ABSTRACT: Severe obesity (body mass index ?35 [calculated as weight in kilograms divided by height in meters squared]) is associated with significant medical comorbidity and increased mortality. Bariatric surgery induces weight loss, the extent of which can vary. Postoperative predictors of weight loss have not been adequately examined.To describe postoperative eating behaviors and weight control and their effects on 3-year change in weight.The Longitudinal Assessment of Bariatric Surgery-2 (LABS-2) study is a multicenter observational cohort study at 10 US hospitals in 6 geographically diverse clinical centers. Adults undergoing first-time bariatric surgical procedures as part of routine clinical care were recruited between 2006 and 2009 and followed up until September 2012. Participants completed detailed surveys regarding eating and weight control behaviors prior to surgery and then annually after surgery for 3 years.Twenty-five postoperative behaviors related to eating behavior, eating problems, weight control practices, and the problematic use of alcohol, smoking, and illegal drugs. Behaviors examined were divided into those that were never present (preoperatively or postoperatively), those that were always present (preoperatively and postoperatively), and those that underwent a healthy change after surgery (development of a positive behavior or omission of a negative behavior).The sample included a total of 2022 participants (median age, 47 years [interquartile range, 38-55 years]; median BMI, 46 [interquartile range, 42-51]; 78% women): 1513 who had undergone Roux-en-Y gastric bypass and 509 who had undergone laparoscopic adjustable gastric banding. If we consider the cumulative effects of the 3 behaviors that explain most of the variability (16%) in 3-year percent weight change following Roux-en-Y gastric bypass, ie, weekly self-weighing, continuing to eat when feeling full more than once a week, and eating continuously during the day, a participant who postoperatively started to self-weigh, stopped eating when feeling full, and stopped eating continuously during the day after surgery would be predicted to lose a mean (SE) of 38.8% (0.8%) of their baseline weight. This average is about 14% greater weight loss compared with participants who made no positive changes in these variables (mean [SE], -24.6% [1.6%]; mean difference, -14.2%; 95% CI, -18.7% to -9.8%; P?