Legalization of Cannabis Does Not Reduce Opioid Prescribing in Patients With Inflammatory Bowel Disease: A Difference-in-Difference Analysis.

INTRODUCTION: Cannabis may provide inflammatory bowel disease (IBD) patients with an alternative to opioids for pain. METHODS: We conducted a difference-in-difference analysis using MarketScan. Changes over time in rates of opioid prescribing were compared in states with legalized cannabis to those without. RESULTS: We identified 6,240 patients with IBD in states with legalized cannabis and 79,272 patients with IBD in states without legalized cannabis. The rate of opioid prescribing decreased over time in both groups and were not significantly different (attributed differential = 0.34, confidence interval -13.02 to 13.70, P = 0.96). DISCUSSION: Opioid prescribing decreased from 2009 to 2016 among patients with IBD in both states with legalized and state without legalized cannabis, similar to what has been observed nationally across a variety of diseases. Cannabis legalization was not associated with a lower rate of opioid prescribing for patients with IBD.

The Importance of Autonomy and Performance Goals in Perceived Workload Among Behavioral Health Providers.

OBJECTIVE: The authors sought to assess workplace characteristics associated with perceived reasonable workload among behavioral health care providers in the Veterans Health Administration. METHODS: The authors evaluated perceived reasonable workload and workplace characteristics from the 2019 All Employee Survey (AES; N=14,824) and 2019 Mental Health Provider Survey (MHPS; N=10,490) and facility-level staffing ratios from Mental Health Onboard Clinical Dashboard data. Nine AES and 15 MHPS workplace predictors of perceived reasonable workload, 11 AES and six MHPS demographic predictors, and facility-level staffing ratios were included in mixed-effects logistic regression models. RESULTS: In total, 8,874 (59.9%) AES respondents and 5,915 (56.4%) MHPS respondents reported having a reasonable workload. The characteristics most strongly associated with perceived reasonable workload were having attainable performance goals (average marginal effect [AME]=0.10) in the AES and ability to schedule patients as frequently as indicated (AME=0.09) in the MHPS. Other AES characteristics significantly associated with reasonable workload included having appropriate resources, support for personal life, skill building, performance recognition, concerns being addressed, and no supervisor favoritism. MHPS characteristics included not having collateral duties that reduce care time, staffing levels not affecting care, support staff taking over some responsibilities, having spirit of teamwork, primary care-mental health integration, participation in performance discussions, well-coordinated mental health care, effective veteran programs, working at the top of licensure, and feeling involved in improving access. Facility-level staffing ratios were not significantly associated with perceived reasonable workload. CONCLUSIONS: Leadership may consider focusing resources on initiatives that support behavioral health providers' autonomy to schedule patients as clinically indicated and develop attainable performance goals.

Relationship between facility number of clinicians and prescribing intensity of psychotropic medications, opioids, and antibiotics in ambulatory practice.

BACKGROUND: Promoting appropriate pharmacotherapy requires understanding the factors that influence how clinicians prescribe medications. While prior work has focused on patient and clinician factors, features of the organizational setting have received less attention, though identifying sources of variation in prescribing may help identify opportunities to improve patient safety and outcomes. OBJECTIVE: To evaluate the relationship between the number of clinicians who prescribe medications in a facility and facility prescribing intensity of six individual medication classes by clinician specialty: benzodiazepines, antipsychotics, antiepileptics, and antidepressants by psychiatrists and antibiotics, opioids, antiepileptics, and antidepressants by primary care clinicians (PCPs). DESIGN: We used 2017 Veterans Health Administration (VHA) administrative data. SUBJECTS: We included patient-clinician dyads of older patients (> 55 years) with an outpatient encounter with a clinician in 2017. Patient-clinician data from 140 VHA facilities were included (n = 13,347,658). Analysis was repeated for years 2014 to 2016. MAIN MEASURES: For each medication, facility prescribing intensity measures were calculated as clinician prescribing intensity averaged over all clinicians at each facility. Clinician prescribing intensity measures included percentage of each clinician's patients prescribed the medication and mean number of days supply per patient among all patients of each clinician. KEY RESULTS: As the number of prescribing clinicians in a facility increased, the intensity of prescribing decreased. Every increase of 10 facility clinicians was associated with a significant decline in prescribing intensity for both specialties for different medication classes: for psychiatrists, declines ranged from 6 to 11%, and for PCPs, from 2 to 3%. The pattern of more clinicians less prescribing was significant across all years. CONCLUSION: Future work should explore the mechanisms that link the number of facility clinicians with prescribing intensity for benzodiazepines, antipsychotics, antiepileptics, antidepressants, antibiotics, and opioids. Facilities with fewer clinicians may need additional resources to avoid unwanted prescribing of potentially harmful or unnecessary medications.

