To account for time factors related to hospitalization, the authors calculated incidence rates of surgical site infection (SSI) instead of cumulative incidence and assessed risk factors for SSI after cardiac surgery. From July 1999 to August 2000, all cardiac surgery patients in the Taipei Veterans General Hospital (Taipei, Taiwan) were invited to join the study. Data were collected by chart review, including information on patient characteristics and potential risk factors at the preoperative, intraoperative, and postoperative stages. The authors employed multiple logistic regression analyses using a generalized logit model to assess associations of interest. SSI incidence rates were 2.5 episodes per 1,000 person-days for the sternum and 3.6 episodes per 1,000 person-days for the leg. After adjustment for covariates, age (in years), gender (female vs. male), New York Heart Association classification (moderate/severe vs. normal/mild), creatinine concentration (mg/dl), and duration of surgery (in minutes) were significantly associated with sternal SSI, while age, peripheral arterial occlusive disease, and length of stay in the intensive care unit (in hours) were significantly associated with leg SSI. In addition to patients' characteristics and health situations, the significant findings for duration of surgery and length of intensive care unit stay indicate that the incidence rate is more appropriate than cumulative incidence for studying risk factors for SSI.