Background: Efforts to promote patient safety and prevent medical errors are becoming increasingly important in the medical field; as such, diagnostic error is being recognized as an important contributor to patient morbidity and mortality. The value of diagnostic decision support (DDS) tools in reducing diagnostic errors is unclear. ‘Isabel’ is a web-based diagnosis checklist system which utilizes the patient's age, gender, and specific keywords to determine a list of diagnoses. Several studies have demonstrated the utility of Isabel in clinical practice, however most have focused on adult diagnostic dilemmas. The utility of DDS in Pediatrics is understudied in the medical literature.
Methods: We sought to examine the performance characteristics of Isabel using challenging published cases. We also sought to determine the impact of level of training on inclusion of correct diagnosis on Isabel’s diagnostic checklist. Using the case based textbook, Pediatric Complaints and Diagnostic Dilemmas, four investigators (senior faculty, junior faculty, pediatric resident, and fourth year medical student) selected five keywords from 25 cases; Investigators were blinded to the confirmed diagnosis. Age, gender and these 5 keywords were run through Isabel. Success rate of Isabel was examined against the confirmed diagnosis on 'Page 1' (top 10 diagnoses) and 'View all' (top 30-40 diagnoses) of Isabel's checklist. Level of training, complexity of keyword phrasing, and success rate by diagnosis were also examined.
Results: 100 observations were made from 4 investigators using the same 25 cases. Isabel correctly included the confirmed diagnosis in 46/100 (45%) of cases on 'Page 1', and 64/100 (64%) of cases on 'View all' (p=0.01). Isabel inquiries from Lower Level of Training (LLT) (medical student and pediatric resident) yielded 19/50 (38%) cases included on 'Page 1' vs. 27/50 (54%) of cases for Higher Level of Training (HLT) (junior and senior faculty) (p=0.1). Isabel's performance on 'View all' increased to 29/50 (58%) for LLT and 35/50 (70%) for HLT (p=0.21). 6 of the 25 (24%)cases were not identified by any of the 4 investigators.
Conclusion: The diagnostic checklist system, Isabel, included the correct diagnosis in 64% of pediatric diagnostic dilemmas. HLT was associated with increased performance by Isabel. This study demonstrates that the web based DDS tool, Isabel, may add value to avoidance of premature closure and reduction of diagnostic errors.
See more of: The 33rd Annual Meeting of the Society for Medical Decision Making