In this webinar, Dr. Shira Maguen presents a series of VA electronic health record (EHR) studies examining the real‑world implementation of evidence‑based psychotherapies (EBPs) for post‑traumatic stress disorder (PTSD) among post‑9/11 Iraq and Afghanistan veterans. The VA launched a national rollout of two EBPs—Prolonged Exposure (PE) and Cognitive Processing Therapy (CPT)—starting in 2005, training thousands of clinicians. Using natural‑language processing to identify CPT and PE sessions in the VHA EHR, the researchers found that 77% of 265,566 veterans received no EBP, only 9% received a minimally adequate dose (≥8 sessions in 24 weeks), and the remainder received 1‑7 sessions. Emulated (mock) trials compared outcomes of veterans receiving CPT or PE to those receiving non‑EBP individual psychotherapy. Initiating CPT yielded a 4.8‑point reduction on the PTSD Checklist (PCL‑4) versus non‑EBP, and completing ≥8 CPT sessions yielded a 6.5‑point reduction. Initiating PE produced a 3‑point reduction, while completing ≥8 PE sessions achieved a 10‑point reduction, meeting the clinically significant threshold. Overall, real‑world improvements were smaller than those reported in randomized controlled trials, likely due to variable engagement, comorbidities, and treatment burden. A logistic regression of 32,780 veterans with at least one EBP and adequate PCL measurements identified predictors of clinically significant improvement (≥10‑point PCL drop): female gender, early engagement (within first year), individual modality, adequate dose, and drug‑dependence diagnosis were associated with higher odds of improvement; Black race, group CPT, and being service‑connected for PTSD were associated with lower odds. Hierarchical Bayesian models on a more recent cohort (2008‑2019) confirmed persistent racial disparities, with Black veterans showing significantly lower odds of improvement across multiple analytic approaches. The findings highlight that only about one‑third of veterans receiving EBPs achieve meaningful symptom reduction, underscoring the need to address low utilization, optimize dose and modality, and eliminate racial inequities in PTSD care within the VA.
Chapters
We're just excited to be able to present this webinar through the support of the Weill Institute for Neurosciences, and I especially want to thank Steve Hauser for giving me the encouragement to go ahead and put this together. Shira McGinn, who will speak first. She's an incredible 15, 20-year researcher at the VA Medical Center in San Francisco studying PTSD. She'll tell you all about that. She's done some incredible real-world research, is kind of, I think, how we think of it. So Shira, would you like to begin?
Thank you. Today I'm going to be talking about some research that we have done at the VA with evidence-based psychotherapy with the electronic health records. I'm going to talk about utilization rates, effectiveness, and some disparities that we found as well. So I'm just going to acknowledge some co-authors and collaborators and start by telling you that the Veterans Health Administration implemented a national rollout for evidence-based psychotherapies to treat PTSD in about— starting in about 2005. This initiative involved training thousands of clinicians nationwide in two evidence-based psychotherapies for PTSD. This was both in prolonged exposure as as well as cognitive processing therapy, which I'll call CPT. So we know that there have been multiple, um, RCTs at this point for both CPT and PE, several meta-analyses, and we know that people experience clinically significant PTSD symptom improvements. But it's really important to understand if veterans on the ground are improving and benefiting from EBPs in an outpatient setting. So we know that in RCTs, people have certain inclusion and exclusion criteria, and we just wanted to really understand if this applied to people who were receiving care at the VA and thought it was a unique opportunity to study that. We also know that several studies have found potential disparities. So I'll show you some data that we found as well, that Black veterans demonstrated less PTSD symptom improvement than their white counterparts following some of these PTSD EBPs specifically. So we wanted to leverage the VHA electronic health record where there's over a decade and now really closer to 15 years of data to evaluate whether veterans with PTSD utilize these evidence-based psychotherapies and whether they improve. So this was kind of a main goal to— and I'm going to share all these data with you in a very condensed way, but I'll let you know where you can get more information if you're interested. So we wanted to identify whether veterans were using CPT PE, right? Because you can have the most amazing therapies, and if people aren't using them, it's not going to get us very far to identify whether veterans who received these treatments improved and whether there were racial disparities. So we started out by looking at 3 different manuscripts where I'll kind of share the results with you. We wanted to look at update— uptake and prevalence of PTSD EBPs, effectiveness and comparative effectiveness. So I'm going to share a