Role of Institutional Support for Evidence-Based Psychotherapy in Satisfaction and Burnout Among Veterans Affairs Therapists.

OBJECTIVE: Burnout is widespread among psychotherapists and leads to negative mental and other health outcomes, absenteeism, and turnover. Job resources, including institutional support for evidence-based practices, can buffer against burnout and may improve satisfaction among therapists. The Veterans Health Administration (VHA) is the nation's largest integrated health system and employs 23,000 therapists, including psychologists, social workers, and counselors. The authors assessed associations between perceived institutional support for evidence-based treatment and satisfaction and burnout among VHA therapists. METHODS: This analysis used data from the VHA's national 2018 Mental Health Provider Survey. Responding therapists (N=5,341) answered questions about the quality of mental health care and job satisfaction. Multilevel logistic regression models were used to predict burnout and satisfaction. The authors tested availability of evidence-based treatment and measurement-based care (MBC) as predictors; analyses were adjusted for therapist workload, demographic characteristics, and potential clustering by facility. RESULTS: VHA therapists had less burnout and more job satisfaction when they perceived receiving institutional support for evidence-based psychotherapy (EBP) and MBC, irrespective of whether the analyses were adjusted for workload. Less difficulty in scheduling EBP was significantly associated with decreased likelihood of burnout (OR=0.83, p<0.001) and increased satisfaction (OR=1.09, p=0.008). Less difficulty ending psychotherapy was significantly associated with decreased likelihood of burnout (OR=0.89, p=0.002) and increased satisfaction (OR=1.12, p=0.004). CONCLUSIONS: Support for evidence-based practices, including EBP and MBC, was closely linked to VHA therapists' satisfaction and burnout. Expanding support for therapists to provide evidence-based treatment may benefit therapists, patients, and the health care system.

Behavioral Health Provider Burnout and Mental Health Care in the Veterans Health Administration.

BACKGROUND: Although many studies assess predictors of provider burnout, few analyses provide high-quality, consistent evidence on the impact of provider burnout on patient outcomes exist, particularly among behavioral health providers (BHPs). OBJECTIVE: To assess the impact of burnout among psychiatrists, psychologists, and social workers on access-related quality measures in the Veterans Health Administration (VHA). DESIGN: This study used burnout in VA All Employee Survey (AES) and Mental Health Provider Survey (MHPS) data to predict metrics assessed by the Strategic Analytics for Improvement and Learning Value, Mental Health Domain (MH-SAIL), VHA's quality monitoring system. The study used prior year (2014-2018) facility-level burnout proportion among BHPs to predict subsequent year (2015-2019) facility-level MH-SAIL domain scores. Analyses used multiple regression models, adjusting for facility characteristics, including BHP staffing and productivity. PARTICIPANTS: Psychologists, psychiatrists, and social workers who responded to the AES and MHPS at 127 VHA facilities. MAIN MEASURES: Four compositive outcomes included two objective measures (population coverage, continuity of care), one subjective measure (experience of care), and one composite measure of the former three measures (mental health domain quality). KEY RESULTS: Adjusted analyses showed prior year burnout generally had no impact on population coverage, continuity of care, and patient experiences of care but had a negative impact on provider experiences of care consistently across 5 years (p < 0.001). Pooled across years, a 5% higher facility-level burnout in AES and MHPS had a 0.05 and 0.09 standard deviation worse facility experiences of care from the prior year, respectively. CONCLUSIONS: Burnout had a significant negative impact on provider-reported experiential outcome measures. This analysis showed that burnout had a negative effect on subjective but not on objective quality measures of Veteran access to care, which could inform future policies and interventions regarding provider burnout.

Characteristics of Facilities With Early and Rapid Ustekinumab Adoption for Patients With Inflammatory Bowel Disease.

