Machine learning-based predictive modeling of decline in intrinsic capacity among migrant older adults with children

Machine learning-based predictive modeling of decline in intrinsic capacity among migrant older adults with children | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Machine learning-based predictive modeling of decline in intrinsic capacity among migrant older adults with children Qinghua Zhang, Yu Wang, Xiaoxiao Hu, Xinuo Yao, Yuhan Yang, Danyan Lu, and 2 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8610961/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 7 You are reading this latest preprint version Abstract Background While migrating to join adult children can provide familial support, it may also pose risks to the intrinsic capacity (IC) of older adults due to disrupted social networks and environmental stressors. To address the paradox, we applied machine learning within a health ecology framework to develop risk predictive models and to rank key drivers of IC decline among migrant older adults with children (MOAC). Methods This multi-center, large-sample cross-sectional study was conducted from December 2022 to September 2023. A total of 3016 MOAC were randomly recruited across three Chinese regions (XX, XX, XX). The health ecology model was operationalized into a structured questionnaire to assess intrinsic capacity and its potential risk factors across five domains. Data were analyzed using interpretable machine learning techniques, including predictor importance ranking and partial dependence plots (PDPs). Results The top ten predictors of IC decline were identified as social network, health literacy, sleep quality, migration reason, economic level, number of chronic diseases, primary source of financial support, prior career, multi-layered medical insurance coverage, and age. The Random Forest model provided the most robust prediction of IC decline (AUC: 0.843; 95% CI: 0.820–0.867). PDPs further delineate the non-linear patterns through which these predictors influence IC decline. Conclusions The findings provide practical and actionable guidance for community healthcare providers. Screening based on non-modifiable risks (advanced age, chronic diseases, etc) enable the early identification of high-risk MOAC individuals. Building on this, tailored interventions can be developed to target modifiable factors such as health literacy and social networks, thereby promoting IC and enhancing quality of life. migrant older adults intrinsic capacity machine learning Figures Figure 1 Figure 2 Figure 3 Figure 4 1 Introduction Throughout the world, the number of migrant older adults is increasing dramatically (Holecki et al., 2020 ), and this trend is particularly prominent in China. The Seventh National Population Census reports 375.82 million older adults, 18.7% of whom are aged 60 or above (Tu et al., 2022 ). This aging trend, combined with heightened internal mobility, has led to sustained migration across regions, thereby giving rise to a specific vulnerable population—migrant older adults with children (MOAC) (Zhi et al., 2021 ). The MOAC are those older adults who reside long-term after moving to where their children live to take care of their grandchildren or to receive care from their children (Muneera et al., 2022 , 2022 ; Stephens et al., 2020 ). After leaving their familiar living environment, the MOAC have to adapt to a new setting and an unfamiliar social network. This adaptation leads to a series of cognitive and psychological challenges, as well as social isolation(H. Li et al., 2022 ), which ultimately negatively impacts their intrinsic capacity (Tian et al., 2021 ). As defined by WHO, intrinsic capacity (IC) is “the composite of all the physical and mental capacities that an individual can draw on at any point in time” (Bautmans et al., 2022 ), whose core domains include sensation, locomotion, vitality, cognition and psychology (Y. Zhou & Ma, 2022 ). While existing research on community-dwelling older adults has identified personal traits, health conditions, and lifestyle habits as key correlates of IC (Ma et al., 2021 ; Muneera et al., 2022 ; Stephens et al., 2020 ), studies focusing specifically on MOAC remain limited. Moreover, the selection of influencing factors in prior studies has often been insufficiently comprehensive, reflecting certain theoretical gaps. To tackle these limitations, this study employs the health ecology model to explore the factors associated with IC decline among MOAC in a more holistic and multidimensional manner. The health ecology model consists of five layers from the inside to the outside, namely demographic factors, health behavior factors, social network factors, living and working conditions factors, and social policy factors (Liang et al., 2023 ). It serves as a theoretical framework that highlights the interactions between health and factors across multiple levels, thereby offering a structured approach to understanding the determinants of individual health. IC is individual and internal in nature, is not static. It evolves dynamically through interaction with the environment and can thus be viewed as the core “individual health asset” within this framework. Given the complexity of IC, the health ecology model provides a useful structure for characterizing an individual’s risk behaviors and helps systematically identify the factors associated with IC decline among MOAC (See Appendix A). Current studies indicate that the rate of IC decline among older