinical records to be used for research purposes. For thepurposes of this study, clinical records were obtained for all menwho had consented to research and had been treated in the program

593COMPARATIVE EFFECTIVENESS OF CPT FOR VETERANS

during the 2 years before and the 2 years after the implementationof CPT in September 2004.

Treatment

CPT. CPT (Resick, 2001), as delivered in this program, wasa 14-session, manualized, trauma-focused form of cognitive-behavioral therapy for PTSD that is based on cognitive theory ofPTSD. Prior to beginning CPT, two initial sessions focused ongathering information about the veteran’s premilitary autobiogra-phy. The first CPT session involved brief psychoeducation aboutPTSD and CPT. The next two sessions focused on writing andreading about the meaning of the traumatic event and beliefs aboutwhy it happened, identifying problematic beliefs associated withthe trauma (“stuck points”), and learning to identify the connectionbetween events, thoughts, and feelings. The next several sessionsinvolved writing and reading a detailed account of the traumaticevent, with a focus on thoughts and feelings associated with thetrauma. During the next few sessions, veterans learned to questionand challenge their self-statements and assumptions and eventuallyto modify maladaptive beliefs related to the trauma. The last fivesessions involved challenging over- generalized beliefs about selfand others in the following specific areas: safety, trust, power/control, esteem, and intimacy. Treatment gains were discussed andconsolidated in the last session. CPT was delivered in a grouptherapy format, typically including four to five patients and twofacilitators.

TAU. TAU, as delivered in this program, was a 15-session,trauma-focused therapy, based on a life-span development model,incorporating elements of CBT. The treatment package was similarto that described in VHA Cooperative Study No. 420 (Foy, Ruzek,Glynn, Riney, & Gusman, 2002). The initial sessions involvedpsychoeducation about PTSD. The majority of the remaining ses-sions focused on reviewing the veteran’s autobiography in a de-velopmental context. Specifically, premilitary and military copingstyles were reviewed, and the impact of these on current function-ing was identified. The final sessions involved one session oftherapist-guided, in-session exposure to the trauma memory perindividual. Groups typically included six to nine patients and twofacilitators.

Therapist training. In September of 2004, three doctoral-level psychologists and one licensed clinical social worker, whoworked on the residential program clinical team and previouslyprovided TAU, participated in an on-site, 2-day training in CPTled by the developer of the treatment, Patricia Resick. The treat-ment manual and all forms/handouts necessary for conducting thetreatment were distributed during the training. After participatingin the training, these clinicians began conducting CPT groups inplace of TAU groups. Weekly telephone case consultation withPatricia Resick and other CPT experts was conducted for approx-imately 1 year following the original training. The training, con-sultation, and entire study period occurred before the publicationof the group CPT manual (Resick et al., 2007) and the nationalroll-out of CPT in VHA.

Over the course of the year, several modifications were made tothe protocol and the written forms to accommodate the needs ofthe program and the patient population. The protocol was length-ened by approximately two sessions to accommodate group mem-bers reading their trauma accounts aloud in the group format. It is

important to note that trauma accounts are typically not read aloudin group CPT. This specific modification was made to preserve theethic of the larger treatment program, which encourages peersupport and group disclosure. In consultation with CPT experts,some of the language in the original manual and documents waschanged to better reflect the trauma experiences or preferences ofthe veterans (i.e., examples pertaining to combat rather than rape,“problematic” rather than “faulty” thinking patterns). In addition,some of the psychoeducation about PTSD and orientation to CBTwas removed, as veterans already learn this information in thetreatment program. A brief discussion of veterans’ premilitaryautobiography was added prior to beginning CPT to provide in-formation about experiences and beliefs that may be related topotential stuck points.

Modifications were also made to the treatment program overallto facilitate the integration of CPT. For example, aspects of othertreatment groups were changed to make them more consistent withCPT (i.e., “5-column” Thought Records already used in the AffectManagement Group were changed to the ABC Sheets/ChallengingBeliefs Worksheets used in CPT). The daily schedule was alsomodified to build in more time for homework. An effort was madeby the clinical team to portray CPT group participation as onecomponent of the overall treatment program.