INTRODUCTION: To examine which facility characteristics, including teamwork, are associated with early or rapid inflammatory bowel disease-related ustekinumab adoption. METHODS: We examined the association between ustekinumab adoption and the characteristics of 130 Veterans Affairs facilities. RESULTS: Mean ustekinumab adoption increased by 3.9% from 2016 to 2018 and was higher in urban compared with rural facilities (ß = 0.03, P = 0.033) and among facilities with more teamwork (ß = 0.11, P = 0.041). Compared with nonearly adopters, early adopters were more likely be high-volume facilities (46% vs 19%, P = 0.001). DISCUSSION: Facility variation in medication adoption provides an opportunity for improving inflammatory bowel disease care through targeted dissemination strategies to improve medication uptake.

Understanding clinician connections to inform efforts to promote high-quality inflammatory bowel disease care.

BACKGROUND: Highly connected individuals disseminate information effectively within their social network. To apply this concept to inflammatory bowel disease (IBD) care and lay the foundation for network interventions to disseminate high-quality treatment, we assessed the need for improving the IBD practices of highly connected clinicians. We aimed to examine whether highly connected clinicians who treat IBD patients were more likely to provide high-quality treatment than less connected clinicians. METHODS: We used network analysis to examine connections among clinicians who shared patients with IBD in the Veterans Health Administration between 2015-2018. We created a network comprised of clinicians connected by shared patients. We quantified clinician connections using degree centrality (number of clinicians with whom a clinician shares patients), closeness centrality (reach via shared contacts to other clinicians), and betweenness centrality (degree to which a clinician connects clinicians not otherwise connected). Using weighted linear regression, we examined associations between each measure of connection and two IBD quality indicators: low prolonged steroids use, and high steroid-sparing therapy use. RESULTS: We identified 62,971 patients with IBD and linked them to 1,655 gastroenterologists and 7,852 primary care providers. Clinicians with more connections (degree) were more likely to exhibit high-quality treatment (less prolonged steroids beta -0.0268, 95%CI -0.0427, -0.0110, more steroid-sparing therapy beta 0.0967, 95%CI 0.0128, 0.1805). Clinicians who connect otherwise unconnected clinicians (betweenness) displayed more prolonged steroids use (beta 0.0003, 95%CI 0.0001, 0.0006). The presence of variation is more relevant than its magnitude. CONCLUSIONS: Clinicians with a high number of connections provided more high-quality IBD treatments than less connected clinicians, and may be well-positioned for interventions to disseminate high-quality IBD care. However, clinicians who connect clinicians who are otherwise unconnected are more likely to display low-quality IBD treatment. Efforts to improve their quality are needed prior to leveraging their position to disseminate high-quality care.

The Association Between Nonsteroidal Anti-Inflammatory Drug Use and Inflammatory Bowel Disease Exacerbations: A True Association or Residual Bias?

INTRODUCTION: Studies suggest that nonsteroidal anti-inflammatory drugs (NSAID) may contribute to inflammatory bowel disease (IBD) exacerbations. We examined whether variation in the likelihood of IBD exacerbations is attributable to NSAID. METHODS: In a cohort of patients with IBD (2004-2015), we used 3 analytic methods to examine the likelihood of an exacerbation after an NSAID exposure. First, we matched patients by propensity for NSAID use and examined the association between NSAID exposure and IBD exacerbation using an adjusted Cox proportional hazards model. To assess for residual confounding, we estimated a previous event rate ratio and used a self-controlled case series analysis to further explore the relationship between NSAID and IBD exacerbations. RESULTS: We identified 15,705 (44.8%) and 19,326 (55.2%) IBD patients with and without an NSAID exposure, respectively. Findings from the Cox proportional hazards model suggested an association between NSAID and IBD exacerbation (hazard ratio 1.24; 95% confidence interval 1.16-1.33). However, the likelihood of an IBD exacerbation in the NSAID-exposed arm preceding NSAID exposure was similar (hazard ratio 1.30; 95% confidence interval 1.21-1.39). A self-controlled case series analysis of 3,968 patients who had both an NSAID exposure and IBD exacerbation demonstrated similar exacerbation rates in the 1 year preceding exposure, 2-6 weeks postexposure, and 6 weeks to 6 months postexposure, but a higher incidence in 0-2 weeks postexposure, suggesting potential confounding by reverse causality. DISCUSSION: While we see an association between NSAID and IBD exacerbations using traditional methods, further analysis suggests this may be secondary to residual bias. These findings may reassure patients and clinicians considering NSAID as a nonopioid pain management option.

