Using natural language processing to assess MDD psychotherapy patterns among Veterans Affairs patients with suicide risk.

Back to table of contents Previous article Next article LettersFull AccessSuicide Prevention in Patient and Nonpatient Populations: In ReplyClaire A. Hoffmire, Ph.D., Janet E. Kemp, R.N., Ph.D., Caitlin Thompson, Ph.D.Claire A. HoffmireSearch for more papers by this author, Ph.D., Janet E. KempSearch for more papers by this author, R.N., Ph.D., Caitlin ThompsonSearch for more papers by this author, Ph.D.Published Online:1 Oct 2015https://doi.org/10.1176/appi.ps.661002AboutSectionsPDF/EPUB ToolsAdd to favoritesDownload CitationsTrack Citations ShareShare onFacebookTwitterLinked InEmail IN REPLY: We thank Dr. Coffey and colleagues for their commentary on our recent study comparing suicide risk for veterans who did and did not use VHA services. Comparing this study with their recent analysis of suicides among Henry Ford Health System members highlights the importance of continued evaluation of suicide prevention activities.The main findings from our analysis were that veterans, particularly female veterans, were at higher and increasing risk of suicide, compared with their nonveteran peers and that use of VHA services mitigated some of this risk. By compiling death certificate records for all known suicide decedents, the Department of Veterans Affairs (VA) archive of state suicide data made direct comparisons between VHA and non-VHA veterans possible for the first time (1–3). Veteran status in this database is carefully validated by using VA and Department of Defense records (2). Validation of veteran status was shown to have an impact on effect estimates in a study evaluating the relationship between veteran status and firearm suicide (3). Death certificate indication of military history is particularly problematic for female and younger veterans (2)—two of the fastest-growing veteran subgroups for which there is an immediate need to better understand suicide risk and trends.The VA has, however, been evaluating suicide risk among VHA-utilizing veterans since implementing the Suicide Prevention Program in 2007 (1). As summarized in our report, VA’s Suicide Prevention Program is poised to reach the broad population of VHA users and provide access to an integrated network of services aimed at reducing risk. Initiatives include awareness and education campaigns such as “The Power of 1,” a 24-hour Veterans Crisis Line, universal training in risk recognition and management for VHA employees, and the placement of suicide prevention coordinators in each VA medical center to support high-risk patients. Recent efforts also include the development of predictive modeling strategies to identify VHA patients at high risk of suicide (4).This strong, multifaceted, and continually evolving program makes the VA a clear leader in suicide prevention. As such, we are pleased that our recent publication shed light on the pressing need to increase awareness of risk among female veterans. Response efforts will include continued research and program evaluation activities conducted collaboratively across the VA Suicide Prevention Program and VA’s Women Veterans Health Care. Comprehensive, high-quality women’s health care is already a top priority within the VA. Every medical center has a women veterans program manager to serve as an administrative leader for women’s health programs and an advocate for women veterans, and a Women Veterans Call Center has been established as a resource for women veterans seeking information on VA benefits and services.Similarly, we recommend that efforts to evaluate gender and age differences in the HFHS population are undertaken. The effectiveness of intervention programs is likely to vary considerably by these factors, and identifying such differences is a key step toward targeting and tailoring programs to prevent additional loss of life.References1 Kemp J, Bossarte R: Suicide Data Report, 2012. Washington, DC, US Department of Veterans Affairs, Mental Health Services, Suicide Prevention Program, Feb 2013Google Scholar2 Hoffmire CA, Piegari RI, Bossarte RM: Misclassification of veteran status on Washington State death certificates for suicides from 1999 to 2008. Annals of Epidemiology 23:298–300, 2013Crossref, Medline, Google Scholar3 Hoffmire CA, Bossarte RM: A reconsideration of the correlation between veteran status and firearm suicide in the general population. Injury Prevention 20:317–321, 2014Crossref, Medline, Google Scholar4 McCarthy JF, Bossarte RM, Katz IR, et al.: Predictive modeling and concentration of the risk of suicide: implications for preventive interventions in the US Department of Veterans Affairs. American Journal of Public Health (Epub ahead of print, June 11, 2015)Google Scholar FiguresReferencesCited byDetailsCited ByRates and Correlates of Suicidality in VA Intensive Case Management Programs4 May 2021 | Community Mental Health Journal, Vol. 58, No. 2JMIR Mental Health, Vol. 5, No. 3 Volume 66Issue 10 October 01, 2015Pages 1120-1121 Metrics PDF download History Published online 1 October 2015 Published in print 1 October 2015

