PREDICT-AVF: The goal of this project is to leverage machine learning methods to develop a tool that can accurately predict outcomes of arteriovenous fistulas (AVFs) in patients with advanced kidney disease based on routinely collected low or zero-cost clinical, anatomical, procedural, and radiological data. We are using data from two rigorously conducted multicenter prospective clinical trial programs to develop our tool: the PATENCY trials (914 patients with radiocephalic AVF) and the National Institutes of Health (NIH) Hemodialysis Fistula Maturation (HFM) Consortium (602 patients with single-stage upper-extremity AVF). Preliminary versions of this calculator to predict readiness of AVF for use and need for reintervention are available online.
EXTINGUISH Target Trial: The aim of this study is to determine if systemic inflammation reduction lowers rates of adverse cardiovascular and limb events in patients with peripheral artery disease (PAD), based on an emulated target trial of Mass General Brigham (MGB)-Medicare linked cohort of PAD patients. Early results from this study have been presented at the 2023 European Society for Cardiovascular Congress in Amsterdam, Netherlands and the American Heart Association 2023 Scientific Sessions in Philadelphia, PA.
MAGNIFY-PAD: This project seeks to develop a model-based algorithm (MAGNIFY-PAD) using machine learning techniques to accurately capture patients with PAD from health databases, which can be used to conduct high-quality health services research in PAD.
Health Services Research in Dialysis Access: We are using the multicenter Vascular Quality Initiative database to study population-based temporal trends and geographic-, center-, and provider-based variation in surgical hemodialysis access creation in North America. In addition, we are using institutional and clinical trial databases to study factors associated with successful low-flow forearm AV access creations (role of pre-operative venogram vein mapping; 2-stage forearm accesses in “no option patients”; effect of intraoperative heparin use on outcomes, etc.).
Target Trial Emulation Using VQI: The aim of this project is to enhance causal inference research using observational data from the Vascular Quality Initiative via application of the Target Trial Emulation framework. We are comparing outcomes of single vs. dual antiplatelet therapy after carotid endarterectomy as an example to highlight application of Target Trial Emulation principles in the VQI.
Dial-In Collaborative: We are leading a multicenter observational cohort comparative effectiveness study of endovascular vs. open arteriovenous fistula creation as member of the Dialysis Access Learning and Innovation (DiAL-In)
Clinical Trials
ACCESS-2 Trial: A Study Testing the Use of a Perivascular Sirolimus Formulation (Sirogen) in ESRD Patients Undergoing AV Fistula Surgery
Humacyte V-012: A Phase 3 Randomized Study to Compare the Efficacy and Safety of the Humacyte Human Acellular Vessel (HAV) With That of an Autogenous Arteriovenous Fistula (AVF) in Female Patients With End-Stage Renal Disease Requiring Hemodialysis