Reinforcement learning evaluation of treatment policies for patients with hepatitis C virus.

BACKGROUND: Evaluation of new treatment policies is often costly and challenging in complex conditions, such as hepatitis C virus (HCV) treatment, or in limited-resource settings. We sought to identify hypothetical policies for HCV treatment that could best balance the prevention of cirrhosis while preserving resources (financial or otherwise). METHODS: The cohort consisted of 3792 HCV-infected patients without a history of cirrhosis or hepatocellular carcinoma at baseline from the national Veterans Health Administration from 2015 to 2019. To estimate the efficacy of hypothetical treatment policies, we utilized historical data and reinforcement learning to allow for greater flexibility when constructing new HCV treatment strategies. We tested and compared four new treatment policies: a simple stepwise policy based on Aspartate Aminotransferase to Platelet Ratio Index (APRI), a logistic regression based on APRI, a logistic regression on multiple longitudinal and demographic indicators that were prespecified for clinical significance, and a treatment policy based on a risk model developed for HCV infection. RESULTS: The risk-based hypothetical treatment policy achieved the lowest overall risk with a score of 0.016 (90% CI 0.016, 0.019) while treating the most high-risk (346.4 ± 1.4) and the fewest low-risk (361.0 ± 20.1) patients. Compared to hypothetical treatment policies that treated approximately the same number of patients (1843.7 vs. 1914.4 patients), the risk-based policy had more untreated time per patient (7968.4 vs. 7742.9 patient visits), signaling cost reduction for the healthcare system. CONCLUSIONS: Off-policy evaluation strategies are useful to evaluate hypothetical treatment policies without implementation. If a quality risk model is available, risk-based treatment strategies can reduce overall risk and prioritize patients while reducing healthcare system costs.

Relationships between work-environment characteristics and behavioral health provider burnout in the Veterans Health Administration.

OBJECTIVE: To identify work-environment characteristics associated with Veterans Health Administration (VHA) behavioral health provider (BHP) burnout among psychiatrists, psychologists, and social workers. DATA SOURCES: The 2015-2018 data from Annual All Employee Survey (AES); Mental Health Provider Survey (MHPS); N = 57,397 respondents; facility-level Mental Health Onboard Clinical (MHOC) staffing and productivity data, N = 140 facilities. STUDY DESIGN: For AES and MHPS separately, we used mixed-effects logistic regression to predict BHP burnout using surveys from year pairs (2015-2016, 2016-2017, 2017-2018; six models). Within each year-pair, we used the earlier year of data to train models and tested the model in the later year, with burnout (emotional exhaustion and/or depersonalization) as the outcome for each survey. We used potentially modifiable work-environment characteristics as predictors, controlling for employee demographic characteristics as covariates, and employment facility as random intercepts. DATA COLLECTION/EXTRACTION METHODS: We included work-environment predictors that appeared in all 4 years (11 in AES; 17 in MHPS). PRINCIPAL FINDINGS: In 2015-2018, 31.0%-38.0% of BHPs reported burnout in AES or MHPS. Work characteristics consistently associated with significantly lower burnout were included for AES: reasonable workload; having appropriate resources to perform a job well; supervisors address concerns; given an opportunity to improve skills. For MHPS, characteristics included: reasonable workload; work improves veterans' lives; mental health care provided is well-coordinated; and three reverse-coded items: staffing vacancies; daily work that clerical/support staff could complete; and collateral duties reduce availability for patient care. Facility-level staffing ratios and productivity did not significantly predict individual-level burnout. Workload represented the strongest predictor of burnout in both surveys. CONCLUSIONS: This study demonstrated substantial, ongoing impacts that having appropriate resources including staff, workload, and supervisor support had on VHA BHP burnout. VHA may consider investing in approaches to mitigate the impact of BHP burnout on employees and their patients through providing staff supports, managing workload, and goal setting.

Adapted time-varying covariates Cox model for predicting future cirrhosis development performs well in a large hepatitis C cohort.

