Statistical Methods
Cohort characteristics were compared using Pearson chi-square (or Fisher’s exact test where cell counts were small) for categorical data, presented as frequencies and percentages. Continuous data, presented as means with standard deviations, were compared using analysis of variance or Wilcoxon rank sum test. Chi-square was used to compare proportion at goal on first ACT draw. T-tests were used to analyze weight bolus and time outcomes. All statistical analyses were conducted using SAS for Windows 9.4 (Cary, NC). For the analysis of therapeutic and or supra-therapeutic proportions, we accounted for variable number of ACT draws per procedure. The number of ACT draws collected ranged from 1 to 14, and consequently the number of therapeutic ACT draws was also variable. Logistic regression was utilized to analyze the average success rates between the usual care and protocol guided cohorts, taking into account the variable number of draws. The odds ratio represents the increase in odds of achieving a value > 300 and > 400 seconds respectively. For those without pre-procedure anticoagulation, the weight-based bolus was plotted against first ACT result. We used a censored regression analysis to account for times recorded at the Hemochron maximum value of 400 seconds, but were likely higher.