Characteristics of U.S. Veterans Who Begin and Complete Prolonged Exposure and Cognitive Processing Therapy for PTSD
Journal of Traumatic StressJune 2014, 27, 265273Characteristics of U.S. Veterans Who Begin and CompleteProlonged Exposure and Cognitive Processing Therapy for PTSDJuliette M. Mott,1,2,3 Sasha Mondragon,1,2 Natalie E. Hundt,1,2,3 Melissa Beason-Smith,1,2Rebecca H. Grady,1,2,3 and Ellen J. Teng1,2,31Michael E. DeBakey Veterans Affairs Medical Center, Houston, Texas, USA2Menninger Department of Psychiatry and Behavioral Sciences at Baylor College of Medicine, Houston, Texas, USA3South Central Mental Illness Research, Education, and Clinical Center, Houston, Texas, USAThis retrospective chart-review study examined patient-level correlates of initiation and completion of evidence-based psychotherapy(EBP) for posttraumatic stress disorder (PTSD) among treatment-seeking U.S. veterans. We identified all patients (N = 796) in a largeVeterans Affairs PTSD and anxiety clinic who attended at least 1 individual psychotherapy appointment with 1 of 8 providers trained inEBP. Within this group, 91 patients (11.4%) began EBP (either Cognitive Processing Therapy or Prolonged Exposure) and 59 patients(7.9%) completed EBP. The medical records of all EBP patients (n = 91) and a provider-matched sample of patients who received anotherform of individual psychotherapy (n = 66) were reviewed by 4 independent raters. Logistic regression analyses revealed that Iraq andAfghanistan veterans were less likely to begin EBP than veterans from other service eras, OR = 0.48, 95% CI= [0.24, 0.94], and veteranswho were service connected for PTSD were more likely than veterans without service connection to begin EBP, OR = 2.33, 95% CI =[1.09, 5.03]. Among those who began EBP, Iraq and Afghanistan veteran status, OR = 0.09, 95% CI = [0.03, 0.30], and a history ofpsychiatric inpatient hospitalization, OR = 0.13, 95% CI = [0.03, 0.54], were associated with decreased likelihood of EBP completion.Given that posttraumatic stress disorder (PTSD) is one of themost common mental health diagnoses among U.S. militaryveterans (Barrera et al., 2014; Cohen et al., 2010), providinghigh-quality PTSD treatment is a critical mission to the Veter-ans Health Administration (VHA). To increase access to care,the VHA implemented large-scale efforts to train clinicians intwo evidence-based psychotherapy (EBP) protocols for PTSD:cognitive processing therapy (CPT) and prolonged exposure(PE). Both are manualized, cognitivebehavioral interventionsthat typically consist of 812 weekly sessions (Foa, Hembree,Juliette Mott is currently affiliated with the National Center for PTSD inWhiteRiver Junction, Vermont USA.This research was supported in part by the Office of Academic Affiliations VAAdvanced Fellowship Program in Mental Illness Research and Treatment, bythe Houston VA HSR&D Center of Excellence (Houston Center for Innova-tions in Quality, Effectiveness and Safety (CIN 13-413), and by a VA ClinicalSciences Research and Development (CSR&D) Career Development Award(#CADE-MHN/F09) awarded to Ellen J. Teng at the Michael E. DeBakeyVeterans Affairs Medical Center. The views expressed reflect those of the au-thors and not necessarily the Department of Veterans Affairs/Baylor Collegeof Medicine.Correspondence concerning this article should be addressed to Juliette Mott,National Center for PTSD, VAMedical Center (116D), 215 NorthMain Street,White River Junction, VT 05009. E-mail: Juliette.Mott@va.govPublished 2014. This article is a US Government work and is in the publicdomain in the USA. View this article online at wileyonlinelibrary.comDOI: 10.1002/jts.21927& Rothbaum, 2007; Resick, Monson, & Chard, 2008). To date,more than 2,300 VHA clinicians have been trained in CPT, andmore than 1,500 have been trained in PE (Eftekhari et al., 2013;Karlin et al., 2010). These dissemination efforts have increasedthe VHAs potential to provide evidence-based care to veteranswith PTSD (Cook, Dinnen, Thompson, Simiola, & Schnurr,2014).To evaluate the impact of EBP dissemination initiatives, sev-eral recent studies have assessed the use of CPT and PE withinthe VHA. Although VHA administrative databases do not con-tain information on the type of psychotherapy delivered, priorstudies have used these databases to determine the number ofpatients who received at least 8 or 9 psychotherapy sessions,and therefore could have potentially completed a full courseof EBP. Consistently, these studies indicate that less than 10%of veterans with PTSD could have feasibly completed an EBP(Mott, Hundt, Sansgiry, Mignogna, & Cully, 2014; Seal et al.,2010). To estimate EBP initiation more precisely, Shiner andcolleagues (2012) reviewed the text of psychotherapy notesfrom more than 1,900 patients across 6 VHA facilities andreported that only 6% of veterans received one or more EBPsessions in the 6 months following enrollment in a PTSD clinic.This suggests that most veterans who receive psychotherapy inspecialized PTSD care settings are receiving treatments otherthan VHA-recommended EBPs.265266 Mott et al.Although the VHA mandates that all patients with PTSDhave access to CPT or PE, the decision of whether a veteran re-ceives one of these treatments is left to the veteran and provider(McHugh & Barlow, 2010). Patients and providers may findlittle empirical guidance for treatment selection, as research in-forming evidence-based treatment