Administrative data quantify hospital infection risk
medwireNews: Administrative hospital data can successfully quantify the incidence of healthcare-associated urinary tract infection (HCA UTI), indicating the data's potential use for risk monitoring and surveillance of this frequent occurrence, show systematic review results.
Of the 96 significant risk factors for HCA UTI identified in the 23 studies reviewed, 65% were also present in the local administrative databases available, the researchers report in the Journal of Hospital Infection.
"The next steps are to investigate the utility and quality of the codes identified by this review in local data," they suggest, noting that, despite clear guidance in the UK, poor-quality coding and issues with variability in coders' skill and experience still exist.
Caroline King, from Imperial College London, UK, and colleagues found a total of 143 risk factors for HCA UTI in the 23 studies, which all reported odds ratios, relative risks, or hazard ratios for such infections among patients admitted to hospital for at least 72 hours.
A total of 96 factors - identified by at least one study - were significant and 43 were independent risk factors for HCA UTI.
In order to examine the importance of these risk factors, the team calculated the population-attributable risk percentages (PAR%) for HCA UTI, which were available for 35 significant risk factors identified by 18 of the studies.
"If the risk factor was removed from the population, the infection incidence would be reduced by PAR%," explain the authors.
The highest PAR% was for urinary catheterization, at 79.3%, followed by transurethral catheterization, at 52.4%, positive meatal culture, at 36.9%, and an American Society of Anesthesiologists' score of more than 2, at 32.7%.
In all, 28 (65%) of the 96 significant risk factors were located in available local hospital administration databases, and 100% of the 14 risk factors identified by multiple studies as independent predictors of HCA UTI (apart from duration of urinary catheterization) were found in local hospital data. These factors included age, female gender, Type 2 diabetes, prior stroke, length of hospital stay, and urinary catheter.
"Electronic surveillance [of the contents of hospital data systems] represents a cost-effective, time-efficient and robust approach to surveillance, although it is subjective to clinical opinion and a changing policy landscape," conclude King and co-authors.
medwireNews (www.medwire-news.md) is an independent clinical news service provided by Springer Healthcare Limited. © Springer Healthcare Ltd; 2012
By Sarah Guy, medwireNews Reporter