Initiation of Assertive Community Treatment Among Veterans With Serious Mental Illness: Client and Program Factors

To optimize methods for identifying patients with diabetes based on computerized records and to obtain best estimates of diabetes prevalence in Department of Veterans Affairs (VA) patients. The VA Diabetes Epidemiology Cohort (DEpiC) is a linked national database of all VA patients since 1998 with data from VA medical visits, Medicare claims, pharmacy and laboratory records, and patient surveys. Using DEpiC, we examined concordance of diabetes indicators, including ICD-9-CM codes (250.xx), prescription drug treatment, HbA(1c) tests, and patient self-report. We determined the optimal criterion for identifying diabetes and used it in estimating diabetes prevalence in the VA. The best criterion was a prescription for a diabetes medication in the current year and/or 2+ diabetes codes from inpatient and/or outpatient visits (VA and Medicare) over a 24-month period. This definition had high sensitivity (93%) and specificity (98%) against patient self-report, and reasonable rates of HbA(1c) testing (75%). HbA(1c) testing alone added few additional cases, and a single diagnostic code added many patients, but without confirmation (reduced specificity). However, including codes from Medicare was critical. Applying this definition for 1998-2000, we identified an average of 500,000 VA patients with diabetes per year. We also estimated high and increasing diabetes prevalence rates of 16.7% in FY1998, 18.6% in FY1999, and 19.6% in FY2000 and an incidence estimated to be approximately 2% per year. Development and evaluation of methodology for analyzing computerized patient data can improve the identification of patients with diabetes. The increasing high prevalence of diabetes in VA patients will present challenges for clinicians and health system management.

Posttraumatic stress disorder (PTSD) is a prevalent psychiatric condition, particularly among US Veterans. PTSD-diagnosed patients are more likely to experience suicidal ideation, suicide attempts and death by suicide when compared to non-PTSD-diagnosed patients. The US Department of Veterans Affairs (VA) emphasizes evidence-based psychotherapy (EBP) for PTSD, including prolonged exposure and cognitive processing therapy. This study focuses on how suicide risk impacts PTSD by evaluating utilization of nondifferentiated psychotherapy and EBP in a national sample of VA patients diagnosed with PTSD who died by suicide in 2017-2018. The study used a dataset of VA patients diagnosed with PTSD who died by suicide and received psychotherapy in the year before death (cases) and patients who had comparable diagnoses, demographics and received psychotherapy during the same interval and remained alive (controls). Cases and controls were matched on suicide risk (high, moderate and low). The study tracked nondifferentiated psychotherapy and EBP and analyzed cases and control utilization rates across risk-tiers. The final sample included high-risk (cases = 171; controls = 2052), moderate-risk (cases = 428; controls = 4280) and low-risk (cases = 53; controls = 529) patients. EBP utilization was markedly low, especially among cases. Higher proportions of moderate- and low-risk controls received EBP and received more sessions than matched cases. Even with VA efforts to promote EBPs, usage remains limited, particularly among patients who die by suicide. Further research is needed to understand utilization barriers and improve EBP delivery to better support PTSD-diagnosed patients and reduce their suicide risk.

The rapid change in healthcare has focused attention on the necessary development of a next-generation electronic health record (EHR) to support system transformation and more effective patient-centered care. The Department of Veterans Affairs (VA) is developing plans for the next-generation EHR to support improved care delivery for veterans. To understand the needs for a next-generation EHR, we interviewed 14 VA operational, clinical and informatics leaders for their vision about system needs. Leaders consistently identified priorities for development in the areas of cognitive support, information synthesis, teamwork and communication, interoperability, data availability, usability, customization, and information management. The need to reconcile different EHR initiatives currently underway in the VA, as well as opportunities for data sharing, will be critical for continued progress. These findings may support the VA's effort for evolutionary change to its information system and draw attention to necessary research and development for a next-generation information system and EHR nationally.

