Predicting functional dependency using machine learning among a middle-aged and older Chinese population

Developing an interpretable machine learning model for screening depression in older adults with functional disability

This study utilized data from the 2020 wave of the China Health and Retirement Longitudinal Study database, selecting 4322 participants aged 60 and above as the study sample. Important predictors of depression in older adults with functional disabilities were identified using LASSO regression, univariate logistic regression, and multivariate logistic regression. Five different machine learning algorithms-Logistic Regression, Random Forest, Gradient Boosting, K-Nearest Neighbors, and Naive Bayes-were employed to construct risk prediction models for depression in older adults with functional disabilities. The results indicated that Sleep duration, Age, Cognitive score, Gender, Residential area, Self-rated health, Arthritis, Gastrointestinal disease, Retirement status, Life satisfaction, Composite pain status, and Physical activity level were significant predictors of depression in this population. Both the Gradient Boosting model (accuracy: 0.69, precision: 0.70, recall: 0.79, F1-score: 0.74, AUC: 0.76) and the Logistic Regression model (accuracy: 0.68, precision: 0.68, recall: 0.79, F1-score: 0.73, AUC: 0.75) demonstrated good performance and strong generalizability. Additionally, SHAP interpretation was applied to the Gradient Boosting model to enhance the explainability of its predictions, while a nomogram was created for the Logistic Regression model to visually represent the predictive process, allowing for more intuitive understanding of the model's output. This study successfully developed a risk prediction model for depression in older adults with functional disabilities through machine learning. It provides a reference tool for community screening and clinical decision-making, helping to identify and manage depression risk within the older adults with functional disabilities.

Construction of disability risk prediction model for the elderly based on machine learning

The study aimed to develop a predictive model using machine learning algorithms, providing healthcare professionals with a novel tool for assessing disability risk in older adults. Data from the 2018 and 2020 waves of the China Health and Retirement Longitudinal Study were utilized, including 3,172 participants aged 65 years and older with no baseline disability. In this study, five machine learning algorithms were employed to construct risk assessment and prediction models for disability in older adults. The Shapley Additive Explanations method was applied to analyze the independent predictors of disability risk. In total, 695 participants (21.9%) were disabled during follow-up. Among the five machine learning models, prediction models constructed using random forest and extreme gradient boosting methods showed superior performance, achieving F1 scores of 0.92 and 0.86 and accuracies of 0.92 and 0.85, respectively. Key predictors of disability risk included self-rated health, education, sleep duration, alcohol consumption, depressive symptoms, hypertension, and arthritis. The Machine learning models for assessing and predicting disability risk in older adults, particularly those developed using RF and XGBoost algorithms, exhibited strong predictive capabilities. These findings highlight the potential of these models for practical application in clinical and public health settings, warranting further exploration and validation.

Association between central obesity and ADL impairment among the middle-aged and elderly population in China based on CHARLS

To explore the associations of central obesity indicators including waist circumference (WC), waist-to-height ratio (WHtR), and weight-adjusted waist index (WWI) with the impairment of basic activities of daily living (BADL) and instrumental activities of daily living (IADL) among middle-aged and elderly population in China. This prospective study used baseline data from 2011 and follow-up data, involving 6440 and 9646 participants, respectively. Binary logistic regression analysis was used to assess the relationships. Restricted cubic spline (RCS) curve was also used to analyze the correlation trends. Stratified analyses were performed to identify potential differences. Receiver operating characteristic curves were plotted to evaluate the predictive value of each indicator. WC (OR = 1.01, 95% CI:1.01-1.02), WHtR (OR = 1.21, 95% CI = 1.09-1.33), and WWI (OR = 1.10, 95% CI:1.02-1.19) were significantly associated with BADL impairment. Only WWI (OR = 1.16, 95%CI:1.09-1.23) was associated with IADL impairment. WC, WHtR and WWI were linearly associated with BADL impairment while WWI was linearly associated with IADL impairment. The risk association between WWI and BADL was stronger in drinking individuals and males. In the participants with a BMI less than 24 kg/m² and who had received a high school education or above, the increase in WWI was accompanied by a more significant risk of IADL impairment. The predictive ability of WWI is higher than that of WC and WHtR, with AUC values of 0.597 and 0.615. WWI, as a comprehensive indicator of central obesity, may be useful in comprehensively identifying the risk of early daily living activity impairment among middle-aged and elderly population.

