Figure 5. Distribution of patients by postconceptual age (weeks) and by logistic regression model. Model 1 was based on the entire patient population, model 2 excluded all anemic patients, and model 3 excluded all anemic patients and those who had an episode of apnea in recovery room. The number of patients at each of the extremes of age is small explaining in part the necessity to extrapolate based on our assumptions. This also shows why conclusions reached at one institution or by a single study would be more subject to error given the small number of patients collected over so many years from four institutions.