884 Surgical Site Infections in Five Brazilian Hospitals: Decennial Trends

Sunday, March 21, 2010
Grand Hall (Hyatt Regency Atlanta)
Edna M. Meireles, MD , Hospital Risoleta Tolentino Neves and Hospital da Baleia, Belo Horizonte, Brazil
Carlos E.F. Starling, MD , Hospital Vera Cruz, Belo Horizonte, Brazil
Estevão U. Silva, MD , Hospital Universitário São José and Hospital Vila da Serra, Belo Horizonte, Brazil
Braulio R.G.M. Couto, MSc , Centro Universitário de Belo Horizonte / UNI-BH, Belo Horizonte, Brazil
Background: Surgical site infection(SSI) is one of the three most frequent Nosocomial infectios in the majority  of publications in this field. The traditional NNIS risk index(1991) is calculated based on the wound class, duration of operation, and American Society of Anesthesiology (ASA) score.

Objective: to evaluate the NNIS risk index and the incidence of SSI infection collected in a period of time that varies from 12 months to fifteen years (jan./94 to dec./08) in five Brazilian hospitals. To identify advantages and main difficulties to use NNIS surgical patient surveillance component in Brazilian hospitals.

Methods: prospective surveillance of SSI according to the protocol of NNIS/CDC - SURGICAL COMPONENT. The concepts of CDC have been used to characterize patients with NI.

Results: we evaluated 150,370 surgical procedures in 5 hospitals. During all period analyzed we diagnosed 8,899 NIs and 3,847 SSI (2.6%). SSI rate according to the NNIS risk index category: risk index 0 = 1.4% (537/37,988); risk index 1 = 4.3% (696/16,196); risk index 2 = 7.8% (235/3016); risk index 3 = 12.4% (25/202). The main advantage for  using NNIS surgical component is the ability to calculate SSI for each professional (surgeon, anesthetist etc) and for specific operative procedure. We calculated SSI infection, global and stratified for ASA score, wound class, duration of surgery and NNIS risk category for all operative procedure defined by the NNIS system. The most difficult is to collect all data from each NNIS surgical patient. For example, we had a high missing data for the ASA score (25%), wound class (33%) and duration of surgery (52%). These missing data varied greatly from hospital to hospital, but it was impossible to obtain the NNIS risk index category for 62% of all procedures analyzed.  This problem was solved in two hospitals, where NNIS risk category missing reduced to almost 0% after 2007. Conclusions: the NNIS surgical patient surveillance component enabled us to calculate SSI rates for each professional and to specific surgical procedures in the hospitals enrolled. The main difficult is to obtain all necessary data. However, with strict involvement of the Hospital Direction, this difficult can be overcome.