decisions based on patientcharacteristics such as age or trauma type is only in the be-ginning stages (Sharpless & Barber, 2011). Prior research hasdemonstrated the effectiveness of CPT and PE in veteran pop-ulations that are diverse with respect to gender (Schnurr et al.,2007), service era (Yoder et al., 2012), trauma type (Rauchet al., 2009), and psychiatric comorbidities (Monson et al.,2006), suggesting that these treatments have the potential forbroad patient reach.Practice guidelines and EBP manuals offer differing per-spectives on which patients are likely appropriate for CPT orPE. The VHA/Department of Defense (DoD) clinical practiceguideline for management of PTSD (2010) endorses PE andCPT at the highest possible level, indicating a strong recom-mendation that clinicians provide the intervention to eligiblepatients (p. 201). PE and CPT therapist manuals, however, in-dicate that not all patients with PTSD are appropriate for thesetreatments. The PEmanual recommends that PE not be initiatedin the presence of imminent threat of suicide or homicide, seri-ous self-injurious behavior, current psychosis, current high riskof being assaulted or insufficient memory of the trauma (Foaet al., 2007). The CPT manual (Resick et al., 2008) cautionsagainst the treatment of patients who are in imminent danger,suffer from severe dissociation or panic attacks that may inter-fere with treatment, and recommends case-by-case decisionsfor patients with substance dependence and self-harming be-haviors. Both manuals indicate that comorbid depression is nota rule-out and note that patients with subthreshold PTSD symp-toms may be appropriate. Thus, these manuals suggest that PEand CPT are not appropriate for all patients with PTSD andmay be useful for patients without a full PTSD diagnosis.Although VHA policies and therapist manuals each provideunique perspectives on which patients should receive EBP forPTSD, a burgeoning body of research has attempted to de-scribe the population of veterans who actually receive thesetreatments. A number of studies have found that veterans whoreceived PE in VHA PTSD clinics were predominately White,male, high-school educated, and service-connected for PTSD(Gros, Yoder, Tuerk, Lozano, & Acierno, 2011; Rauch et al.,2009; Tuerk, Yoder, Ruggiero, Gros, & Acierno, 2010; Yoderet al., 2012). This mirrors the general population of veteranswho seek treatment (not necessarily EBP) for PTSD (Seal et al.,2010; Spoont, Murdoch, Hodges, & Nugent, 2010). Given thatthese studies did not compare directly the characteristics of pa-tients who received EBP to those who received an alternativetreatment, the degree to which observed patient characteristicsare specific to those who engage in EBP remains unclear.Further examination of EBP patients clinical profileswithparticular attention to factors that are likely to influence EBPreferrals, such as comorbid pathology, suicidality, and mentalhealth treatment historyis needed. Identification of patientcharacteristics associated with an increased likelihood of EBPengagement can enhance knowledge ofwhich patients are likelyto participate in these therapies and, importantly, may aid in theidentification of underserved populations who are not currentlyaccessing these effective treatments. Toward this end, this studyexamined a sample of veterans who received individual psy-chotherapy from a provider trained in CPT or PE to examinepatient-level correlates of EBP initiation and completion.MethodParticipants and ProcedureAll procedures were approved by the local institutional reviewboard and by the Veterans Affairs (VA) Research and Devel-opment Committee. This study included both provider and pa-tient participants. We identified nine EBP-trained providers in aPTSD and anxiety specialty clinic in a large VA medical centerwho had previously completed an optional multiday trainingin CPT, PE, or both. Of the nine providers identified, eightexpressed willingness to participate. The provider sample in-cluded licensed psychologists (n = 4), psychology fellows (n =2), and social workers (n = 2). All eight providers were trainedin CPT and seven were trained in PE. Five providers were VAcertified in at least one EBP (four in CPT, and five in PE). Weincluded providers who attended an EBP trainingworkshop, butwere not fully certified because the completion of the additionaltraining components required for certification (e.g., follow-upconsultation) may potentially be impacted by a variety of fac-tors (e.g., busy work schedule, inability to complete consul-tation within specified time frame, patient dropout), and doesnot necessarily reflect the providers receptiveness or attitudetoward EBP. Certified and noncertified providers were similarwith respect to the proportion of patients in their caseloads whobegan EBP, 2(1, N = 8)= 1.21, p = .271 and completed EBP2(1, N = 8) = 2.52, p = .112.The length of time between providers first EBP training anddata extraction ranged from 12 to 48 months. After training,two providers saw between 10 and 20 patients, four providerssaw between 30 and 90 patients, and two providers saw morethan 200 patients. Providers began EBP ( 1 EBP session) withbetween 2 and 40 patients, and the proportion of patients withineach providers caseload who began EBP ranged from 3.4% to69.6% (M = 20.6%, SD = 16.2%). Length of time since EBPtraining (r = .06, p = .888) and total caseload (r = .30, p =.470) were unrelated