Reference: Middleton, B.; Kovack, J.; & Tovar, M. Cost-Effectiveness of a Medical Decision-Support System: A Pilot Study. Knowledge Systems Laboratory, July, 1991.
Abstract: Decision-support systems (DSS) have been shown to assist doctors in making diagnoses in difficult cases. It has not been shown that use of DSS leads to a more cost-effective diagnostic work-up. We performed a historical prospective case-control analysis comparing time to correct diagnosis and cost of diagnostic tests between doctors in a university hospital ward team practice (controls) and a physician assisted by QMR (study subject). From 20 diagnostic dilemma referrals in one month and 18 NEJM CPCs 8 cases were eligible for abstraction by day to determine findings, diagnostic tests, and working differential diagnosis for each day of the hospital course. Case findings by day were presented to the study subject for analysis with QMR. A two-tailed matched pair t-test was used to determine if the time to arrival at the correct diagnosis in days and the diagnostic test costs to arrive at the correct diagnosis by day was significantly different between the study and control groups. No significant difference was found in time to arrival at the correct diagnosis between the two groups. On average, however, the time to arrival at the correct diagnosis was reduced by 2.4 days with use of QMR. The t statistic for diagnostic test charges was significantly different on days one and four. The overall t statistic for net diagnostic charges was not statistically significant. We conclude use of QMR early in the hospital course by an experienced physician-user may lead to reduced net charges and early diagnosis in diagnostically challenging cases.