BACKGROUND: Patients with hepatitis C virus (HCV) frequently remain at risk for cirrhosis after sustained virologic response (SVR). Existing cirrhosis predictive models for HCV do not account for dynamic antiviral treatment status and are limited by fixed laboratory covariates and short follow up time. Advanced fibrosis assessment modalities, such as transient elastography, remain inaccessible in many settings. Improved cirrhosis predictive models are needed. METHODS: We developed a laboratory-based model to predict progression of liver disease after SVR. This prediction model used a time-varying covariates Cox model adapted to utilize longitudinal laboratory data and to account for antiretroviral treatment. Individuals were included if they had a history of detectable HCV RNA and at least 2 AST-to-platelet ratio index (APRI) scores available in the national Veterans Health Administration from 2000 to 2015, Observation time extended through January 2019. We excluded individuals with preexisting cirrhosis. Covariates included baseline patient characteristics and 16 time-varying laboratory predictors. SVR, defined as permanently undetectable HCV RNA after antiviral treatment, was modeled as a step function of time. Cirrhosis development was defined as two consecutive APRI scores > 2. We predicted cirrhosis development at 1-, 3-, and 5-years follow-up. RESULTS: In a national sample of HCV patients (n = 182,772) with a mean follow-up of 6.32 years, 42% (n = 76,854) achieved SVR before 2016 and 16.2% (n = 29,566) subsequently developed cirrhosis. The model demonstrated good discrimination for predicting cirrhosis across all combinations of laboratory data windows and cirrhosis prediction intervals. AUROCs ranged from 0.781 to 0.815, with moderate sensitivity 0.703-0.749 and specificity 0.723-0.767. CONCLUSION: A novel adaptation of time-varying covariates Cox modeling technique using longitudinal laboratory values and dynamic antiviral treatment status accurately predicts cirrhosis development at 1-, 3-, and 5-years among patients with HCV, with and without SVR. It improves upon earlier cirrhosis predictive models and has many potential population-based applications, especially in settings without transient elastography available.

Trends in Incident Varenicline Prescribing Among Veterans Following the US Food and Drug Administration Drug Safety Warnings.

Objective: To evaluate national trends in incident varenicline and nicotine replacement therapy (NRT) prescribing among Department of Veterans Affairs (VA) beneficiaries before and after US Food and Drug Administration (FDA) warnings regarding neuropsychiatric side effects with varenicline use.Methods: All adult VA patients identified as smokers from 2007 to 2019 (N = 3,600,947) were determined and monthly counts of new varenicline and NRT users were calculated. An interrupted time-series analysis estimated the effect of the FDA warnings on varenicline and NRT prescribing overall and among Veterans with and without mental health disorders.Results: The incident use rate of varenicline decreased from a peak of 6.2 per 1,000 veteran smokers in October 2007 to 1.0 by July 2009 following the first FDA warning (pre-warning monthly slope = -0.27; P = .03). New NRT use increased from 10.7 per 1,000 veteran smokers in October 2007 to a peak of 12.6 per 1,000 in July 2009 (slope change = 0.71; P = .01), suggesting potential substitution. Following removal of the FDA boxed warning in December 2016, varenicline prescribing increased but did not return to pre-warning levels by December 2019. Among veterans with and without mental health disorders, varenicline use decreased 90% and 88%, respectively, following the first FDA warning, and both groups had comparable rates of new NRT use.Conclusions: Following the first FDA warning, incident use of varenicline declined significantly among veterans both with and without mental health disorders. Despite removal of the FDA boxed warning in December 2016, new use of varenicline had not returned to pre-warning levels 3 years following the removal.

Employment After Vocational Rehabilitation Predicts Decreased Health Care Utilization in Veterans With Mental Health Diagnoses.

