In reply
We appreciate the comments of Drs Yachimski and Poulose concerning our article. We did indeed perform a priori power sample size calculations in the design of our study. Our protocol that was used continuously throughout the study proposed 4 separate primary end points: efficacy of stone removal, length of hospital stay, cost of index hospitalization, and morbidity/mortality of the index procedure. For sample size calculation, we chose 1 of these, ie, length of stay. We estimated a priori that only 50% of the patients in the ERCP/S group would be discharged within 2 days, whereas 85% of the patients in the LCBDE group would be so discharged. We calculated in our protocol that to achieve an α of 0.05 (2-tailed) and a β of .10, we would need a sample size of 100 patients per group. We reasoned that, given the volume of laparoscopic cholecystectomies performed in our institution, we would be able to randomize these patients within a few years. Our study continued for many more years than we had anticipated. We therefore terminated the study after 6 years, as stated in the article. In the discussion section of the article, we do mention the possibility of a type II error in our study, considering our sample of 112 patients. In retrospectively calculating sample size based on the numbers achieved, if we accept a 98% stone clearance in the ERCP/S and an 88% clearance in the LCBDE groups with an α error of 5% and a β error level of 10%, the sample size needed would be 110 for each group.