Data Analyses

Data were analyzed using SPSS for Windows, Version 17.0.Independent-samples t tests and chi-square tests were used toexamine potential cohort differences on demographic and baselinesymptom variables. Analyses of covariance (ANCOVAs) control-ling for intake symptoms and cohort differences were used toexamine differences in changes from intake to discharge betweentreatment groups. Of note, period of military service was notincluded as an additional covariate in ANCOVAs because of thehigh degree of collinearity between age and period of service,rs(183) � .69, p � .001; this variable was not related to outcome.Also, information about therapy group membership was not avail-able for the TAU cohort, but the intraclass correlation coefficientfor therapy group membership in the CPT cohort was very low(� .001) for all measures except the SCL-6, which was low (.11).This suggests that the results of a mixed effect regression includingtherapy group would likely yield the same results as the ANCOVApresented.

A two-step process was employed to evaluate clinically significantimprovement allowing us to classify individuals into four categories:recovered, improved, unchanged, or deteriorated. These categorieswere determined by (a) a combination of whether the level of func-tioning at discharge was sufficiently improved so that the individualwas no longer a member of the intake population (Jacobson, Follette,& Revenstorf, 1984) and (b) a reliable change index (Jacobson &Truax, 1991).

Missing data. Of the 197 patients with baseline PCL data,166 (84.3%) provided discharge PCL data. There is a rich litera-ture on inference in the presence of missing data, and particularlyin the context of missing follow-up data in longitudinal studies(Hedeker & Gibbons, 1997; Little, 1993; Rubin, 1976; Schafer,1997; Schafer & Graham, 2002). To address the missing follow-updata and potential attrition bias as recommended by this literature,a multiple imputation strategy was used and supplemented by

594 ALVAREZ ET AL.

sensitivity analyses (Daniels & Hogan, 2009; Schafer, 1997; Scha-fer & Graham, 2002).

For these analyses, all baseline characteristics and intake itemswere used to impute missing items at follow-up. Although patientswith complete data were more likely to have a higher BDI atbaseline than those with missing follow-up data, the imputationmodel adjusts for these differences when making the imputations.Ten data sets were imputed with draws from the theoreticallycomplete data distribution using the Multivariate Imputation byChained Equations program (R package Version 1.14; NetherlandsOrganization for Applied Scientific Research, 2005) implementedin the R statistical language (R Foundation for Statistical Comput-ing, 2005). All analyses were conducted on the imputed data setsand pooled to derive the final statistics and inferences presentedhere.

Results

Cohort Differences

The only statistically significant differences between the twocohorts (see Table 1) were age, period of military service, andwhether the veteran received disability compensation for PTSD;veterans in the CPT cohort were younger, more likely to haveserved in Iraq/Afghanistan or periods other than Vietnam, and lesslikely to receive disability compensation for PTSD compared toveterans in the TAU cohort. None of these variables were signif-icantly associated with outcome.

Comparative Effectiveness of CPT Versus TAU

Intake and discharge scores for veterans in CPT and TAU arelisted in Table 2 for descriptive purposes. ANCOVAs comparingthe discharge scores for CPT and TAU participants controlling forintake scores and cohort differences (see Table 3) indicated that

veterans treated with CPT evidenced more symptom improvementthan TAU on PTSD symptoms (total score and all subscales),depression symptoms, psychological quality of life, coping, andpsychological distress. There was no significant difference bytreatment group reported for physical or social quality of life.

Clinically Significant Improvement

In the CPT cohort, 16.3% of participants were classified asrecovered, 41.3% as improved, 31.7% as unchanged, and 10.6% asdeteriorated based upon PCL scores. In the TAU cohort, 4.3%were recovered, 37.6% as improved, 38.7% as unchanged, and19.4% as deteriorated. Men treated with CPT groups were signif-icantly more likely to be classified as recovered on the PCL, �2(1,N � 197) � 7.48, p � .006, and significantly more likely to beclassified as recovered or improved, �2(1, N � 197) � 4.93, p �.032. In addition, men treated with CPT groups were significantlyless likely than men treated with TAU groups to report a PCL �50 at discharge (73% vs. 89%, respectively), �2(1, N � 197) �8.37, p � .004.

Variables Associated With Outcome in CPT Cohort

Analyses indicated that treatment effects in the CPT cohort werenot significantly influenced by race/ethnicity, marital status, edu-cation, income, period of military service, branch of militaryservice, disability compensation, or whether the individual wasseeking disability compensation. Older age was significantly as-sociated with greater reductions in psychological distress reportedon the SCL-6 at discharge, � � –.21, t(104) � –2.22, p � .027.

