Nosocomial infection surveillance in Belgian ICUs: aim and methodology of the feedback

Introduction

The risk for nosocomial infecions (NI) in ICU's is five to ten times higher than in the other hospital care units, with increasing mortality, morbidity, length of stay and costs. Previous studies demonstrated a significant decrease in NI after the set up of effective surveillance programs. If the surveillance is performed in a multicentric setting with representative sample, the results can be dysplayed as a whole but also aggregated for each individual ICU (IICU). The production of meaningful feedback reports may help IICU to evaluate its own situation as compared to that of the whole studied group (WS).

Objectives

The aim of the study was to provide IICU with its own incidence of Nosocomial Pneumonia (NPN) and Blood Stream Infection (NBSI), and their mortality rate after controlling for the risk factors. The risk factors (RF) are the case mix at admission (severity of illness = SAPSII score and the pathologies), the exposition to devices at risk (mechanical ventilation and central venous catheters) and the length of ICU stay. These 3 last RF are counfounding variables depending on the intensisty of ICU resources' use.

Material and methods

At the end of 1995, all Belgian ICUs were invited to join a voluntary prospective NI surveillance network which was based on the criteria defined in the HELICS project. Participation was for one are two 3-months periods, starting in January 1996. During the first two trimesters of 1996, 64 different ICU's, from 28% of the Belgian acute care hospitals, have joined the study. A total of 8475 patients were observed. Data included: (i) patient characteristics at admission: administrative data, type of admission, SAPSII score, prior surgery or antibiotics, impaired immunity and infection at entry; (ii) during ICU stay, daily exposure to mechanical ventilation, central venous catheters and treatments; (iii) if a NPN and/or a NBSI occured during ICU stay, diagnostic criteria and microbiological data; (iv) discharge data: vital status and date.

Results

Two kinds of feedbacks were produced for IICU with results grouped by trimester: descriptive and analytical. The descriptive feedbacks reported firstly, the observed results of the WS and IICU concerning the case mix at admission, the intensity of devices use, length of ICU stay, mortality and NI rates and secondly, the distribution of those variables throughout the WS as well as the localisation of IICU, expressed in percentile, within each distribution. The analytical feedbacks produced specific standardised rates (SR) for the intensity of devices use and length of stay weighted for the case mix, and the SR for mortality and NI rates weighted for the ICU resources use and the case mix. Moreover, those SR may vary widely as a consequence of hazard, specially for small ICUs, so that confidence interval (CI) must be calculated for IICU specific SR.