Rothman-Keller model for risk prediction of motoric cognitive risk syndrome in older Chinese adults: A proof-of-concept study

Abstract

Objective

The risk factors for motoric cognitive risk syndrome (MCR) remains unclear, and a lack of predictive studies on the risk of developing MCR. This proof-of-concept study aims to investigate the risk factors of MCR, and develop and internally validate a prediction system to assess the risk of MCR in older Chinese adults.

Methods

Through performing a meta-analysis to obtain the risk factors and pooled risk ratio of MCR, the Rothman-Keller model which could be used to predict risk level of MCR was constructed. A total of 5020 participants from the CHARLS were used to validate discrimination and calibration of model. Predictive performance was assessed using AUC-ROC, accuracy, sensitivity, and specificity.

Results

The pooled prevalence of MCR was 12.0 %. The results of the meta-analysis showed that there are eleven significant risk factors of MCR, including age, sex, marital status, diabetes, stroke, depression, falls, physical inactivity, social isolation, fear responses, and hospitalization, with RR_i of 3.00, 1.29, 1.21, 1.53, 2.90, 2.13, 2.95, 1.46, 0.79, 1.55, and 1.31, respectively. Rothman-Keller model performed well to predict MCR risk with an AUC-ROC of 0.631 (95 %CI: 0.610–0.653), and the score was well calibrated (chi-square test for goodness-of-fit = 5.981, P = 0.649).

Conclusions

MCR is highly prevalent in China, caused by the complex interaction of genetics, comorbidities, and lifestyle factors. This study serves as proof-of-concept that Rothman-Keller model performs a good calibration and fair discrimination, which can be used as a simple, time-saving, and cost-effective mean to assess and monitor the risk level of MCR.