Implementation Variables Associated With Outcome

Examination of the therapist case consultation variable revealedthat men who participated in CPT groups during the year of expert

Table 2Mean Intake and Discharge Scores for Men in CPT and TAU

Measure

CPT (n � 104) TAU (n � 93)

Intake Discharge Intake Discharge

M SD M SD M SD M SD

PCL 64.05 11.14 55.50 12.78 66.13 10.01 62.12 10.26PCL-B 17.77 4.36 16.89 4.45 18.28 4.56 18.45 3.71PCL-C 26.55 5.08 21.89 5.66 27.86 4.30 25.16 5.00PCL-D 19.73 3.72 16.71 4.29 19.99 3.24 18.51 3.70BDI 25.41 9.17 18.98 10.31 26.11 9.41 22.43 10.34WHOQOL-Physical 11.24 2.87 11.87 2.77 11.08 2.66 11.61 2.15WHOQOL-Psychological 9.73 2.51 11.40 2.68 9.75 2.36 10.47 2.31WHOQOL-Social 9.51 3.21 11.03 3.27 9.35 3.30 10.93 2.69COPE-Avoidant 10.50 4.07 9.62 3.05 12.62 4.65 12.77 4.20COPE-Positive 26.57 6.56 29.06 5.93 25.42 6.48 24.96 6.43SCL-6 21.27 4.72 16.20 4.94 21.46 4.87 18.52 5.11

Note. CPT � cognitive processing therapy; TAU � treatment as usual; PCL � Posttraumatic Stress Disorder Checklist; PCL-B � Posttraumatic StressDisorder Checklist, reexperiencing symptoms; PCL-C � Posttraumatic Stress Disorder Checklist, avoidance/numbing symptoms; PCL-D � PosttraumaticStress Disorder Checklist, hyperarousal symptoms; BDI � Beck Depression Inventory; WHOQOL-Physical � World Health Organization Quality of Life,Physical subscale; WHOQOL-Psychological � World Health Organization Quality of Life, Psychological subscale; WHOQOL-Social � World HealthOrganization Quality of Life, Social subscale; COPE-Avoidant � Brief COPE, Avoidant Coping subscale; COPE-Positive � Brief COPE, Positive Copingsubscale; SCL-6 � six-item index of the Symptom Checklist.

595COMPARATIVE EFFECTIVENESS OF CPT FOR VETERANS

consultation reported significantly greater symptom improvementon two secondary outcome measures: the BDI, F(1, 102) � 6.34,p � .018, and the SCL-6, F(1, 102) � 9.58, p � .004. Meanpre–post effect sizes on the BDI were large (d � 0.94) in the firstyear and small (d � 0.38) in subsequent years. Mean pre–posteffect sizes on the SCL-6 were large (d � 1.41) in the first year andmedium (d � 0.65) in subsequent years. There were no statisticallysignificant differences on the primary outcome measure of PTSDseverity or any other outcome measures between the year ofintensive consultation and subsequent years. However, there werereductions in the effect sizes of improvements in PTSD (large tomedium) and quality of life (large to medium or medium to small)from Year 1 to Year 2. Effect sizes for improvements in copingincreased slightly or remained the same.

Discussion

The results of this investigation suggest that veterans treated inthe CPT cohort during residential treatment for PTSD demon-strated significantly more improvement, and more clinically sig-nificant improvement, than a prior cohort treated with TAUtrauma-focused group therapy. These findings indicate that effec-tive implementation of evidence-based treatment packages such asCPT is superior to TAU and suggest that CPT improved patientoutcomes in this residential treatment program.

The findings in the current study are important because theyrepresent some of the first evidence of greater effectiveness ofCPT relative to a TAU comparison condition including trauma-focused psychotherapy delivered by experienced clinicians. Priorresearch has compared CPT to waitlist or minimal attention controlgroups (Chard, 2005; Monson et al., 2006; Resick et al., 2002,1992). In the current study, patients in both the TAU and CPTcohorts received an intensive treatment package in the residentialprogram that included approximately 450 hr of present-centeredgroup treatment (i.e., cognitive therapy, communication skills,psychoeducation, process groups) as well as approximately 25–30

hr of trauma-focused therapy. It is therefore compelling that im-plementing a manualized, evidence-based intervention such asCPT produced an improvement in program outcomes. Given re-cent evidence about the active ingredients of CPT (Resick et al.,2008), it is possible that the direct challenging of dysfunctionalposttraumatic beliefs in CPT (not present in TAU) contributed toimproved outcomes.

