Appropriate empiric antibiotic choices in health care associated urinary tract infections in urology departments in Europe from 2006 to 2015: A Bayesian analytical approach applied in a surveillance study

BACKGROUND: Health care associated urinary tract infections (HAUTI) is a common complicating factor of urological practice. It is unclear what the appropriate empirical antibiotic choices are and how infection control policies (ICP) influence this. The aim of this study is to use probabilistic approaches towards the problem. That is to determine the chances of coverage of empirical antibiotic choices in HAUTIs and their annual trends in Europe. In addition, the impact of departmental self-reported compliance with catheter management and regulated usage of prophylactic antibiotics policies was tested. The estimated chances of coverage of antibiotics and further probabilistic calculations are carried out using the Global Prevalence of Infections in Urology (GPIU) annual surveillance study European data. METHODS: GPIU is a multi-state annual prevalence study conducted in urology departments to detect patients with HAUTIs, using the Center for Disease Control (CDC) definitions and antimicrobial resistance (AMR). In this analysis; the European cohort from 2005 to 2015 was used. The estimated chance of coverage for each antibiotic choice in HAUTIs was calculated using the Bayesian Weighted Incidence Syndromic Antibiogram (WISCA) approach. Annual trend of the overall cohort and number of appropriate antibiotic choices were estimated. Departments were compared according to their self-reported compliance to ICPs to determine if there was an impact on chances of coverage and appropriate antibiotic choices. RESULTS: We estimated that in most study years less than half of the single agent antibiotics and all combination options were appropriate for empirical treatment of HAUTIs. Departments with compliance to both ICPs were estimated to have 66%(2006) to 44% (2015) more antibiotic choices compared to departments with complete lack of compliance to the ICPs. In our estimates departments with adherence to a single policy was not superior to departments with complete lack of adherence to ICPs. CONCLUSIONS: Most single agent choices had limited coverage for HAUTIs and combination choices had improved chance of coverage. Optimal antibiotic selection decision should be part of decision experiments and tested in local surveillance studies. Departments with self-reported compliance to ICPs have more antibiotic choices and details of the compliance should be evaluated in future studies. The analysis herein showed that over the 10-year course there was no clear time trend in the chances of coverage of antibiotics (Bayesian WISCA) in European urology departments.