INTRODUCTION: Although the benefits of employment for veterans with mental health conditions are well-known, the effect of veterans' employment on a health system has not been evaluated. The purpose of this study was to evaluate the effect of veterans' employment (versus unemployment) on subsequent health care utilization in the Veterans Health Administration (VHA). MATERIALS AND METHODS: This study used a sample of 29,022 veterans with mental health and substance use disorders who were discharged from VHA's employment services programs between fiscal years 2006 and 2010. Veterans' employment status (employed/unemployed) upon discharge from VHA employment programs was ascertained from program discharge forms and linked with VHA administrative health care utilization data for the subsequent 1- and 5-year periods. RESULTS: Multivariable ordinary least-squares and logistic regression models adjusted for site clustering and covariates indicated that employment (versus unemployment) predicted less health care utilization 1 year and 5 years post-discharge from employment services, including fewer outpatient mental health visits, homelessness services visits, employment services visits, primary care visits, and lower odds of mental health hospitalizations, mental health or vocational rehabilitation residential stays, and medical hospitalizations. Employment did not predict emergency department visits. CONCLUSIONS: VHA's investment in employment services for veterans with mental health and substance use disorders could reduce health care utilization system wide.

Association of Anti-Tumor Necrosis Factor Therapy With Mortality Among Veterans With Inflammatory Bowel Disease.

Importance: Inflammatory bowel disease (IBD) is commonly treated with corticosteroids and anti-tumor necrosis factor (TNF) drugs; however, medications have well-described adverse effects. Prior work suggests that anti-TNF therapy may reduce all-cause mortality compared with prolonged corticosteroid use among Medicare and Medicaid beneficiaries with IBD. Objective: To examine the association between use of anti-TNF or corticosteroids and all-cause mortality in a national cohort of veterans with IBD. Design, Setting, and Participants: This cohort study used a well-established Veteran's Health Administration cohort of 2997 patients with IBD treated with prolonged corticosteroids (≥3000-mg prednisone equivalent and/or ≥600 mg of budesonide within a 12-month period) and/or new anti-TNF therapy from January 1, 2006, to October 1, 2015. Data were analyzed between July 1, 2019, and December 31, 2020. Exposures: Use of corticosteroids or anti-TNF. Main Outcomes and Measures: The primary end point was all-cause mortality as defined by the Veterans Health Administration vital status file. Marginal structural modeling was used to compare associations between anti-TNF therapy or corticosteroid use and all-cause mortality. Results: A total of 2997 patients (2725 men [90.9%]; mean [SD] age, 50.0 [17.4] years) were included in the final analysis, 1734 (57.9%) with Crohn disease (CD) and 1263 (42.1%) with ulcerative colitis (UC). All-cause mortality was 8.5% (n = 256) over a mean (SD) of 3.9 (2.3) years' follow-up. At cohort entry, 1836 patients were new anti-TNF therapy users, and 1161 were prolonged corticosteroid users. Anti-TNF therapy use was associated with a lower likelihood of mortality for CD (odds ratio [OR], 0.54; 95% CI, 0.31-0.93) but not for UC (OR, 0.33; 95% CI, 0.10-1.10). In a sensitivity analysis adjusting prolonged corticosteroid users to include patients receiving corticosteroids within 90 to 270 days after initiation of anti-TNF therapy, the OR for UC was statistically significant, at 0.33 (95% CI, 0.13-0.84), and the OR for CD was 0.55 (95% CI, 0.33-0.92). Conclusions and Relevance: This study suggests that anti-TNF therapy may be associated with reduced mortality compared with long-term corticosteroid use among veterans with CD, and potentially among those with UC.

The Use of Readily Available Longitudinal Data to Predict the Likelihood of Surgery in Crohn Disease.

BACKGROUND: Although imaging, endoscopy, and inflammatory biomarkers are associated with future Crohn disease (CD) outcomes, common laboratory studies may also provide prognostic opportunities. We evaluated machine learning models incorporating routinely collected laboratory studies to predict surgical outcomes in U.S. Veterans with CD. METHODS: Adults with CD from a Veterans Health Administration, Veterans Integrated Service Networks (VISN) 10 cohort examined between 2001 and 2015 were used for analysis. Patient demographics, medication use, and longitudinal laboratory values were used to model future surgical outcomes within 1 year. Specifically, data at the time of prediction combined with historical laboratory data characteristics, described as slope, distribution statistics, fluctuation, and linear trend of laboratory values, were considered and principal component analysis transformations were performed to reduce the dimensionality. Lasso regularized logistic regression was used to select features and construct prediction models, with performance assessed by area under the receiver operating characteristic using 10-fold cross-validation. RESULTS: We included 4950 observations from 2809 unique patients, among whom 256 had surgery, for modeling. Our optimized model achieved a mean area under the receiver operating characteristic of 0.78 (SD, 0.002). Anti-tumor necrosis factor use was associated with a lower probability of surgery within 1 year and was the most influential predictor in the model, and corticosteroid use was associated with a higher probability of surgery. Among the laboratory variables, high platelet counts, high mean cell hemoglobin concentrations, low albumin levels, and low blood urea nitrogen values were identified as having an elevated influence and association with future surgery. CONCLUSIONS: Using machine learning methods that incorporate current and historical data can predict the future risk of CD surgery.

