Teams mounting separate clinical trials at Vanderbilt University Medical Center sought to form research cohorts on two rare, life-threatening adverse drug reactions: drug-induced torsade de pointes (an arrythmia) and Stevens-Johnson syndrome/toxic epidermal necrolysis.
In a paper in the Journal of the American Medical Informatics Association, Sarah DeLozier, Joshua Denny, MD, MS, and colleagues report an automated solution for identifying candidates for these trials.
While study coordinators continued to search manually for candidates, the team, with approval from the institution’s research ethics board, set up immediate automated keyword searches on clinical notes signed by medical trainees or nurse practitioners. For notes that matched these queries, a detailed alert was automatically sent via email to study coordinators and was entered automatically in a project database.
Compared to manual chart review, more than twice the number of drug reactions were found with the automated system, leading to rapid securing of patient consent and speeding the completion of enrollment for these trials.
Also on the study were Peter Speltz, Jason Brito, Leigh Anne Tang, Janey Wang, MS, MEng, Joshua Smith, PhD, Dario Giuse, PhD, Elizabeth Phillips, MD, Kristina Williams, MSN, Teresa Strickland, RN, Giovanni Davogustto, MD, and Dan Roden, MD. The study was supported in part by the National Institutes of Health (GM115305).