adults varies from 19.23% to 89.3% (Ramírez-Vélez et al., 2019 ; Stolz et al., 2022a ). This decline is linked to adverse health outcomes such as falls, frailty, functional impairment, longer hospital stays, and increased mortality risk (Hernandes et al., 2024 ; Lee et al., 2023 ; S. Liu et al., 2021 ; Salinas-Rodríguez et al., 2022 ). Therefore, establishing a predictive model for IC decline is essential to accurately identify the MOAC with high-risk of IC decline. Machine learning (ML) technology has provided advanced analytical methods for developing predictive systems. It enables more accurate classification and prediction by analyzing complex interactions within multi-dimensional datasets, while overcoming the limitations of traditional models that rely on linear assumptions (Tokodi et al., 2020 ). Additionally, ML can enhance predictive accuracy by assessing feature importance when handling high dimensional data (Stafford et al., 2020 ). The MOAC population, characterized by the dual vulnerability of aging and mobility, is expanding rapidly. Despite facing complex physiological, psychological, and social challenges, this group remains significantly understudied, and their risks of IC decline have been largely overlooked. To bridge this gap, this study evaluates and compares multiple ML algorithm—including Random Forest (RF), Artificial Neural Network (ANN), Classification and Regression Tree (CART), and Logistic Regression (Logit)—to develop an interpretable predictive model for early identification of IC decline among MOAC. The findings offer practical and actionable guidance for community healthcare providers, enabling early identification of high-risk individuals and informing the development of tailored interventions to support IC maintenance and quality of life improvement among MOAC. 2 Methods 2.1 Study design and participants A multi-stage cluster random sampling method was adopted in China from December 2022 to September 2023. Firstly, Zhejiang Province, Jiangsu Province, and Shanghai Municipality were selected as primary sampling units based on their level of economic development. Secondly, three cities/districts were selected from each region as secondary sampling units: the XX Province sample included XX, XX, and XX cities; the XX Province sample included XX, XX, and XX cities; and the XX Municipality sample included XX New Area, XX District, and XX District. Thirdly, two subordinate towns were randomly selected from each city/district as tertiary sampling units. Finally, five communities were selected from each of the tertiary sampling units, totaling 90 communities. Due to the scattered distribution of the MOAC, we adopted a combination of purposive and snowball sampling methods to identify participants in local communities and expand the sample size. All participants should meet the following inclusion criteria: (1) Age ≥ 60 years; (2) Following their children from original residence to a different area in order to take care of their grandchildren or to receive care from their children; (3) Having lived there at least six months after crossing district/county/province boundaries were included; (4) Having voluntarily signed the informed consent form. The exclusion criteria included serious physical diseases, mental disorders or cognitive disorders. The data were collected by seven trained geriatric care professionals who were assessed by a professor in geriatric psychological care. According to Kendall’s sample estimation method, the sample size is at least 10–20 times the number of variables. Since the study’s questionnaire had 23 variables, the benchmark sample size for simple random sampling should be at least 230–460. Given that an intraclass correlation coefficient (ICC) of approximately 0.0304 was identified in the preliminary small-sample survey conducted by the research team, the required sample size for cluster sampling calculated according to the formula n = n 0 [1+(m-1) ρ ] ranges from 461 to 921. Considering the 20% invalid recovery rate (Burmeister & Aitken, 2012 ), the final sample size ranges from 577 to 1152. Ultimately, approximately 1000 participants were surveyed in each primary sampling unit, with a total of 3055 MOAC enrolled in the study. After excluding samples with missing values, 3016 MOAC remained in the dataset for further analysis. 2.2 Outcome variables IC encompasses five domains: locomotion, cognition, psychology, sensory, and vitality. IC was measured by the multidimensional domain evaluation tool which was recommended by the WHO for integrated care of older adults. Evaluation criteria for IC: A score of 0 is assigned if the ability in a certain field decline, and a score of 1 if it remains normal. The total score of the IC composite index ranges from 0 to 5, with a higher score indicating better IC. The evaluation criteria are detailed in Appendix B. The total IC score was obtained by summing the scores from these five domains, categorizing IC into two groups: (a) the maintenance group (total IC score = 5) and (b) the decline group (total IC score < 5) (B. Zhou et al., 2025a ). The evaluation tools for each domain are as follows: (1) Self-reported auditory and visual functions Self-reported auditory and visual functions were used to measure the sensory domain. Participants were asked whether they had any visual or hearing impairments and if these impairments affect their daily lives. (2) The Short Physical Performance Battery (SPPB) The Short Physical Performance Battery (SPPB), originally