The DISCIPLINE Team
The DISCIPLINE team is comprised of Project and Core Leaders who are faculty at the Geisel School of Medicine at Dartmouth and Vanderbilt University Medical Center. This established partnerships builds on strong existing collaborations and capitalizes on the teams demonstrated strengths in clinical informatics (Brown, Matheny), healthcare AI (Matheny, Davis, Al-Garadi), and implementation science (Brown, Lord). The team has collectively studied clinical informatics using healthcare AI approaches and implementation science for over 12 years, evidenced by many notable publications in the field.
Jeremiah Brown, PhD, MS
Jeremiah Brown, PhD, MS, is a Professor of Epidemiology, Biomedical Data Science, and Health Policy & Clinical Practice, and founding Director of the Center for Implementation Science at the Geisel School of Medicine at Dartmouth. Dr. Brown will lead the Admin Core and Project 1. Dr. Brown is the current Study Section Chair for the Science for Implementation in Health and Healthcare (SIHH) at the NIH. Dr. Brown is an expert in implementation science, developing clinical risk models using novel healthcare AI approaches, and was the PI directing three multi-institutional risk modeling grants (R01 HL119664, R01 HL130828, R01 HL157130) modeling clinical endpoints in heart surgery, acute myocardial infarction hospitalization, and COPD. Dr. Brown is the contact PI directing with Dr. Matheny as co-PI R01 HL157130 “DeepCOPD: Development and Implementation of Deep Learning to Predict and Prevent COPD Health Care Encounters” focused on developing and implementing healthcare AI HEALTHCARE AI TOOLKITSs in EHRs for ambulatory COPD. Dr. Brown is the contact PI directing with Drs. Matheny and Davis as co-PI R01DK113201 “IMPROVE AKI: A Cluster-Randomized Trial of Team-Based Coaching Interventions to IMPROVE Acute Kidney Injury” implementing an informatics dashboard, surveillance, and team-based coaching in the national VA. Dr. Brown was the site-PI for Dr. Matheny’s VA HSR&D grant (IIR 11-292) in developing machine learning risk models for acute kidney injury. Dr. Brown also serves as the PI directing the national implementation sciences grant R01 AI155752 “The BASIC trial: Improving implementation of evidence-based approaches and surveillance to prevent bacterial transmission and infection,” which involves team-based coaching and implementation of a surveillance dashboard to monitor and track bacteria transmission and 90-day surgical site infections. These studies form the basis for the work proposed in Project 1 and Project 3 to develop, evaluate, and implement healthcare AI toolkits for infection among pediatric and adult heart surgery patients.
Michael Matheny, MD, MS, MPH
Michael Matheny, MD, MS, MPH, an Associate Professor of Medicine, Biostatistics, and Biomedical Informatics at Vanderbilt University, will provide overall project oversight for Vanderbilt, as well as Core Lead for the HAI2C. Dr. Matheny will provide expertise to develop and validate the DeepCOPD predictive toolkits using statistical methods, machine learning, and deep learning approaches for time-to-event COPD related outcomes. Dr. Matheny was the PI of the VA HSR&D IIR 11-292, in which the team developed the risk models and adapted the prospective surveillance tools to the VA national electronic health record and VA Clinical Assessment Reporting and Tracking (CART) quality improvement database for predicting post-procedural AKI and conducting institutional variation detection. This included currently completing the development and integration of natural language processing to extract risk factors not obtainable in the structured data. Dr. Matheny has collaborated with Dr. Brown on other grants relating to AKI, NLP, and automated surveillance of AKI. Dr. Matheny was a co-PI with Dr. Brown on R01 HL130828 for improving prediction through multiple machine learning approaches for 30-day readmission after acute myocardial infarction leveraging both structured and unstructured EMR data elements. Dr. Matheny is currently co-PI with Dr. Brown on R01 DK113201 “IMPROVE AKI: A Cluster-Randomized Trial of Team-Based Coaching Interventions to IMPROVE Acute Kidney Injury,” a national implementation clinical trial including the evaluation of an automated surveillance reporting dashboard. Dr. Matheny is a co-PI directing with Dr. Brown and Davis as co-PI R01 HL157130 “DeepCOPD: Development and Implementation of Deep Learning to Predict and Prevent COPD Health Care Encounters” focused on developing and implementing healthcare AI toolkits in EHRs for ambulatory COPD. These studies form the basis for the work proposed in Project 2 and Project 4 focusing on the development and application of healthcare AI algorithms and dashboards into routine clinical practice.
