MDRO pose a challenge to hospitals, adversely impacting patient outcomes and resulting in increased healthcare costs (15). Active patient screening allows hospitals to identify MDRO colonised patients and potentially implement controls such as isolation to reduce the risk of transmission to other patients, but there is little evidence on the optimal screening strategy in contexts where universal screening and/or isolation is not feasible (16–20). Here we demonstrate that active patient screening detects additional MDRO colonised patients that would otherwise have not been identified, with incremental gains for additional patients screened.

Previous studies have reported variable success with screening, with MDRO detection rates ranging between 27–87%(21,22). In our hospital, asymptomatic screening identified 465 of the 628 (74%) MDRO colonised or infected patients while the remainder were identified after developing clinical infections. Although impossible to know how many transmission events and additional MDRO cases were averted through early detection and isolation, the absence of asymptomatic screening would have missed detecting the majority of MDRO cases resulting in significantly more opportunities for transmission. The comparison between “standard risk” wards (asymptomatic screening not routinely undertaken) and “high-consequence” wards (patients screened weekly and isolated if MDRO are detected) is also notable. The MDRO incidence rate on “standard risk” wards was nearly three times the incidence on “high-consequence” wards, and had an associated transmission rate 14 times the rate on “high-consequence” wards, suggesting occult transmission of MDRO occurs frequently in the absence of targeted control measures such as case identification and isolation, and that those isolation measures including contact precautions were effective in reducing transmission.

Although several studies suggest that rates of MDRO-related healthcare-associated infection remain unaffected after discontinuing isolation practices for some MDRO such as VRE (23–25), there are are clear differences between MDRO and among different patient cohorts in their propensity for invasive infection following colonisation (26). The justification for screening and isolation in “high-consequence” patient cohorts (higher risk of invasive disease due to MDRO), and/or “high-prevalence” patient cohorts (higher risk of MDRO of significance such as carbapenemase-producing organisms and Candidozyma auris) in our institution follows this rationale. Furthermore, assessing the impact of controls such as screening and contact precautions with MDRO clinical infections as an outcome uses an indirect outcome measure, with multiple confounders influencing the transition from colonisation to clinical infection, and negates the ability to accurately measure the impact of interventions on preventing MDRO transmission.

Whole-genome sequencing (WGS) of bacterial isolates has been shown to enhance surveillance and outbreak investigations for MDRO (7,11) and is increasingly used for hospital infection prevention and control, but is reliant upon samples from identified cases (27,28). As shown in this study, sole reliance upon isolates from invasive infections (e.g. blood cultures) will miss the majority of cases required to infer transmission pathways to control outbreaks. Similarly, based on our data, an absence of MDRO transmission after discontinuing screening practices may actually represent missed detection of uncontrolled transmission. Despite the resource requirements for prospective WGS, recent studies have projected significant cost savings through integrated WGS surveillance (29). However as shown in our study, these cost savings can only be achieved through sufficient screening practices to identify MDRO cases and transmission events.

Some limitations should be noted. The data are derived from a cohort of patients in a single healthcare network which may limit extrapolation to other healthcare settings. However, comparing transmission is difficult without adjusting for confounding differences between healthcare facilities in screening and isolation practices and hospital infrastructure – factors difficult to control without large cluster-randomised crossover trials which are impractical for accurately. While previous studies have attempted to estimate broader impacts on population-level MDRO incidence (30), we focused on a smaller dataset with high-specificity definitions for transmission requiring both genomic and epidemiological criteria. Our reported rates of MDRO and transmission incidence are likely to underestimate the true rates due to imperfect sensitivity of MDRO screening and omission of cases detected outside our healthcare network, though these factors are likely to have been applied consistently across the study. Our data were limited to three indicator MDRO for pragmatic reasons, though we would anticipate similar findings for other multidrug-resistant Enterobacterales and enterococci. Additionally, our data report MDRO incidence in a setting where asymptomatic screening occurs regularly in many areas of the hospital and detected cases are isolated – actual numbers of MDRO cases may have been higher than in our scenarios if screening and isolation did not occur.

Our study provides real-world data to guide the relative gains from different screening approaches when implemented with isolation practices. Ultimately, the decision to employ a screening and isolation approach to MDRO will be influenced by the cost-benefit equation with different levels of benefit for different MDRO threats.