Evidence-Based and Evidence-Informed Treatments: A Naturalistic Study of the Impact of Treatment Type on Engagement in Posttraumatic Stress Disorder

Posttraumatic stress disorder (PTSD) is a debilitating disorder with lifetime prevalence rates between 6.9 and 8.3% in the US.^1,2 The deleterious effects of PTSD have been well-documented, including negative impacts on mental health,^3–5 overall functioning,^6–8 and quality of life.^9,10 PTSD is also associated with increased morbidity, mortality, and suicidal thoughts and behaviors.^11–19 Veterans, in particular, are at higher risk for developing PTSD, relative to the general population.^5,20–23 In fact, a meta-analysis of PTSD in post-9/11 Veterans estimates prevalence rates as high as 23%,²¹ and studies among Vietnam-era Veterans have estimated lifetime prevalence rates between 10% and 30%.^24–26 As such, they represent an at-risk population in need of targeted attention and intervention.

The development and study of evidence-based psychotherapies (EBPs) have revolutionized PTSD treatment and resulted in multiple efficacious treatments, including cognitive processing therapy (CPT),²⁷ prolonged exposure (PE) therapy,²⁸ eye movement desensitization and reprocessing (EMDR),²⁹ and written exposure therapy (WET).³⁰ These treatments have been extensively tested, broadly disseminated, and included in clinical practice guidelines, including Veterans Affairs (VA) and Department of Defense (DoD) guidelines that govern clinical care within VA settings.^31–34

Although these treatments are highly effective, there is significant variability in treatment response: patients often do not experience clinically meaningful change and/or retain their diagnosis after treatment completion,^35,36 and many drop out of treatment.^36,37 Furthermore, multiple studies have suggested that EBPs may have lower completion rates and may not be as robustly effective in naturalistic settings, relative to randomized controlled trials (RCTs).^38,39 Even when Veterans do have access to EBPs, they may not select them.⁴⁰ As a result, many clinics offer other evidence-informed, nontrauma-focused interventions for PTSD,⁴¹ such as Skills Training in Affective and Interpersonal Regulation (STAIR),^42,43 acceptance and commitment therapy (ACT),⁴⁴ and mindfulness-based stress reduction (MBSR). While not focused specifically on the traumatic event, these interventions address critical aspects of psychosocial functioning impacted by trauma, as well as transdiagnostic skills training (eg, interpersonal functioning, emotion regulation, acceptance, and mindfulness). Importantly, many of these interventions have preliminary empirical support for their efficacy,^45–48 but due to the current limited number of studies, VA/DoD clinical practice guidelines have determined there to be insufficient evidence to recommend for or against their use in the treatment of PTSD. No study to our knowledge has examined outcomes from these treatments within a naturalistic clinical setting that includes both EBPs for PTSD (eg, CPT, PE, EMDR, WET) along with other evidence-informed interventions (eg, STAIR, ACT, MBSR).

The current study examines the impact of trauma-focused EBPs listed in the VA/DoD Clinical Practice Guidelines during the time of the study³⁴ (hereafter referred to as EBPs) and other evidence-informed, non–trauma-focused interventions (hereafter referred to as EIPs) on PTSD treatment outcomes. We hypothesized that Veterans would be more likely to engage, attend a greater number of sessions, and receive an adequate dose of EIPs, relative to EBPs. We also hypothesized that they would be more likely to engage, attend a greater number of sessions, and receive an adequate dose of individual therapy relative to group therapy. Importantly, this study is not an RCT; Veterans in this study have presented to a trauma clinic and are free to select the therapy within which they would like to participate, in collaboration with their therapist. As such, this study is critically able to examine these outcomes within the context of routine clinical care. However, by the nature of this naturalistic study, we are unable to specifically address reasons for treatment discontinuation or outcomes beyond treatment engagement, receipt of an adequate dose, and number of sessions attended.

METHODS

Participants

This retrospective study includes 480 Veterans who presented to and were enrolled in the Washington, DC, VA Medical Center (VAMC) Trauma Services Program (TSP) during the 2019 calendar year. TSP is a specialty mental health outpatient clinic providing a variety of psychological treatments to Veterans with PTSD. To be enrolled in the clinic, Veterans complete an assessment and treatment planning session, during which a clinician conducts a diagnostic interview to determine that PTSD is the primary untreated diagnosis. The veteran must either (a) report a Criterion A trauma during the time of their military service, although the trauma does not have to be “military related” or (b) report a Criterion A trauma prior to their military service that was subsequently exacerbated by their experiences within the military. Finally, Veterans must be interested in and available to engage with weekly or biweekly, time-limited therapy to target trauma-related symptoms.

