The accuracy of fully automated algorithms for surveillance of healthcare-associated urinary tract infections in hospitalized patients

Introduction

Healthcare-associated infections (HCAIs) annually affect millions of patients, are a major burden for the healthcare system, and are associated with prolonged hospital stay, increased morbidity, mortality, and costs [1e3]. Healthcareassociated urinary tract infections (HA-UTIs) account for nearly 20%ofallHCAIs,affecting nearly870,000 patients yearly in Europe [3]. A significant proportion of HCAIs can be prevented [1]. Therefore, to allocate necessary resources and evaluate the effect of interventions, continuous surveillance with feedback to healthcare personnel and stakeholders is important [4,5]. Much HCAI surveillance is currently based on time-consuming and resource-intensive manual review of patient records, which is also prone to subjective interpretation and surveillance bias [6e8]. With the use of electronic health records (EHRs), there is increasing access to detailed electronic health data. This digitalization allows automated surveillance systems to replace manual approaches and to generate standardized and continuous surveillance data [9]. However, surveillance algorithms needtobethoroughly validated before being implemented in a clinical setting. In this study, the aim was to develop a fully automated rulebasedsurveillancealgorithmusingEHRdataforthedetectionof HA-UTI in hospitalized patients, and validate it against manual record review according to the HA-UTI definitions of, primarily, the EuropeanCentre for Disease Prevention and Control (ECDC) and, secondly, the US Centers for Disease Control and Prevention (CDC). To demonstrate a possible use-case, the bestperforming algorithm was used to determine HA-UTI incidence during a three-year period in all hospitalized patients.

Read Complete Report Here.