(HealthDay News) — Patient characteristics extracted from their electronic health records can be used to refine risk prediction for hospital readmission after percutaneous coronary intervention, according to a study published online August 18 in Circulation: Cardiovascular Quality and Outcomes.
Jason H. Wasfy, MD, MPhil, from Massachusetts General Hospital in Boston, and colleagues matched readmitted to non-readmitted patients in a 1:2 ratio by risk of readmission in an effort to improve ability to predict 30-day readmission after percutaneous coronary intervention. Unstructured and structured data were extracted from the electronic health record for 888 readmitted patients and 1,776 non-readmitted patients.
The researchers found that cases and controls differed significantly with respect to interpreter (7.9 and 5.3%, respectively; P=0.009), emergency department visits (1.12 and 0.77, respectively; P<0.001), homelessness (3.2 and 1.6%, respectively; P=0.007), anticoagulation (33.9 and 22.1%, respectively; P<0.001), atrial fibrillation/flutter (32.7 and 28.9%, respectively; P=0.045), presyncope/syncope (27.8 and 21.3%, respectively; P<0.001), and anxiety (69.4 and 62.4%, respectively; P<0.001) in univariate analyses. There were independent associations for anticoagulation, emergency department visits, and anxiety with readmission.
“These results have the potential to improve risk assessment for early readmission after percutaneous coronary intervention, and emphasize the role of unmeasured confounders in risk assessment,” the authors write.
Several authors disclosed financial ties to the pharmaceutical and medical technology industries.