mock RCT that we did. And then what are the predictors of whether people improve in these treatments? Again, looking particularly for disparities. For those of you who are interested, we published this article, which I'm going to share more information about in just a second. So Our research questions are of veterans with PTSD who are receiving psychotherapy, how many actually received an EBP in the period from 2004 to 2014, which is the largest integrated VA national healthcare system, US integrated healthcare system. And then we also wanted to look at, are veterans receiving minimally adequate EBP treatment, right? So this is defined as at least 8 sessions in 24 weeks. So we know that many people start and then don't finish up. And we wanted to, again, wanted to make sure that they got a dose that was adequate for us to say they had really tried it. So we wanted to focus on Iraq and Afghanistan veterans because those are the veterans who most recently came back from war. And so we looked at those who were receiving care between 2001, 2015. They had to be diagnosed with PTSD and have that in their chart. And from there, we narrowed it down to a population with— that had actually received some kind of psychotherapy. So that was about a group of 265,566 veterans. So from there, we wanted to look at whether they had gotten an EBP. And I'm not going to have time to go into this, but we actually use natural language processing there. Natural language processing to identify those who had received PE or CPT. And what we found was that 77.2% did not get any EBPs, that about 9.1% got the minimally adequate dose, and 13.8% got between 1 and 7 EBP sessions. So you can see really a minority are getting that minimally adequate dose of any EBPs. And I'm gonna show you, um, the numbers here by CPT and PE. So you can see that, um, the numbers are even smaller for PE. So 1.9% got 8 or more sessions, 4.8% got 1 to 7. And again, the numbers for CPT are just slightly higher with 7.4% getting 8 or more and 10.9% getting 1 to 7 sessions, and the vast majority not getting any EBPs. So when we take all these data together, we find that less than 10% of post-9/11 veterans got minimally adequate EBPs. There's several possible reasons that I'd like to throw out there that we can maybe come back to in the discussion, but several veterans might be on pharmacotherapy, they might be getting alternative treatments, especially if they have comorbid psychiatric conditions. There may be administrative barriers. And also, as we know, there's a huge cognitive and emotional load associated with EBPs, and that may be a barrier for many veterans who are just not really willing to either think about their trauma in an active way. They wanna put that in a box and not come back to it, and it's very difficult to open that back up. So many veterans have challenges with balancing other demands. These are, young, relatively younger veterans who may have families, work, school, and they may find the therapy too stressful or have concerns that the treatment isn't going to work for them. So now I want to share with you, you know, so we answered how many people are getting it. Now I want to share with you data on are these therapies effective if we look at a mock clinical trial. So we know that there have been RCTs, but again, we really wanted to narrow in on the veterans who are getting clinical care. So we used emulated trials to determine real-world effectiveness. And we wanted to compare those who received CPT individual to an individual comparison group, to those who received PE to an individual— as well to an individual therapy comparison group. So we have a lot of data on people who have received individual therapy that is not CPT or PE. So I'm going to share with you data here. I'm going to cut to the chase. So this was CPT versus— a non-EBP individual psychotherapy. So I'm gonna share with you data on those who initiated CPT as well as completed. So first you can see here, if we go to the first figure, that those who initiated CPT, CPT was associated with a 4.8 PTSD Checklist improvement, and that's the PCL-4 since that's the data that we have for most veterans. During this time period. So CPT was associated with a 4.8 PCL improvement compared to those initiating a non-EBP. So again, it's not— I will just say that a clinically meaningful improvement is more like 10 points, or in some cases even 15 or 20 points, right? So those are kind of the, the marks that we tend to look at. So those who completed 8 or more sessions of CPT had a 6.5 for PCL greater improvement than those completing 8 or more sessions of a non-EBP. And here are the numbers for PE. So, those initiating PE had a 3-point PCL improvement, and those completing 8 or more sessions of PE had a 10-point PCL improvement. So, you can see that for those who engaged in 8 or more sessions of PE, they did meet kind of that clinically significant threshold, but that was really the only group that met that. We also know, as you saw from the data, that people are less likely to engage in PE because it's a more— you have to read about, you have to share your trauma over and over again and be recorded and listen to