to the number of patients who received atleast one EBP session from the provider.The patient sample included all patients (n = 796) who at-tended at least one individual outpatient psychotherapy sessionwith any of the eight EBP-trained providers since the providerstraining (for providers trained in both CPT and PE, we used theearliest training date), identified through the VA ComputerizedPatient Record System. All patients received treatment withtheir EBP-trained provider between 2008 and 2012. Per clinicJournal of Traumatic Stress DOI 10.1002/jts. Published on behalf of the International Society for Traumatic Stress Studies.Veterans in Evidence-Based Therapy for PTSD 267policy, patients were required to have an anxiety disorder di-agnosis; it was not required that clinic patients meet criteriafor PTSD. Upon enrollment in the PTSD and anxiety clinic, allpatients completed a comprehensive mental health assessmentconsisting of a semistructured diagnostic interview created forthe purpose of the evaluation and expressed intent to initiatepsychotherapy.Providers indicated which of their patients received at leastone session of individual protocol-driven CPT and/or PE(providers maintained a list of EBP patients for administrativepurposes and could consult this list). Patients who began EBP(i.e., received at least one session of individual CPT or PE) weredeemed EBP initiators and patients who began a non-EBP (i.e.,received a form of individual psychotherapy other than CPT orPE) were deemed non-EBP initiators. Examples of individualpsychotherapies offered and anxiety that were for the purposesof this study classified as non-EBP included supportive psy-chotherapy, relaxation, and psychoeducation. Among the 796patients who received at least one individual psychotherapy ses-sion with an EBP-trained provider, 91 patients (11.4%) wereidentified as EBP initiators; all others were identified as non-EBP initiators (n = 705; 88.6%).To obtain more detailed information on patients character-istics and use of psychotherapy, we reviewed the 91 medicalrecords of EBP initiators and a randomly selected provider-matched sample of 66 non-EBP initiators. Althoughwe initiallyintended to include an equal number of EBP and non-EBP ini-tiators from each provider, three providers had administeredEBP to more than half of their patients. Data extraction wasbased on best-practice standards (Gilbert, Lowenstein, Koziol-McLain, Barta, & Steiner, 1996).MeasuresA coding manual developed for the study included operationaldefinitions, data extraction instructions, and decision rules re-garding missing data and conflicting patient records. A dataextraction form was designed to record patient demographicvariables including age, gender, race, marital status, educationlevel, annual income, employment status, military service era,religion, and PTSD service connection (i.e., disability ratingthat affords veterans increased access to VHA services andmonthly disability payments) and clinical variables includingnumber and type of psychotherapy visits (CPT, PE, or non-EBP), prior inpatient hospitalizations, prior group psychother-apy, Global Assessment of Functioning (GAF) scores, delayedtherapy with EBP-trained provider (>6 months between in-take and therapy initiation), and suicide risk (extracted fromthe comprehensive mental health assessment prior to the startof therapy). We also extracted mental health diagnoses (PTSD,other anxiety disorder, depressive disorder, bipolar disorder,substance use disorder, psychotic disorder, personality disor-der); these diagnoses were not mutually exclusive, and it wasnot possible to discern which was the primary diagnosis.Prior to chart review, four independent raters completed VAelectronic medical record training and were trained by the firstauthor in the use of the data extraction form. To establish initialinterrater reliability, raters independently extracted data for thesame six training cases (not eligible for study inclusion) anddemonstrated acceptable agreement on demographic and clini-cal variables (> .80). Extracted data were entered into an elec-tronic database. Throughout the project, raters attended weeklymeetings to review coding rules and resolve disagreements.A fifth rater re-extracted data from a random 20% of patientsand achieved good interrater agreement with original raters onclinical and demographic variables from double-coded cases( = .871.00).Demographic and clinical characteristics for the chart re-view sample (n = 157) are displayed in Table 1. Chart reviewdata were also used to identify those patients who completeda full course of either CPT or PE. Participants were defined asEBP completers if they received at least seven EBP sessions(verified via chart review) and their provider indicated thatthey completed the full EBP protocol. We selected a 7-sessionbenchmark based on prior research indicating that CPT pa-tients meet good end-state criteria in an average of 7.5 sessions(Galovski, Blain, Mott, Elwood, & Houle, 2012) and because a7-session cutoff has been used previously to define PE comple-tion (Tuerk et al., 2012). A psychotherapy session was consid-ered EBP if the provider identified the session as PE or CPT inthe note (e.g., This was session 2 of PE) or described specificelements of PE (imaginal exposure, in vivo exposure) or CPT(impact statement, ABC sheets, trauma account, challengingbeliefs worksheet). Given increasing