Machine Learning-Based Prediction for 4-Year Risk of Metabolic Syndrome in Adults: A Retrospective Cohort Study

Abstract

Machine learning (ML) techniques have emerged as a promising tool to predict risk and make decisions in different medical domains. We aimed to compare the predictive performance of machine learning-based methods for 4-year risk of metabolic syndrome in adults with the previous model using logistic regression. Patients and Methods: This was a retrospective cohort study that employed a temporal validation strategy. Three popular ML techniques were selected to build the prognostic models. These techniques were artificial neural networks, classification and regression tree, and support vector machine. The logistic regression algorithm and ML techniques used the same five predictors. Discrimination, calibration, Brier score, and decision curve analysis were compared for model performance. Results: Discrimination was above 0.7 for all models except classification and regression tree model in internal validation, while the logistic regression model showed the highest discrimination in external validation (0.782) and the smallest discrimination differences. The logistic regression model had the best calibration performance, and ANN also showed satisfactory calibration in internal validation and external validation. For overall performance, logistic regression had the smallest Brier score differences in internal validation and external validation, and it also had the largest net benefit in external validation. Conclusion: Overall, this study indicated that the logistic regression model performed as well as the flexible ML-based prediction models at internal validation, while the logistic regression model had the best performance at external validation. For clinical use, when the performance of the logistic regression model is similar to ML-based prediction models, the simplest and more interpretable model should be chosen.

The integrated common-sense model of illness self-regulation: predicting healthy eating, exercise behaviors, and health among individuals at risk of metabolic syndrome

“…The exclusion criteria were: pregnant and breastfeeding women. Our team had developed a web-based calculator to predict the 4-year risk of developing MetS based on the prediction model [6][7][8][9], and the webpage calculator can be found at https:// msypr edict. shiny apps.…”

“…(2) timeline; (3) personal control; (4) treatment control; (5) identity; (6) concern; (7) coherence; and (8) emotional response. Concern and emotional response can be merged to form an item named emotional representations [26].…”

“…The results suggested that existing prediction models for MetS had a high risk of bias in their methodological quality. Consequently, our team had developed and validated a prediction model for the 4-year risk of MetS in adults using logistic regression, which had satisfactory discrimination, calibration, brier score, and clinical utility [6][7][8][9]. After screening individuals with high risk of MetS using the above prediction model, it is important to help the target population change their healthy eating and exercise behaviors, which are also primary interventions for the management of MetS [10].…”

Automated Machine Learning to Develop Predictive Models of Metabolic Syndrome in Patients with Periodontal Disease

“…In a retrospective cohort study published in [50], Zhang et al employ machine learning techniques to predict the 4-year risk of metabolic syndrome in adults. The research focuses on leveraging computational methods for risk assessment over an extended period, offering insights into the potential of machine learning in predicting the development of metabolic syndrome.…”

Introducing AI to the molecular tumor board: one direction toward the establishment of precision medicine using large-scale cancer clinical and biological information

“…1 ) [ 51 , 52 ]. Supervised learning is then broadly classified into regression and classification problems, and regression models are used as models for disease onset prediction and prognostic prediction [ 53 – 55 ]. Classification by supervised learning is currently the most widely studied method, especially in medical image analysis, and is practically applied in clinical practice [ 33 , 56 – 63 ].…”