developed by Guralnik (Guralnik et al., 1994 ) and subsequently translated into Chinese by Li (W. Li et al., 2023 ), was used as the measurement tool for the locomotor domain. This scale assessed lower limb function in older adults and evaluates three key dimensions: balance tests (3 items), chair stand tests (1 item), and gait speed tests (1 item). Each test is scored from 0 to 4, with a total score ranging from 0 to 12. Higher scores indicate better locomotor function. The Cronbach’s α coefficient for SPPB in this study was 0.81. (3) The Mini Nutritional Assessment (MNA) The Mini Nutritional Assessment (MNA), originally developed by Guigoz (Guigoz, 2006 ) and later translated into Chinese by Lei (Lei et al., 2009 ), was employed to measure the vitality domain. This scale consists of four dimensions with a total of 18 items: anthropometric measurements (3 items), general health status (6 items), dietary intake (7 items), and subjective health assessment (2 items). Higher scores indicate better nutritional status. The Cronbach’s α coefficient for MNA in this study was 0.71. (4) The Mini-Cog The Chinese version of the Mini-Cog developed and translated into Chinese by Borson (Borson et al., 2000 ) is a brief cognitive test that involves an assessment of an older person’s ability to recall three words and draw a clock. In this study, the Cronbach’s α coefficient for Mini-cog was 0.89. (5) 15 - item Geriatric Depression Scale (GDS-15) The Chinese version of the 15-item Geriatric Depression Scale (GDS-15), originally developed by Yesavage (Yesavage et al., 1982 ) and adapted for Chinese populations by Zhang (C. Zhang et al., 2022 ), was applied to measure the psychological domain. This scale evaluates the emotional state of participants over the past week which includes unhappiness (5 items), apathy and anxiety (5 items), loss of hope and morale (3 items), loss of social activity and memory (2 items). The total score ranges from 0 to 15, with higher scores indicating more severe depression. Cronbach’s α coefficient for GDS-15 in this study was 0.75. 2.3 Candidate predictors To identify initial predictors associated with IC, we conducted a literature review and applied the health ecology model. This model conceptualizes influences on health across five concentric layers: demographic factors, health behaviors, social networks, living/working conditions, and social policies. To initially identify predictors associated with IC, a thorough review of the relevant literature was performed, and the theory of health ecology model was also applied. This model consists of five layers from the inside to the outside, namely the demographic factors, health behavior factors, social network factors, living and working conditions factors, and social policy factors (Fan Tao et al., 2012 ). The factors across these five layers interact dynamically, ultimately shaping human health outcomes. This theoretical model led to the inclusion of 23 baseline variables in the analytical framework, encompassing five demographic factors (gender, age, household registration, current residence, number of children); five health behavior factors (health literacy, smoking, drinking, sleep quality, number of chronic diseases); two social network factors (marital status, social network); ten living and working condition factors (education, previous career, reemployment, primary source of financial support in the host location, economic level, migration scope, migration reason, migration status, migration duration, stay willingness); and one social policy factor, that is the multi-layered medical insurance coverage (whether through portable benefits from their place of origin, local schemes in the destination, or both) (see Appendix A). The health literacy scale used in this study is the 2015 Chinese Resident Health Literacy Questionnaire, further developed by the National Health and Family Planning Commission (Chen, Shuqi, 2017). It covers three dimensions, with the distribution of questions as follows: 24 items assessing basic health knowledge and concepts, 18 items on healthy lifestyles and behaviors, and 14 items measuring basic health skills. The social network was assessed by The Elderly Individual Social Network Scale which is developed by Tao Shengsheng and his research team from Anhui Medical University in 2019 (Tao, Shengsheng, 2019) was adopted. It includes 32 items across four dimensions: health guardianship and emotional support, social participation, social interaction, and social support. 2.4 Statistical analyses Characteristics of participants by the maintenance in IC group and the IC decline group were compared by using chi-square test for categorical variables and t-test for continuous variables. Important risk factors were identified by importance matrix plot of the RF model, with importance of each factor ranking in descending order. After features were selected by importance matrix plot, this study adopted RF, ANN, CART and Logit algorithms to construct predictive models for predicting risk of decline in intrinsic capacity. The dataset was randomly split into a training set (2/3 participants) in which models were conducted, and a validation set (1/3 participants) in which model performance was assessed, by using area under the ROC curve (AUC), classification accuracy (ACC), balanced error rate (BER), false positive rate (FPR), positive predictive value (PPV) and negative predictive value (NPV) as the metrics to evaluate of predictive models. Furthermore, PDPs of each predictor were drawn to help explain the relationship between each predictor and risk of IC decline. To further validate this finding, we conducted additional sensitivity analyses using different feature selection methods, confirming the robustness of our results. Data analyses were performed with SPSS (27.0 version) and R (version 4.4.3). 