Development and validation of a risk prediction model for dyslipidemia in community-dwelling middle-aged and older adults in China: a nationwide survey.

Background: The incidence of dyslipidemia as a risk factor for many serious diseases is increasing year by year. This study aimed to construct and visualize a risk prediction model for dyslipidemia in middle-aged and older adults. Method: The subjects of our study are derived from CHARLS. Participants were allocated to training and validation groups in a 7:3 ratio at random. To identify potential predictors of dyslipidemia, we employed univariate analysis, lasso regression, and multivariate binary logistic regression analyses. A nomogram was constructed based on logistic regression results, and a ROC curve was used to evaluate its predictive performance. The accuracy and discriminatory capability were assessed using calibration curve analysis, while the net clinical benefit rate was evaluated through decision curve analysis (DCA). Results: Our study included a total of 12,589 participants, of which 1,514 were detected with dyslipidemia syndrome. Model construction: Based on the results of the logistic regression analysis of the training set, six variables were selected to construct the model, which were ranked in order of importance as comorbid hypertension, comorbid diabetes, waistline, comorbid digestive disease, place of abode, and comorbid liver disease. The ROC curve results indicated that the prediction model exhibited moderate discriminatory ability (AUC > 0.7). Additionally, the calibration curve confirmed the model's strong predictive accuracy. The decision curve analysis (DCA) illustrated a positive net benefit associated with the prediction model. Conclusions: The prediction model of dyslipidemia risk in middle-aged and older adults constructed in this study has good efficacy and helps to screen high-risk groups.