Learn about the Cardiovascular Outcomes team's current projects:
Information Extraction from EMRs to Predict Readmission following Acute Myocardial Infarction (NHLBI R01)
This is a two-site study that will examine the ability of information in electronic medical records (EMR) to predict hospital readmission. There are existing predictive models that utilize structured fields from the EMR, but are limited in power due to fragmented data entry and storage, poor compliance in filling out all structured fields, and inability to take advantage of unstructured information in narrative notes. This study seeks to apply novel Natural Language Processing (NLP) techniques to extract data, including symptoms, treatments, procedures, diagnoses, social risk factors, and functional status, from these narrative notes. We hypothesize that combining structured variables form the EMR with NLP-derived variables will improve our ability to predict 30-day readmission following acute myocardial infarction.
We will map potential AMI risk factors for 30-day readmission using structured variables from Epic EMR systems and novel NLP-derived variables at both sites (Dartmouth-Hitchcock medical Center in Lebanon, NH and University of Utah Hospital in Salt Lake City, UT). We will then develop predictive models in tandem at our two sites. These models will be cross-validated at the other site, and the final model will be validated for portability using non-Epic EMR systems.
Dartmouth College & Dartmouth-Hitchcock Medical Center
Jeremiah Brown, MS, PhD
Amarendra Das, MD, PhD
Todd MacKenzie, MSc, PhD
Paul Thompson, M.Libr, PhD
Nathaniel Niles II, MD
University of Utah
Wendy Chapman, PhD
Ramkiran Gouripeddi, MBBS, MS
Bruce Bray, MD
Rashmee Shah, MS, MD
Vanderbilt University and VA Tennessee Valley Healthcare System
Michael Matheny, MS, MPH, MD
Novel Biomarkers to Predict Readmission in Pediatric and Adult Heart Surgery (NHBLI R01)
This project is supported through an R01 grant from the National Heart, Lung, and Blood Institute (NHLBI). This is a multicenter retrospective study to determine if predictive modeling for hospital readmission can be improved by the addition of novel biomarker data. Specifically, the study will investigate whether the addition of peri-operative biomarkers of cardiac injury (ST2, B-type natriuretic peptide, cardiac troponin T), renal injury (cystatin C), and non-specific inflammation (galectin-3, cytokines) will improve upon clinical model prediction for 30-day readmission or death for adult and pediatric patients.
We will use a cohort from the Northern New England Cardiovascular Disease Study Group (NNECDSG) for model derivation. For external validation, we will use the Translational Research Investigating Biomarkers in Early Acute Kidney Injury (TRIBE) cohort at Yale University for adult patients and a cohort from the Johns Hopkins Taussig Children’s Heart Center in Baltimore for pediatric patients.
Yale University – Program of Applied Translational Research
Chirag Parikh, MD, PhD, FACP
Heather Thiessen Philbrook
Steven Coca, MS, DO
Northern New England Cardiovascular Disease Study Group
Jeremiah Brown, MS, PhD
Todd Mackenzie, PhD
Johns Hopkins University
Allen Everett, MD
Marshall Jacobs, MD (retired)
Society of Thoracic Surgeons
Jeffrey Jacobs, MD
David Shahian, MD
University of Michigan
Sara Pasquali, MD
Donny Likosky, MS, PhD
Cincinnati Children’s Hospital Medical Center
Samir Shah, MD, MSCE
Prasad Devarajan, MD
Multi-vessel coronary artery disease patients have multiple treatment options, each with tradeoffs. Option Grid™ decision aids for clinical encounters are tools that enable patients and physicians discuss and compare treatment options. You can learn more about them on the Option Grid's website. We developed an Option Grid decision aid for patients with multi-vessel coronary artery disease with the help of healthcare providers from across Northern New England. We’re currently evaluating the tool for improvements in patient knowledge, patient decisional conflict, and shared decision making.
We work closely with researchers from The Preference Lab, Northern New England Cardiovascular Disease Study Group, The Option Grid Collaborative, and with clinical teams at individual sites including Dartmouth-Hitchcock Medical Center, Maine Medical Center, Albany Stratton VA Medical Center, and Albany Medical Center.
If you have any questions about this project, please contact Elizabeth Nichols.
One of our team’s current areas of focus is the Medicare Hospital Readmissions Reduction Program (HRRP). In particular, we are looking at how various characteristics of hospitals and their local areas and patient populations are related to the readmission penalties that they incur. For our study, we have put together a database using publicly available information from various sources, including the Centers for Medicare and Medicaid Services, the Bureau of Labor Statistics, the U.S. Census Bureau, and the Dartmouth Atlas. We hope to identify risk factors for penalties on the hospital level.
The Dartmouth Institute for Health Policy and Clinical Practice
Carrie Colla, MA, PhD
A. James O’Malley, MS, PhD
Acute kidney injury (AKI) is a common complication following operations and may occur during or following a hospital stay. When AKI occurs it leads to a higher risk of immediate and long-term consequences including a more rapid progression of kidney disease, need for renal replacement therapy or transplant, and mortality.
In this section we report on the epidemiology of AKI and outcomes and risk prediction models in patients undergoing cardiac catheterization (angiography or angioplasty) and cardiac surgery.
We are conducting interventions in cardiac catheterization laboratories across the United States to prevent AKI from occurring after patient exposure to radio-contrast (also known as contrast-induced nephropathy). We report on our initial findings from Northern New England and the interventions we have found to helpful in the prevention of AKI.
Team-based coaching and surveillance interventions to IMPROVE Acute Kidney Injury
Jeremiah R. Brown, PhD, MS
Mark J. Sarnak, MD
David J. Malenka, MD
Richard Solomon, MD
R. Brooks Robey, MD, PhD