A novel nomogram for predicting successful weight loss following diet and exercise intervention in people with obesity

Abstract

Abstract Purpose Obesity is a global health challenge. However, achieving successful weight loss remains challenging. Therefore, this study aims to identify potential factors for weight loss failure by analyzing pre-weight loss data. Methods We utilized data encompassing records of 2577 people with obesity who visited weight management clinics from 2013 to 2022, with 1276 having at least a 3-month follow-up visit. Data preprocessing involved selecting 1276 patients with follow-up data. After dietary and exercise interventions, 580 participants achieved successful weight loss. We then divided the participants into two groups to analyze their baseline, those who lost weight and those who did not. Results Statistical analysis was conducted using RStudio, 13 predictor variables were identified based on LASSO and logistic regression, and age emerged as the most influential predictor. A nomogram for predicting weight loss success was then developed. The nomogram demonstrated good predictive performance (AUC = 0.807) and clinical applicability, as validated by internal validation methods. Decision curve analysis (DCA) also demonstrated the nomogram's clinical utility in predicting weight loss success. Conclusion We developed a nomogram prediction model for successful weight loss. The nomogram is easy to use, highly accurate, and has excellent effect discrimination and calibration capabilities.