Surveillance of surgical-site infections: the world coming together?

Surgical-site infections (SSIs) are associated with substantial risk of morbidity, resulting in prolongation of hospital stay and monetary costs attributable to the infection. Reporting surgeon-specific SSI rates back to the surgeon has been shown in a number of studies to reduce SSI rates significantly, with appropriate cautions." Because of this, several countries have recently examined and reported their SSI experience." This issue of the Journal brings another article for consideration. This article describes the first year of the Dutch national surveillance of SSIs and risk factors. The results, while preliminary, show many of the same findings seen in other studies: • There are multiple risk factors for SSIs. The Dutch study found age, preoperative length of stay, wound class, anesthesia score, and duration of surgery to be independently important risk factors for SSIs when all procedures were pooled together for analysis. • Risk factors vary by procedure. The nature and relative importance of the risk factors differed when individual procedures were examined in the Dutch PREZIES (Preventie van Ziekenhuisinfecties door Surveillance [Prevention of Nosocomial Infections Through Surveillance]) Project. • SSIs are very costly to the healthcare system. The Dutch study found an attributable increased length of stay of 8.2 days for patients with SSIs. In any nation, an additional week or so of hospitalization is expensive. It should not be surprising that so many countries are attempting to monitor this infection. However, the efforts of Geubbels and colleagues point out the difficulties with which all countries struggle in doing so. These include monitoring a national representation based upon a probability sampling of procedures; cooperation of hospitals and, in particular, surgeons; adequate sample size for certain surgical procedures; accuracy of data, especially with regard to postdischarge surveillance; and, ultimately, validity of comparisons with other SSI rates. Despite the difficulties, the Dutch study specifically points out the feasibility of such a surveillance system. The study also confirmed that many of the risk factors from the Basic SSI Risk Index used in the Centers for Disease Control and Prevention (CDC)'s National Nosocomial Infections Surveillance (NNIS) System were risk factors in The Netherlands. Further, there was close correlation of the 75th percentile of duration of surgery between the Dutch system and the NNIS system for similar group procedures. In general, if differences existed, they were minor, with the NNIS durations tending to be longer (Table). The differences may be due simply to smaller sample size and thus poorer estimates for the 75th percentiles in the PREZIES Network compared to the NNIS System; the distributions of the specific ICD-9 codes may have been different for the procedures (which are grouped codes in both systems); the actual measure of duration of surgery may have differed slightly between The Netherlands and the United States; or Dutch surgeons may have been slightly faster at operating. The reasons for the differences in duration of surgery are not any clearer than the reasons for differences in the SSI rates. Thus, even for the simplest of variables, duration of surgery, comparisons across surveillance systems are difficult; only where enormous differences occur can conclusions be drawn. For example, a comparison of surveillance systems for SSIs in Belgium and The Netherlands found an extraordinary difference in the use of antimicrobial prophylaxis for herniorrhaphies (3.7% in The Netherlands vs 41.9% in Belgium). This observation led directly to closer examination of antimicrobial prophylaxis in Belgium.