Sharon Davis, PhD
Sharon Davis, PhD, Assistant Professor of Biomedical Informatics at Vanderbilt University, will provide overall project oversight and be the site co-lead for Vanderbilt. Dr. Davis is a biomedical informatician with formal statistical training, who focuses on the development and maintenance of predictive models to support practical, implementable clinical prediction tools. She is an expert in machine learning and has developed a suite of generalizable and customizable methods supporting data-driven model updating for regression and machine learning. Dr. Davis’ research involves the design and evaluation of a new statistical test for selecting model updating approaches, a method of continuous real-time calibration assessment, and a calibration drift detector to alert users to changes in model accuracy. Dr. Davis is a co-investigator on multiple federal grants including NIDDK R01DK113201 (Brown, Matheny, Solomon), VA HSR&D IIR-16-072 (Virani), NEST-R2-B5 (Matheny), and NHLBI R01HL149948 (Matheny, Resnic). Dr. Davis is a co-PI directing with Dr. Brown and Matheny as co-PI R01 HL157130 “DeepCOPD: Development and Implementation of Deep Learning to Predict and Prevent COPD Health Care Encounters” focused on developing and implementing healthcare AI toolkits in EHRs for ambulatory COPD. These studies form the basis for the work proposed in Project 1 and Project 3 focusing on learning prediction systems evaluating healthcare AI algorithm model drift and recalibration over time.
Sarah Lord, PhD
Sarah Lord, PhD, a clinical-developmental psychologist and Associate Professor of Psychiatry, Biomedical Data Sciences, and Pediatrics at Geisel School of Medicine will lead the CDS Imp-C Core. Dr. Lord is also part of the leadership team of the recently established Dartmouth Center for Implementation Science (Director: Jeremiah Brown, PhD, Epidemiology). Dr. Lord’s research focuses on development, evaluation, and implementation of evidence-based approaches to treatment of substance use and mental health conditions, with particular focus on leveraging digital technologies as platforms for implementation in diverse systems of care. Current projects include a comparative effectiveness study of medication treatment for opioid use disorder delivered in the context of maternity care (integrated) versus specialty care (Referral) on outcomes for women and infants (PCORI), a study to develop and evaluate a yoga-mindfulness intervention for pregnant women with opioid use disorder (NCCIH), and an implementation study to adapt a digital decisional support tool and care management platform for supported employment to facilitate linkages to community resources for individuals with serious psychiatric and co-occurring substance use disorders (NIDILRR). The theme of Dr. Lord’s work is meeting people where they are, which provides substantial benefit to an AI innovation-focused team. As a clinician-researcher with a background in industry, Dr. Lord brings a stakeholder-centered implementation science and community engagement perspective to her work. These studies form the basis for the work proposed in Projects 1 to 4 focusing on translation and adoption of healthcare AI toolkits into routine clinical practice.
Mohammed Al-Garadi, Phd
Mohammed Al-Garadi,PhD, Research Assistant Professor of Biomedical Informatics at Vanderbilt University, specializes in NLP/ML/DL in healthcare. His research focuses on developing NLP/ML-driven solutions to address healthcare issues, leveraging unstructured data from diverse sources like clinical records and public health data. He has developed several NLP models tailored for medical text, including systems for medical entity recognition, clinical document classification, patient risk prediction, medication extraction, and report summarization. Dr. Al-Garadi is involved in multiple NIH R01 grants, including R01HL157130 (Brown, Sharon, Matheny), R01HL166305 (Girotra), R01HS027417 (Colborn, Louise), and VA HSR&D including, I01HX002917 (Navaneethan), and I01HX003473 (Oh). The commonality of these projects is that Dr. Al-Garadi seeks to lead NLP development in these projects, such as COPD exacerbation prediction, medication information extraction, and analyzing information related to health equity. This work forms the basis for Project 4, which he is the PI for, and Project 2, in which he is leading the NLP developing AI-based models.
Prior Collaborations
