Real-Time Measures of ICU Strain and their Impact on Patient Outcomes: A Systematic Review

Background ICU strain is a growing challenge due to increasing demand for limited healthcare resources. Identifying real-time strain metrics validated against clinically relevant outcomes is essential to support monitoring and intervention. Objective To identity ICU strain metrics measured in real-time and associated with clinically relevant outcomes Methods We conducted a systematic review searching Ovid Medline, Embase, CINAHL, and Web of Science for studies published between January 2016 and December 2024. We included studies that (1) were conducted in ICU settings (2) described strain metrics measured at a maximum of 24-hour intervals (3) evaluated impact on clinical outcomes. Screening and extraction were performed by 2 reviewers, and quality was assessed using the Newcastle-Ottawa Scale. Analysis was descriptive. Results From 1099 studies, 34 met inclusion criteria. All studies were at low risk of bias. 83 strain metrics mapped onto 5 domains – turnover, nursing workload, occupancy, staffing (medical), and staffing (nursing). 76 outcome measures were mapped onto 11 broad concepts (e.g.hospital mortality). 116 strain-outcome analyses were performed, of which 11 types of outcome measures were assessed. 54% (63/116) of analyses demonstrated worse patient outcomes with rising strain. Strain metrics consistently associated with adverse outcomes were nursing staffing (8/9) and workload (6/8), whilst occupancy (33/63) and patient turnover (7/16) demonstrated moderate association. The Activity Index, a composite score of occupancy, workload and staffing, was consistently associated with worse outcomes. Conclusion Nursing staffing, nursing workload, occupancy and turnover rates are real-time ICU strain metrics associated with adverse outcomes and could be incorporated into a real-time measurement of ICU strain. Future research should validate metrics for prospective monitoring and prediction modelling, as well as integration into decision support systems.