The use of clinical coding data for the surveillance of healthcare-associated urinary tract infections in Australia

• There are a number of limitations to using administrative data for healthcare associated urinary tract infection surveillance.
• Clinical coding data for identifying cases of healthcare associated urinary tract infection is very unreliable.
• Coding data alone should not be relied upon for the surveillance of healthcare urinary tract infections.

IntroductionGiven the trends in antimicrobial resistance, particularly for Gram-negative organisms, the surveillance of urinary tract infections (UTIs) has the potential to become increasingly important in the future. Whilst considering accuracy and efficiency, we undertook a cohort study in a large Australian health district to inform future discussions around surveillance approaches to healthcare-associated UTIs (HAUTI).MethodsA retrospective cohort study in eight hospitals was conducted to examine the clinical coding data of all patients hospitalised for more than two days over a four-and-half-year period. These data were compared to a conservative laboratory-based HAUTI definition.ResultsThe data from 162,503 patient admissions were examined. During the study period, 2821 of the admitted patients acquired a HAUTI. Of those patients identified as having a laboratory-diagnosed HAUTI, 29.3% had a clinical code relating to a UTI.ConclusionThe clinical coding data used to identify cases of HAUTI is very unreliable as a significant proportion of cases were not identified. To ensure the efficient and effective use of resources, a range of approaches should be considered in the event of HAUTI surveillance being required.