recognition that some pa-tients require sessions beyond the traditional protocol length toexperience meaningful change in PTSD symptoms (Galovskiet al., 2012), provider verification of EBP completion statusoffered increased assurance of treatment completion. Patientswho began EBP, but failed to meet either completion criterion,were deemed EBP dropouts. Figure 1 displays the number ofCPT and PE completers within the EBP initiator sample. FourEBP dropouts met one criterion for EBP completion, but notthe other; three EBP dropouts had seven or more EBP sessions,but their provider did not consider them an EBP completer, andconversely, one EBP dropout was defined as an EBP completerby the provider, but attended only five EBP sessions.Excluding patients who were more than 3 standard devia-tions above the mean with respect to number of psychotherapyvisits (n = 3), EBP initiators attended a mean of 16.07 (SD= 11.34) total therapy sessions with their provider; on aver-age, 9.92 (SD = 6.02) of these were EBP sessions. Most EBPinitiators (70.3%) received a mixture of EBP and non-EBP ses-sions from their provider. EBP completers attended an averageof 17.50 (SD = 10.40) total therapy sessions including 13.17(SD = 4.45) EBP sessions whereas EBP dropouts attended anaverage of 13.47 (SD = 12.62) total therapy sessions includ-ing 4.03 (SD = 3.46) EBP sessions. CPT dropouts most com-monly discontinued treatment after three CPT sessions, and PEpatients most commonly discontinued after two PE sessions.Journal of Traumatic Stress DOI 10.1002/jts. Published on behalf of the International Society for Traumatic Stress Studies.268 Mott et al.Table 1Total Sample and Difference Between EBP Initiators and non-EBP Initiators on Demographic and Clinical VariablesFull chart review EBP initiators Non-EBP initiatorssample (n = 157) (n = 91) (n = 66)Variable n or M % or SD n or M % or SD n or M % or SD 2 or tPTSD diagnosis 138 87.9 86 94.5 52 78.8 18.32*Non-PTSD anxiety disorder 50 31.8 26 28.6 24 36.4 1.07Depressive disorder 100 63.7 58 63.7 42 63.6 6 months delayed therapy 104 66.2 70 77.0 34 51.5 10.33*Suicide risk 23 14.6 10 11.0 13 19.6 2.48PTSD service connection 116 73.9 75 82.4 41 62.1 8.17*Ethnicitya (White) 98 64.9 53 60.9 45 70.3 1.43Gender (male) 142 90.4 83 91.2 59 89.4 0.15Educationa (> high school) 93 64.6 50 59.5 43 71.7 2.26Employmenta (employed) 63 41.7 42 46.2 21 31.8 3.13Marital statusa (married) 86 55.8 56 61.5 30 45.5 3.57Era (OEF/OIF/OND) 79 50.6 37 40.6 42 63.6 8.71*Religiona 119 84.4 70 84.3 49 84.5 6 months between intake and therapy initiation),and clinician-rated suicide risk. Candidate predictors for EBPcompletion were identical except that EBP type (CPT/PE) wasincluded as a predictor and four clinical variables were not in-cluded due to low (3%; bipolar disorder, psychotic disorder,Journal of Traumatic Stress DOI 10.1002/jts. Published on behalf of the International Society for Traumatic Stress Studies.Veterans in Evidence-Based Therapy for PTSD 269Table 2Differences Between EBP Completers and EBP Dropouts on Demographic and Clinical VariablesEBP completers (n = 58) EBP dropouts (n = 33)Variable n or M % or SD n or M % or SD 2 or tNon-PTSD anxiety disorder 16 27.6 10 30.3 0.08Depressive disorder 37 63.8 21 63.6 6 months delayed EBP 50 86.2 20 60.6 7.79*Suicide risk 6 10.3 4 12.1 0.05EBP type (PE) 46 79.3 21 63.6 3.79PTSD service connection 46 79.3 29 87.9 1.11Ethnicitya (White) 31 56.4 22 68.8 1.30Gender (male) 55 94.8 28 84.8 2.61Educationa (> high school) 37 67.3 13 44.8 3.97*Employmenta (employed) 24 42.9 18 53.6 1.46Marital statusa (married) 38 66.7 18 54.5 1.31Service era (OEF/OIF/OND) 14 24.1 23 69.7 18.09*Religiona 46 85.2 24 82.8 0.77Age (years) 55.53 13.67 41.87 15.94 4.63*GAFa 55.48 6.25 55.28 5.87 0.13Annual income ($) 38,000 29,000 33,000 22,000 0.75Note. The ns for categorical variables represent the number of patients who affirmatively endorsed the parenthetically noted category. EBP = evidence-basedpsychotherapy; GAF = Global Assessment of Functioning; OEF/OIF/OND = Operation Enduring Freedom/Operation Iraqi Freedom/Operation New Dawn; PE =prolonged exposure.aVariable has missing values from no more than 10% of participants. All other variables had no missing values.*p < .05, two-tailed.personality disorder) or high (95%; PTSD) base rates amongEBP initiators.To limit the number of predictors in the logistic models, bi-variate analyses (2, t tests) were used to compare EBP initia-tors to non-EBP initiators and EBP completers to EBP dropoutswith respect to candidate predictors. Variables significant atp < .10 were entered into logistic regression models predictingEBP initiation and EBP completion. Nonsignificant predictorswere removed from the model using backward selection untilall predictors were significant at an level of .05. NagelkerkesR2 was calculated to assess improvement over the null model.ResultsTo examine patient factors associated with EBP initiation, weused bivariate analyses (see Table 1) to compare EBP ini-tiators to non-EBP initiators and identify those variables onwhich these groups differed significantly (p < .10). EBP ini-tiators were older and more likely than non-EBP initiators tobe employed, married, non-OEF/OIF/OND (Operation Endur-ing Freedom/Operation Iraqi Freedom/Operation New Dawn)era, and service-connected for PTSD. EBP initiators were alsomore likely to meet criteria for PTSD (five patients withoutPTSD