Hospital unit colonization pressure and nosocomial acquisition of drug susceptible and drug resistant pathogens

BackgroundHospitalized patients are at risk for developing hospital acquired infections. Active surveillance for bacterial colonization is effective at preventing infection but is resource-intensive and limited to high-risk units and a subset of high-risk pathogens. Colonization pressure (CP) for common pathogens has been associated with hospital acquired infection and can be calculated in real-time using data in the electronic health record. We aimed to assess the impact of CP on nosocomial acquisition for a range of drug susceptible and drug resistant pathogens, across an entire hospital system. MethodsWe conducted a retrospective matched cohort study of all inpatients admitted to a large regional integrated healthcare system between May 2015 and July 2024 who stayed in one room during the 30 day observation period. Cases had target organisms detected in any clinical or surveillance culture taken between 3 and 30 days after admission into their first room. Controls were matched on demographics, length of stay, prior surgery and 14 classes of antibiotic exposure. Our outcome was nosocomial acquisition of 11 common pathogens spread across enteric, skin and environmental niches. We applied conditional logistic regression and XGBoost to model nosocomial acquisition using as covariates the Elixhauser comorbidity index and CP, defined as the time-weighted prevalence of an organism in ward co-occupants over the previous 6 months. CP was calculated for 9 organism sets, including the Enterobacterales, ESBL Enterobacterales, vancomycin susceptible and resistant Enterococcus spp, C. difficile, methicillin susceptible Staphylococcus aureus (MSSA), methicillin resistant S. aureus (MRSA) and drug susceptible and drug resistant (DR) P. aeruginosa (PsA). FindingsOur pooled cohort included 14,923 cases matched to 28,480 controls. Hospital acquisition occurred four times more frequently for drug susceptible versus drug resistant organisms. Baseline characteristics were well matched between cases and controls. The strongest positive associations were between CPC. difficile and nosocomial acquisition of C. difficile (+32.5%, 95% CI +21.9% to +44.0%), CPESBL and ESBL K. pneumoniae (+29.4%, 95% CI +11.3% to +50.6%), and CPPsA-DR with drug resistant P. aeruginosa (+28.6%, 95% CI +14.0% to +45.0%). Among the skin flora, CPMSSA was associated with a +12.1% (95% CI +9.9% to +14.4%) increase in the odds of nosocomial acquisition of MSSA, CPMRSA was associated with a +6.7% (95% CI +1.3% to +12.5%) increase in the odds of MRSA. Negative associations were observed between organisms inhabiting different niches, including CPMSSA and ESBL K. pneumoniae (-7.9%, 95% CI -15.1% to -0.2%), and CPPsA-DR and vancomycin susceptible E. faecalis (-10.0%, 95% CI -15.6% to -4.0%). InterpretationA hospitalized patients odds of nosocomially acquiring a potential pathogen is associated with its prevalence among that patients ward co-occupants, regardless of the organisms drug resistance profile. Further research is necessary to understand the role of passive surveillance of CP for preventing infection. FundingLWS was supported by the NLM (2T15LM007092-31). ZW was supported with institutional funding from the Department of Population Medicine. TRP was supported by AHRQ (K08-HS030118). SK was supported by AHRQ (grant no. K08 HS027841-01A1).