3 Results 3.1 Sociodemographic characteristics of participants A total of 3055 participants completed the questionnaire, with 39 excluded for missing values, resulting in a 98.7% recovery rate and a final sample of 3016 MOAC. Among these MOAC, 1593 (52.8%) were in the IC decline group. Baseline characteristics of the study population are shown in Table 1 . The mean age of all MOAC was 67.22 ± 6.95. Table 1 Sociodemographic characteristics of the decline intrinsic capacity among the MOAC Characteristics Total the decline group the maintenance group P Total 3016(100) 1423(47.2) 1593(52.8) Gender 0.326 Male 1404(46.6) 117(47.8) 100(40.0) Female 1612(53.4) 128(52.2) 150(60.0) Age <0.001 60–69 2243(74.4) 876(61.6%) 1367(85.8%) 70–79 627(20.8) 418(29.4%) 209(13.1%) ≥ 80 146(4.8) 129(9.1%) 17(1.1%) Education <0.001 ≤Primary school 1744(57.8) 901(63.3%) 843(52.9%) Middle school and high school 1207(40.0) 495(34.8%) 712(44.7%) ≥College 65(2.2) 27(1.9%) 38(2.4%) Marital status <0.001 Married 2772(91.9) 1242(87.3%) 1530(96%) Unmarried 18(0.6) 13(0.9%) 5(0.3%) Divorced 31(1.0) 16(1.1%) 15(0.9%) Bereaved 195(6.5) 152(10.7%) 43(2.7%) Number of children <0.001 0 12(0.4%) 11(0.8%) 1(0.1%) 1 759(25.2%) 322(22.6%) 437(27.4%) 2 1401(46.5%) 608(42.7%) 793(49.8%) ≥ 3 844(28%) 482(33.9%) 362(22.7%) Household registration 0.006 Rural 2400(79.6%) 1102(77.4%) 1298(81.5%) Urban 616(20.4%) 321(22.6%) 295(18.5%) Current residence 0.026 Rural 53(1.8) 33(2.3%) 20(1.3%) Urban 2963(98.2) 1390(97.7%) 1573(98.7%) Migration scope 0.055 Cross the district 315(10.4) 158(11.1%) 157(9.9%) Cross the county 556(18.4) 238(16.7%) 318(20%) Cross the province 2145(71.1) 1027(72.2%) 1118(70.2%) Migration reason <0.001 Look after the younger generation 1345(44.6%) 477(33.5%) 868(54.5%) Provide for the aged 426(14.1%) 277(19.5%) 149(9.4%) Therapy 126(4.2%) 108(7.6%) 18(1.1%) Family reunions 853(28.3%) 446(31.3%) 407(25.5%) Others 266(8.8%) 115(8.1%) 151(9.5%) Migration status 0.004 Personal migration 2327(77.2%) 358(25.2%) 331(20.8%) Migrate with spouse 689(22.8%) 1065(74.8%) 1262(79.2%) Migration duration 0.011 0.5 to 1 (including 1 year) 424(14.1) 212(14.9%) 212(13.3%) 1 to 3(including 3 years) 667(22.1) 342(24%) 325(20.4%) 3 to 6 (including 6 years) 659(21.9) 316(22.2%) 343(21.5%) 6 to 9 (including 9 years) 342(11.3) 142(10%) 200(12.6%) >9 924(30.6) 411(28.9%) 513(32.2%) Stay willingness 0.838 Yes 1923(63.8) 910(63.9%) 1013(63.6%) No 1093(36.2) 513(36.1%) 580(36.4%) Previous career <0.001 Personnel in public institutions 419(13.9) 261(18.3%) 158(9.9%) Self-employed workers 248(8.2) 96(6.7%) 152(9.5%) Non-agricultural workers 1215(40.3) 412(29%) 803(50.4%) Peasant 1131(37.5) 653(45.9%) 478(30%) Others 3(0.1) 1(0.1%) 2(0.1%) Reemployment <0.001 Yes 666(22.1) 215(15.1%) 451(28.3%) No 2350(77.9) 1208(84.9%) 1142(71.7%) Primary source of financial support in the host location <0.001 Pension 635(21.1%) 350(24.6%) 285(17.9%) Child maintenance 969(32.1%) 534(37.5%) 435(27.3%) Labor income 1061(35.2%) 387(27.2%) 674(42.3%) Old-age insurance 315(10.4%) 132(9.3%) 183(11.5%) Others 36(1.2%) 20(1.4%) 16(1%) Economic level <0.001 Needy 226(7.5) 189(13.3%) 37(2.3%) Below the average 685(22.7) 419(29.4%) 266(16.7%) Medium 1717(56.9) 636(44.7%) 1081(67.9%) Above the average 334(11.1) 149(10.5%) 185(11.6%) Affluent 54(1.8) 30(2.1%) 24(1.5%) Health literacy 32.78 ± 12.05 40.75 ± 11.87 <0.001 Social participation 6.47 ± 3.99 8.65 ± 3.57 <0.001 Social network 91.58 ± 16.80 110.39 ± 23.84 <0.001 Smoking 0.969 Yes 764(25.3) 1063(74.7%) 1189(74.6%) No 2252(74.7) 360(25.3%) 404(25.4%) Drinking 0.171 Yes 869(28.8) 1030(72.4%) 1117(70.1%) No 2147(71.2) 393(27.6%) 476(29.9%) Sleep quality <0.001 Good 1502(49.8) 420(29.5%) 1082(67.9%) Commonly 1174(38.9) 731(51.4%) 443(27.8%) Poor 316(10.5) 253(17.8%) 63(4%) Very poor 24(0.8) 19(1.3%) 5(0.3%) Multi-layered medical insurance coverage <0.001 Employee medical insurance 541(17.9) 302(21.2%) 239(15%) Resident medical insurance 399(13.2) 155(10.9%) 244(15.3%) New rural insurance 1774(58.8) 835(58.7%) 939(58.9%) Other insurance 83(2.8) 26(1.8%) 57(3.6%) No insurance 219(7.3) 105(7.4%) 114(7.2%) Number of chronic diseases <0.001 0 1391(41.6) 464(32.6%) 927(58.2%) 1 1035(34.3) 528(37.1%) 507(31.8%) ≥ 2 590(19.6) 431(30.3%) 159(10%) 3.2 Feature selection This study employed the importance matrix plot, a wrapper algorithm that leveraged the random forest algorithm to rank variable importance, for feature selection. The importance matrix plot employs the Gini index as the metric for split point identification—an indicator representing the probability of misclassifying a randomly selected sample in the dataset. Factors that have a significant impact on the decline in intrinsic abilities among the MOAC were screened out using the importance matrix plot, including social network, health literacy, sleep quality, migration reason, economic level, number of chronic diseases, age, primary source of financial support in the host location, previous career, multi-layered medical insurance coverage (see Fig. 1 ). 