A rise in bloodstream infections on a hospital ward can signal a need to review infection prevention practices. Research by UMC Utrecht PhD candidate Manon Brekelmans shows how routinely collected clinical data can reveal such patterns automatically. Her thesis also identifies what hospitals need to achieve automated surveillance and present understandable feedback information to support improvements in patient safety.
Hospital acquired infections (HAI) are defined as infections that occur while receiving healthcare, in hospitals or other healthcare facilities. HAI are seen as the most frequent adverse event during healthcare delivery. In Europe, an estimated 8 percent of hospitalized patients acquire at least one HAI. HAI can serve as a quality indicator as half of all HAI are estimated to be preventable by infection prevention and control programs, including surveillance of HAI.
Monitoring of HAI often requires infection prevention professionals to review patient records manually. This is labor-intensive, time-consuming and can produce inconsistent results. PhD candidate Manon Brekelmans, MSc (Department of Medical Microbiology and Infection Control, UMC Utrecht and Department of Epidemiology and Surveillance, National Institute for Public Health and the Environment (RIVM), Bilthoven) investigated how electronic health records and laboratory data could support a more standardized and automated surveillance, focusing on bloodstream infections and surgical site infections, with particular attention to validated, scalable and interoperable surveillance solutions that enable large-scale implementation.
Manon Brekelmans, MSc
Cover PhD thesis
In a study in four Dutch hospitals, it was feasible to apply a fully automated algorithm that identified hospital-onset bacterial or fungal bloodstream infections, starting at least two days after admission. Hospital-wide rates ranged from 1.0 to 1.9 episodes per 1,000 patient-days, as compared to 8.2–12.5 episodes per 1,000 patient-days in intensive care units. These rates are within the same range as reported in previous studies.
The system revealed changes over time and differences between wards and microorganisms. Such signals could prompt infection prevention teams to investigate potential sources and review practices such as catheter or wound care. Surveillance alone, however, cannot establish whether or not an infection would have been preventable.
Manon Brekelmans also highlighted that infection rates cannot be compared fairly without accounting for differences in patient populations, care complexity and blood-culture practices. She explored risk-adjustment models and visual tools, including funnel plots, to help distinguish unusual infection patterns from expected variation. These approaches offer a starting point for more informative benchmarking, but they require further refinement.
The PhD thesis also explored how automated surveillance could be scaled up. Existing data from the RIVM’s national antimicrobial resistance surveillance system (ISIS-AR) could help identify hospital-onset bloodstream infections caused by recognized pathogens with little additional workload from hospitals. More complete admission-date reporting and careful interpretation of local laboratory practices are essential.
In a separate study evaluating the implementation of automated surveillance demonstrated that in only one out of five participating hospitals semi-automated surveillance for surgical site infections after hip or knee replacement was implemented within the study period. Barriers for timely implementation included limited IT capacity, complex data extraction and unclarity in organization of local project management. According to Brekelmans, an improved implementation strategy has since supported nine successful implementations, with more than 20 hospitals working towards adoption.
Manon Brekelmans on the implications of her research: “My practical message is that automated surveillance of HAI needs more than an algorithm. Standardized data, clinical validation, clear project leadership, IT support and collaboration between hospitals are all critical. The next step is to determine whether better surveillance actually leads to fewer healthcare-associated infections.”
Manon Brekelmans, PhD (1998, ‘s Hertogenbosch) defended her PhD thesis on November 3, 2026 at Utrecht University. The thesis is titled “Automated surveillance of healthcare associated infections – Towards large-scale implementation.” Supervisors were Prof. Jan Kluijtmans, MD PhD (Department of medical Microbiology, UMC Utrecht) and Prof. Marc Bonten, MD PhD (Research Program Epidemiology of Infectious Diseases, UMC Utrecht). Co-supervisors were Maaike van Mourik, MD PhD (Department of Medical Microbiology, UMC Utrecht) and Stephanie van Rooden, PhD (Department of Epidemiology and Surveillance, National Institute for Public Health and the Environment, Bilthoven). Manon will continue her career as an infection control practitioner in training in the Elisabeth-Tweesteden Hospital in Tilburg.
