Options in bariatric surgery: modeled decision analysis supports Roux-en-Y gastric bypass and sleeve gastrectomy as the treatments of choice
S. Keogh; J. C. Bolger; S. Brady; A. Rodgers; M. Arumugasamy; W. B. Robb
Year of publication
Surg Obes Relat Dis
BACKGROUND: Obesity is a chronic disease associated with significant morbidity and mortality. Bariatric surgery has been shown to significantly reduce both morbidity and mortality. Numerous surgical strategies exist, but the most frequently used worldwide are adjustable gastric banding, sleeve gastrectomy (SG), and Roux-en-Y gastric bypass (RYGB). It is not clear which of these strategies provides the optimal quality-of-life pay-off. OBJECTIVE: Modeled decision analysis allows comparison of different treatment interventions allowing for plausible differences in input variables. This facilitates establishment of the optimal intervention under numerous conditions. SETTING: University Hospital, Ireland. METHODS: Modeled decision analysis was performed from the patient's perspective comparing best medical therapy, adjustable gastric banding, SG, and RYGB. Input variables were calculated based on previously published decision analyses and a systematic search of obesity-related literature. Utilities were based on previously published studies. One-way sensitivity analysis was performed. Sensitive variables underwent 3-way analysis. RESULTS: The optimal treatment strategy in the base case was RYGB with a quality-adjusted life-year payoff (QALY) of 1.53 QALYs at 2 years postprocedure. Sleeve gastrectomy provided 1.49 QALYs. Medical therapy and adjustable gastric banding provided .98 and .96 QALYs, respectively. Rate of complications in RYGB and the utility of SG and RYGB proved sensitive. If complication rates are high, SG becomes the optimal strategy. Sensitive thresholds were established for the utility of SG and RYGB at .804 and .78, respectively. CONCLUSION: SG and RYGB offer similar outcomes in terms of QALY payoffs. Decision making should be in line with institutional and patient preference.