Tjallie van der Kooi

as well as device use (Dutch PPS data, not shown) so HAI rates can be expected to increase. As demographic distribution differs between e.g. regions[50] this might not affect all hospitals and regions at the same rate. Understanding trends requires insight into these developments. The present possibilities in information technology has facilitated surveillance and will continue to do so. Part of the patient and procedure data are nowadays conveniently retrieved from the electronic hospital data. Reduced infection rates in combination with the ongoing digitalisation of patient documentation and growing possibilities to retrieve data from electronic patient records has led to an increased interest in semi‐automation of HAI surveillance, in both individual hospitals and the national surveillance programmes [51, 52]. Automated surveillance including automated identification of HAI or of patients at high risk of HAI reduces the workload required for surveillance and may increase inter‐ rater reliability in some settings [52]. Automated SSI surveillance in knee and hip arthroplasty is now being introduced in the Dutch surveillance programme [53, 54]. Likewise, with CRBSI, reduced rates have led to an increased interest in automation [55, 56], but in many hospitals the required data are not yet fully recorded in the electronic patient records. In search of other options to reduce the workload of surveillance in the Dutch network, possibilities to simplify the CRBSI surveillance have been evaluated. Analysis of recent data demonstrate that most hospitals rate comparably when evaluating CRBSI rates per 1000 CVC days or as percentages of CVCs (not shown). With the apparent low rates, some of the risk factors can be omitted without jeopardising case‐mix adjustment. In 2023 the CRBSI surveillance protocol was changed accordingly. Although reducing the time needed to perform CRBSI surveillance might convince some hospitals to start or continue participation, it may be more promising to widen the surveillance scope, leading to a larger numerator and at the same time furthering automated data collection. An outcome that might meet these demands is hospital‐acquired bloodstream infection, often called hospital‐onset bloodstream infection (HOB), though they are not strictly the same. HOBs are microbiologically‐confirmed BSIs that are associated with hospital stay (usually defined as developing ≥ 48 hours after admission). Microbiologically‐confirmed BSIs can easily be retrieved from electronic databases and are less subject to debate. Based on larger numbers of infectious episodes, the results also allow more discrimination between hospitals [57]. The major problem is to balance clinical relevance and interpretability, for example by selecting primary bloodstream infection only which implies determining the focus of infection, against practical feasibility. Studies and surveillance programmes have made various choices in this regard [55, 57‐61]. As the 10 243 General Discussion

RkJQdWJsaXNoZXIy MTk4NDMw