Multimodal integration of liquid biopsy and radiology for the noninvasive diagnosis of gallbladder cancer and benign disorders

Gallbladder cancer (GBC) frequently mimics gallbladder benign lesions (GBBLs) in radiological images, leading to preoperative misdiagnoses. To address this challenge, we initiated a prospective, multicenter clinical trial (ChicCTR2100049249) and proposed a multimodal, non-invasive diagnostic model to distinguish GBC from GBBLs. A total of 301 patients diagnosed with gallbladder-occupying lesions (GBOLs) from 11 medical centers across 7 provinces in China were enrolled and divided into a discovery cohort and an independent external validation cohort. An artificial intelligence (AI)-based integrated model, GBCseeker, is created using cell-free DNA (cfDNA) genetic signatures, radiomic features, and clinical information. It achieves high accuracy in distinguishing GBC from GBBL patients (93.33% in the discovery cohort and 87.76% in the external validation cohort), reduces surgeons' diagnostic errors by 56.24%, and reclassifies GBOL patients into three categories to guide surgical options. Overall, our study establishes a tool for the preoperative diagnosis of GBC, facilitating surgical decision-making.

An explainable machine learning-based prediction model for sarcopenia in elderly Chinese people with knee osteoarthritis

BACKGROUND

Sarcopenia is an age-related progressive skeletal muscle disease that leads to loss of muscle mass and function, resulting in adverse health outcomes such as falls, functional decline, and death. Knee osteoarthritis (KOA) is a common chronic degenerative joint disease among elderly individuals who causes joint pain and functional impairment. These two conditions often coexist in elderly individuals and are closely related. Early identification of the risk of sarcopenia in KOA patients is crucial for developing intervention strategies and improving patient health.

METHODS

This study utilized data from the China Health and Retirement Longitudinal Study (CHARLS), selecting symptomatic KOA patients aged 65 years and above and analyzing a total of 95 variables. Predictive factors were screened via least absolute shrinkage and selection operator (LASSO) regression and logistic regression. Eight machine learning algorithms were employed to construct predictive models, with internal cross-validation and independent test validation performed. The final selected model was analyzed via the SHapley Additive exPlanations (SHAP) method to enhance interpretability and clinical applicability. To facilitate clinical use, we developed a web application based on this model ( http://106.54.231.169/ ).

RESULTS

The results indicate that six predictive factors-body mass index, upper arm length, marital status, total cholesterol, cystatin C, and shoulder pain-are closely associated with the risk of sarcopenia in KOA patients. CatBoost demonstrated excellent overall performance in both calibration analyses and probability estimates, reflecting accurate and dependable predictions. The final results on the independent test set (accuracy = 0.8902; F1 = 0.8627; AUC = 0.9697; Brier score = 0.0691) indicate that the model possesses strong predictive performance and excellent generalization ability, with predicted probabilities closely aligning with actual occurrence rates and thereby underscoring its reliability.

CONCLUSION

From the perspective of public health and aging, this study constructed an interpretable sarcopenia risk prediction model on the basis of routine clinical data. This model can be used for early screening and risk assessment of symptomatic KOA patients, assisting health departments and clinicians in the early detection and follow-up of relevant populations, thereby improving the quality of life and health outcomes of elderly individuals.

Risk prediction of functional disability among middle-aged and older adults with arthritis: A nationwide cross-sectional study using interpretable machine learning

BACKGROUND

Arthritis is a common chronic disease among middle-aged and older adults and is strongly related to functional decline.

METHODS

The research sample and data were derived from the China Health and Retirement Longitudinal Study (CHARLS) 2015. We employed the least absolute shrinkage and selection operator (LASSO) and multifactor logistic regression analysis to identify features for model construction. We proposed six machine learning (ML) predictive models. The optimal model was selected using various learning metrics and was further interpreted using the SHapley Additive exPlanations (SHAP) method.

RESULTS

A total of 5111 subjects were included in the analysis, of which 1955 developed functional disability. Among the six models, XGBoost showed the best performance, achieving a test set area under the curve (AUC) of 0.74. SHAP analysis ranked the features by their contribution as follows: waist circumference, handgrip strength, self-reported health status, age, body pains, depression, history of falls, sleeping duration, and availability of care resources. SHAP dependence plots indicated that individuals over 60 with increased waist circumference (>85 cm), short sleeping duration (<5 h), and lower handgrip strength (<25 kg) had a higher probability of functional disability.

