The Health-Economic Impact of a Prospectively Validated Severity of Illness Predictor for Adult Critical Care

Abstract

Background: The ViSIG system, a clinical decision support system for adult intensive care units, was previously shown to be associated with significant reductions in length of stay, readmissions, and duration of mechanical ventilation. However, the health-economic benefits of ViSIG have yet to be fully understood. Infections and acute kidney injury are two measures of clinical significance with key financial implications, and their reduction could further strengthen the published association of ViSIG use with improved outcomes. Further, it is important to demonstrate the economic benefit of ViSIG for its widespread adoption by ICUs. Objective: To investigate the potential health-economic benefits associated with using a previously validated clinical decision support system, ViSIG. Study Design and Methods: In this study, we analyzed data from a previous analysis of ViSIG. The study cohort consisted of six ICUs at two hospitals, with a total of 2,256 admissions in the 'Blinded' Phase and 1,890 admissions in the 'Visible' Phase. We compared the frequency of new infections and Acute Kidney Injury (AKI) between the two phases and used multivariable mixed models to adjust for each patient’s severity of illness. To estimate the economic benefit of ViSIG, we calculated the mean total charges per patient for each phase and the change in patient throughput to arrive at a net gain/loss annual total charges per bed. Results: In the Visible Phase, infections declined by 40.5% (p<0.001). When adjusted for patient severity of illness, the Visible Phase had an odds ratio of 0.52 (95% confidence interval = 0.43, 0.62). There was a 24.5% relative decrease in the occurrence of AKI (p<0.001) and a 35% adjusted decrease from Blinded to Visible Phase (p<0.001). The mean total charges were $59,886 and $55,821 for the Blinded and Visible Phases, respectively. However, due to a reduced ICULOS in the Visible Phase, there was a 14.2% increase in admissions during the 90-day period. This resulted in an estimated mean increased revenue of $3,872 per patient at ICUs using ViSIG. Conclusion: ViSIG was associated with a reduction in two major clinical events: new infections and progression to AKI. This lessened clinical burden may be associated with reduced ICULOS, which resulted in increased patient throughput and, thus, increased revenues. Future studies need to measure specific clinician actions to characterize a cause-and-effect relationship.