A nomogram model for the risk prediction of type 2 diabetes in healthy eastern China residents: a 14-year retrospective cohort study from 15,166 participants

Abstract

Background Risk prediction models can help identify individuals at high risk for type 2 diabetes. However, no such model has been applied to clinical practice in eastern China. Aims This study aims to develop a simple model based on physical examination data that can identify high-risk groups for type 2 diabetes in eastern China for predictive, preventive, and personalized medicine. Methods A 14-year retrospective cohort study of 15,166 nondiabetic patients (12–94 years; 37% females) undergoing annual physical examinations was conducted. Multivariate logistic regression and least absolute shrinkage and selection operator (LASSO) models were constructed for univariate analysis, factor selection, and predictive model building. Calibration curves and receiver operating characteristic (ROC) curves were used to assess the calibration and prediction accuracy of the nomogram, and decision curve analysis (DCA) was used to assess its clinical validity. Results The 14-year incidence of type 2 diabetes in this study was 4.1%. This study developed a nomogram that predicts the risk of type 2 diabetes. The calibration curve shows that the nomogram has good calibration ability, and in internal validation, the area under ROC curve (AUC) showed statistical accuracy (AUC = 0.865). Finally, DCA supports the clinical predictive value of this nomogram. Conclusion This nomogram can serve as a simple, economical, and widely scalable tool to predict individualized risk of type 2 diabetes in eastern China. Successful identification and intervention of high-risk individuals at an early stage can help to provide more effective treatment strategies from the perspectives of predictive, preventive, and personalized medicine.

Short communication: unique metabolic signature of proliferative retinopathy in the tear fluid of diabetic patients with comorbidities — preliminary data for PPPM validation

“…A paradigm shift from reactive medical services to costeffective and patient-friendly PPPM approaches as proposed by the European Association for Predictive, Preventive and Personalised Medicine (EPMA) in the year 2009 [5][6][7][8][9][10][11][12] is pivotal to reverse current trends in DM care. Numerous research efforts have been made to promote PPPM strategies for DM risk prediction and personalised prevention at the level of pre-diabetes [13,14]. Still, evidence-based predictive and preventive approaches are strongly underrepresented in daily medical practice in primary, secondary and tertiary DM care and specifically for prevention of DM complications in case of a positive diagnosis.…”

Short communication: Unique metabolic signature of proliferative retinopathy in the tear fluid of diabetic patients with comorbidities – Preliminary data for PPPM validation

“…A paradigm shift from reactive medical services to cost-effective and patient-friendly PPPM approaches as proposed by the European Association for Predictive, Preventive and Personalised Medicine (EPMA) in the year 2009 [5][6][7][8][9][10][11][12] is pivotal to reverse current trends in DM care. Numerous research efforts have been made to promote PPPM strategies for DM risk prediction and personalised prevention at the level of prediabetes [13,14]. Still, evidence-based predictive and preventive approaches are strongly underrepresented in daily medical practice in primary, secondary and tertiary DM care and speci cally for prevention of DM complications in case of a positive diagnosis.…”

A feature optimization study based on a diabetes risk questionnaire

“…Xu et al collected 15,166 health examination data from eastern China and constructed a multifactorial regression model (LASSO regression and logistic regression) to predict the risk of type 2 diabetes. The statistical accuracy of the model was evaluated using the area under the ROC curve (AUC), which reached 0.865, indicating good predictive performance (15). Gollapalli et al ( 16) used a dataset from a hospital in Saudi Arabia, which included 897 hospitalized patients with 10 unique features.…”