Detection of Temporal Clusters of Healthcare-Associated Infections or Colonizations with Pseudomonas aeruginosa in Two Hospitals: Comparison of SaTScan and WHONET Software Packages.

Abstract

The identification of temporal clusters of healthcare-associated colonizations or infections is a challenge in infection control. WHONET software is available to achieve these objectives using laboratory databases of hospitals but it has never been compared with SaTScan regarding its detection performance. This study provided the opportunity to evaluate the performance of WHONET software in comparison with SaTScan software as a reference to detect clusters of Pseudomonas aeruginosa. A retrospective study was conducted in two French university hospitals. Cases of P. aeruginosa colonizations or infections occurring between 1st January 2005 and 30th April 2014 in the first hospital were analyzed overall and by medical ward/care unit. Poisson temporal and space-time permutation models were used. Analyses were repeated for the second hospital on data from 1st July 2007 to 31st December 2013 to validate WHONET software (in comparison with SaTScan) in another setting. During the study period, 3,946 isolates of P. aeruginosa were recovered from 2,996 patients in the first hospital. The incidence rate was 89.8 per 100,000 patient-days (95% CI [87.0; 92.6]). Several clusters were observed overall and at the unit level and some of these were detected whatever the method used. WHONET results were consistent with the analyses that took patient-days and temporal trends into account in both hospitals. Because it is more flexible and easier to use than SaTScan, WHONET software seems to be a useful tool for the prospective surveillance of hospital data although it does not take populations at risk into account.