Purpose: Previous attempts to define the learning curve for HALC have used a method in which the operative experience is partitioned into groups of cases and then operative (OR) time and morbidity are compared. This simple method may not accurately describe the learning curve for HALC. We hypothesized that there are non-linear relationships between operative experience, OR time, and quality-related short-term outcomes for HALC. We used a validated statistical tool, the Sigmoid Emax model, to evaluate these relationships. The results of this model were compared to those of a simple partition of the HALC experience into groups based on operative experience.Methods: A prospective HALC database was used to analyze consecutive segmental and total HALC performed by a single surgeon with no prior HALC experience. Using sigmoid Emax models, the relationship between the HALC case number, OR time, conversion rate, morbidity, length of hospital stay (LOS), readmission, and reoperation were demonstrated. Total improvement capability (TIC), defined as significant improvement between the initial and final average values of the measured variables, were calculated. The number of HALC cases required to achieve 90% of the TIC was accepted as the learning curve. A separate comparison of the same variables, in consecutive groups of 25 or 50 HALC, was also performed.Results: From December 2005 to August 2008, 187 HALC were performed, including 61 right, 61 sigmoid, and 65 total colectomies. The indication for HALC was neoplasia, diverticulitis, and IBD in 56%, 25%, and 18%, respectively. 90% TIC was achieved for OR time only and occurred at 32 cases. OR time decreased from a mean of 191±73 to 165±57 minutes before and after 90% TIC was achieved. Sigmoid Emax models for conversion (6%), operative and postoperative morbidity(8 and 30%), median LOS 4 (2-32) days, readmission (7%), and reoperation (5%) produced plateau-like curves and 90% TIC was not achieved which indicates that these measures remained similar throughout the HALC experience. Comparison of the same measures in HALC groups of 25 or 50 cases was notable only for a significant decrease in operative time (190 to 169 min) after 50 HALC were performed. Conclusions: The learning curve for HALC is 32-50 cases when operative time is the measure of learning. For quality-related measures, we found no learning curve for HALC. Our results indicate that a novice can perform HALC with good quality-related outcomes and that only operative speed will improve with experience.