those recordings. So it's fairly intensive and do— and as well as do a number of other exposures for those of you who are familiar. So it's quite an intensive treatment. So overall, we found that while patients engaging in EBPs experienced some improvements, The improvements absolutely did not reach levels observed in EBPs that were administered through randomized controlled trials. So patients in routine clinical practice may be more variable in their engagement, their ambivalence, their motivation, and they may be more skeptical in treatments. They also may have other clinical factors such as substance use disorders. They may have higher suicidal ideation. A lot of— sometimes people who have these conditions are excluded from trials. So it's important to, to kind of note that the patients that we— there were— we didn't exclude people. They could have lots of comorbidities. And so these are the numbers from general clinical charts. So given that most patients do not currently receive PE or CPT, there's a substantial improvement here to improve PTSD outcomes at the VA. So for my final set of slides, I'm going to talk about what leads people to improve. So what are the demographic, military, and clinical factors associated with clinically significant improvement in evidence-based psychotherapies for PTSD? For this, for this group of slides I'm going to share, we looked at 32,780 Iraq and Afghanistan war veterans who had at least one EBP, and they also had to have adequate PTSD symptom measurements, so they had to have a PCL. So what share with you is that overall, um, 67.5% did not demonstrate clinically significant PTSD improvement, um, via the PCL, and 32.5% did. So what, um, given that, what was associated with the likelihood of PTSD symptom improvement— again, we're defining that as at least 10-point improvement on the PCL— and what was associated with a decreased likelihood? So this is from a random effects logistic regression depression. You can see that female gender was associated with more improvement. So if you first kind of look at the arrow on the top, those are the factors that are significantly associated with more improvement. So if you were female, if you engaged in early— so within the first year of being engaged in psychotherapy, if you got CPT or PE, that's early versus delayed is longer than that. If you've got a dose of 8 or more sessions in 16 weeks, if you got PE if you got CPT individual versus group. And interestingly, if you had a diagnosis of drug dependence, you were also more likely to improve. Now, those factors associated with less improvement included race. So if you were Black, you improved less. If you only had CPT group, you improved less. And if you were service-connected, meaning disabled for PTSD, you also improved less. So then we also followed that up. This is hot off the presses more recently. We looked at veterans who engage in care between 2008 and 2019. We ran some hierarchical Bayesian logistic regressions. And I won't get into all the nitty-gritty about this, but we wanted to follow up the racial disparity finding. We wanted to see if those receiving care more recently, if that racial disparity was still present. And we found that indeed it was. So here instead instead of just calculating odds ratios, which you can see in that first figure, you can see that absolutely the odds ratios are below 1. So Black race is associated with less improvement. We also wanted to kind of look at the effect sizes specifically. And so we calculated Cohen's d, and then we defined a regional— a region of practical equivalence, which is also called a ROPE. And we wanted to see— you can see here that we wanted to make sure that Black race fell outside of the rope, and indeed you can see that it does here. And so in the parsimonious model, Black race was associated with lower odds of PTSD symptom improvement when we looked at it a number of different ways. So to summarize these last points, in our national sample, only 32.5% of veterans who received an EBP over a 15-year period evidenced clinically significant PTSD improvement. Women were more were more likely to improve. Black veterans were less likely to improve. The timing of EBP was important. The dose was important. And whether people got individual treatment versus group was also important. Group CPT improved less, and those who were service-connected or disabled for PTSD were also less likely to improve. So we know that despite national rollouts, EBP are routinely, relatively infrequently used in VA. Using real-world clinical data, we know that although EBPs are statistically better than treatment as usual, rates of improvements are lower than what we expect. So while factors like dose modality and timing are associated with improvement, there may be racial disparities that we absolutely need to work to better understand why these are happening and how we can improve those to, to close that gap and create more equity. And improve that disparity as well. So I'll stop there, and I will just acknowledge my funding sources and really appreciate your time and attention and look forward to kind of delving more into all of these topics in the discussion.