3.3 Model training and evaluation By using factors identified important by importance matrix plot, predictive models were developed by using RF, ANN, CART and Logit method in training dataset, and predictive accuracy of those models in test dataset is presented in Fig. 2 and Table 2 . As RF model performed the best, we developed the predictive model for predicting risk of IC decline among the MOAC by RF model, with mtree of 5 and ntree of 500. For model performance, the AUC was 0.843 (95% CI : 0.820–0.867), the sensitivity was 0.772, and the specificity was 0.772, suggesting that the predictive model had a good prediction effect (see Fig. 3 ). Table 2 Comparison of predictive accuracy of four machine learning models in test dataset. MODEL SENSITIVITY SPECIFICITY ACC BER FRP PPV NPV RF 74.47 77.97 76.32 23.78 22.03 75.11 77.38 Logistic 71.94 76.65 74.43 25.71 23.35 73.33 75.37 CART 73.84 72.50 73.13 26.83 27.50 70.56 75.64 ANN 73.00 76.84 75.02 25.08 23.16 73.77 76.12 3.4 Elucidating Feature-Prediction Associations PDPs, an agnostic method, reveal how changes in a single feature influence prediction, aiding in understanding the marginal effect of a variable on the predicted outcome of an ML model. The result presented the non-linear effect of each predictor on the risk of IC decline after adjusting for all other factors. Social network and economic level roughly showed a V-shaped curve with the risk of IC decline (Fig. 4 ). As health literacy increases, the risk of IC decline also decreases. Both the increase in the number of chronic diseases and advancing age elevate the risk of IC decline, while the deterioration of sleep quality also contributes to a higher risk of IC decline. Differences in migration reasons, primary source of financial support in the host location, previous career and multi-layered medical insurance coverage are associated with variations of the IC decline among the MOAC. Among them, MOAC who migrated to care for the younger generation, labor income, non-agricultural workers, and those with employee multi-layered medical insurance coverage exhibited the mildest IC decline. While those who migrated for therapeutic purposes, pension, personnel in public institutions, no insurance faced the highest risk of IC decline. Among them, MOAC who migrated to care for the younger generation, individuals with labor income, non-agricultural workers, and those covered by employee multi-layered medical insurance coverage exhibited the mildest IC decline. In contrast, those who migrated for therapeutic purposes, retirees relying on pensions, personnel in public institutions, and the uninsured faced the highest risk of IC decline. 4 Discussion Based on multi-center, large-sample data (N = 3,016), this study found that over half of the MOAC have experienced a IC decline, which leads to worsening function, lower quality of life, and poor health outcomes (Stolz et al., 2022b ; J. Yu et al., 2021 ; Y. Zhao et al., 2024 ). Therefore, this study conducted further analysis to identify key factors influencing IC decline. This was the first time that ML methods had been applied to predict the risk of IC decline among the MOAC. Based on the five dimensions of the health ecology model, namely the demographic factors, health behavior factors, social network factors, living and working conditions factors, and social policy factors, this study incorporated 23 variables. Meanwhile, the importance matrix plot was employed to identify the top 10 predictors with the most significant impact on the risk of IC decline. Four predictive models were developed in this study, and among them, the RF model showed the best predictive performance and clinical utility. Furthermore, after adjusting for all other factors, PDPs were employed to analyze the non-linear influence of these top predictors on the risk of IC decline, revealing the marginal effects of changes. This study employed an importance matrix plot to identify key influencing factors. The analysis revealed that social network, health literacy, sleep quality, migration reason, economic level, number of chronic diseases, primary source of financial support in the host location, previous career, multi-layered medical insurance coverage and age are the top ten most important risk factors for IC decline, which indicates that favorable individual health and social characteristics are beneficial for preventing IC decline. The selection of these influencing factors contributes to enhancing the credibility of practical ML applications (Degenhardt et al., 2019 ). In a comparison of four ML algorithms, the RF model achieved the highest accuracy in predicting IC decline, supporting its potential as a community-level screening tool for healthcare providers to identify who may be at risk of IC decline among the MOAC. RF models are like “black boxes” where only the inputs and outputs are visible, while the internal processes are obscured by complex calculations. PDPs clarify the relationship between a specific feature and the prediction outcome by keeping all other features constant. This method extends the interpretation of feature effects beyond simple linear regression and is applicable to more complex models (Q. Zhao & Hastie, 2021 ). PDPs effectively depict the relationship between the response variable and independent variables, regardless of whether the relationship is linear, monotonic, or complex. They show how the response variable changes as the value of an independent variable changes, while accounting for the average effects of all other independent variables in the model (Boehmke & Greenwell, 2019 ). Therefore, the PDPs method provides a clear and