In the general population, bariatric surgery facilitates sustained weight loss and remission of comorbidities (e.g., CVD). The effectiveness among Medicaid beneficiaries, a population with a disproportionately high burden of obesity, remains unclear. We sought to determine if patterns of weight change following bariatric surgery differed in patients with Medicaid compared to commercial insurance.
Data were obtained from the Longitudinal Assessment of Bariatric Surgery, an observational cohort study of adults undergoing bariatric surgery (2006-2009) at one of 6 geographically diverse centers. We used group-based trajectory models (PROC TRAJ, SAS) to identify discrete groups of individuals with similar weight trajectories over 5-years; we considered cubic trajectory models with between 1 and 6 groups. We selected the 6-group model based on Bayesian Information Criteria, visual fit and statistical significance of parameter estimates. We tested demographic predictors against trajectory group membership; Medicaid (n=174) vs. Commercial (n=991) was our primary predictor of interest.
Medicaid participants had higher mean weight at baseline compared to the commercially insured (307.9 vs. 291.2 lbs), but lost a similar amount of weight over time (74.8 vs 85.0 lbs). All groups experienced the most weight loss in year 1 (Figure). Group 6 was the heaviest at baseline and began to regain weight after year 1; groups 1-5 continued to lose weight through year 2. Group’s 1 & 2 weight remained stable over 5 years while groups 3-6 began to slowly regain weight. Medicaid beneficiaries were more likely to be in group 4 (OR 3.2, 95% CI: 1.2, 8.6), group 5 (OR 5.7, 95% CI: 2.0, 16.1) and group 6 (OR 7.9, 95% CI: 2.5, 25.1) compared to group 1.
Medicaid beneficiaries experience substantial long term weight loss. However, they had higher baseline weight and tended to be in groups with less long-term weight loss compared to the commercially insured. Focusing on tailored interventions to specific patient groups could help maintain weight loss.
Statistical trajectory models for phonetic classification