began EBP, they may or may not have had subthresh-old PTSD symptoms given that PTSD was not a requirementfor clinic enrollment), to have received prior group psychother-apy for PTSD, and to have received delayed treatment withtheir EBP provider. EBP initiators and non-EBP initiators didnot differ significantly on ethnicity, gender, education, religion,income, prior inpatient stays, GAF, suicide risk, or comorbiddiagnoses. Two variables remained significant in the multivari-ate model predicting EBP initiation (Table 3): OEF/OIF/ONDveterans were less likely to begin EBP, and veterans who wereservice connected for PTSD were more likely to begin EBP.Nagelkerkes R2 for the overall model was .23, suggesting amoderate-to-large degree of improvement over the null model.To examine patient factors associated with EBP com-pletion, we used bivariate analyses (see Table 2) to com-pare EBP completers to EBP dropouts. EBP completerswere older, more likely to be non-OEF/OIF/OND veter-ans, and more likely to have received post high schooleducation than EBP dropouts. EBP completers were alsoless likely to have a history of psychiatric inpatient treat-ment and were more likely to have received prior grouppsychotherapy and delayed EBP treatment. No significantdifferences between EBP completers and EBP dropouts weredetected for ethnicity, gender, employment status, marital sta-tus, PTSD service connection, religion, income, GAF, suicideJournal of Traumatic Stress DOI 10.1002/jts. Published on behalf of the International Society for Traumatic Stress Studies.270 Mott et al.Table 3Patient-Level Predictors of EBP Initiation and Completion95% CIOutcome Variable SE Wald 2 OR Lower UpperBegan EBP PTSD service connection 0.85 0.39 4.70* 2.33 1.09 5.03OEF/OIF/OND veteran 0.74 0.35 4.54* 0.48 0.24 0.94Constant 0.12 0.39 0.10 1.13 Completed EBP Prior inpatient stay 2.05 0.73 5.02*** 0.13 0.03 0.54OEF/OIF/OND veteran 2.44 0.62 15.37*** 0.09 0.03 0.30Constant 2.25 0.52 18.12*** 9.52 Note. This table lists only those variables that were retained in the final model. EBP= evidence-based psychotherapy; degree of freedom= 1 for all predictor variables;OR = odds ratio; CI = confidence interval; OEF/OIF/OND = Operation Enduring Freedom/Operation Iraqi Freedom/Operation New Dawn; PTSD = posttraumaticstress disorder.*p < .05. **p < .01. ***p < .001.risk, comorbid diagnoses, or EBP type. Two variables remainedsignificant in the model predicting EBP completion (Table 3):OEF/OIF/OND veterans and those with a prior psychiatric in-patient stay were less likely to complete EBP. NagelkerkesR2 for the overall model predicting EBP completion was .42,indicating a high level of model fit.DiscussionThis study examined initiation and completion of EBP forPTSD in treatment-seeking veterans. Among our sample of pa-tients who received individual therapy from a provider trainedin EBP, 11.4% began EBP. This is some what higher thanShiner et al.s (2012) recent estimate of EBP use across sixVA PTSD clinics (6.3%) and is likely accounted for by thefact that Shiner and colleagues examined EBP initiation amongall clinic patients, whereas this study examined EBP initiationamong psychotherapy users treated by EBP-trained clinicians.Comparatively more patients received PE (8.5%) than CPT(3.0%), possibly reflecting treatment preferences of patients orproviders. Although this study examined patients who receivedindividual EBP (and did not examine receipt of group EBP),the PTSD and anxiety clinic where the study was conductedoffered CPT in both group and individual format, whereasPE was offered exclusively as an individual treatment. Con-sequently, some patients may have opted to receive CPT in agroup format, which may account for the lower frequency of in-dividual CPT initiation. Overall, our data suggest that althoughlarge-scale initiatives to disseminate CPT and PE are mak-ing inroads, most patients are not engaging in gold-standardinterventions.This study also examined veterans course of psychother-apy to investigate psychotherapy dose and attrition from EBP.The observed dropout from PE (30.9%) falls within the ob-served dropout range from previous PE trials conducted withveterans (21%38%; Schnurr et al., 2007; Tuerk et al., 2011;Tuerk, et al., 2010). The observed dropout from CPT (50.0%)exceeds previously reported CPT dropout in veterans (15%35%; Chard, Schumm, Owens, & Cottingham, 2010; Monsonet al., 2006; Morland, Hynes, Mackintosh, Resick, & Chard,2011; Surs, Link-Malcolm, Chard, Ahn, & North, 2013), andmay be biased by the relatively small sample who began CPT(n = 24). Importantly, our findings do not reflect the clinic-wide CPT dropout, as we did not account for patients whoreceived CPT in a group format. Additionally, lack of providertraining may have impacted attrition given that some studyproviders did not complete all elements of VA EBP train-ing. There was also a large range in the number of individ-ual patients providers saw after training and the proportionof EBP initiators within providers caseloads, suggesting thatproviders likely varied in their experience and expertise deliv-ering EBP. Although not examined directly in this study, theimpact of these provider variables on EBP completion warrantsinvestigation.Regarding treatment dose, EBP initiators received more psy-chotherapy sessions than non-EBP initiators. EBP treatmentsmay promote continued engagement, or alternatively, patientswho are likely