robust way to interpret complex models because they are easy to implement and simple to explain. The study visualized the results via PDPs, revealing that social network and economic level exhibit a V-shaped curve in relation to the risk of IC decline. The risk of IC decline decreases with the increase in social network (Guo et al., 2024 ) (Sun et al., 2025 ); however, 175 points is a turning point, beyond which a larger social network leads to an increase in the risk of IC decline. This may be attributed to either excessive reliance on others within the social network or excessive social engagement that leads to mental and physical exhaustion, both of which contribute to a IC decline (Arafat & Thoma, 2024 ; Qin et al., 2024 ). Economic level showed a similar pattern: risk declined until a medium level was reached, then rose among affluent groups When the economic level was below the medium level, the higher the economic level, the lower the risk of IC decline. When the economic level was above average and affluent, the risk of IC decline increased instead (Salinas-Rodríguez et al., 2024 ). Within the non-modifiable dimensions of the health ecology model, age, previous career, number of chronic diseases and migration reasons were identified as key factors influencing the risk of IC decline among the MOAC. Specifically, MOAC who are elder (R. Yu et al., 2023 , 2024 ), non-agricultural workers, (C.-Y. Li et al., 2002 ), have multiple chronic conditions (M. Zhang et al., 2025 ), or migrated for medical therapy (Willen et al., 2021 ) face a higher risk of IC decline. Therefore, community primary healthcare providers should prioritize these populations for early screening to enable timely identification and intervention of related risks. Within the health ecology model, health literacy stands out as the foremost modifiable factor protective against IC decline, the finding consistent with prior research (Q. Liu et al., 2024 ). Given this primary status, community healthcare providers should adopt a proactive strategy that integrates health literacy promotion into social activities and health education programs, aiming to improve health outcomes for the MOAC. This study also analyzed other modifiable factors, specifically the primary source of financial support in the host location, access to multi-layered medical insurance coverage, and sleep quality. The results indicate that MOAC with poorer sleep quality, reliance on self-earned income, or lack of adequate multi-layered medical insurance coverage face a heightened risk of IC decline, which aligns with existing evidence (Kang et al., 2025 ; Rodriguez & Saenz, 2022 ; Yang et al., 2025 ; B. Zhou et al., 2025b ). This elevated risk likely stems from an interconnected decline: age and multimorbidity impair sleep and physiological resilience, while insufficient financial support and inadequate insurance coverage exacerbate psychological vulnerability (Arokiasamy et al., 2021 ), collectively driving the risk of IC decline. To summarize, this study compared several ML algorithms and identified the RF model as having the optimal predictive performance. After screening out the top ten key predictors of IC decline among the MOAC population, the study further used PDPs to make up for the inherent “black-box” limitation of the RF model. This approach clearly delineated the core impact trends of these ten factors on IC decline, thereby providing a robust scientific foundation for formulating precise intervention strategies to improve the IC among the MOAC. Regarding non-modifiable risk factors, this study identifies five high-risk subgroups of MOAC requiring prioritization by community primary healthcare providers: older individuals, those with multimorbidity, migrants for medical therapy, have a previous career as non-agricultural workers, and those without adequate multi-layered medical insurance coverage in the host location. Furthermore, guided by the health ecology model, a multilevel intervention pathway is proposed: a) At the individual level, healthcare providers can strengthen health literacy and improve sleep quality. b) At the social/community level, tailored activities should be designed to expand meaningful social networks and enhance social engagement. c) At the family level, family members are encouraged to support MOAC’s living and working conditions, such as stabilizing economic status and diversifying income sources. Through such coordinated, multidimensional interventions, IC decline among the MOAC population can be effectively delayed, supporting sustained improvements in their quality of life. 5 Limitations This study has several limitations. First, the cross-sectional design limits the ability to establish causal relationships among the identified risk factors for IC decline. Future longitudinal studies are needed to further explore these causal links. Second, while recent research has identified promising biomarkers associated with IC (Fang et al., 2025 ; Guyonnet et al., 2025 ), this study did not incorporate such biological measures. Subsequent research should aim to identify accessible biomarkers for comprehensively assessing IC decline and integrate them to improve the accuracy of predictive models. 