Evaluation of a Model to Target High-risk Psychiatric Inpatients for an Intensive Postdischarge Suicide Prevention Intervention.

Importance: The months after psychiatric hospital discharge are a time of high risk for suicide. Intensive postdischarge case management, although potentially effective in suicide prevention, is likely to be cost-effective only if targeted at high-risk patients. A previously developed machine learning (ML) model showed that postdischarge suicides can be predicted from electronic health records and geospatial data, but it is unknown if prediction could be improved by adding additional information. Objective: To determine whether model prediction could be improved by adding information extracted from clinical notes and public records. Design, Setting, and Participants: Models were trained to predict suicides in the 12 months after Veterans Health Administration (VHA) short-term (less than 365 days) psychiatric hospitalizations between the beginning of 2010 and September 1, 2012 (299 050 hospitalizations, with 916 hospitalizations followed within 12 months by suicides) and tested in the hospitalizations from September 2, 2012, to December 31, 2013 (149 738 hospitalizations, with 393 hospitalizations followed within 12 months by suicides). Validation focused on net benefit across a range of plausible decision thresholds. Predictor importance was assessed with Shapley additive explanations (SHAP) values. Data were analyzed from January to August 2022. Main Outcomes and Measures: Suicides were defined by the National Death Index. Base model predictors included VHA electronic health records and patient residential data. The expanded predictors came from natural language processing (NLP) of clinical notes and a social determinants of health (SDOH) public records database. Results: The model included 448 788 unique hospitalizations. Net benefit over risk horizons between 3 and 12 months was generally highest for the model that included both NLP and SDOH predictors (area under the receiver operating characteristic curve range, 0.747-0.780; area under the precision recall curve relative to the suicide rate range, 3.87-5.75). NLP and SDOH predictors also had the highest predictor class-level SHAP values (proportional SHAP = 64.0% and 49.3%, respectively), although the single highest positive variable-level SHAP value was for a count of medications classified by the US Food and Drug Administration as increasing suicide risk prescribed the year before hospitalization (proportional SHAP = 15.0%). Conclusions and Relevance: In this study, clinical notes and public records were found to improve ML model prediction of suicide after psychiatric hospitalization. The model had positive net benefit over 3-month to 12-month risk horizons for plausible decision thresholds. Although caution is needed in inferring causality based on predictor importance, several key predictors have potential intervention implications that should be investigated in future studies.

Documentation of Screening for Firearm Access by Healthcare Providers in the Veterans Healthcare System: A Retrospective Study.

INTRODUCTION: Presence of a firearm is associated with increased risk of violence and suicide. United States military veterans are at disproportionate risk of suicide. Routine healthcare provider screening of firearm access may prompt counseling on safe storage and handling of firearms. The objective of this study was to determine the frequency with which Veterans Health Administration (VHA) healthcare providers document firearm access in electronic health record (EHR) clinical notes, and whether this varied by patient characteristics. METHODS: The study sample is a post-9-11 cohort of veterans in their first year of VHA care, with at least one outpatient care visit between 2012-2017 (N = 762,953). Demographic data, veteran military service characteristics, and clinical comorbidities were obtained from VHA EHR. We extracted clinical notes for outpatient visits to primary, urgent, or emergency clinics (total 105,316,004). Natural language processing and machine learning (ML) approaches were used to identify documentation of firearm access. A taxonomy of firearm terms was identified and manually annotated with text anchored by these terms, and then trained the ML algorithm. The random-forest algorithm achieved 81.9% accuracy in identifying documentation of firearm access. RESULTS: The proportion of patients with EHR-documented access to one or more firearms during their first year of care in the VHA was relatively low and varied by patient characteristics. Men had significantly higher documentation of firearms than women (9.8% vs 7.1%; P < .001) and veterans >50 years old had the lowest (6.5%). Among veterans with any firearm term present, only 24.4% were classified as positive for access to a firearm (24.7% of men and 20.9% of women). CONCLUSION: Natural language processing can identify documentation of access to firearms in clinical notes with acceptable accuracy, but there is a need for investigation into facilitators and barriers for providers and veterans to improve a systemwide process of firearm access screening. Screening, regardless of race/ethnicity, gender, and age, provides additional opportunities to protect veterans from self-harm and violence.