Reply to: “Treatment of veterans with hepatitis C in the United States Department of Veterans Affairs”

Comparison of Risk-Adjusted 30-Day Postoperative Mortality and Morbidity in Department of Veterans Affairs Hospitals and Selected University Medical Centers: Vascular Surgical Operations in Women

The risk is uncertain, so patients should make an informed decision.

Effects of Hospital Procedure Volume and Resident Training on Clinical Outcomes and Resource Use in Radical Retropubic Prostatectomy Surgery in the Department of Veterans Affairs

Best Practices: Supporting Clinicians in the Development of Best Practice Innovations in Education

The quality of mental health care provided by the U.S. Department of Veterans Affairs (VA) was compared with care provided to a comparable population treated in the private sector. Two cohorts of individuals with mental disorders (schizophrenia, bipolar disorder, posttraumatic stress disorder, major depression, and substance use disorders) were created with VA administrative data (N=836,519) and MarketScan data (N=545,484). The authors computed VA and MarketScan national means for seven process-based quality measures related to medication evaluation and management and estimated national-level performance by age and gender. In every case, VA performance was superior to that of the private sector by more than 30%. Compared with individuals in private plans, veterans with schizophrenia or major depression were more than twice as likely to receive appropriate initial medication treatment, and veterans with depression were more than twice as likely to receive appropriate long-term treatment. Findings demonstrate the significant advantages that accrue from an organized, nationwide system of care. The much higher performance of the VA has important clinical and policy implications.

109 Background: Machine learning tools based on EHR data hold promise to help avoid unnecessary risks associated with lung cancer and its treatment. Additionally, molecular genetic profiling is becoming an integral tool for clinicians to individualize treatment for lung cancer. However, relatively few survival models have been built that integrate this data in individualized predictive models. Here, we combine real-world EHR and tumor sequencing data from the VA Precision Oncology Data Repository (PODR) to build accurate individualized survival predictions in newly-diagnosed NSCLC patients. Methods: We identified a cohort of 356 VA patients newly diagnosed with NSCLC for whom EHR, cancer registry, and targeted tumor sequencing data is available in PODR. We defined 41 features reflecting 15 baseline clinical and demographic characteristics from the EHR and registry, such as age, race, stage, histology, and therapy. We also defined features reflecting 206 clinically actionable somatic variants. We selected 5 important variants for inclusion in the model, as well as the total number of mutations. We trained a random forests algorithm to predict 1-year survival. Precision, recall, and area under the ROC curve (AUC) were assessed using 5-fold cross validation. Results: Mean age at diagnosis was 66 years. The majority of patients had late stage disease (15% stage I, 6% II, 15% III, 44% IV), and 59% of patients received systemic therapy. 45% died within 1 year of diagnosis, and 55% survived past 1 year. Our predictive model for 1-year survival achieves strong results. Cross-validated AUC is 0.79 (SD 0.08), precision is 0.79 (SD 0.07), recall is 0.74 (SD 0.07), suggesting that the trained model combining clinical and genomic features is effective at predicting 1-year survival. Conclusions: By integrating real-world EHR and sequencing data, we built a highly accurate predictive model of 1-year survival in NSCLC patients at the VA. Such a model, after ongoing validation in a larger cohort, offers the ability to make individualized predictions that could inform patient care to improve outcomes.

