Higher sensitivity, lower specificity than patient safety indicators based on discharge coding
TUESDAY, Aug. 23 (HealthDay News) -- Natural language process analysis of electronic medical records has higher sensitivity and lower specificity for identifying postoperative complications than patient safety indicators based on discharge coding for patients undergoing surgical procedures, according to a study published in the Aug. 24/31 issue of the Journal of the American Medical Association.
Harvey J. Murff, M.D., M.P.H., from the Veterans Affairs Medical Center in Nashville, Tenn., and colleagues investigated the use of a natural language processing-based approach to identify postoperative complications within a comprehensive electronic medical record for 2,974 patients who underwent surgical procedures. Data for postoperative acute renal failure requiring dialysis, sepsis, deep vein thrombosis, pulmonary embolism, myocardial infarction (MI), and pneumonia were collected through a medical record review. Sensitivity and specificity of the natural language processing approach to identify complications was evaluated and compared to the performance with patient safety indicators that use discharge coding information.
The investigators found that the proportion of each event was 2 percent for postoperative acute renal failure requiring dialysis, 0.7 percent for pulmonary embolism, 1 percent for deep vein thrombosis, 7 percent for sepsis, 16 percent for pneumonia, and 2 percent for MI. The proportion of correctly identified cases between natural language process and patient safety indicators was 82 versus 38 percent for acute renal failure, 59 versus 46 percent for venous thromboembolism, 64 versus 5 percent for pneumonia, 89 versus 34 percent for sepsis, and 91 versus 89 percent for MI. Natural language processing and patient safety indicators were very specific for these diagnoses.
"Natural language processing analysis of electronic medical records to identify postoperative complications had higher sensitivity and lower specificity compared with patient safety indicators based on discharge coding," the authors write.
Full Text (subscription or payment may be required)
Editorial (subscription or payment may be required)