CONCLUSION

This study presents an interpretable machine learning-based model for the early detection of functional disability in patients with arthritis and informs the development of care strategies aimed at delaying functional disability in this population.

U-shaped association between sleep duration and frailty in Chinese older adults: a cross-sectional study

Objective

As the population ages, understanding the association between sleep patterns and physical frailty in older adults is crucial for formulating effective health interventions. This study aimed to explore the relationship among nap time, nighttime sleep duration, and physical frailty in older Chinese individuals; establish recommended sleep times; and provide a scientific and reasonable basis for the prevention and management of frailty in older adults.

Methods

On the basis of the 2020 China Health and Retirement Longitudinal Study database, demographic information, health data, and lifestyle information of the research subjects were obtained. A total of 5,761 survey participants were included, and logistic regression and restricted cubic splines were used to explore the association between sleep duration and frailty.

Results

In our cross-sectional analysis, the duration of napping in older adults did not show a significant correlation with frailty. The optimal nighttime sleep interval for older adults was 7-8 h, and the maximum health benefit was achieved when nighttime sleep reached 7.5 h. Compared with older adults in China who slept 6-8 h at night, those with a sleep duration of <6 h (OR = 1.58, 95% CI: 1.36-1.82) were more likely to be frail. After adjusting for all covariates such as smoking, multimorbidity, self-rated health, social events, education level, and frequency of physical activity, we found no interaction between gender and age concerning sleep duration.

Conclusion

The potential correlation between nighttime sleep duration and frailty in older adults is basically U-shaped. Older Chinese adults with a moderate nighttime sleep duration of 7-8 h exhibited the lowest likelihood of frailty than their counterparts. The duration of napping is not related to the likelihood of frailty in older people. Thus, the importance of sufficient nighttime sleep for the health of older adults must be emphasized.

Predicting the decline of physical function among the older adults in China: A cohort study based on China longitudinal health and longevity survey (CLHLS)

BACKGROUND

As the arrival of healthy aging, maintaining physical function (PF) in older adults is crucial for their health, so it is necessary to detect the decline of PF among them and take intervention measures.

METHODS

We construct eight machine learning models to predict declines of PF in this study. The performance of the models was tested by Area Under Curve (AUC), sensitivity, specificity, accuracy, precision-recall (PR) curve and calibration degree. Decision Curve Analysis (DCA) curve was used to evaluate their discrimination ability and clinical practicability.

RESULTS

There were 2,017 participants in this study. We found that logistic regression models performed the best, with AUC, sensitivity, specificity and accuracy of 0.803, 0.698, 0.761 and 0.744 respectively, and its DCA curve, calibration degree and PR curve also performed well.

CONCLUSION

Logistic regression can be used as optimal model to identify the risk of PF decline among older adults in China.

Interpretable machine learning for identifying overweight and obesity risk factors of older adults in China

OBJECTIVE

To estimate the importance of risk factors on overweight/obesity among older adults by comparing different predictive model.

METHODS

Survey data from 400 older individuals in China was employed to assess the impacts of four domains of risk factors (demographic, health status, physical activity and neighborhood environment) on overweight/obesity. Six machine learning algorithms were utilized for prediction, and SHapley Additive exPlanations (SHAP) was employed for model interpretation.

RESULTS

The CatBoost model demonstrated the highest performance among the prediction models for overweight/obesity. Gender, transportation-related physical activity and road network density were top three important features. Other significant factors included falls, cardiovascular conditions, distance to the nearest bus stop and land use mixture.

CONCLUSION

Insufficient physical activity, denser road network and incidents of falls increased the likelihood of older adults being overweight/obese. Strategies for preventing overweight/obesity should target transportation-related physical activity, neighborhood environments, and fall prevention specifically.

Machine learning algorithms to predict mild cognitive impairment in older adults in China: A cross-sectional study

OBJECTIVE

This study aimed to explore the predictive value of machine learning (ML) in mild cognitive impairment (MCI) among older adults in China and to identify important factors causing MCI.