Using Administrative Data to Predict Suicide After Psychiatric Hospitalization in the Veterans Health Administration System.

There is a very high suicide rate in the year after psychiatric hospital discharge. Intensive postdischarge case management programs can address this problem but are not cost-effective for all patients. This issue can be addressed by developing a risk model to predict which inpatients might need such a program. We developed such a model for the 391,018 short-term psychiatric hospital admissions of US veterans in Veterans Health Administration (VHA) hospitals 2010-2013. Records were linked with the National Death Index to determine suicide within 12 months of hospital discharge (n=771). The Super Learner ensemble machine learning method was used to predict these suicides for time horizon between 1 week and 12 months after discharge in a 70% training sample. Accuracy was validated in the remaining 30% holdout sample. Predictors included VHA administrative variables and small area geocode data linked to patient home addresses. The models had AUC=.79-.82 for time horizons between 1 week and 6 months and AUC=.74 for 12 months. An analysis of operating characteristics showed that 22.4%-32.2% of patients who died by suicide would have been reached if intensive case management was provided to the 5% of patients with highest predicted suicide risk. Positive predictive value (PPV) at this higher threshold ranged from 1.2% over 12 months to 3.8% per case manager year over 1 week. Focusing on the low end of the risk spectrum, the 40% of patients classified as having lowest risk account for 0%-9.7% of suicides across time horizons. Variable importance analysis shows that 51.1% of model performance is due to psychopathological risk factors accounted, 26.2% to social determinants of health, 14.8% to prior history of suicidal behaviors, and 6.6% to physical disorders. The paper closes with a discussion of next steps in refining the model and prospects for developing a parallel precision treatment model.

Development and Assessment of a Crosswalk Between ICD-9-CM and ICD-10-CM to Identify Patients with Common Pain Conditions.

Effective management of patients with pain requires accurate information about the prevalence, outcomes, and co-occurrence of common pain conditions. However, the transition from ICD-9-CM to ICD-10-CM diagnostic coding in 2015 left researchers without methods for comparing the prevalence of pain conditions before and after the transition. In this study, we developed and assessed a diagnostic framework to serve as a crosswalk between ICD-9-CM and ICD-10-CM diagnosis codes for common pain-related health conditions. We refined existing ICD-9-CM definitions for diagnostic clusters of common pain conditions consistent with the US National Pain Strategy and developed corresponding ICD-10-CM definitions. We then assessed the stability of prevalence estimates and associated patient socio-demographic features of each diagnostic cluster during 1-year periods before and after the transition to ICD-10-CM in 3 US health care systems using electronic health records data for in-person encounters. Prevalence estimates and socio-demographic characteristics were similar before and after the transition. The Pain Condition ICD-9-CM to ICD-10-CM Crosswalk includes a full spectrum of common pain conditions to enable prevalence estimates of multiple and chronic overlapping pain conditions. This allows the tool to serve as a foundation for a broad array of pain-related health services research utilizing electronic databases. PERSPECTIVE: This article details the development and assessment of the Pain Condition ICD-9-CM to ICD-10-CM Crosswalk, a diagnostic framework for assessing pain condition prevalence across the ICD-9-CM to ICD-10-CM transition. This framework can serve as a standardized tool for research on pain conditions, including health services and epidemiologic research.

Dual Use of Department of Veterans Affairs and Medicare Benefits on High-Risk Opioid Prescriptions in Veterans Aged 65 Years and Older: Insights from the VA Musculoskeletal Disorders Cohort.