Regarding the sustainability of implementation of CPT, it isencouraging that veterans continued to report improvements inPTSD symptoms, coping style, and quality of life during the timeperiod when therapists were participating in weekly, case consul-tation with experts and during the time period after this consulta-tion was discontinued. However, it should be noted that analyseswere not powered to detect lack of differences, and although notstatistically significantly different, the effect sizes of improve-ments in PTSD and quality of life in the time period withoutconsultation were smaller. The effect sizes of improvements incoping increased slightly or stayed the same. In addition, reduc-tions in symptoms of depression and psychological distress (whichis highly associated with depression) were significantly largerduring the period of consultation. This suggests that there may besome potential benefit of continued consultation as it may increasetreatment adherence and improve outcomes. In particular, BDIoutcomes observed in Year 2 of CPT may reflect a need for moreexpert supervision regarding Socratic questioning of coredepression-related beliefs.

Observed reductions in PCL scores from intake to dischargewere not as large as those reported in a randomized controlled trialof CPT with veterans (Monson et al., 2006), and unfortunately, thelack of interview measures in this study did not allow for anaccurate assessment of PTSD diagnosis before and after treatment.The proxy equivalent in PCL score (below 50) indicated moreveterans may have retained a diagnosis than in prior research(Chard et al., 2010; Monson et al., 2006). However, it should benoted that a prior study with Vietnam veterans suggests that the

Table 3ANCOVAs Comparing Discharge Scores for CPT and TAU Participants

Measure F(3, 193) p b 95 % CI

PCL 15.32 .000 6.25 [3.06, 9.44]PCL-B 6.01 .017 1.38 [0.25, 2.51]PCL-C 15.88 .000 2.97 [1.41, 4.52]PCL-D 12.94 .000 2.00 [0.89, 3.12]BDI 7.50 .008 4.02 [1.07, 6.96]WHOQOL-Physical 0.77 .399 0.29 [�0.39, 0.96]WHOQOL-Psychological 9.85 .003 1.05 [0.36, 1.73]WHOQOL-Social 0.79 .420 0.35 [�0.51, 1.22]COPE-Avoidant 26.61 .000 2.72 [1.65, 3.79]COPE-Positive 22.60 .000 4.23 [2.40, 6.05]SCL-6 14.72 .000 2.63 [1.24, 4.03]

Note. ANCOVAs � analyses of covariance; CPT � cognitive processing therapy; TAU � treatment as usual;PCL � Posttraumatic Stress Disorder Checklist; PCL-B � Posttraumatic Stress Disorder Checklist, reexperi-encing symptoms; PCL-C � Posttraumatic Stress Disorder Checklist, avoidance/numbing symptoms; PCL-D �Posttraumatic Stress Disorder Checklist, hyperarousal symptoms; BDI � Beck Depression Inventory; WHOQOL-Physical � World Health Organization Quality of Life, Physical subscale; WHOQOL-Psychological � World HealthOrganization Quality of Life, Psychological subscale; WHOQOL-Social � World Health Organization Quality ofLife, Social subscale; COPE-Avoidant � Brief COPE, Avoidant Coping subscale; COPE-Positive � Brief COPE,Positive Coping subscale; SCL-6 � six-item index of the Symptom Checklist.

596 ALVAREZ ET AL.

PCL may underestimate change in PTSD symptoms during treat-ment relative to the Clinician Administered PTSD Scale (CAPS;Forbes et al., 2001). Importantly, these improvements were stillclinically significant and similar to PCL results reported in priordescriptive studies of CPT implemented in Veterans Affairs clinicsettings (Chard et al., 2010; Shia, Grattan, & Galovski, 2009). Thedifferences in outcomes from prior published studies may be dueto differences in measurement (self-report vs. interview), differ-ences in implementation of CPT (such as treatment fidelity orreading trauma accounts in group format), additional psychother-apy delivered in the treatment program, or in differences betweenpatient populations. For example, Chard et al. (2010) found similarreductions in PCL scores for Iraq and Afghanistan veterans com-pared to older Vietnam veterans, but more Vietnam veteransretained their diagnosis of PTSD. Future research should continueto explore the role of age and period of military service ontreatment outcome in CPT and other treatments for PTSD as thepopulation of veterans seeking services for PTSD changes. Inaddition, reexperiencing symptoms of PTSD appeared to show lessimprovement, relative to avoidance/numbing and hyperarousalsymptoms. A potential explanation for this finding, or smallereffect sizes overall compared to randomized trials, may be thegroup therapy format. In particular, the reading of trauma accountsin group may reduce the amount of time for Socratic questioning,which may have negatively impacted outcomes. However, thisstudy was not designed to investigate the effects of modificationsto the CPT protocol on outcomes. Therefore, this remains anempirical question and may warrant further investigation.

