Interventions to improve early recognition and management of sepsis in emergency departments: A systematic review and meta-analysis.

Objectives

Timely recognition and management of sepsis in emergency departments (ED) is critical to improving patient outcomes. We conducted a systematic review and meta-analysis to evaluate the impact of targeted interventions on sepsis care in ED.

Methods

Studies published between 2001 and 2024 were included if they assessed interventions to improve sepsis recognition and/or management. Risk of bias was assessed using the ROBINS-I quality appraisal tools by Cochrane. Our primary outcome was the posterior probability of any intervention effect, using a Bayesian approach. Outcomes of interest included rates of compliance with sepsis guidelines, patient mortality, and time to administration of intravenous antibiotics and fluids. Both frequentist and Bayesian meta-analyses were performed, and heterogeneity was assessed using the I2 statistic.

Results

Fifty-six studies were included; all were before-and-after designs, except a single randomised controlled trial. Bayesian analyses estimated a posterior probability >99% of any improvement in compliance, and a posterior probability of 97% of any mortality reduction, in the context of sceptical priors, for targeted sepsis interventions. Pooled random-effects showed higher compliance during intervention periods (Risk Ratios(RR) =1.71, 95% Confidence Intervals(CI) 1.39-2.15; I2 = 93%) and lower mortality (RR =0.80, 95% CI 0.73-0.86; I2 = 46%), when compared to non-intervention periods. Time to treatment was also estimated to be reduced for antibiotics administration (Mean Difference (MD) = -32.5 min, 95% CI -44.7 to -20.2) and fluid administration (MD = -20.4 min, 95% CI -37.6 to -3.1). Posterior probabilities suggested a 90% likelihood of any time reduction, though estimates of reduced antibiotic times were sensitive to analysis that accounted for risk-of-bias assumptions.

Conclusions

AND IMPLICATIONS TO CLINICAL PRACTICE Targeted ED interventions improve compliance with sepsis guidelines and reduce mortality. While gains in treatment timeliness were modest, likely reflecting prioritisation of sicker patients, evidence supports multimodal, context-specific strategies strengthening early sepsis management. Given heterogeneity of studies, future research should evaluate specific intervention components in well-designed, adequately powered randomised controlled trials. REGISTRATION PROSPERO ID: CRD42024560232.