Does pretreatment microbiological testing translate into lower multidrug-resistant organism infection risk? A patient-level target-trial emulation of hospitalization episodes

Objectives To assess whether pretreatment microbiological testing is associated with subsequent multidrug-resistant organism (MDRO) hospital-acquired infection. Methods We conducted a retrospective target-trial emulation using linked hospital data from a tertiary hospital in China (May 2018 to December 2025). Episodes were classified by whether qualifying microbiological testing had been recorded by the first antimicrobial use. The primary outcome was incident postindex MDRO hospital-acquired infection. Propensity-score overlap weighting estimated risk differences per 10,000 episodes and risk ratios in the overlap population. Results Among 859,135 episodes, 234,100 (27.2%) had qualifying testing recorded by index; 1029 events occurred. After overlap weighting, risks were 11.29 vs 11.27 per 10,000 episodes, with a risk difference of 0.02 (95% CI, −1.63 to 1.67) and a risk ratio of 1.002 (95% CI, 0.866-1.159). Average treatment effect in the overlap population-aligned doubly robust and prespecified sensitivity analyses were broadly consistent with the primary estimate. A post hoc analysis restricted to infection-directed episodes similarly showed no protective association. Conclusion In the primary overlap population, qualifying microbiological testing recorded by antimicrobial initiation had a near-null association with subsequent MDRO hospital-acquired infection. This finding concerns an index-time diagnostic process rather than downstream result-guided stewardship action.