Improving Ventilator-Associated Event Surveillance with Dynamic Device-Collected Data

Abstract

Abstract Objective: To understand the feasibility of automated surveillance, we compared the results of a surveillance algorithm to identify ventilator-associated events (VAE) to the current method of manual chart review and data entry. Methods and Methods: This is a retrospective cohort study of patients (n = 654) receiving mechanical ventilation in the hospital ICU between 01/01/2018 and 12/31/2019. A computerized surveillance algorithm was developed and retrospectively. Algorithm-identified ventilator days and VAEs were compared to those reported during the same period for quality improvement. Results: The algorithm identified 2,473 ventilator days and 41 VAEs among 39 patients. Quality improvement reports documented a similar number of days (n = 1,776, p = 0.14) and VAEs (n = 24, p = 0.13). Overall, VAE rates per 1,000 ventilator days identified by the algorithm and reported by quality improvement were similar (20.2 vs. 13.5, respectively, p = 0.46) Discussion: Algorithm-identified ventilator surveillance measures were no different from those identified by manual review and data entry. Conclusion: Triaging VAE surveillance with automated surveillance is feasible and could reduce the time and economic burden of manual chart review.