to adhere to treatment may engage selectivelyin EBP. One unexpected finding was the high proportion ofpatients who received a combination of EBP and non-EBPsessions. Although patients who began EBP received an av-erage of 10 EBP sessions, which aligns with the traditionallength of CPT or PE, they also participated in an average ofseven non-EBP sessions. Providers may be engaging patientsin preparatory work prior to EBP, continuing therapy after EBPto address residual symptoms, or interspersing non-EBP ses-sions. This finding comports well with recent data suggestingthat VA providers frequently report integrating CPT or PE withanother treatment framework (Cook et al., 2014). Additional re-search is needed to identify how delivery of non-EBP sessionsbefore, throughout, or following EBP may impact treatmentengagement or effectiveness.Within this treatment-seeking sample, OEF/OIF/OND vet-erans were less likely than their counterparts from otherJournal of Traumatic Stress DOI 10.1002/jts. Published on behalf of the International Society for Traumatic Stress Studies.Veterans in Evidence-Based Therapy for PTSD 271service eras to begin EBP. Unique aspects of EBP may deterCPT or PE initiation in this population, functioning as a secondlayer of barriers beyond the well-recognized factors that inhibittreatment initiation in general (e.g., distance to care, stigma;Ouimette et al., 2011, Vogt, 2011). Consistent with previous re-ports that OEF/OIF/OND veterans are more likely to terminateprematurely fromCPTandPE (Chard et al., 2010; Jeffreys et al.,2014), OEF/OIF/OND veterans were also less likely than veter-ans of other eras to complete EBP. This finding, however, is notunique to EBP and may be an artifact of the well-documentedchallenges in retaining returning veterans in treatment (Sealet al., 2010; Stecker, Fortney, Hamilton, & Ajzen, 2007).Continued attention to outreach and interventions to increaseacceptability of EBP among recently returning veterans isneeded.Findings also indicated that PTSD service connection wasassociated with increased likelihood of EBP initiation. Veteranswho receive mental health disability benefits tend to experienceconsiderably more impairment than those who do not (Stein,Anderson, Lassor, & Freidmann, 2006), and clinicians maybe more apt to begin EBP with those veterans who presentwith significant PTSD symptoms and functional impairment(both criteria for approval of a PTSD disability claim). Serviceconnection also promotes access to healthcare by determiningthe patients priority status in VA, and care related to a service-connected disability is exempt from copay. Relatedly, veteranswho are service-connected are often not able or eligible towork,and may be more likely to have the time to commit to weeklyEBP sessions.We suspect that previous psychiatric inpatient treatment,which was associated with decreased likelihood of EBP com-pletion, may be a proxy for veterans preparedness to partici-pate EBP. Although empirical research supports CPT and PE assafe and effective treatments, both require veterans to disclosedetails of their trauma, and are thus emotionally demandingprotocols. A previous mental health inpatient stay may reflecta level of clinical severity or instability predictive of poor re-sponse to these intensive treatments. Moreover, regardless ofthe patients actual ability to tolerate the treatments, an inpatienthistory may impact the providers confidence in the patientsability to tolerate the affective intensity often associated withCPT and PE.Overall, we found that veterans who begin EBP are diversewith respect to clinical and demographic features. Consistentwith VHA policy that all veterans with PTSD must have accessto CPT or PE, these treatments are being administered to a rangeof veterans, including those who possess some of the potentialcontraindicators identified in the CPT and PE manuals (e.g.,substance use disorders, high suicide risk). Although there isincreasing evidence that EBP for PTSD may be effective forthose with comorbid substance use disorders, research is sparsewith respect to other potential contraindicators. Thus, the reachof these treatments in clinical practice may be growing morerapidly than the research base supporting the use of EBP forPTSD in these populations.This study has several limitations, including the use of datafrom patient medical records, which are subject to the valid-ity of patients reports and the accuracy of clinicians records.We also relied on provider report for initial identification ofEBP patients (subsequently confirmed via chart review), andproviders may have varying perceptions of what constitutesparticipation in a CPT or PE session. It is also possible thatparticipants may have received EBP prior to the study time-frame or from a non-VA provider. Thus, there may be patientsnot identified as EBP initiators who received aspects of CPTor PE. In the absence of treatment outcome data, we definedEBP completion based on an a priori benchmark (seven EBPsessions) and provider report. Early responders who improvedand therefore terminated after fewer than seven sessions arenot captured in our completer sample. Although receipt of EBPwas verified through medical record documentation indicatingdelivery of core elements of EBP, we did not otherwise as-sess clinician fidelity to EBP protocols and it is unknown howtherapist adherence may have impacted study findings.The methods used