OBJECTIVE: To examine the association of dual use of both Veterans Health Administration (VHA) and Medicare benefits with high-risk opioid prescriptions among Veterans aged 65 years and older with a musculoskeletal disorder diagnosis. DATA SOURCES/STUDY SETTING: Data were obtained from the VA Musculoskeletal Disorder (MSD) cohort and national Medicare claims data from 2008 to 2010. STUDY DESIGN: We conducted a retrospective analysis of Veterans enrolled in Medicare to examine the association of dual use with long-term opioid use (>90 days of prescription opioids/year) and overlapping opioid prescriptions. Multivariable logistic regression was performed adjusting for demographic and clinical characteristics. DATA COLLECTION/EXTRACTION METHODS: We identified 21,111 Veterans enrolled in Medicare who entered the MSD cohort in 2008 and received an opioid prescription in 2010. We linked VHA data with Medicare claims data to identify opioid prescriptions for these Veterans in 2010. PRINCIPAL FINDINGS: As compared to Veterans who used only VHA or Medicare, Veterans with dual use of VHA and Medicare were significantly more likely to be prescribed long-term opioid therapy (OR = 4.61 (95 percent CI 4.05-5.25) and were also found to have higher median number of opioid prescriptions and higher odds of overlapping opioid prescriptions in 1 year. Patients reporting moderate-to-severe pain, non-white-race/ethnicity, and higher scoring on the Charlson comorbidity index had significantly higher odds of long-term opioid prescriptions. CONCLUSIONS: Among Veterans aged 65 years or older, dual use of both VHA and Medicare was associated with higher odds of long-term opioid therapy. Our findings suggest there may be benefit to combining VHA and non-VHA electronic health record data to minimize exposure to high-risk opioid prescribing.

Posttraumatic stress disorder diagnosis and gender are associated with accelerated weight gain trajectories in veterans during the post-deployment period.

BACKGROUND: Veterans are disproportionately affected by overweight/obesity and growing evidence suggests that post-deployment is a critical period of accelerated weight gain. OBJECTIVE: We explored the relationship between posttraumatic stress disorder (PTSD) diagnosis, gender, and post-deployment weight trajectories among U.S. Operations Iraqi Freedom, Enduring Freedom, and New Dawn veterans. DESIGN: We used Veterans Affairs electronic health record data from 248,089 veterans (87% men) who, after their last deployment, had at least one medical visit between October 2001 and January 2009 and more than one BMI recorded through September 2010. We analyzed repeated BMI measurements using linear mixed models, with demographics, PTSD and other relevant psychiatric diagnoses as predictors. RESULTS: At the first recorded BMI, veterans' median age was 29, and 59% of women and 77% of men were overweight/obese. They had a median of 6 BMI measurements during a median follow-up of 2.4 years. Controlling for potential confounders, women with a PTSD diagnosis had a yearly BMI growth rate of 0.11 kg/m2 (95% CI 0.09 to 0.13, p < 0.001) higher than women without PTSD. For men, the corresponding PTSD effect was also significant, but slightly lower: 0.07 kg/m2 ((95% CI 0.05 to 0.09, p < 0.001); women-men difference: 0.03 (95% CI 0.01 to 0.06) kg/m2, p = 0.006). CONCLUSIONS: The post-deployment period is critical for weight gain, particularly for veterans diagnosed with PTSD and women veterans with PTSD. Efforts are needed to engage post-deployment veterans in weight management services, and to determine whether tailored recruitment/treatment interventions will reduce disparities for veterans with PTSD.

Multiple Sources of Prescription Payment and Risky Opioid Therapy Among Veterans.

BACKGROUND: Opioid overdose and other related harms are a major source of morbidity and mortality among US Veterans, in part due to high-risk opioid prescribing. OBJECTIVES: We sought to determine whether having multiple sources of payment for opioids-as a marker for out-of-system access-is associated with risky opioid therapy among veterans. RESEARCH DESIGN: Cross-sectional study examining the association between multiple sources of payment and risky opioid therapy among all individuals with Veterans Health Administration (VHA) payment for opioid analgesic prescriptions in Kentucky during fiscal year 2014-2015. MEASURES: Source of payment categories: (1) VHA only source of payment (sole source); (2) sources of payment were VHA and at least 1 cash payment [VHA+cash payment(s)] whether or not there was a third source of payment; and (3) at least one other noncash source: Medicare, Medicaid, or private insurance [VHA+noncash source(s)]. Our outcomes were 2 risky opioid therapies: combination opioid/benzodiazepine therapy and high-dose opioid therapy, defined as morphine equivalent daily dose ≥90 mg. RESULTS: Of the 14,795 individuals in the analytic sample, there were 81.9% in the sole source category, 6.6% in the VHA+cash payment(s) category, and 11.5% in the VHA+noncash source(s) category. In logistic regression, controlling for age and sex, persons with multiple payment sources had significantly higher odds of each risky opioid therapy, with those in the VHA+cash having significantly higher odds than those in the VHA+noncash source(s) group. CONCLUSIONS: Prescribers should examine the prescription monitoring program as multiple payment sources increase the odds of risky opioid therapy.