Procedure

This study is a retrospective chart review; the study team extracted data between January 2021 and August 2023, overseen by the PI and research coordinator. Identified Veterans were retrospectively followed for 1 year following their intake appointment, using the VA’s Computerized Patient Recording System (CPRS) and Joint Legacy Viewer (JLV). All procedures were approved by the DC VAMC Institutional Review Board and Research and Development Committee.

Measures

Demographic characteristics. Age, sex, self-reported race/ethnicity, and military service variables (eg, branch, era of service) were extracted from CPRS or JLV. Military service variables were included due to potential differences in mental and physical health (including rates of PTSD),^49–54 as well as in health behaviors and stigma.^51,55

Trauma-related characteristics. All trauma-related characteristics were extracted from the CPRS note associated with the Veteran’s intake appointment in the TSP. The Veteran’s identified index trauma was coded using the categories from the Life Events Checklist for DSM-5.⁵⁶ Whether or not a Veteran reported experiencing multiple traumas, deployment trauma, and military sexual trauma was coded as present, absent, or missing from the note. Whether or not the veteran reported minority stress was coded as present and index trauma-related, present and not index-trauma related, not present, or missing from the note.

Other mental health diagnoses. DSM-5 diagnoses of depression and substance use disorder were extracted from the CPRS note associated with the Veteran’s intake appointment in TSP and coded as present or absent, based on the diagnostic determination of the provider conducting the assessment.

Primary predictors. Treatment type was defined as either EBP or EIP. EBPs included PE, CPT, EMDR, and WET. EIPs included STAIR, mindfulness-based stress reduction (MSBR), acceptance and commitment therapy (ACT), dialectical behavior therapy (DBT), race-based stress and trauma (RBST), and coping skills and supportive therapy. Within person, an individual could have participated in more than 1 treatment. Each treatment per person was therefore split into “episodes of care,” defined as all of the sessions associated with a single treatment approach within person. Treatment type was extracted from visit notes in CPRS.

Treatment setting was defined as either individual therapy only, group therapy only, or both. Each episode of care was defined as either individual or group. Only person-level analyses include the “both” category, as it was not possible for a person to receive both group and individual treatment within a single episode of care. Treatment setting was extracted from visit notes in CPRS.

Primary outcomes. Number of sessions was collected via visit notes in CPRS, and only included full sessions occurring within TSP. Both number of sessions associated with the episode of care and total number of sessions across episodes of care within the follow-up period were coded.

Engagement was defined as having attended at least 2 sessions and was coded as binary. This more stringent definition was intentionally chosen, as a single session may better reflect treatment initiation, but not necessarily engagement; attendance of a second session—by definition—displays a willingness to at least return following treatment initiation.

Adequate dose was defined as attending at least 8 sessions within an episode of care, with the exception of WET (3 sessions). It should be noted that there are conflicting findings and a substantial lack of consensus on a definition of “adequate dose” within the literature. This includes studies showing that the median effective dose of PE is 4 sessions,⁵⁷ the median effective dose of CPT is 8 sessions,⁵⁸ and—on average—at least 10 sessions of CPT are needed to fall below clinical cutoff on the PCL.⁵⁹ There are also multiple preexisting definitions of EBP adequate dose and/or completion, to include at least 7 sessions,⁶⁰ 8 sessions over varying time frames,^61–69 9 sessions in 15 weeks,^70,71 and needing to have a final session note and/or completing all sessions of a protocol.^59,72,73 Within that context, we selected our definition based on prior convention for defining adequate course of treatment as 8 sessions^61–69 and then applied this definition consistently across protocols. In deference to WET’s substantially different length of treatment (eg, 5 sessions), we applied the most similar percentage of sessions we could (eg, completion of at least 3 sessions). Of note, examinations of our data revealed that no participants attended exactly 3 sessions of WET; as such, this cut point functionally behaves as at least 4 sessions within our sample.

Statistical Analyses

Statistical analyses were conducted at 2 levels: (1) episode of care and (2) person. We used logistic or linear regression models to examine person and episode of care-level differences in treatment outcomes of interest for all analyses, with the exception of person-level setting differences (χ² and analysis of variance). The fully adjusted models covaried for age, sex, race, branch of service, index trauma, report of experiencing multiple traumas, and number of episodes of care.

Post hoc logistic and linear regressions were conducted to examine differences between individual-only and group-only treatment settings on all person-level outcomes.

All analyses were conducted in R statistical package.