Assessment of a Deep Learning Model to Predict Hepatocellular Carcinoma in Patients With Hepatitis C Cirrhosis.

Importance: Deep learning, a family of machine learning models that use artificial neural networks, has achieved great success at predicting outcomes in nonmedical domains. Objective: To examine whether deep learning recurrent neural network (RNN) models that use raw longitudinal data extracted directly from electronic health records outperform conventional regression models in predicting the risk of developing hepatocellular carcinoma (HCC). Design, Setting, and Participants: This prognostic study included 48 151 patients with hepatitis C virus (HCV)-related cirrhosis in the national Veterans Health Administration who had at least 3 years of follow-up after the diagnosis of cirrhosis. Patients were identified by having at least 1 positive HCV RNA test between January 1, 2000, to January 1, 2016, and were followed up from the diagnosis of cirrhosis to January 1, 2019, for the development of incident HCC. A total of 3 models predicting HCC during a 3-year period were developed and compared, as follows: (1) logistic regression (LR) with cross-sectional inputs (cross-sectional LR); (2) LR with longitudinal inputs (longitudinal LR); and (3) RNN with longitudinal inputs. Data analysis was conducted from April 2018 to August 2020. Exposures: Development of HCC. Main Outcomes and Measures: Area under the receiver operating characteristic curve, area under the precision-recall curve, and Brier score. Results: During a mean (SD) follow-up of 11.6 (5.0) years, 10 741 of 48 151 patients (22.3%) developed HCC (annual incidence, 3.1%), and a total of 52 983 samples (51 948 [98.0%] from men) were collected. Patients who developed HCC within 3 years were older than patients who did not (mean [SD] age, 58.2 [6.6] years vs 56.9 [6.9] years). RNN models had superior mean (SD) area under the receiver operating characteristic curve (0.759 [0.009]) and mean (SD) Brier score (0.136 [0.003]) than cross-sectional LR (0.689 [0.009] and 0.149 [0.003], respectively) and longitudinal LR (0.682 [0.007] and 0.150 [0.003], respectively) models. Using the RNN model, the samples with the mean (SD) highest 51% (1.5%) of HCC risk, in which 80% of all HCCs occurred, or the mean (SD) highest 66% (1.2%) of HCC risk, in which 90% of all HCCs occurred, could potentially be targeted. Among samples from patients who achieved sustained virologic response, the performance of the RNN models was even better (mean [SD] area under receiver operating characteristic curve, 0.806 [0.025]; mean [SD] Brier score, 0.117 [0.007]). Conclusions and Relevance: In this study, deep learning RNN models outperformed conventional LR models, suggesting that RNN models could be used to identify patients with HCV-related cirrhosis with a high risk of developing HCC for risk-based HCC outreach and surveillance strategies.

Predictors of Veterans Affairs Health Service Utilization by Women Veterans during Pregnancy.

