Influence of sociodemographic variables on weight loss outcomes up to 3-years following primary bariatric surgery

Background: Our group has previously demonstrated that low socioeconomic status (SES) independently predicts ≤ 25th percentile weight-loss following bariatric surgery (BS). Given that sociodemographic metrics can be separated into income, education, and race, we sought to investigate how each metric independently impacted weight loss following BS.

Methods: Patients from a single academic institution who underwent bariatric surgery from 2014 to 2016 were retrospectively reviewed. Patients were stratified by income (low/high), education (≤ high school/ ≥ college), and race (black/white) then compared using univariate analysis. Variables significant on univariate analyses were subsequently used for a greedy 1:3 propensity score match with a caliper of 0.2. After matching, groups were balanced on demographics, social/medical/psychological history, and surgery type. Percent excess body weight loss for each post-operative time point was compared using appropriate univariate analyses. A p-value ≤ 0.05 was considered statistically significant.

Results: 571 patients were included. Unmatched race analysis demonstrated black patients were significantly younger (p = 0.05), single (p < 0.0001), in a lower income bracket (p < 0.0001), and experienced less weight loss at 2- (p = 0.01), 6- (p = 0.007), 12- (p = 0.008) and 24- (p = 0.007) months post-op. After matching, black patients continued to experience less weight loss at 2- (p = 0.01) and 6- (p = 0.03) months, which trended at 1 year (p = 0.06). Initial income analysis demonstrated patients in the low-income group (LIG) were more likely to be black (p < 0.0001), have ≤ high school education (p = 0.004), a higher preoperative BMI (p = 0.008), and lower postoperative weight loss at 2- (p = 0.001), 6- (p = 0.01), and 12- (p = 0.04) months after surgery. After matching, no differences were observed up to 3-years post-op. Analysis of education demonstrated no effect on weight loss in both unmatched and matched analyses.

Conclusion: Unmatched analysis demonstrated that low income and race impact short-term weight loss after BS. After matching, however, race, not socioeconomic status, predicted weight loss outcomes up to 1-year.