Validation of Data in the Veteran Health Administration Electronic Medical Record for Identification of Tobacco Use.

Purpose: Tobacco use is not commonly represented as computable information in the electronic health record (EHR). We developed an algorithm in the Veterans Health Administration (VHA) to identify tobacco ever-use among Veterans.

Methods: We used the VHA corporate data warehouse to develop an algorithm comprised of multiple data types (health factors [semi-structured template data entry and decision support tools], billing, orders, medication, and encounter codes) to identify tobacco ever-use (current or former) versus never use.

Automated Extraction of Mortality Information From Publicly Available Sources Using Large Language Models: Development and Evaluation Study.

Background: Mortality is a critical variable in health care research, especially for evaluating medical product safety and effectiveness. However, inconsistencies in the availability and timeliness of death date and cause of death (CoD) information present significant challenges. Conventional sources such as the National Death Index and electronic health records often experience data lags, missing fields, or incomplete coverage, limiting their utility in time-sensitive or large-scale studies.

Unlocking efficiency in real-world collaborative studies: a multi-site international study with one-shot lossless GLMM algorithm.

The widespread adoption of real-world data has given rise to numerous healthcare-distributed research networks, but multi-site analyses still face administrative burdens and data privacy challenges. In response, we developed a Collaborative One-shot Lossless Algorithm for Generalized Linear Mixed Models (COLA-GLMM), the first-ever algorithm that achieves both lossless and one-shot properties. COLA-GLMM ensures accuracy against the gold standard of pooled data while requiring only summary statistics and completes within a single communication round, eliminating the usual back-and-forth overhead.

A Statistical Framework to Detect and Quantify Operator-Learning Curves in Medical Device Safety Evaluation.

Importance: Safety issues leading to patient harm and significant costs have been identified in several post-market medical devices. Recently, powerful learning effects (LE) have been documented in numerous medical devices. Correctly attributing safety signals to learning or device effects allows for appropriate corrective actions and recommendations to improve patient safety.

Neural network models for predicting readmission among patients undergoing peripheral vascular intervention using electronic health record data and clinical registry data.

Objectives: To determine whether neural network models based on electronic health record (EHR) data can match and augment the performance of models based on clinical registry data in predicting readmission after peripheral vascular intervention (PVI).

Design: Observational cohort study.

Setting: Vascular Quality Initiative registry and INSIGHT Clinical Research Network EHR data from multiple academic institutions in New York City.

Real-time automated billing for tobacco treatment: developing and validating a scalable machine learning approach.

Objectives: To develop CigStopper, a real-time, automated medical billing prototype designed to identify eligible tobacco cessation care codes, thereby reducing administrative workload while improving billing accuracy.

Materials And Methods: ChatGPT prompt engineering generated a synthetic corpus of physician-style clinical notes categorized for CPT codes 99406/99407. Practicing clinicians annotated the dataset to train multiple machine learning (ML) models focused on accurately predicting billing code eligibility.

Predicting Heart Failure Outcomes Using Patient-Reported Health Status: Real-World Validation of the KCCQ-12.

Background: The Kansas City Cardiomyopathy Questionnaire-12 (KCCQ-12), a patient-reported outcome measure for adults with heart failure, is associated with hospitalizations and mortality in clinical trials. Curated data sets from controlled trials differ substantially from pragmatic data collected from real-world settings, however, and few data exist on the KCCQ-12's predictive utility in clinical practice.

Objectives: This study sought to evaluate the predictive utility of the KCCQ-12 for hospitalizations and mortality when administered during outpatient heart failure care.

Sustained Improvements After Intervention to Prevent Contrast-Associated Acute Kidney Injury: A Randomized Controlled Trial.

Background: In the IMPROVE AKI (A Cluster-Randomized Trial of Team-Based Coaching Interventions to Improve Acute Kidney Injury) trial, a combination of team-based coaching and data-driven surveillance dashboards reduced the odds of AKI following cardiac catheterization by 46%. The objective of this study was to determine if improvements in AKI outcomes would be sustained after completion of the active intervention.

Methods And Results: A 2×2 factorial cluster-randomized trial with an 18-month active intervention phase (October 2019-March 2021) and an 18-month sustainability phase (April 2021-September 2022) conducted among cardiac catheterization laboratories in 20 Veterans Affairs sites.

Cost-Effectiveness of Team-Based Coaching With Surveillance for Prevention of Acute Kidney Injuries.

Importance: More than 10% of US patients undergoing endovascular procedures experience contrast-associated acute kidney injuries (AKIs), resulting in increased costs and health deficits. Prevention protocols reduce AKIs, but uptake and adherence vary greatly, and the cost-effectiveness of these interventions is unknown.

Objective: To analyze the cost-effectiveness of 4 implementation interventions for AKI prevention in patients undergoing cardiac catheterizations.

Long-Term Outcomes of Peripheral Artery Disease in Veterans: Analysis of the Peripheral Artery Disease Long-Term Survival Study (PEARLS).

Background: Contemporary research in peripheral artery disease (PAD) remains limited due to lack of a national registry and low accuracy of diagnosis codes to identify patients with PAD.

Methods: Leveraging a novel natural language processing system that identifies PAD with high accuracy using ankle-brachial index and toe-brachial index values, we created a registry of 103 748 patients with new-onset PAD in the Veterans Health Administration. Study end points include mortality, cardiovascular events (hospitalization for acute myocardial infarction or stroke) and limb events (hospitalization for critical limb ischemia or major amputation) and were identified using Veterans Affairs and non-Veterans Affairs encounters.