Comparing quality of care between large health care systems is important for health systems management. This study compared measures of the quality of pharmacotherapy for patients with major depression across a sample of patients from the Department of Veterans Affairs (VA) and the private sector. In this observational study, all patients who were given a new prescription for an antidepressant and a diagnosis of major depression in the VA during fiscal year 2000 were identified by using administrative data (N=27,713). In the private sector, a similar sample of patients were identified by using Medstat's MarketScan database (N=4,852). For both groups, measures of the quality of antidepressant pharmacotherapy were constructed. These measures were compared across the two groups by using logistic regression models. Controls for age, gender, comorbid disorders, and initial antidepressant drug prescribed were included in some models. Although the populations had different demographic and clinical characteristics, differences in the quality measures between the two systems were few, with the VA slightly outperforming the private sector in the prescription of antidepressants during the acute phase of treatment, the first 84 days (84.7 compared with 81 percent) and during the maintenance phase of treatment, the first 181 days (53.9 compared with 50.9 percent). Patient characteristics that were associated with quality measures included being older, being female, and having a comorbid diagnosis of substance use disorder, bipolar disorder, or anxiety or adjustment disorder. Both systems had relatively high rates of adherence to pharmacotherapy guidelines. Even though the populations in the two systems were different, adjusting the analyses for clinical characteristics did little to change the measured differences between the two systems.

To reduce suicides among Veterans, the Department of Veterans Affairs (VA) has designated suicide risk assessments for Veterans who screen positive for depression or post-traumatic stress disorder as a national performance goal. Many VA Medical Centers (VAMCs) are using brief suicidal ideation screens, administered in non-mental health ambulatory care settings, as the first step in the assessment process. To explore Veterans' perceptions of the suicide screening and risk assessment process, the barriers and facilitators to disclosing suicidal thoughts, and perceptions of possible consequences of revealing suicidal thoughts. Investigators recorded one semi-structured interview with each Veteran. Transcripts were analyzed using a modified grounded theory approach. Thirty-four Operation Enduring Freedom/Operation Iraqi Freedom (OEF/OIF) Veterans who screened positive for suicidal ideation in non-mental health ambulatory care settings in 2009 and 2010. Veterans accepted the need to assess suicide risk. They increasingly experienced attempts to suppress and avoid thoughts of suicide as burdensome and exhausting. Despite this, Veterans often failed to disclose severe and pervasive suicidal thoughts when screened because: (1) they considered suicidal thoughts as shameful and a sign of weakness; (2) they believed suicidal thoughts were private and not to be divulged to strangers; (3) they worried that disclosure would lead to unwanted hospitalization or medication recommendations; and (4) the templated computer reminder process was perceived as perfunctory and disrespectful. In contrast, admitting and discussing thoughts of suicide with a health provider who focused on building a relationship, demonstrated genuineness and empathy, offered information on the rationale for suicide risk assessment, and used straightforward and understandable language, all promoted trust that resulted in more honest disclosure of suicidal thoughts. In ambulatory care settings, both provider behaviors and system modifications may lead to more honest disclosure of suicidal thoughts.

Identifying predictors of improvement amongst patients receiving routine treatment for post-traumatic stress disorder (PTSD) could provide information about factors that influence the clinical effectiveness of guideline-concordant care. This study builds on prior work by accounting for delivery of specific evidence-based treatments (EBTs) for PTSD while identifying potential predictors of clinical improvement using patient-reported outcomes measurement. Our sample consisted of 2643 US Department of Veterans Affairs (VA) outpatients who initiated treatment for PTSD between 2008 and 2013 and received at least four PTSD checklist (PCL) measurements over 12weeks. We obtained PCL data as well as demographic, diagnostic, and health services use information from the VA corporate data warehouse. We used latent trajectory analysis to identify classes of patients based on PCL scores, then determined demographic, diagnostic, and treatment predictors of membership in each class. Patients who met our PCL-based inclusion criteria were far more likely than those who did not receive EBTs. We identified two latent trajectories of PTSD symptoms. Patients in the substantial improvement group (25.9%) had a mean decrease in PCL score of 16.24, whereas patients in the modest improvement group improved by a mean of 8.09 points. However, there were few differences between the groups, and our model to predict group membership was only slightly better than chance (area under the curve [AUC]=0.55). Of the 64 covariates we tested, the only robust individual predictor of improvement was gender, with men having lower odds of being in the substantial improvement group compared with women (odds ratio [OR] 0.76; 95% confidence interval [CI] 0.58-0.96). VA patients with PTSD can realize significant improvement in routine clinical practice. Although available medical records-based variables were generally insufficient to predict improvement trajectory, this study did indicate that men have lower odds of substantial improvement than women.