RESULTS

Overall, 480 Veterans completed an intake interview and were enrolled within TSP. In the year following intake, 71.87% (n = 345) engaged with treatment and 45.42% (n = 218) received an adequate dose of at least 1 treatment, with Veterans attending an average of 8.40 (median = 6) sessions. Of those that engaged in treatment, 63.19% received an adequate dose of at least 1 treatment. Veterans averaged 1.39 treatment episodes of care, with 70.83% (n = 340) attending only 1 episode, 21.88% (n = 105) attending 2 episodes, and 7.29% (n = 35) attending greater than 2 episodes. Full sample characteristics are reported in Table 1.

Approximately half of the Veterans planned to and/or engaged with at least 1 EBP (45.83%, n = 220). Treatment was most frequently planned and/or delivered in group-only format (49.38%, n = 237), although 27.50% (n = 132) planned to and/or received treatment in individual-only format, and 14.17% (n = 68) planned to and/or received treatment in both individual and group format.

Examinations by Episode of Care

There was no significant difference in engagement between those that planned to and/or received EBP vs EIP in either adjusted or unadjusted models (OR = 1.26, P =.27; OR = 1.39, P = .14, respectively). Those who planned to and/or received EBP were significantly more likely to receive an adequate dose, relative to those that planned to and/or received EIP in both unadjusted and adjusted models (OR = 1.44, P = .03; OR = 1.44, P = .04, respectively). Those who planned to and/or received EBP attended a significantly greater number of sessions per episode of care (7.60) relative to those that planned to and/or received EIP (6.00) in both unadjusted (b = 1.6, P = .001) and adjusted models (b = 1.64, P = .001).

There was no significant difference in terms of engagement, receipt of an adequate dose, or number of attended sessions between those who planned to and/or received group versus individual treatment in both unadjusted (OR = 1.06, P = .75; OR = .64, P = .12; b = −.38, P = .42, respectively) and adjusted models (OR = 1.06, P = .78; OR = 1.38, P = .08; b = −.38, P = .46, respectively). See Table 2.

Examinations by Person

Rates of engagement differed based on treatment type in unadjusted (OR = 1.76, P = .03) but not adjusted models (OR = 1.52, P = .17). Rates of receipt of an adequate dose differed based on treatment type in unadjusted (OR = 1.76, P = .004) but not adjusted models (OR = 1.52, P = .17). Finally, those who planned to and/or received at least 1 EBP attended a significantly greater number of total sessions (11.82) relative to those that planned to and/or received EIP (7.31) in both unadjusted (b = 4.51, P <.001) and adjusted models (b=1.69, P=.04).

There was a significant difference in terms of engagement (χ²₂= 10.16, P = .006), receipt of an adequate dose (χ²₂ = 11.95, P = .002), and number of attended sessions (F_2,434 = 29.99, P < .001) depending on method of treatment delivery (eg, individual-only, group-only, or both individual and group). Because those who planned to and/or received both individual and group therapy—by definition—planned to and/or attended more than 1 treatment epoch, they were excluded from post hoc sensitivity analyses designed to examine differences between those who planned to and/or received individual only versus group only treatment. These analyses revealed no significant difference in terms of engagement or receipt of an adequate dose between those who planned to and/or received individual-only or group-only treatment in both unadjusted (OR = 1.28, P = .32; OR = 0.96, P = .86, respectively) and adjusted models (OR = 1.0, P = .99; OR = 0.92, P = .75, respectively). Furthermore, those who planned to and/or received individual-only treatment attended a significantly greater number of total sessions (9.14) relative to those that planned to and/or received group-only treatment (7.32) in unadjusted (b = −1.81, P = .02) but not adjusted models (b = −1.48, P = .06). See Tables 3 and 4.

DISCUSSION

Overall, just over 70% of Veterans who were enrolled in TSP engaged with treatment, and just under half received an adequate dose of treatment (eg, attended at least 8 sessions of any treatment, except WET [3 sessions]). This rate is substantially lower than completion rates reported in many RCTs, despite differences in operational definitions, many of which are more stringent than those employed here (ie, completion of all sessions of a protocol, completion of at least 10 sessions of a protocol). In fact, recent meta-analyses of psychological treatments for PTSD have reported dropout rates within RCTs of 16%,³⁷ although greater dropout rates have been reported for guideline-recommended treatments (eg, 13–30%)^37,74 and among Veterans (eg, 16–30%).^39,75,76 Our findings therefore support prior work that suggest lower rates of adequate dose and/or completion in naturalistic settings, relative to RCTs,³⁸ although it should be noted that some individual RCTs reflect comparable dropout rates among Veterans (eg, as high as 46.6% for CPT⁷⁷ and 47–55.8% for PE^77,78).