6 Conclusion This study demonstrates that the RF model provides superior accuracy in predicting IC decline among the MOAC. Guided by the health ecology model, communities healthcare providers should pay more attention to screen the high-risk individuals based on key non-modifiable factors, including advanced age, multimorbidity, a non‑agricultural work history, and migrant for medical reasons. Furthermore, tailored multilevel interventions targeting modifiable determinants should be designed to improve IC, with a focusing on behavioral and interpersonal factors such as health literacy, social networks, sleep quality, diversified income sources, and adequate multi-layered medical insurance coverage in the host location for MOAC. Declarations Declaration of Conflicting Interests All authors certify that they have no affiliations with or involvement in any organization or entity with any financial interest or non-financial interest in the subject matter or materials discussed in this manuscript. Funding information This work was supported by the National Social Science Fund of China (22BGL251). Consent Informed consent was obtained from all individual participants included in the study participants signed informed consent regarding publishing their data. This study did not involve participant identity images or other personal or clinical details that would compromise anonymity. Ethics approval Approval was obtained from the Institutional Review Board of Huzhou University (IRB No. 2022-03-05). The procedures used in this study adhere to the tenets of the Declaration of Helsinki. Availability of Data and Materials All data generated or analyzed during this study are included in this manuscript. The data that support the findings of this study are available from the corresponding author upon reasonable request. Clinical trial number Not applicable. Consent for publication Not Applicable. References Arafat, D., & Thoma, P. (2024). Impairments of Sociocognitive Functions in Individuals with Behavioral Addictions: A Review Article. Journal of Gambling Studies , 40 (2), 429–451. https://doi.org/10.1007/s10899-023-10227-w Arokiasamy, P., Selvamani, Y., Jotheeswaran, A. T., & Sadana, R. (2021). Socioeconomic differences in handgrip strength and its association with measures of intrinsic capacity among older adults in six middle-income countries. Scientific Reports , 11 (1), 19494. https://doi.org/10.1038/s41598-021-99047-9 Bautmans, I., Knoop, V., Amuthavalli Thiyagarajan, J., Maier, A. B., Beard, J. R., Freiberger, E., Belsky, D., Aubertin-Leheudre, M., Mikton, C., Cesari, M., Sumi, Y., Diaz, T., Banerjee, A., & WHO Working Group on Vitality Capacity. (2022). WHO working definition of vitality capacity for healthy longevity monitoring. The Lancet. Healthy Longevity , 3 (11), e789–e796. https://doi.org/10.1016/S2666-7568(22)00200-8 Boehmke, B., & Greenwell, B. (2019). Interpretable Machine Learning. In Hands-On Machine Learning with R . Chapman and Hall/CRC. Borson, S., Scanlan, J., Brush, M., Vitaliano, P., & Dokmak, A. (2000). The mini-cog: A cognitive “vital signs” measure for dementia screening in multi-lingual elderly. International Journal of Geriatric Psychiatry , 15 (11), 1021–1027. https://doi.org/10.1002/1099-1166(200011)15:11%253C1021::aid-gps234%253E3.0.co;2-6 Burmeister, E., & Aitken, L. M. (2012). Sample size: How many is enough? Australian Critical Care: Official Journal of the Confederation of Australian Critical Care Nurses , 25 (4), 271–274. https://doi.org/10.1016/j.aucc.2012.07.002 Chen, Shuqi. (2017). Study on Change Tendency of Health Literacy and ItsInfluencing Factors among Rural Residents in KunshanCity [Master’s Thesis, Soochow University]. https://kns.cnki.net/KCMS/detail/detail.aspx?dbcode=CMFD&dbname=CMFD201701&filename=1016212474.nh Degenhardt, F., Seifert, S., & Szymczak, S. (2019). Evaluation of variable selection methods for random forests and omics data sets. Briefings in Bioinformatics , 20 (2), 492–503. https://doi.org/10.1093/bib/bbx124 Fan Tao, Cao Qian, Jiang Lulu, & Xu Zhengzi. (2012). Explaining the Influencing Factors on Chronic Diseases of the Elderly Using Health Ecological Model. Chinese General Practice , 15 (1), 33–36, 40. Fang, J., Tan, J., Chen, X., Hua, K., Ding, Y., Wu, Y., Tan, L., Hua, Q., Shao, S., Liu, T., & Hua, Y. (2025). Association between homocysteine levels and decreased intrinsic capacity in Chinese community-dwelling older adults: A cross-sectional study. Medicine , 104 (32), e43771. https://doi.org/10.1097/MD.0000000000043771 Guigoz, Y. (2006). The Mini Nutritional Assessment (MNA) review of the literature—What does it tell us? The Journal of Nutrition, Health & Aging , 10 (6), 466–485; discussion 485-487. Guo, W., Meng, L., Han, J., Yang, B., Sun, J., Guo, Y., Wu, J., & Liu, Y. (2024). Intrinsic capacity and its association with predictors among Chinese empty nest older adults in communities: A latent class analysis. BMC Geriatrics , 24 (1), 1008. https://doi.org/10.1186/s12877-024-05583-0 Guralnik, J. M., Simonsick, E. M., Ferrucci, L., Glynn, R. J., Berkman, L. F., Blazer, D. G., Scherr, P. A., & Wallace, R. B. (1994). A short physical performance battery assessing lower extremity function: Association with self-reported disability and prediction of mortality and nursing home admission. Journal of Gerontology , 49 (2), M85-94. https://doi.org/10.1093/geronj/49.2.m85 Guyonnet, S., Hooper, C., Cantet, C., Coley, N., Andrieu, S., Vellas, B., & IHU HealthAge INSPIRE/Open Science study group. (2025). The INSPIRE-T longitudinal observational study centred on intrinsic capacity: Baseline data. The Journals of Gerontology. Series A, Biological Sciences and Medical Sciences , glaf181. https://doi.org/10.1093/gerona/glaf181 Hernandes, E. C. R., Aliberti, M. J. R., Guerra, R. O., Ferriolli, E., & Perracini, M. R. (2024). Intrinsic capacity and hospitalization among older adults: A nationally representative cross-sectional study. European Geriatric Medicine , 15 (3), 843–852. https://doi.org/10.1007/s41999-024-00933-y Holecki, T., Rogalska, A., Sobczyk, K., Woźniak-Holecka, J., & Romaniuk, P. (2020). Global Elderly Migrations and Their Impact on Health Care Systems. Frontiers in Public Health , 8 , 386. https://doi.org/10.3389/fpubh.2020.00386 Kang, Y., Shi, H., Zhang, J., Meng, X., Zhang, C., Shen, J., & Zhang, P. (2025). The Bidirectional Relationship Between Intrinsic Capacity and Catastrophic Health Expenditure in China: A Longitudinal Study. The Journals of Gerontology. Series A, Biological Sciences and Medical Sciences , 80 (3), glae301. https://doi.org/10.1093/gerona/glae301 Lee, W.