JMIR Publications

ABSTRACT

Background:

Visual impairment (VI) is a prevalent global health issue, affecting over 2.2 billion people worldwide, with nearly half of the Chinese population aged 60 and above being affected. Early detection of high-risk VI is essential for preventing irreversible vision loss among Chinese middle-aged and elderly adults. While machine learning (ML) algorithms exhibit significant predictive advantages, their application in predicting VI risk among the general middle-aged and elderly population in China remains limited.

Objective:

We aimed to predict VI and identify its determinants using ML algorithms.

Methods:

We used 19,047 participants from four waves of the China Health and Retirement Longitudinal Study (CHARLS) that were conducted between 2011 and 2018. To envisage the prevalence of VI, we generated a geographical distribution map. Additionally, we constructed a model using indicators of self-reported questionnaire, physical examination, and blood biomarkers as predictors. Multiple ML algorithms, including gradient boosting machine (GBM), dynamic random forest (DRF), generalised linear model (GLM), deep learning (DL), and stacked ensemble, were used for prediction. We plotted receiver operating characteristics (ROC) and calibration curves to assess the predictive performance. Variable importance analysis was used to identify key predictors.

Results:

Among 19,047 participants, 33.9% suffered from VI. Qinghai, Chongqing, Anhui, and Sichuan showed the highest VI rates, while Beijing and Xinjiang had the lowest. GLM, GBM, and stacked ensemble achieved acceptable area under curve values of 0.706, 0.710, and 0.715, respectively, with the stacked ensemble performing best. Key predictors included hearing impairment, self-expectation of health status, pain, age, hand grip strength, depression, night sleep duration, haemoglobin, high-density lipoprotein cholesterol, and arthritis or rheumatism.

Conclusions:

Nearly one-third of middle-aged and elderly adults in China had VI. The prevalence of VI shows regional variations, but no distinct east-west or north-south distribution differences. ML algorithms demonstrate accurate predictive capabilities for VI. The combination of prediction models and variable importance analysis provides valuable insights for the early identification and intervention of VI among Chinese middle-aged and elderly adults.