To the Editor.
—The articles by Dr Liu and colleagues1 and Dr Edmonson and colleagues2 provide new and better insight into whether early discharge of newborns affects readmission rate. The accompanying Commentary by Dr Braveman and colleagues3 cast these papers in the context of "evidence-based practice—good science, good judgment." We would like to extend the discussion from an evidence-based medicine perspective. We agree with the careful critique of Braveman et al that the findings, while, at first blush, appear contradictory, may not be disparate. We especially agree that the findings of "negative" studies be stated in terms of power to detect differences.In the article by Liu et al,1 a large number of cases were excluded because of the inability to classify them as either early or late discharges. This intermediate group, using contrast variables, could have been retained in the logistic regression models. This would