Tjallie van der Kooi

decreased from 8.2 to 6.3 days and the use per patient remained approximately the same.²¹,²² Because of this development in hospital stay, fewer SSIs are found when no postdischarge surveillance is carried out. For the other infection types, a shorter hospital stay results in a weaker and more dependent hospital population – needing more devices which leads to more NIs. It is therefore possible that the prevalence of NI and of device use has changed over time, without a real change in risk on a per patient basis or vice versa. A difference in case‐mix over the course of time or between institutions or countries could partly be overcome by including a comorbidity index. Other studies showed comorbidity indices to be signicantly associated with the risk of acquiring an NI.⁴,²⁰ We have not recorded comorbidities, which hampers case‐mix adjustment. The higher NI prevalence for increasing length of stay (until the survey day) is most likely to be at least partly a consequence of sicker patients and/or those with a nosocomial infection staying in hospital longer. In the reference data fed back to the hospitals and published on the PREZIES website, case‐mix adjustment is attained by stratication. Prevalence surveys are valuable because they can give an overall picture, including all NI types and, with that, estimate the total burden of NI. A drawback of prevalence surveys in general is the overestimation of NI occurrence because more vulnerable patients and those with an NI stay in hospital longer and are thus ‘oversampled’. However, although these data do not present the actual risk per patient of developing an NI, they are valuable for creating insight into high risk populations. Hospitals that, in their feedback report, detect higher than average infection prevalences in certain populations, can decide to implement interventions or, when causes are unclear, choose to analyse this patient population further. As an option to the prevalence survey, hospitals could start recording more detailed data on antibiotic use and/or urinary catheter use and many hospitals have now done so. In addition, the majority of Dutch hospitals are now focusing on the reduction of SSI and CR‐BSI, within a national patient safety programme. The large interhospital range in device use and prevalence of infections suggests that improvements are possible. This study presents the rst national estimates of the prevalence of NI in hospitals in The Netherlands. The prevalences stated are the best available estimate of the prevalence of NIs at national level. The use of devices and antibiotics as well as the prevalence of NI appeared to be comparable to that in other European countries. The results provide insight into high risk patients and could lead to the start of in‐depth investigations and/or targeted interventions for reducing the burden of NIs. ACKNOWLEDGEMENTS The authors wish to thank all infection control practitioners, microbiologists, and other 5 99 Prevalence of nosocomial infections: first four national studies

RkJQdWJsaXNoZXIy MTk4NDMw