These results have important implications for the disseminationand implementation of evidenced-based, manualized treatmentpackages in existing VHA treatment settings. These representsome of the first data indicating that evidence-based treatments forPTSD can be disseminated, adopted, and sustained effectively byexisting staff over the long-term. Furthermore, this study providessupport that evidence-based treatments for PTSD such as CPT canbe successfully integrated into existing residential program struc-tures, to the benefit of the existing program and perhaps even themanualized treatment package itself. In this case, minor modifica-tions and adjustments were made to the delivery of CPT and to theoverall treatment program that appeared to preserve the effective-ness of CPT and to improve the overall standard of care in thetreatment program. In addition, information gathered by CPTexperts during case consultation with this program contributed tothe knowledge base about treatment of veterans with CPT. Thisstudy also provides promising evidence that CPT can be deliveredin a mixed trauma type, group therapy format and that it iseffective for complex cases typically seen in VHA residential care.

There are several additional issues that are relevant to theinterpretation of these results. This investigation was not a ran-domized controlled trial and lacks the experimental methodologyand rigorous controls necessary to conclude that the improvementsin clinical outcomes between cohorts were due to the implemen-tation of CPT. Although the quasi-experimental design does notprovide absolute proof of causality, it may offer more informationabout the implementation and effectiveness of evidence-basedtreatments that is generalizable to real-world clinical settings.Indeed, some of the improved outcomes reported by the latercohort of veterans may have been due to the overall improvementsin the residential treatment program motivated by the implemen-

tation of CPT (i.e., updating of cognitive therapy group to beconsistent with CPT, more time for homework). This is an impor-tant point and suggests that implementation of evidenced-based,manualized treatment packages may be a catalyst for improvingother aspects of the treatment program. It should be noted, how-ever, that other than these changes related to the implementationand integration of CPT, other changes to the residential treatmentprogram (e.g., other groups, staffing) were minimal over the studyperiod. This suggests that the observed improvements in outcomesare associated with CPT group and related changes.

This study did not include any measures of treatment adherenceor therapist competency that are typically included in randomized,clinical trials. Whereas these measures would have allowed formore certainty that the therapy delivered in this treatment programwas adherent to the manual, the reality of most clinical settings isthat administering these measures as a part of routine practice isunlikely. Importantly, the results suggest that regardless of strictadherence, the treatment delivered was effective and associatedwith improved outcomes relative to the treatment delivered previ-ously.

As mentioned above, this study did not include a clinician-administered measure of PTSD, such as the CAPS, because thisassessment was not a part of routine clinical care in the setting.Clinician-administered assessment is often difficult to implementdue to staffing constraints in clinical settings, and future studiesand implementation efforts should endeavor to find ways to buildin this type of assessment. Reporting on self-report measures maybe influenced by demographic variables, symptoms, personality,and compensation-seeking (Enns, Larsen, & Cox, 2000; Frueh etal., 2003), and in this study, the method of administration of thePCL did not allow for certainty that the participants were reportingsymptoms related to a specific Criterion A event. Examination ofa clinician-rated measure of PTSD may have allowed for a moreaccurate assessment of symptoms and functioning and better com-parison of this study to others in the literature. However, theliterature is inconsistent regarding the correspondence between thePCL and CAPS; in contrast to research cited above (Forbes et al.,2001), a recent study examining two PTSD treatment studiesconducted with veterans found no difference in symptom rating onthe PCL compared to the CAPS (Monson et al., 2008).

Finally, the results of this study may not generalize to outpatientsettings or other patient populations such as women veterans. Thisstudy was also unable to report on patients who did not participatein TAU or CPT in this setting (although criteria for participating ina trauma-focused group did not change when CPT was imple-mented).

