Abstract
To compare surgical site infection (SSI), the risk-adjusted index is the standardized infection ratio (SIR). A statistical test for SIR is proposed using software provided by the Centers for Disease Control and Prevention (CDC). Comparison of observed SSI number (SSIO) and expected SSI number (SSIE) uses the Risk Index Category for adjustment. In the formula SIR=SSIO/SSIE, SSIE is obtained as the total sum of SSI probabilities for the respective category stratification. The test ratio approximates to the normal distribution by z-transformation. If SSIE≧5, the z-test is indicated. If SSIE<5, either the binomial test or the Poisson test is recommended using, for example, Epi InfoTM 6 provided by CDC. Multiple tests repeated for the category stratifications may lead to the Type I error, which the test on SIR would offset.
