Prediction Model of All-cause Death Based on Balance Ability Among Middle-aged and Older Chinese Adults of Overweight and Obesity

Background: Advances in studies using body indicators to predict death risk. Estimating the balance ability of death risk in middle-aged and older Chinese adults with overweight and obesity is still challenging.

Methods: A retrospective analysis of the data from the China Health and Retirement Study from January 2011 to December 2018. A total of 8,632 participants were randomly divided into 7:3 a training group and a verification group, respectively. Univariable Cox analysis was used to prescreen 17 potential predictors for incorporation in the subsequent multivariable Cox analysis. Nine variables were included in the nomogram finally and validated with concordance index (C-index), calibration plots, Hosmer-Lemeshow test, and internal validation population.

Results: 287 participants were death in the training group. One hundred and thirteen participants were death in the verification group. A total of nine indicators were included in the modeling group, including gender, age, marriage, hypertension, diabetes, stroke, ADL, IADL, and balance ability to establish a prediction model. The nomogram predicted death with a validated concordance index of (C-index = 0.77, 95% CI: 0.74-0.80). The inclusion of balance ability variables in the nomogram maintained predictive accuracy (C-index = 0.77, 95% CI: 0.73-0.82). The calibration curve graph and Hosmer-Lemeshow test ( > 0.05 for both the modeling group and the verification group) showed the model has a good model consistency.

Conclusion: In the present study, we provide a basis for developing a prediction model for middle-aged and older people with overweight and obesity. In most cases, balance ability is more reversible than other predictors.