Journal of Medical Internet Research - International Scientific Journal for Medical Research, Information and Communication on the Internet

Timely stroke diagnosis and intervention are necessary considering its high prevalence. Previous studies have mainly focused on stroke prediction with balanced data. Thus, this study aimed to develop machine learning models for predicting stroke with imbalanced data in an elderly population in China. Data were obtained from a prospective cohort that included 1131 participants (56 stroke patients and 1075 non-stroke participants) in 2012 and 2014, respectively. Data balancing techniques including random over-sampling (ROS), random under-sampling (RUS), and synthetic minority over-sampling technique (SMOTE) were used to process the imbalanced data in this study. Machine learning methods such as regularized logistic regression (RLR), support vector machine (SVM), and random forest (RF) were used to predict stroke with demographic, lifestyle, and clinical variables. Accuracy, sensitivity, specificity, and areas under the receiver operating characteristic curves (AUCs) were used for performance comparison. The top five variables for stroke prediction were selected for each machine learning method based on the SMOTE-balanced data set. The total prevalence of stroke was high in 2014 (4.95%), with men experiencing much higher prevalence than women (6.76% vs. 3.25%). The three machine learning methods performed poorly in the imbalanced data set with extremely low sensitivity (approximately 0.00) and AUC (approximately 0.50). After using data balancing techniques, the sensitivity and AUC considerably improved with moderate accuracy and specificity, and the maximum values for sensitivity and AUC reached 0.78 (95% CI, 0.73–0.83) for RF and 0.72 (95% CI, 0.71–0.73) for RLR. Using AUCs for RLR, SVM, and RF in the imbalanced data set as references, a significant improvement was observed in the AUCs of all three machine learning methods (p < 0.05) in the balanced data sets. Considering RLR in each data set as a reference, only RF in the imbalanced data set and SVM in the ROS-balanced data set were superior to RLR in terms of AUC. Sex, hypertension, and uric acid were common predictors in all three machine learning methods. Blood glucose level was included in both RLR and RF. Drinking, age and high-sensitivity C-reactive protein level, and low-density lipoprotein cholesterol level were also included in RLR, SVM, and RF, respectively. Our study suggests that machine learning methods with data balancing techniques are effective tools for stroke prediction with imbalanced data.

Background: Long-term care (LTC) needs for the elderly have become increasingly crucial policy concerns in rapidly aging Asia, especially in China, the most populous nation. However, very few studies have examined the cohort differences in terms of their existing and expected utilization of LTC services, above all urban–rural differences. This study aims to evaluate the differences of LTC current status and needs between urban–rural areas and age groups, and to identify influencing factors causing the different LTC needs. Methods: The data come from the Chinese Longitudinal Health Longevity Survey in 2014. A total of 7192 home-based elderly aged ≥65 years by multistage sampling were enrolled. The Andersen Model was applied to categorize the influential factors into three components including predisposing, enabling and needs. Multivariate logistic regression analysis was used to analyze the influential factors of the three levels of LTC needs. Results: A total of 6909 valid sample sizes were included in this study. The overall LTC needs of the elderly showed a rapidly increasing trend among which older people had the highest needs for bathing (27.29%) and toileting (15.8%). It was also demonstrated the aged cohort between urban and rural exerted an impact on all aspects of LTC status and needs to varying degrees (p < 0.05). Compared with urban areas, the LTC needs for the elderly in rural areas was more vigorous, but the supply was seriously inadequate. The elderly who were older, living in rural areas, unmarried, non-farming, with low income, in poor health and having less autonomy had higher anticipated needs for LTC services (OR > 1, p < 0.01). Compared with the young-old in rural areas, the young-old in urban areas were prone to live alone (OR = 1.61, p < 0.01). The elderly who were older, living in rural areas, farming, with low income, lonely and depressed had higher anticipated needs for community-based services (1 < OR < 1.69, p < 0.05). Conclusions: The aged cohort in urban–rural distinction were facing an increasing need for immediate care due to the inadequate support being provided, especially among rural elderly. The oldest elderly in rural areas had higher LTC needs, and different levels of needs were affected by age, economic level, family support, health status and other related effects. This study provides evidence-based recommendation for further improving the construction and development of the LTC system in China.

Objective: The objective of this study was to examine whether loneliness was associated with the risk of developing dementia in Chinese older adults and whether the association was moderated by gender. Method: A 3-year cohort study was conducted using data from the 2008/2009 and 2011/2012 waves of the Chinese Longitudinal Healthy Longevity Survey (CLHLS). Multiple logistic regression was used to analyze the relationship between loneliness and dementia. The interaction between loneliness and gender was also evaluated. Results: At 3-year follow-up, 393 of the 7867 participants had dementia. Loneliness was associated with dementia (odds ratio (OR) = 1.31, 95% confidence interval (CI) = 1.11–1.56) after adjustment for sociodemographic characteristics, lifestyle, and baseline health status. A significant interaction between loneliness and gender was also found (OR = 0.81, 95% CI = 0.65–0.99). Conclusion: Loneliness increased the risk of developing dementia among people aged 65 years and older in China. Moreover, the effect of loneliness on dementia risk varied by gender. Specifically, men who felt lonely were more likely to suffer from dementia than women.

This article used the nationally representative Chinese Longitudinal Healthy Longevity Survey to explore the associations between living arrangements and health among older adults. Living arrangements were stratified into six categories. Health was measured by self-rated health, activities of daily living (ADL) disability, and cognitive impairment. Random-effects ordered probit regressions were applied. The results indicated that coresidence had a positive effect on self-rated health compared with living alone. After introducing psychological well-being, the health differences observed in living with a spouse and living with both spouse and children were not significant. Participants with each of the living arrangement were more likely to have a higher rate of cognitive impairment and ADL disability than those living alone. Living arrangements were associated with older adults’ health. Psychological well-being was a key factor in this association, which may result from living with a spouse, and could contribute to the self-rated health of older adults.

This study aimed to investigate the patterns and predictive factors of healthcare-seeking behavior among older Chinese adults. A sample of 10,914 participants aged ≥60 years from the 2011, 2013 and 2015 China Health and Retirement Longitudinal Study (CHARLS) was included. The bivariate analyses and Heckman selection model was used to identify predictors of healthcare-seeking behavior. Results shows that the utilization rate of outpatient services increased from 21.61% in 2011 to 32.41% in 2015, and that of inpatient services increased from 12.44% to 17.68%. In 2015, 71.93% and 92.18% chose public medical institutions for outpatient and inpatient services, 57.63% and 17.00% chose primary medical institutions. The individuals who were female, were younger, lived in urban, central or western regions, had medical insurance, had poor self-rated health and exhibited activity of daily living (ADL) impairment were more inclined to outpatient and inpatient services. Transportation, medical expenses, the out-of-pocket ratio and the urgency of the disease were associated with provider selection. The universal medical insurance schemes improved health service utilization for the elderly population but had little impact on the choice of medical institutions. The older adults preferred public institutions to private institutions, preferred primary institutions for outpatient care, and higher-level hospitals for hospitalization.