Estimating alcohol content of traditional brew in Western Kenya using culturally relevant methods: the case for cost over volume.

Traditional homemade brew is believed to represent the highest proportion of alcohol use in sub-Saharan Africa. In Eldoret, Kenya, two types of brew are common: chang'aa, spirits, and busaa, maize beer. Local residents refer to the amount of brew consumed by the amount of money spent, suggesting a culturally relevant estimation method. The purposes of this study were to analyze ethanol content of chang'aa and busaa; and to compare two methods of alcohol estimation: use by cost, and use by volume, the latter the current international standard. Laboratory results showed mean ethanol content was 34% (SD = 14%) for chang'aa and 4% (SD = 1%) for busaa. Standard drink unit equivalents for chang'aa and busaa, respectively, were 2 and 1.3 (US) and 3.5 and 2.3 (Great Britain). Using a computational approach, both methods demonstrated comparable results. We conclude that cost estimation of alcohol content is more culturally relevant and does not differ in accuracy from the international standard.

Drug toxicity, HIV progression, or comorbidity of aging: does tipranavir use increase the risk of intracranial hemorrhage?

The US Food and Drug Administration has issued a warning that tipranavir may be associated with increased risk of intracranial hemorrhage. We studied 2 large cohorts to estimate the background rate of intracranial hemorrhage and compared it with rates reported among persons who had been exposed to tipranavir.

Determinants of hormone therapy discontinuation among female veterans nationally.

PURPOSE: The growing presence of female veterans within the Department of Veterans Affairs (VA) health care system highlights the need to assess the quality of and access to gender-specific care for menopause. We assessed the use of hormone therapy (HT) among female veterans before and after the release of the Women's Health Initiative clinical trial results and evaluated whether the structure of women's health care services within the VA system affects the use of HT. METHODS: We identified all female veterans using HT in 2001 by using the VA pharmacy benefits management database and administrative data. Subjects identified as using HT in 2001 were evaluated to determine estrogen use status in 2003 and 2004. We calculated the change in HT use over time and performed multivariate analyses to identify patient and utilization determinants of HT discontinuation. RESULTS: In 2001, 36,222 female veterans used HT. By 2004, 23,924 (66%) had discontinued HT. Subjects who had used a VA women's clinic or were younger (40-54 years of age) were significantly less likely to discontinue HT. However, Hispanic ethnicity, African American race, and clinical diagnoses such as heart disease and mastectomy were significantly associated with discontinuation. CONCLUSION: Discontinuation rates in the VA system parallel those in the private sector. However, patients with any use of VA women's clinics were less likely to discontinue HT, indicating a practice setting variation that may indicate either more specific care or differential implementation of the new HT guidelines. Further research is warranted to assess whether a disparity occurs according to practice setting (or provider factors) with rapid shifts in guidelines.

Mental health care in juvenile detention facilities: a review.

Juvenile detention facilities have come under increasing legal pressure to provide mental health services to detainees, and mental health clinicians may be asked to design and implement programs in detention facilities. However, there is little consensus on what types of services should be provided, and virtually no data on the effectiveness of such services in a detention setting. The objective of this article is to provide an overview of the existing literature on mental health services in juvenile detention and to make suggestions about future research needs. Specifically, it highlights the tension surrounding the provision of mental health care in juvenile detention, presents data on the prevalence of psychiatric problems in detention settings and what types of services are currently provided, and draws on the larger child and adolescent mental health literature to suggest what types of services might be most appropriate for juvenile detention settings. We conclude that, although there are some suggestions of promising interventions that may be appropriate, much more research, specifically in detention settings, is needed to determine their effectiveness.