Kidney Week 2025 Early Program - Advances in Research Conference 2025: Generative Artificial Intelligence in Kidney Disease Research and Management-Omics in the Clinics

Video Transcription

Video Summary

Kasei Green describes how the University of Colorado’s Center for Personalized Medicine operationalizes “omics in the clinic” through large-scale, preemptive pharmacogenomic (PGx) testing integrated into the electronic health record (Epic). The goal is to deliver the right drug and dose for each patient by using genetic variants that affect medication metabolism and response. Rather than relying on clinicians to look up results in a genomic tab, the program uses tiered clinical decision support: severe, interruptive alerts for high-risk situations (e.g., clopidogrel nonresponse genotypes) and non-interruptive inline warnings for lower-severity issues to reduce alert fatigue.<br /><br />Using a biobank of ~270,000 participants, the team has returned genomic indicators to over 100,000 people and generated over a million results, finding that ~90% have at least one actionable PGx phenotype. Thousands of patients have had actionable alerts during prescribing, without overwhelming support staff—helped by “one-click” recommended alternatives and a consult line.<br /><br />They also implemented PGx as standard-of-care in GI oncology with rapid turnaround (under 5 days for chemo-related genes), aiming to generate real-world outcomes data to convince insurers to reimburse testing. Green emphasizes lessons for AI deployment: workflow fit, stakeholder buy-in, monitoring compliance, and avoiding alert fatigue.

Asset Subtitle

Casey Greene

Meta Tag

Module ARC

Speaker Casey Greene

Keywords

preemptive pharmacogenomics

clinical decision support alerts

Epic electronic health record integration

actionable PGx phenotypes

biobank genomic return of results

GI oncology pharmacogenomic testing

alert fatigue reduction