Predicting frailty risk in adults exposed to parental violence during childhood: a nomogram development and...

Develop and validate a predictive model to assess the risk of frailty in adulthood among individuals exposed to parental violence during childhood.

Data from the China Health and Retirement Longitudinal Study (CHARLS), this predictive model was analyzed. A total of 17 variables, including age, gender, region, and smoking status, were considered. The cohort was randomly split into a training set and a validation set at a ratio of 70 to 30%. LASSO regression with 10-fold cross-validation was utilized for dimensionality reduction to identify the most predictive variables, which were subsequently incorporated into a multivariate regression model to identify frailty-related factors. Evaluate the agreement between the model's predicted probabilities and the actual observed probabilities using the calibration curve. The model's performance was evaluated using the area under the curve (AUC) and decision curve analysis (DCA).

A total of 3,651 respondents from the CHARLS database who had witnessed parental violence during childhood were included in the final analysis. After performing LASSO regression, four factors were selected: age, educational attainment, self-perceived health, and mental health. The AUC values of training and validation sets were of 0.875(95% CI: 0.858-0.892)and 0.892(95% CI༚0.869-0.915). The corrected calibration curve indicated that reasonable calibration accuracy was achieved by the model. DCA demonstrated that the model offers a significant net clinical benefit.

The study introduces a nomogram-based predictive model that can be used to assess the risk of frailty in individuals exposed to parental violence during childhood.