Background This study evaluated whether natural language processing (NLP) of psychotherapy note text provides additional accuracy over and above currently used suicide prediction models. Methods We used a cohort of Veterans Health Administration (VHA) users diagnosed with post-traumatic stress disorder (PTSD) between 2004–2013. Using a case-control design, cases (those that died by suicide during the year following diagnosis) were matched to controls (those that remained alive). After selecting conditional matches based on having shared mental health providers, we chose controls using a 5:1 nearest-neighbor propensity match based on the VHA's structured Electronic Medical Records (EMR)-based suicide prediction model. For cases, psychotherapist notes were collected from diagnosis until death. For controls, psychotherapist notes were collected from diagnosis until matched case's date of death. After ensuring similar numbers of notes, the final sample included 246 cases and 986 controls. Notes were analyzed using Sentiment Analysis and Cognition Engine, a Python-based NLP package. The output was evaluated using machine-learning algorithms. The area under the curve (AUC) was calculated to determine models' predictive accuracy. Results NLP derived variables offered small but significant predictive improvement (AUC = 0.58) for patients that had longer treatment duration. A small sample size limited predictive accuracy. Conclusions Study identifies a novel method for measuring suicide risk over time and potentially categorizing patient subgroups with distinct risk sensitivities. Findings suggest leveraging NLP derived variables from psychotherapy notes offers an additional predictive value over and above the VHA's state-of-the-art structured EMR-based suicide prediction model. Replication with a larger non-PTSD specific sample is required.

Objective: Despite long-standing interest in posttraumatic stress disorder (PTSD) and opioid use disorder (OUD) comorbidity, there is a paucity of data on the prevalence of OUD in patients with PTSD. Therefore, there is limited understanding of the use of medications for OUD in this population. We determined the prevalence of diagnosed OUD and use of medications for OUD in a large cohort of patients with PTSD. Methods: We obtained administrative and pharmacy data for Veterans who initiated PTSD treatment in the Department of Veterans Affairs (VA) between 2004 and 2013 (n=731,520). We identified those with a comorbid OUD diagnosis (2.7%; n=19,998) and determined whether they received a medication for OUD in the year following their initial clinical PTSD diagnosis (29.6%; n=5,913). Using logistic regression, we determined the predictors of receipt of OUD medications. Results: Comorbid OUD diagnoses increased from 2.5% in 2004 to 3.4% in 2013. Patients with comorbid OUD used more health services and had more comorbidities than other patients with PTSD. Among patients with PTSD and comorbid OUD, use of medications for OUD increased from 22.6% to 35.1% during the same time period. Growth in the use of buprenorphine (2.0% to 22.7%) was accompanied by relative decline in use of methadone (19.3% to 12.7%). Patients who received buprenorphine were younger and more likely to be rural, white, and married. Patients who received methadone were older, urban, unmarried, from racial and ethnic minorities, and were more likely to see substance abuse specialists. While use of naltrexone increased (2.8% to 8.6%), most (87%) patients who received naltrexone also had an alcohol use disorder. Controlling for patient factors, there was a substantial increase in the use of buprenorphine, a substantial decrease in the use of methadone, and no change in use of naltrexone across years. Conclusions: OUD is an uncommon but increasing comorbidity among patients with PTSD. Patients entering VA treatment for PTSD have their OUD treated with opioid agonist treatments in large and increasing numbers. There is a need for research both on the epidemiology of OUD among patients with PTSD and on screening for OUD.

Introduction: Published research indicates that posttraumatic stress disorder (PTSD) is associated with increased mortality. However, causes of death among treatment-seeking PTSD patients remain poorly characterized. The study objective was to describe causes of death among PTSD patients to inform preventive interventions for this treatment population. Methods: A retrospective cohort study was conducted of all veterans who initiated PTSD treatment at any Department of Veterans Affairs Medical Center from fiscal year 2008-2013. The primary outcome was mortality within the first year after treatment initiation. In 2018, the collected data was analyzed to determine leading causes of death. For the top 10 causes, standardized mortality ratios (SMRs) were calculated from age-and sex-matched mortality tables of the U.S. general population. Results: 491,040 veterans were identified who initiated PTSD treatment. Mean age was 48.5 years (+/− 16.0 years), 90.7% were male, and 63.5% were of white race. In the year following treatment initiation, 1.1% (5,215/491,040) died. All-cause mortality was significantly higher for veterans with PTSD compared to the U.S. population (SMR 1.05, 95% CI: 1.02-1.08, p<0.001). Veterans with PTSD had a significant increase in mortality from suicide (SMR 2.52, CI: 2.24-2.82, p<0.001), accidental injury (SMR 1.99, CI: 1.83-2.16, p<0.001) and viral hepatitis (SMR 2.26, CI: 1.68-2.93, p<0.001) compared to the U.S. population. Of those dying from accidental injury, more than half died of poisoning (52.3%, 325/622).

Following a FY15 mandate, EMR templates documenting EBP delivery were widely used by therapists working in VA residential PTSD programs. EBP receipt measured using EMR templates was consistent with therapist self-report of EBT delivery. There were several patient-level predictors of EBP receipt and therapist-level predictors of EBP delivery. However, therapists most likely to deliver EBPs were clustered at a limited number of sites.

Extant literature suggests that patient-therapist gender matching may be associated with psychotherapy retention. We examined this relationship in a national cohort of Veterans (n = 506,471) initiating psychotherapy for posttraumatic stress disorder (PTSD) using multivariate logistic regression models. Overall, women were retained in psychotherapy at higher rates than men. When patient and therapist factors as well as practice patterns are considered, gender match between female patients with PTSD and female therapists was not a positive predictor of psychotherapy retention. Contrary to our expectations, gender match between male patients with PTSD and male therapist was a negative predictor of psychotherapy retention.