in this retrospective chart review studydid not allow for examination of the process by which patientsand providers selected a course of EBP or non-EBP treatment,nor examination of which system, clinic, or therapist factorsimpacted EBP initiation or completion. Providers in this studyattended a voluntary EBP training, suggesting that they werereceptive to administering EBP, and our study period coincidedwith widespread efforts to disseminate CPT and PE throughoutVHA. Data were collected, however, prior to the release of aVHAmandate that the offer of CPT or PE must be documentedin the medical records of all patients with a primary PTSD di-agnosis (VHA, 2012), and it is unknown whether all patientswere offered EBP. It is possible that clinicians offered EBP toa subset of patients based on their preferences or clinical judg-ment. Although not the focus of this study, it is important toacknowledge that provider factors such as theoretical orienta-tion, experience administering EBP, and attitudes towards EBPmay also influence the likelihood that a patient will begin orcomplete EBP. Further research is needed to disentangle theinfluence of the multilevel factors impacting utilization of EBPfor PTSD, and to better characterize the nature of treatmentdecision-making processes surrounding the initiation of thesetreatments in routine clinical care (e.g., how patients are iden-tified as appropriate for EBP, how EBP and non-EBP optionsare introduced and/or discussed).All veterans in this study were treatment seeking, allowingus to expand upon previous literature by examining factors thatincrease likelihood of initiating EBP as opposed to alternativetreatments. Our data suggest that rates of CPT and PE initiationand completion remain low, and indicate that patient demo-graphic and clinical characteristics are associated with likeli-hood of EBP initiation and completion. That EBP for PTSD isbeing provided to veterans with complex and comorbid clinicalprofiles, high clinician-rated suicide risk, GAF scores below50, and less than a high school education indicates that the dis-semination of EBP for PTSD throughout the VA system hasJournal of Traumatic Stress DOI 10.1002/jts. Published on behalf of the International Society for Traumatic Stress Studies.272 Mott et al.resulted in broad treatment reach. We eagerly anticipate futurestudies that extend existing knowledge of when and to whomthese treatments are administered. Identification of veteran sub-groups not currently accessing EBP can inform future researchon the causes of these disparities, and spark efforts to increaseengagement within these populations. Such research will ad-vance our ability to provide timely and high-quality treatmentto veterans with PTSD.ReferencesBarrera, T. L., Mott, J. M., Hundt, N. E., Mignogna, J., Yu, H. J.,Stanley, M. A., & Cully, J. A. (2014). Diagnostic specificity andmental health service utilization among veterans with newly diag-nosed anxiety disorders. General Hospital Psychiatry, 36, 192198.doi:10.1016/j.genhosppsych.2013.10.013Chard, K. M., Schumm, J. A., Owens, G. P., & Cottingham, S. M. (2010).A comparison of OEF and OIF veterans and Vietnam veterans receiv-ing cognitive processing therapy. Journal of Traumatic Stress, 23, 2532.doi:10.1002/jts.20500Cohen, B. E., Gima, K., Bertenthal, D., Kim, S., Marmar, C. R., & Seal, K.H. (2010). Mental health diagnoses and utilization of VA non-mental healthmedical services among returning Iraq and Afghanistan veterans. Journal ofGeneral Internal Medicine, 25, 1824. doi:10.1007/s11606-009-1117-3Cook, J. M., Dinnen, S., Thompson, R., Simiola, V., & Schnurr, P. P. (2014).Changes in implementation of two evidence-based psychotherapies forPTSD in VA residential treatment programs: A national investigation. Jour-nal of Traumatic Stress, 27, 17. doi:10.1002/jts.21902Eftekhari, A., Ruzek, J. I., Crowley, J. J., Rosen, C. S., Greenbaum, M. A., &Karlin, B. E. (2013). Effectiveness of national implementation of prolongedexposure therapy in Veterans Affairs care. JAMA Psychiatry, 70, 949955.doi:10.1682/JRRD.2011.08.0152Foa, E. B., Hembree, E. A., & Rothbaum, B. O. (2007). Prolonged exposuretherapy for PTSD: Emotional processing of traumatic experiences. NewYork, NY: Oxford University Press.Galovski, T. E., Blain, L. M., Mott, J. M., Elwood, L., & Houle, T. (2012).Manualized therapy for PTSD: Flexing the structure of cognitive process-ing therapy. Journal of Consulting and Clinical Psychology, 80, 968981.doi:10.1037/a0030600Gilbert, E. H., Lowenstein, S. R., Koziol-McLain, J., Barta, D. C., & Steiner,J. (1996). Chart reviews in emergency medicine research: Where are themethods? Annals of Emergency Medicine, 27, 305308. doi:10.1016/S0196-0644(96)70264-0Gros, D. F., Yoder, M., Tuerk, P. W., Lozano, B. E., & Acierno, R. (2011).Exposure therapy for PTSD delivered to veterans via telehealth: Predictorsof treatment completion and outcome and comparison to treatment deliveredin person. Behavior Therapy, 42, 276283. doi:10.1016/j.beth.2010.07.005Jeffreys, M. D., Reinfeld, C., Nair, P. V., Garcia, H. A., Mata-Galan, E.,& Rentz, T. O. (2014). Evaluating treatment of posttraumatic stress dis-order with cognitive processing therapy and prolonged exposure therapyin a VHA specialty clinic. Journal of Anxiety Disorders, 28, 108114.doi:10.1016/j.janxdis.2013.04.010Karlin, 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, 