BACKGROUND: Researchers have examined predictors of Veterans Affairs (VA) service use by women veterans in general, but less is known about predictors of VA service use by pregnant veterans. This study examined characteristics associated with planned and actual VA service use by pregnant veterans. METHODS: This study includes data from 510 pregnant veterans enrolled in the Center for Maternal and Infant Outcomes Research in Translation Study. Women veterans completed phone interviews during their first trimester and at 3 months postpartum. The Center for Maternal and Infant Outcomes Research in Translation surveys assessed medical and mental health conditions, VA health care use, trauma history, and pregnancy complications. We conducted bivariate and multivariable logistic regression models assessing planned and actual use of VA services during pregnancy. RESULTS: Lifetime post-traumatic stress disorder (odds ratio [OR], 1.73; 95% confidence interval [CI], 1.11-2.69) and history of military sexual trauma (OR, 1.85; 95% CI, 1.19-2.87) were significantly associated with planned VA service use in multivariable models. Lifetime diagnoses of anxiety (OR, 1.78; C.I., 1.15-2.75) were associated with an increased likelihood of actual VA use during pregnancy, whereas Hispanic ethnicity (OR, 0.59; 95% CI, 0.36-0.96), younger age (OR, 0.95; 95% CI, 0.91-0.99), and access to private health insurance (OR, 0.55; 95% CI, 0.37-0.84) were associated with a decreased likelihood of actual VA service use during pregnancy. CONCLUSIONS: Results emphasize the association between high-risk mental health characteristics and specific demographic characteristics with VA service use among pregnant veterans. Study findings highlight a continued need for women's health care at the VA, as well as the availability of VA providers knowledgeable about perinatal health issues, and informed community providers regarding women veterans' health.

Nurse staffing and healthcare-associated infections in a national healthcare system that implemented a nurse staffing directive: Multi-level interrupted time series analyses.

BACKGROUND: In 2010, the Veterans Health Administration Office of Nursing Services (VHA ONS) issued a Staffing Methodology (SM) Directive, standardizing the method of determining appropriate nurse staffing for VHA facilities. OBJECTIVES: To assess associations between the Directive, nurse staffing trends, and healthcare-associated infections. RESEARCH DESIGN: We conducted multi-level interrupted time series analyses of nurse staffing trends and the rates of two healthcare-associated infections before and after implementation of the Directive, October 1, 2008 - June 30, 2014. SUBJECTS: Acute care, critical care, mental health acute care, and longterm care nursing units (called Community Living Centers, CLC in VHA) among 285 VHA facilities were included in nurse staffing trends analyses, while acute and critical care units in 123 facilities were used in the analysis of infection rates. MEASURES: Monthly rates were calculated at the facility unit level and included nursing hours per patient day (NHPPD) for all nursing personnel and number of catheter-associated urinary tract infections (CAUTI) and central line-associated bloodstream infections (CLABSI) per 1000 device days. RESULTS: Nursing hours per patient day increased in both time periods. However, the differential change in rate of nursing hours per patient day following implementation of the Directive was not statistically significant. On average, we found a statistically significant decrease of 0.05 unit in the post-Directive central line-associated bloodstream infection rates associated with a unit increase in nursing hours per patient day. CONCLUSIONS: System-wide implementation of Staffing Methodology may be one contributing factor impacting patient outcomes.

Predictors of Long-Term and High-Dose Use of Zolpidem in Veterans.

OBJECTIVE: Prescriptions for sedative hypnotics are routinely initiated and renewed to treat insomnia, despite evidence supporting nonpharmacologic treatments as comparable and more favorable over time. We used national Veterans Health Administration data to assess patient characteristics associated with high-dose and long-term zolpidem use. METHOD: The study included outpatients with new zolpidem prescriptions (January 1, 2013, to June 3, 2014). We defined high-dose use as use of doses above those recommended in the 2013 FDA safety warning (> 5 mg for women, > 10 mg for men) and defined long-term use as at least 180 days of continued supply. We fit separate logistic regression models by sex to evaluate how patient factors, adjusting for facilities, predicted high-dose and long-term use. RESULTS: Of 139,525 new zolpidem users, < 1% of men and 41% of women used high doses within 180 days of initiation, and 20% continued to use zolpidem long-term. Prior-year use of other sleep medications was associated with both high-dose and long-term use. Substance abuse/dependence was associated with high-dose use in women (odds ratio = 1.20, P < .001). Although long-term use was less likely in those over the age of 85 years, about 1 in 5 users aged 65 to 85 continued long-term. In both sexes, individuals of Hispanic ethnicity and nonwhite races were less likely to use long-term, whereas those with ICD-9-CM-defined psychiatric and sleep disorder diagnoses were more likely to use long-term. CONCLUSIONS: Zolpidem use at a higher-than-recommended dose was common in women who were new zolpidem users. In both sexes, 1 in 5 users continued to use zolpidem for at least 180 days. Efforts to improve access to effective nonpharmacologic treatment alternatives may benefit from attention to subpopulations with higher risk of high-dose and long-term use.