While we are unable to investigate potential reasons for treatment dropout, recent qualitative studies illuminate several possible explanations, including practical logistical barriers (eg, conflicts with existing responsibilities at work, school, or caretaking; transportation; scheduling), acute distress within the context of treatment (eg, the treatment is uncomfortable), perceptions regarding treatment (eg, if the Veteran believes the treatment cannot work or feels they are not responding, or if symptoms are more burdensome and not being addressed), perceived improvements in symptoms, therapeutic relationship with provider, and stigma.^69,73,79,80 Critically, all of these explanations contain specific nuances that can lead to differential decisions. For example, investigations of treatment dropout have consistently found both perceived worsening in symptoms and acute distress within the context of treatment are factors that influence decision making regarding treatment continuation.^69,81 However, they found divergent treatment paths regarding this distress; specifically, while most Veterans in the study reported increases in at least some of their symptoms, they differed in the meaning attached to those changes such that completers were more likely to view increases as either normal or temporary within the context of treatment whereas noncompleters viewed it as an indication that the treatment was not working.⁷³ Furthermore, there is a strong body of research indicating that perceived improvements in symptoms may also result in treatment dropout—but with potentially very different clinical outcomes.⁸² Future studies that build upon the current research to carefully examine possible reasons for treatment non-engagement and discontinuation within naturalistic settings are desperately needed.

With regards to EBP and EIP comparisons, Veterans attended a greater number of sessions and were more likely to receive an adequate dose of an EBP episode of care, relative to EIP episodes. When examined by person, Veterans that planned to and/or attended any EBP did not differ significantly from those who planned to and/or attended exclusively EIP in terms of rates of engagement or receipt of an adequate dose; they did, however, attend a greater number of therapy sessions overall. These findings were contrary to both our hypotheses and prior literature that suggest higher dropout rates among EBPs, as compared to EIPs.^{37,41,74,83,84} The exact reason for this discrepancy is unknown, but it is possible that Veterans in this study had greater expectancies regarding completing an EBP, as they all presented to a trauma-focused clinic. It is also possible that providers within this specialty clinic hold a preference for EBPs due to more extensive training in these interventions, CBT theoretical backgrounds, and/or a preference to engage in treatments with strong support in the literature; such a bias could impact greater engagement and session attendance rates in these interventions.⁸⁵ That said, overall, these findings importantly suggest that Veterans in PTSD specialty clinics can not only tolerate, but may be more likely to attend a greater number of EBP sessions, relative to EIPs.

Rates of engagement and receipt of an adequate dose, as well as number of treatment sessions attended, did not differ between group and individual therapy. Prior literature regarding the impact of treatment delivery method on dropout rates and session attendance has been mixed, with some finding no differences,³⁷ some finding lower dropout/greater attendance for treatment delivered in individual format compared to group format,³⁹ and still others finding the opposite.⁸⁶ Importantly, outcomes examined (eg, number of sessions attended, completion, adequate dose, dropout), study populations (eg, Veteran, civilian, focus on different types of trauma), and setting (eg, naturalistic, RCT) can vary greatly between studies, which may contribute to disparate results. Given this, it may be particularly important that clinicians work with patients to consider both patient preferences for treatment delivery and individual-level factors that may be relevant to success in group contexts (eg, motivation, interpersonal skills, comfort with disclosure in group settings).⁸⁷

This study builds a foundation from which future work can occur. While we importantly were able to examine treatment engagement and receipt of an adequate dose within a naturalistic setting, we were unable to examine other critical outcomes (eg, symptom reduction), as this study spanned the beginning of the COVID-19 pandemic, which resulted in substantial reductions in the collection of measures such as the PTSD Checklist for DSM-5 (PCL-5) during the transition to telehealth. Furthermore, even outside of the context of a global pandemic, there are significant challenges to implementing measurement-based care in many clinical settings, despite extensive evidence showing its utility.^88–90 That said, uptake of measurement-based care continues to grow, and recommendations have been included in clinical practice guidelines³¹; future studies that examine outcomes such as reductions in PTSD symptoms and improvements in functioning and quality of life within naturalistic settings are needed.

We also are unfortunately unable to examine outcomes among those who did not engage or receive an adequate dose of treatment, as they did not respond to multiple attempts at outreach and were therefore lost to follow-up. While this is the nature of a naturalistic study, future work that is intentionally designed to prospectively follow individuals enrolled in active clinics and/or retrospectively identify individuals who drop out of treatment is needed. Such studies would importantly build upon the work already being done within the context of RCTs and a few prior VA studies.^69,73,79,80

Finally, although our results suggest that EBPs result in greater rates of receipt of an adequate dose and greater session attendance, on average, these results were far from homogenous. As such, future research is needed to understand for whom, and under what conditions each treatment might be most effective. Such work on moderation has already begun, although results have been mixed.^91–95 This may be due to the fact that the vast majority of these studies have been completed as secondary data analyses; as such, studies specifically designed to examine questions of moderation are particularly needed.

