Excess resource use and cost of drug-resistant infections for six key pathogens in Europe: a systematic review and Bayesian meta-analysis

Background: Quantifying the resource use and cost of antimicrobial resistance establishes the magnitude of the problem and drives action. Objectives: Assessment of resource use and cost associated with infections with six key drug-resistant pathogens in Europe. Methods: A systematic review and Bayesian meta-analysis. Data sources: MEDLINE (Ovid), Embase (Ovid), Econlit databases, and grey literature for the period 1 January 1990, to 21 June 2022. Study eligibility criteria: Resource use and cost outcomes (including excess length of stay, overall costs, and other excess in or outpatient costs) were compared between patients with defined antibiotic-resistant infections caused by carbapenem-resistant (CR) Pseudomonas aeruginosa and Acinetobacter baumannii, CR or third-generation cephalosporin Escherichia coli (3GCREC) and Klebsiella pneumoniae, methicillin-resistant Staphylococcus aureus, and vancomycin-resistant Enterococcus faecium, and patients with drug-susceptible or no infection. Participants: All patients diagnosed with drug-resistant bloodstream infections (BSIs). Interventions: NA. Assessment of risk of bias: An adapted version of the Joanna Briggs Institute assessment tool, incorporating case-control, cohort, and economic assessment frameworks. Methods of data synthesis: Hierarchical Bayesian meta-analyses were used to assess pathogen-specific resource use estimates. Results: Of 5969 screened publications, 37 were included in the review. Data were sparse and heterogeneous. Most studies estimated the attributable burden by, comparing resistant and susceptible pathogens (32/37). Four studies analysed the excess cost of hospitalization attributable to 3GCREC BSIs, ranging from -€ 2465.50 to € 6402.81. Eight studies presented adjusted excess length of hospital stay estimates for methicillin-resistant S. aureus and 3GCREC BSIs (4 each) allowing for Bayesian hierarchical analysis, estimating means of 1.26 (95% credible interval [CrI], −0.72 to 4.17) and 1.78 (95% CrI, −0.02 to 3.38) days, respectively. Conclusions: Evidence on most cost and resource use outcomes and across most pathogen-resistance combinations was severely lacking. Given the importance of this evidence for rational policymaking, further research is urgently needed.