Demographic factors as determinant of selection of sleeve gastrectomy or gastric bypass for surgical weight loss.

ABSTRACT

INTRODUCTION: Use of sleeve gastrectomy (SG) for weight loss has grown exponentially; however, clear indications for SG versus Roux-en-Y gastric bypass (RNYGB) are lacking. Certain populations may be more likely to undergo SG due to its simpler technique and without clear clinical indications. We aim to examine underlying predictors of patients undergoing SG vs RNY across a single state. METHODS: We queried the Colorado All Payers Claim Database for patients undergoing laparoscopic SG or RNY. Patient-level variables included patient demographics, comorbidities, distance traveled for surgery, and distressed communities index (DCI), a zip code-based measure of socioeconomic status. Hospital-level variables included annual bariatric surgery volume, academic status, and whether hospitals were a bariatric Center of Excellence. We performed mixed-effects logistic regression adjusting for demographics, insurance coverage, and comorbidities to compare odds of undergoing SG vs RNY, with a random effect for hospital. RESULTS: 5,017 patients were included with 3,042 (60.6%) undergoing SG and 1,975 (39.4%) undergoing RNY. On multivariable analysis, patients with a high DCI were not more likely to undergo a SG (OR 1.18, CI 0.89-1.55, p = 0.25). However, patients who underwent surgery at hospitals serving the greatest proportion of those from highly distressed communities were significantly more likely to undergo SG (OR 4.22, CI 1.38-12.96, p = 0.01). Patients managed at Bariatric Centers of Excellence were less likely to undergo SG (OR 0.22, CI 0.07-0.62, p = 0.005). Patients with higher BMI, diabetes, or GERD were all more likely to undergo RNY. CONCLUSION: While patients with high DCI were more likely to undergo SG on univariate analysis, these associations disappeared after addition of a hospital-level random effect, suggesting that disparities may be due access to surgeons or systems with preference for one procedure. However, hospitals serving a higher proportion of high-DCI patients are more likely to utilize SG.