Abstract: Objective To establish a prediction model of cardiovascular disease(CVD)in people with dual sensory impairment based on the data of CHARLS database. Methods A cohort from China Longitudinal Study on Health and Retirement(CHARLS)in 2011 was included. The baseline population was people over 45 years of age with dual sensory impairment and no CVD, and the follow-up CVD in 2013, 2015, 2018 and 2020 was used as the outcome variable. LASSO regression and Logistic regression were used to screen the best predictors in the model, and a prediction model was constructed. Then, the research population was randomly divided into training set and verification set at the ratio of 75% to 25%. The prediction performance is evaluated by ROC curve, calibration curve and decision curve analysis. Results A total of 5866 participants with double sensory impairment were included, and the incidence of CVD was 16%(938/5866). Lasso regression and multivariate logistic regression selected age, sex, sleep duration, dyslipidemia, nephropathy, waist circumference, BMI and diabetes as predictors to build a prediction model, and the formula was: 0.027 40×age-0.052 52×gender-0.073 61×sleep duration+0.153 26×BMI-0.372 05×dyslipidemia -0.057 37×diabetes -0.252 30×nephropathy-0.014 67×waist circumference. The AUC value of the model was 0.729 in the training set and 0.692 in the verification set. Conclusion In this study, a prediction model was established to evaluate the risk of CVD in people with double sensory impairment.

Key words: Sensory impairment, Cardiovascular disease, Influencing factor, Prediction model