Emerging algorithmic bias: fairness drift as the next dimension of model maintenance and sustainability.

Objectives: While performance drift of clinical prediction models is well-documented, the potential for algorithmic biases to emerge post-deployment has had limited characterization. A better understanding of how temporal model performance may shift across subpopulations is required to incorporate fairness drift into model maintenance strategies.

Materials And Methods: We explore fairness drift in a national population over 11 years, with and without model maintenance aimed at sustaining population-level performance.

Detecting Opioid Use Disorder in Health Claims Data With Positive Unlabeled Learning.

Accurate detection and prevalence estimation of behavioral health conditions, such as opioid use disorder (OUD), are crucial for identifying at-risk individuals, determining treatment needs, monitoring prevention and intervention efforts, and recruiting treatment-naive participants for clinical trials. The availability of extensive health data, combined with advancements in machine learning (ML) frameworks, has enabled researchers to employ various ML techniques to predict or identify OUD within patient health data. Ideally, we could directly estimate the prevalence, or the proportion of a population with a condition over time.

Adapting a Risk Prediction Tool for Neonatal Opioid Withdrawal Syndrome.

Background: The American Academy of Pediatrics recommends up to 7 days of observation for neonatal opioid withdrawal syndrome (NOWS) in infants with chronic opioid exposure. However, many of these infants will not develop NOWS, and infants with seemingly less exposure to opioids may develop severe NOWS that requires in-hospital pharmacotherapy. We adapted and validated a prediction model to help clinicians identify infants at birth who will develop severe NOWS.

Risk of aortic aneurysm or dissection following use of fluoroquinolones: a retrospective multinational network cohort study.

Background: Fluoroquinolones (FQs) are commonly used to treat urinary tract infections (UTIs), but some studies have suggested they may increase the risk of aortic aneurysm or dissection (AA/AD). However, no large-scale international study has thoroughly assessed this risk.

Methods: A retrospective cohort study was conducted using a large, distributed network analysis across 14 databases from 5 countries (United States, South Korea, Japan, Taiwan, and Australia).

Capacity assessment for EHR-based medical device post-market surveillance for synthetic mid-urethral slings among women with stress urinary incontinence: a NEST consortium study.

Objectives: To evaluate the feasibility for use of electronic health record (EHR) data in conducting adverse event surveillance among women who received mid-urethral slings (MUS) to treat stress urinary incontinence (SUI) in five health systems.

Design: Retrospective observational study using EHR data from 2010 through 2021. Women with a history of MUS were identified using common data models; a common analytic code was executed at each site.

Embracing Generative Artificial Intelligence in Clinical Research and Beyond: Opportunities, Challenges, and Solutions.

To explore threats and opportunities and to chart a path for safely navigating the rapid changes that generative artificial intelligence (AI) will bring to clinical research, the Duke Clinical Research Institute convened a multidisciplinary think tank in January 2024. Leading experts from academia, industry, nonprofits, and government agencies highlighted the potential opportunities of generative AI in automation of documentation, strengthening of participant and community engagement, and improvement of trial accuracy and efficiency. Challenges include technical hurdles, ethical dilemmas, and regulatory uncertainties.

Augmenting the Hospital Score with social risk factors to improve prediction for 30-day readmission following acute myocardial infarction.

Background: Hospital Score is a well-known and validated tool for predicting readmission risk among diverse patient populations. Integrating social risk factors using natural language processing with the Hospital Score may improve its ability to predict 30-day readmissions following an acute myocardial infarction.

Methods: A retrospective cohort included patients hospitalized at Vanderbilt University Medical Center between January 1, 2007, and December 31, 2016, with a primary index diagnosis of acute myocardial infarction, who were discharged alive.

National Survey of Lung Cancer Screening Eligibility in United States Veterans.

Introduction: Lung cancer screening is underutilized, especially in rural areas where lung cancer mortality is high. Approximately 11.2% of the U.

Artificial Intelligence In Health And Health Care: Priorities For Action.

The field of artificial intelligence (AI) has entered a new cycle of intense opportunity, fueled by advances in deep learning, including generative AI. Applications of recent advances affect many aspects of everyday life, yet nowhere is it more important to use this technology safely, effectively, and equitably than in health and health care. Here, as part of the National Academy of Medicine's Vital Directions for Health and Health Care: Priorities for 2025 initiative, which is designed to provide guidance on pressing health care issues for the incoming presidential administration, we describe the steps needed to achieve these goals.

Predicting Nephrotoxic Acute Kidney Injury in Hospitalized Adults: A Machine Learning Algorithm.

Rationale And Objective: Acute kidney injury (AKI) is a common complication among hospitalized adults, but AKI prediction and prevention among adults has proved challenging. We used machine learning to update the nephrotoxic injury negated by just-in time action (NINJA), a pediatric program that predicts nephrotoxic AKI, to improve accuracy among adults.

Study Design: A retrospective cohort study.

A machine learning framework to adjust for learning effects in medical device safety evaluation.

Objectives: Traditional methods for medical device post-market surveillance often fail to accurately account for operator learning effects, leading to biased assessments of device safety. These methods struggle with non-linearity, complex learning curves, and time-varying covariates, such as physician experience. To address these limitations, we sought to develop a machine learning (ML) framework to detect and adjust for operator learning effects.