This study has several limitations. First, as mentioned above, missing data impacted our ability to examine certain outcomes (eg, symptom reduction); it also meant we were not able to control for co-occurring mental health conditions (eg, depression, substance use). Symptom severity scores, in particular, may have been impacted by the shift into telehealth driven by the COVID-19 pandemic. Further, this shift may also have impacted session attendance, although the exact nature of that confound is difficult to determine, and there have been mixed findings from studies examining the impact of the transition in care during the onset of the pandemic.^96,97 Second, we compared many different treatments and determined “adequate dose” based on prior convention, with the exception of WET; that said, there is substantial variability in this definition within the literature and it is possible that other definitions would have yielded different results. This may be particularly true for examinations of WET, as the protocol itself is of a substantially shorter duration. Future data-driven examinations to support a shared definition of “adequate dose” are desperately needed, but unfortunately beyond the scope of the present study. Third, these data were collected as part of a naturalistic study within an active specialty PTSD clinic within the VA; our examination was therefore limited to those therapies offered within this clinic at the time of data collection. Some therapies notably were not offered, including interpersonal psychotherapy (IPT) and person centered therapy (PCT), both of which have shown efficacy in treating PTSD.^78,94 Fourth, this study compared EBPs and EIPs, which conflates treatments from different theoretical perspectives (eg, exposure-based [PE and WET] and cognitive [CPT]) and unfortunately does not allow for a more specific investigation of the impact of those theoretical perspectives on treatment engagement, session attendance, and receipt of an adequate dose of treatment. Future, more specific, work in this area is needed. Fifth, all data were collected from medical record charts; as such, any data that were not collected as a part of routine care and entered into the medical record could not be included within this study design (eg, those associated with psychotherapy research protocols). Finally, given the specialized nature of this setting, results may not generalize to other clinics within the VA or outside of the VA system. Despite these limitations, this study derives considerable strength from its size, representation of commonly underrepresented racial groups, and the fact that we were able to examine treatment outcomes within the context of routine care and in a naturalistic setting.

CONCLUSIONS

In this naturalistic study, over 70% of Veterans who were enrolled in a specialty PTSD clinic engaged with treatment, but just under half received an adequate dose, supporting prior work suggesting that dropout rates are higher in naturalistic settings relative to RCTs. Furthermore, our findings often contrasted with those of RCTs regarding the impact of treatment type (eg, EBPs, EIPs) and delivery modality (eg, group or individual) on treatment outcomes of interest. As such, this study importantly contributes to the existing literature and highlights the need for additional studies that examine PTSD treatment outcomes within the context of routine care.

Article Information

Published Online: May 14, 2025. https://doi.org/10.4088/JCP.24m15567
© 2025 Physicians Postgraduate Press, Inc.
Submitted: August 15, 2024; accepted February 24, 2025.
To Cite: Arenson M, Crone B, Cortell R, et al. Evidence-based and evidence-informed treatments: a naturalistic study of the impact of treatment type on engagement in posttraumatic stress disorder. J Clin Psychiatry 2025;86(2):24m15567.
Author Affiliations: Washington DC VA Medical Center, Washington, DC (Arenson, Cortell); Boston University Chobanian and Avedisian School of Medicine, Boston, Massachusetts (Arenson); VA Boston Healthcare System, Boston, Massachusetts (Arenson); Fort Sam Houston, San Antonio, Texas (Crone); the Trauma Resilience and Education Center of Greater Washington, DC, Washington, DC (Carlin).
Corresponding Author: Melanie Arenson, PhD, Boston University Chobanian and Avedisian School of Medicine, 150 S Huntington Ave, Boston, MA 02130 ([email protected]).
Relevant Financial Relationships: The authors report no financial or other relationship relevant to the subject of this article.
Funding/Support: None.

Clinical Points

• Little is known about comparative outcomes for evidence based psychotherapies (EBPs) versus other evidence informed treatments for PTSD in a naturalistic setting.
• Veterans in PTSD specialty clinics not only can tolerate EBP but may be more likely to attend a greater number of sessions, relative to other evidence-informed treatments.

Editor’s Note: We encourage authors to submit papers for consideration as a part of our Focus on Psychotherapy section. Please contact John C. Markowitz, MD, at Psychiatrist.com/contact/markowitz or Rachel C. Vanderkruik, PhD, at Psychiatrist.com/contact/vanderkruik.