The results of this study indicate that dissemination efforts suchas those employed by VHA to train providers in evidence-basedtreatments and to implement these treatments into clinics andprograms may be well-worth the effort and cost because they areassociated with considerably improved patient outcomes. Al-though more research is needed, these findings suggest that CPTmay be acceptable and well-tolerated by the complex patientsoften treated in residential settings. Implementing CPT in residen-tial settings may increase cost-effectiveness if more patients areimproved and utilization is decreased; future studies should exam-ine the rate of rehospitalization and outpatient treatment utilizationafter treatment with CPT. This study also illustrates the utility ofpatient outcomes monitoring for examining the impact of decisions

597COMPARATIVE EFFECTIVENESS OF CPT FOR VETERANS

to modify treatment programming. Certainly, there is still room forimprovement, as substantial numbers of veterans continue to ex-perience significant symptoms even after treatment with CPT in aresidential program; however, it appears that replacing older treat-ment packages with therapies such as CPT may give veterans abetter chance at clinically significant symptom reduction and im-proved functioning and quality of life.

References

American Psychiatric Association. (1994). Diagnostic and statistical man-ual of mental disorders (4th ed.). Washington, DC: Author.

Beck, A. T., Steer, R. A., & Garbin, M. G. (1988). Psychometric propertiesof the Beck Depression Inventory: Twenty-five years of evaluation.Clinical Psychology Review, 8, 77–100. doi:10.1016/0272-7358(88)90050-5

Beck, A. T., Ward, C. H., Mendelsohn, M., Mock, J., & Erbaugh, J. (1961).An inventory for measuring depression. Archives of General Psychiatry,4, 561–571.

Carver, C. S. (1997). You want to measure coping but your protocol’s toolong: Consider the brief COPE. International Journal of BehavioralMedicine, 4, 92–100. doi:10.1207/s15327558ijbm0401_6

Chard, K. M. (2005). An evaluation of cognitive processing therapy for thetreatment of posttraumatic stress disorder related to childhood sexualabuse. Journal of Consulting and Clinical Psychology, 73, 965–971.doi:10.1037/0022-006X.73.5.965

Chard, K. M., Schumm, J. A., Owens, G. P., & Cottingham, S. M. (2010).A comparison of OEF and OIF veterans and Vietnam veterans receivingcognitive processing therapy. Journal of Traumatic Stress, 23, 25–32.

Daniels, M., & Hogan, J. (2009). Missing data in longitudinal studies:Strategies for Bayesian modeling and sensitivity analysis. New York,NY: Chapman & Hall.

Derogatis, L. R., Lipman, R. S., & Covi, L. (1973). The SCL-90: Anoutpatient psychiatric rating scale–Preliminary report. Psychopharma-cology Bulletin, 9, 13–28.

Dohrenwend, B. P., Turner, J. B., Turse, N. A., Adams, B. G., Koenen,K. C., & Marshall, R. (2006, August 18). The psychological risks ofVietnam for U.S. veterans: A revisit with new data and methods.Science, 313, 979–982. doi:10.1126/science.1128944

Enns, M. W., Larsen, D. K., & Cox, B. J. (2000). Discrepancies betweenself and observer ratings of depression: The relationship to demographic,clinical, and personality variables. Journal of Affective Disorders, 60,33–41. doi:10.1016/S0165-0327(99)00156-1

Foa, E. B., Dancu, C. V., Hembree, E. A., Jaycox, L. H., Meadows, E. A.,& Street, G. P. (1999). A comparison of exposure therapy, stressinoculation training, and their combination for reducing posttraumaticstress disorder in female assault victims. Journal of Consulting andClinical Psychology, 67, 194–200. doi:10.1037/0022-006X.67.2.194

Foa, E. B., Rothbaum, B. O., Riggs, D. S., & Murdock, T. B. (1991).Treatment of posttraumatic stress disorder in rape victims: A comparisonbetween cognitive-behavioral procedures and counseling. Journal ofConsulting and Clinical Psychology, 59, 715–723. doi:10.1037/0022-006X.59.5.715

Forbes, D., Creamer, M., & Biddle, D. (2001). The validity of the PTSDChecklist as a measure of symptomatic change in combat-related PTSD.Behaviour Research and Therapy, 39, 977–986. doi:10.1016/S0005-7967(00)00084-X

Foy, D. W., Ruzek, J. I., Glynn, S. M., Riney, S. J., & Gusman, F. D.(2002). Trauma focus group therapy for combat-related PTSD: Anupdate. Journal of Clinical Psychology/In Session: Psychotherapy inPractice, 58, 907–918.

