Improved Survival in Critically Ill Patients After Implementing A Visual Decision Support System

David H Berger, MD, MHCM, FACS; John Holcomb, MD, FACS; John Stroh, MD; James P Herlihy, MD; Panagiotis Kougias, MD, MSc, FACS

Presented by Baylor St. Luke’s Medical Center at the American College of Surgeons Clinical Congress

Background

Improvement in clinical outcomes after implementation of electronic health records (EHRs) has not occurred. Actionable clinical data is not easily visible. We hypothesized that implementation of a real time bedside clinical visualization system would decrease hospital mortality. 

Methods

A web-based clinical surveillance and decision support system was installed Oct 2017. Results from Oct–Dec 2017 were not visible to the clinicians (baseline data, BD). Data visible to the care team (Jan–June 2018) comprise the intervention data (ID). The National Early Warning Score (NEWS) was used to identify patients at high risk (≥ 7) of adverse outcomes. The automated system continuously displayed updated vital signs, laboratory data, bundles of care and NEWS to all members of the care team. Hospital mortality, ICU admission, hospital and ICU length of stay (hLOS and iLOS) and demographics were recorded. The Chi-squared test was used to evaluate significance. 

Results

There were 3090 BD and 8928 ID patients evaluated. Median age (60 years), gender (45% male) and mean NEWS (3) were the same for both groups. There were 343 patients with NEWS ≥ 7 in BD group and 1311 in the ID group. Time in the ED, ICU admissions, hLOS and iLOS were similar between groups. Hospital mortality decreased from 10.8% in BD to 7.2% in ID, (p =0.03).  

Conclusion

Implementation was associated with a 33% decrease in hospital mortality. Integration of a continuously available decision support system can help the care team identify the sickest patients in real time and improve quality of care.