Quantifying excess length of postoperative stay attributable to infections: a comparison of methods.

To quantify the net effect of deep surgical site infection (DSSI) on postoperative stay (POS) among patients who had undergone open heart surgery, and to assess the comparability of two methods, two observational studies were conducted: one on a retrospective cohort of 701 operated patients, and the other on a cohort of 31 infected patients versus a cohort of uninfected patients, with 1:1 matching. In addition to DSSI, a further three factors were identified by multivariate analysis as independent POS-related predictor variables. After internal validation of the multivariate model, excess POS attributable to DSSI amounted to 20.7 days (95% confidence interval [CI] 16.7-24.9). In contrast, excess length of stay attributable to DSSI among the matched pairs who survived infection (22) totaled 14.3 days (95% CI 3.2-25.4) and 26.5 days (mean and median differences). Multivariate techniques may prove a more appropriate and reliable analysis than matched-pair comparisons for the purpose of evaluating the extra stay and cost attributable to the nosocomial infections.