Frueh, B. C., Elhai, J. D., Gold, P. B., Monnier, J., Magruder, K. M.,Keane, T. M., & Arana, G. W. (2003). Disability compensation seeking

among veterans evaluated for posttraumatic stress disorder. PsychiatricServices, 54, 84–91. doi:10.1176/appi.ps.54.1.84

Hedeker, D., & Gibbons, R. D. (1997). Application of random-effectspattern-mixture models for missing data in longitudinal studies. Psycho-logical Methods, 2, 64–78. doi:10.1037/1082-989X.2.1.64

Hoge, C. W., Castro, C. A., Messer, S. C., McGurk, D., Cotting, D. I., &Koffman, R. L. (2004). Combat duty in Iraq and Afghanistan, mentalhealth problems, and barriers to care. The New England Journal ofMedicine, 351, 13–22. doi:10.1056/NEJMoa040603

Jacobson, N. S., Follette, W. C., & Revenstorf, D. (1984). Psychotherapyoutcome research: Methods for reporting variability and evaluatingclinical significance. Behavior Therapy, 15, 336–352. doi:10.1016/S0005-7894(84)80002-7

Jacobson, N. S., & Truax, P. (1991). Clinical significance: A statisticalapproach to defining meaningful change in psychotherapy research.Journal of Consulting and Clinical Psychology, 59, 12–19. doi:10.1037/0022-006X.59.1.12

Karlin, B. E., Ruzek, J. I., Chard, K. M., Eftekhari, A., Monson, C. M.,Hembree, E. A., . . . Foa, E. B. (2010). Dissemination of evidence-basedpsychological treatments for posttraumatic stress disorder in the Veter-ans Health Administration. Journal of Traumatic Stress, 23, 663–673.doi:10.1002/jts.20588

Kilbourne, A. M., Neumann, M. S., Pincus, H. A., Bauer, M. S., & Stall,R. (2007). Implementing evidence-based interventions in health care:Application of the replicating effective programs framework. Implemen-tation Science, 2, 42. doi:10.1186/1748-5908-2-42

Kulka, R. A., Schlenger, W. E., Fairbank, J. A., Hough, R. L., Jordan,B. K., Marmar, C. R., . . . Grady, D. A. (1990). Trauma and the VietnamWar generation: Report of findings from the National Vietnam VeteransReadjustment Study. New York, NY: Brunner/Mazel.

Lee, R. M., & Liu, H.-T. T. (2001). Coping with intergenerational familyconflict: Comparison of Asian American, Hispanic, and EuropeanAmerican college students. Journal of Counseling Psychology, 48, 410–419. doi:10.1037/0022-0167.48.4.410

Little, R. J. A. (1993). Pattern-mixture models for multivariate incompletedata. Journal of the American Statistical Association, 88, 125–133.doi:10.2307/2290705

Miller, W. R., Yahne, C. E., Moyers, T. B., Martinez, J., & Pirritano, M.(2004). A randomized trial of methods to help clinicians learn motiva-tional interviewing. Journal of Consulting and Clinical Psychology, 72,1050–1062. doi:10.1037/0022-006X.72.6.1050

Milliken, C. S., Auchterlonie, J. L., & Hoge, C. W. (2007). Longitudinalassessment of mental health problems among active and reserve com-ponent soldiers returning from the Iraq war. JAMA: Journal of theAmerican Medical Association, 298, 2141–2148. doi:10.1001/jama.298.18.2141

Monson, C. M., Gradus, J. L., Young-Xu, Y., Schnurr, P. P., Price, J. L.,& Schumm, J. A. (2008). Change in posttraumatic stress disorder symp-toms: Do clinicians and patients agree? Psychological Assessment, 20,131–138. doi:10.1037/1040-3590.20.2.131

Monson, C. M., Schnurr, P. P., Resick, P. A., Friedman, M. J., Young-Xu,Y., & Stevens, S. P. (2006). Cognitive processing therapy for veteranswith military-related posttraumatic stress disorder. Journal of Consult-ing and Clinical Psychology, 74, 898 –907. doi:10.1037/0022-006X.74.5.898

Netherlands Organization for Applied Scientific Research. (2005). Multi-variate Imputation by Chained Equations (R package Version 1.14)[Computer software]. Delft, Netherlands: Author.