METHODS

In this study, 6434 older adults were selected based on the data of the China Health and Elderly Care Longitudinal Survey (CHARLS) in 2020, and the dataset was subsequently divided into the training set and the test set, with a ratio of 6:4. To construct a prediction model for MCI in older adults, six ML algorithms were used, including logistic regression, KNN, SVM, decision tree (DT), LightGBM, and random forest (RF). The Delong test was used to compare the differences of ROC curves of different models, while decision curve analysis (DCA) was used to evaluate the model performance. The important contributions of the prediction results were then used to explain the model by the SHAP value.The Matthews correlation coefficient (MCC) was calculated to evaluate the performance of the models on imbalanced datasets. Additionally, causal analysis and counterfactual analysis were conducted to understand the feature importance and variable effects.

RESULTS

The area under the ROC curve of each model range from 0.71 to 0.77, indicating significant difference (P < 0.01). The DCA results show that the net benefits of LightGBM is the largest within various probability thresholds. Among all the models, the LightGBM model demonstrated the highest performance and stability. The five most important characteristics for predicting MCI were educational level, social events, gender, relationship with children, and age. Causal analysis revealed that these variables had a significant impact on MCI, with an average treatment effect of -0.144. Counterfactual analysis further validated these findings by simulating different scenarios, such as improving educational level, increasing age, and increasing social events.

CONCLUSION

The ML algorithm can effectively predict the MCI of older adults in China and identify the important factors causing MCI.

Development, validation and economic evaluation of a machine learning algorithm for predicting the probability of kidney damage in patients with hyperuricaemia: protocol for a retrospective study

INTRODUCTION

Accurate identification of the risk factors is essential for the effective prevention of hyperuricaemia (HUA)-related kidney damage. Previous studies have established the efficacy of machine learning (ML) methodologies in predicting kidney damage due to other chronic diseases. Nevertheless, a scarcity of precise and clinically applicable prediction models exists for assessing the risk of HUA-related kidney damage. This study aims to accurately predict the risk of developing HUA-related kidney damage using a ML algorithm, which is based on a retrospective database.

METHODS AND ANALYSIS

This retrospective study aims to collect clinical data on outpatients and inpatients from the Sichuan Provincial People's Hospital, China, covering the period from 1 January 2018 to 31 December 2021 with a focus on patients diagnosed with 'hyperuricaemia' or 'gout'. Predictive models will be constructed using techniques such as data imputation, sampling, feature selection and ML algorithms. This research will evaluate the predictive accuracy, interpretability and fairness of the developed models to determine their clinical applicability. The net benefit and net saving will be calculated to gauge the economic value of the model. The most effective model will then undergo external validation and be made available as an online predictive tool to facilitate user access.

ETHICS AND DISSEMINATION

The Ethics Review Committee at Sichuan Provincial People's Hospital granted approval for the ethical review of this study without requiring informed consent. The findings of the study will be disseminated in a peer-reviewed journal.

Dynapenic abdominal obesity and incident functional disability: Results from a nationwide longitudinal study of middle-aged and older adults in China

BACKGROUND

There is little epidemiological evidence on the relationship of dynapenic abdominal obesity (DAO) and the development of functional disability, particularly in Asian populations. We aimed to investigate the association of DAO with new-onset functional disability in Chinese adults.

METHODS

A total of 7881 participants aged ≥45 years from China Health and Retirement Longitudinal Study (CHARLS) in 2011 and 2015 were included in the study. Dynapenia and abdominal obesity were respectively defined based on handgrip strength (<28 kg for male and <18 kg for female) and waist circumference (≥ 90 cm for male and ≥85 cm for female). The sample was divided into four groups: non-dynapenic/non-abdominal obesity (ND/NAO), non-dynapenic/abdominal obesity (ND/AO), dynapenic/non-abdominal obesity (D/NAO) and dynapenic/abdominal obesity (D/AO). Functional status was assessed by basic activities of daily living (BADL) or instrumental activities of daily living (IADL). Logistic regression model was used to explore the longitudinal association between dynapenic abdominal obesity and incident functional disability.

RESULTS

After a 4-year follow-up, 1153 (14.6 %) developed BADL disability and 1335 (16.9 %) developed IADL disability. The multivariable-adjusted odds ratios (95 % CIs) for the D/AO versus ND/NAO were 2.21 (1.61-3.03) for BADL disability, and 1.68 (1.23-2.30) for IADL disability. In addition, DAO was associated with an increased risk for functional dependency severity (odds ratio, 2.08 [95 % CI, 1.57-2.75]).

CONCLUSIONS

DAO was significantly associated with greater risk of functional disability among Chinese middle-aged and older adults. Our findings indicated that interventions targeted DAO might be effective in the primary prevention of functional disability.