663673.doi:10.1002/jts.20588McHugh, R. K., & Barlow, D. H. (2010). The dissemination and implementa-tion of evidence-based psychological treatments: A review of current efforts.American Psychologist, 65, 7384. doi:10.1037/a0018121Monson, C. M., Schnurr, P. P., Resick, P. A., Friedman, M. J., Young-Xu,Y., & Stevens, S. P. (2006). Cognitive processing therapy for veterans withmilitary-related posttraumatic stress disorder. Journal of Consulting andClinical Psychology, 74, 898907. doi:10.1037/0022-006X.74.5.898Morland, L.A., Hynes, A. K., Mackintosh, M. A., Resick, P. A., & Chard,K. M. (2011). Group cognitive processing therapy delivered to veteransvia telehealth: A pilot cohort. Journal of Traumatic Stress, 2, 465469.doi:10.1002/jts.20661Mott, J. M., Hundt, N. E., Sansgiry, S., Mignogna, J., & Cully, J. A. (2014).Changes in psychotherapy use among veterans with depression, anxiety, andPTSD. Psychiatric Services, 65, 106112. doi:10.1176/appi.ps.201300056Ouimette, P., Vogt, D.,Wade,M., Tirone, V., Greenbaum,M.A., Kimerling, R.,. . . Rosen, C. S. (2011). Perceived barriers to care among veterans healthadministration patients with posttraumatic stress disorder. PsychologicalServices, 8, 212223. doi:10.1037/a0024360Rauch, S., Defever, E., Favorite, T., Duroe, A., Garrity, C., Martis, B., &Liberzon, I. (2009). Prolonged exposure for PTSD in a Veterans HealthAdministration PTSD clinic. Journal of Traumatic Stress, 22, 6064.doi:10.1002/jts.20380Resick, P. A., Monson, C. M., & Chard, K. M. (2008). Cognitive processingtherapy: Veteran/military version. Washington, DC: Department of VeteransAffairs. Retrieved from http://depts.washington.edu/hcsats/PDF/research/Cognitive%20Processing%20Therapy%20Manual%208.08.pdfSchnurr, P. P., Friedman, M. J., Engel, C. C., Foa, E. B., Shea, M. T., Chow, B.K., . . . Bernardy, N. (2007). Cognitive behavioral therapy for posttraumaticstress disorder in women: A randomized controlled trial. Journal of theAmerican Medical Association, 297, 820830. doi:10.1001/jama.297.8.820Seal, K. H., Maguen, S., Cohen, B., Gima, K. S., Metzler, T. J., Ren, L.,. . . Marmar, C. R. (2010). VA mental health services utilization in Iraqand Afghanistan veterans in the first year of receiving new mental healthdiagnoses. Journal of Traumatic Stress, 23, 516. doi:10.1002/jts.20493Sharpless, B. A., &Barber, J. P. (2011). A clinicians guide to PTSD treatmentsfor returning veterans. Professional Psychology: Research and Practice, 42,815. doi:10.1037/a0022351Shiner, B., DAvolio, L. W., Nguyen, T. M., Zayed, M. H., Young-Xu, Y.,Desai, R. A., . . . Watts, B. V. (2012). Measuring use of evidence basedpsychotherapy for posttraumatic stress disorder. Administration and Pol-icy in Mental Health and Mental Health Services Research, 40, 311318.doi:10.1007/s10488-012-0421-0Spoont, M., Murdoch, M., Hodges, J., & Nugent, S. (2010). Treatment receiptby veterans after a PTSD diagnosis in PTSD, mental health, or general med-ical clinics. Psychiatric Services, 61, 5863. doi:10.1176/appi.ps.61.1.58Stecker, T., Fortney, J., Hamilton, F., Sherbourne, C. D., & Ajzen, I. (2010).Engagement in mental health treatment among veterans returning from Iraq.Patient Preference and Adherence, 4, 4549. doi:10.2147/PPA.S7368Stein, M. D., Anderson, B. J., Lassor, J., & Friedmann, P. D. (2006).Receipt of disability payments by substance users: Mental and phys-ical health correlates. American Journal of Addiction, 15, 160165.doi:10.1080/10550490500528514Surs, A., Malcolm, J., Chard, K., Ahn, C., & North, C. (2013). A random-ized clinical trial of cognitive processing therapy for veterans with PTSDrelated to military sexual trauma. Journal of Traumatic Stress, 26, 2837.doi:10.1002/jts.21765Tuerk, P. W., Yoder, M., Grubaugh, A., Myrick, H., Hamner, M., & Acierno,R. (2011). Prolonged exposure therapy for combat-related posttraumaticJournal of Traumatic Stress DOI 10.1002/jts. Published on behalf of the International Society for Traumatic Stress Studies.Veterans in Evidence-Based Therapy for PTSD 273stress disorder: An examination of treatment effectiveness for veterans ofthe wars in Afghanistan and Iraq. Journal of Anxiety Disorders, 25, 397403.doi:10.1016/j.janxdis.2010.11.002Tuerk, P. W., Yoder, M., Ruggiero, K. J., Gros, D. F., & Acierno, R. (2010). Apilot study of prolonged exposure therapy for posttraumatic stress disorderdelivered via telehealth technology. Journal of Traumatic Stress, 23, 116123. doi:10.1002/jts.20494Veterans Health Administration. (2012). Local implementation of evidence-based psychotherapies for mental and behavioral health conditions.(VHA Handbook 1160.05). Washington, DC: Government PrintingOffice.Veterans Health Administration and Department of Defense. (2010). VA/DoDclinical practice guideline for the management of post-traumatic stress [Ver-sion 2.0]. Washington, DC: Veterans Health Administration, Department ofDefense.Vogt, D. (2011). Mental health-related beliefs as a barrier to service use formilitary personnel and veterans: A review. Psychiatric Services, 62, 135142. doi:10.1176/appi.ps.62.2.135Yoder, M., Tuerk, P. W., Price, M., Grubaugh, A. L., Strachan, M., Myrick,H., & Acierno, R. (2012). Prolonged exposure therapy for combat-relatedposttraumatic stress disorder: Comparing outcomes for veterans of differentwars. Psychological Services, 9, 1625. doi:10.1037/a0026279Journal of Traumatic Stress DOI 10.1002/jts. Published on behalf of the International Society for Traumatic Stress Studies.