Resick, P. A. (2001). Cognitive processing therapy: Generic version. St.Louis, MO: University of Missouri—St. Louis.

Resick, P. A., Galovski, T. E., O’Brien Uhlmansiek, M., Scher, C. D.,Clum, G. A., & Young-Xu, Y. (2008). A randomized clinical trial todismantle components of cognitive processing therapy for posttraumaticstress disorder in female victims of interpersonal violence. Journal of

598 ALVAREZ ET AL.

Consulting and Clinical Psychology, 76, 243–258. doi:10.1037/0022-006X.76.2.243

Resick, P. A., Monson, C. M., & Chard, K. M. (2007). Cognitive process-ing therapy: Veteran/military version. Washington, DC: Department ofVeterans Affairs.

Resick, P. A., Nishith, P., Weaver, T. L., Astin, M. C., & Feuer, C. A.(2002). A comparison of cognitive-processing therapy with prolongedexposure and a waiting condition for the treatment of chronic posttrau-matic stress disorder in female rape victims. Journal of Consulting andClinical Psychology, 70, 867–879. doi:10.1037/0022-006X.70.4.867

Resick, P. A., & Schnicke, M. K. (1992). Cognitive processing therapy forsexual assault victims. Journal of Consulting and Clinical Psychology,60, 748–756. doi:10.1037/0022-006X.60.5.748

R Foundation for Statistical Computing. (2005). R: A language and envi-ronment for statistical computing [Computer software]. Vienna, Austria:Author.

Rosen, C. S., Drescher, K. D., Moos, R. H., Finney, J. W., Murphy, R. T.,& Gusman, F. (2000). Six- and ten-item indexes of psychologicaldistress based on the Symptom Checklist-90. Assessment, 7, 103–111.doi:10.1177/107319110000700201

Rubin, D. B. (1976). Inference and missing data. Biometrika, 63, 581–592.doi:10.1093/biomet/63.3.581

Schafer, J. L. (1997). Analysis of incomplete multivariate data. New York,NY: Chapman & Hall.

Schafer, J. L., & Graham, J. W. (2002). Missing data: Our view of the stateof the art. Psychological Methods, 7, 147–177. doi:10.1037/1082-989X.7.2.147

Schnurr, P. P., Friedman, M. J., Foy, D. W., Shea, M. T., Hsieh, F. Y.,

Lavori, P. W., . . . Bernardy, N. C. (2003). Randomized trial of trauma-focused group therapy for posttraumatic stress disorder. Archives ofGeneral Psychiatry, 60, 481–489. doi:10.1001/archpsyc.60.5.481

Shia, S., Grattan, J., & Galovski, T. (2009, November). Group CPT in amixed trauma veteran population: Implementation strategies and effec-tiveness. Poster presented at the meeting of the International Society forTraumatic Stress Studies, Atlanta, GA.

Skevington, S. M., Lotfy, M., & O’Connell, K. A. (2004). The WorldHealth Organization’s WHOQOL-BREF quality of life assessment: Psy-chometric properties and results of the international field trial. A reportfrom the WHOQOL group. Quality of Life Research, 13, 299–310.doi:10.1023/B:QURE.0000018486.91360.00

Veterans Health Administration Office of Public Health and Environmen-tal Hazards. (2009). Analysis of VA health care utilization among U.S.global war on terrorism (GWOT) veterans. Washington, DC: U.S.Department of Veteran Affairs.

Weathers, F., Litz, B., Herman, D., Huska, J., & Keane, T. (1993). ThePTSD Checklist (PCL): Reliability, validity, and diagnostic utility. Paperpresented at the annual convention of the International Society forTraumatic Stress Studies, San Antonio, TX.

Zappert, L. N., & Westrup, D. (2008). Cognitive processing therapy forposttraumatic stress disorder in a residential treatment setting. Psycho-therapy: Theory, Research, Practice, Training, 45, 361–376. doi:10.1037/0033-3204.45.3.361

Received September 8, 2010Revision received May 9, 2011

Accepted May 23, 2011 �

599COMPARATIVE EFFECTIVENESS OF CPT FOR VETERANS