-J., Peng, L.-N., Lin, M.-H., Loh, C.-H., Hsiao, F.-Y., & Chen, L.-K. (2023). Intrinsic capacity differs from functional ability in predicting 10-year mortality and biological features in healthy aging: Results from the I-Lan longitudinal aging study. Aging , 15 (3), 748–764. https://doi.org/10.18632/aging.204508 Lei, Z., Qingyi, D., Feng, G., Chen, W., Hock, R. S., & Changli, W. (2009). Clinical study of mini-nutritional assessment for older Chinese inpatients. The Journal of Nutrition, Health & Aging , 13 (10), 871–875. https://doi.org/10.1007/s12603-009-0244-1 Li, C.-Y., Wu, S. C., & Sung, F.-C. (2002). Lifetime principal occupation and risk of cognitive impairment among the elderly. Industrial Health , 40 (1), 7–13. https://doi.org/10.2486/indhealth.40.7 Li, H., Pang, M., Wang, J., Xu, J., & Kong, F. (2022). Effects of Health Service Utilization and Informal Social Support on Depression, Anxiety, and Stress among the Internal Migrant Elderly following Children in Weifang, China. International Journal of Environmental Research and Public Health , 19 (22), 14640. https://doi.org/10.3390/ijerph192214640 Li, W., Rao, Z., Fu, Y., Schwebel, D. C., Li, L., Ning, P., Huang, J., & Hu, G. (2023). Value of the short physical performance battery (SPPB) in predicting fall and fall-induced injury among old Chinese adults. BMC Geriatrics , 23 (1), 574. https://doi.org/10.1186/s12877-023-04290-6 Liang, H., Zheng, J., & Sun, Y. (2023). Prevalence and predictive risk factors of self-stigmatization in patients with rare diseases: Based on health ecology model. The International Journal of Social Psychiatry , 69 (6), 1541–1545. https://doi.org/10.1177/00207640231168023 Liu, Q., Li, X., Hu, M., Zhao, Y., Wu, S., & Feng, H. (2024). Factors influencing the self-management ability among older adults experiencing intrinsic capacity decline: A cross-sectional study. Frontiers in Aging Neuroscience , 16 , 1456167. https://doi.org/10.3389/fnagi.2024.1456167 Liu, S., Kang, L., Liu, X., Zhao, S., Wang, X., Li, J., & Jiang, S. (2021). Trajectory and Correlation of Intrinsic Capacity and Frailty in a Beijing Elderly Community. Frontiers in Medicine , 8 , 751586. https://doi.org/10.3389/fmed.2021.751586 Ma, L., Chhetri, J. K., Zhang, L., Sun, F., Li, Y., & Tang, Z. (2021). Cross-sectional study examining the status of intrinsic capacity decline in community-dwelling older adults in China: Prevalence, associated factors and implications for clinical care. BMJ Open , 11 (1), e043062. https://doi.org/10.1136/bmjopen-2020-043062 Muneera, K., Muhammad, T., & Althaf, S. (2022). Socio-demographic and lifestyle factors associated with intrinsic capacity among older adults: Evidence from India. BMC Geriatrics , 22 (1), 851. https://doi.org/10.1186/s12877-022-03558-7 Qin, C., Li, Y., Wang, T., Zhao, J., Tong, L., Yang, J., & Liu, Y. (2024). Too much social media? Unveiling the effects of determinants in social media fatigue. Frontiers in Psychology , 15 , 1277846. https://doi.org/10.3389/fpsyg.2024.1277846 Ramírez-Vélez, R., Correa-Bautista, J. E., García-Hermoso, A., Cano, C. A., & Izquierdo, M. (2019). Reference values for handgrip strength and their association with intrinsic capacity domains among older adults. Journal of Cachexia, Sarcopenia and Muscle , 10 (2), 278–286. https://doi.org/10.1002/jcsm.12373 Rodriguez, F. S., & Saenz, J. (2022). Working in old age in Mexico: Implications for Cognitive Functioning. Ageing and Society , 42 (11), 2489–2509. https://doi.org/10.1017/s0144686x2100012x Salinas-Rodríguez, A., Fernández-Niño, J. A., Rivera-Almaraz, A., & Manrique-Espinoza, B. (2024). Intrinsic capacity trajectories and socioeconomic inequalities in health: The contributions of wealth, education, gender, and ethnicity. International Journal for Equity in Health , 23 (1), 48. https://doi.org/10.1186/s12939-024-02136-0 Salinas-Rodríguez, A., González-Bautista, E., Rivera-Almaraz, A., & Manrique-Espinoza, B. (2022). Longitudinal trajectories of intrinsic capacity and their association with quality of life and disability. Maturitas , 161 , 49–54. https://doi.org/10.1016/j.maturitas.2022.02.005 Stafford, I. S., Kellermann, M., Mossotto, E., Beattie, R. M., MacArthur, B. D., & Ennis, S. (2020). A systematic review of the applications of artificial intelligence and machine learning in autoimmune diseases. NPJ Digital Medicine , 3 , 30. https://doi.org/10.1038/s41746-020-0229-3 Stephens, C., Allen, J., Keating, N., Szabó, Á., & Alpass, F. (2020). Neighborhood environments and intrinsic capacity interact to affect the health-related quality of life of older people in New Zealand. Maturitas , 139 , 1–5. https://doi.org/10.1016/j.maturitas.2020.05.008 Stolz, E., Mayerl, H., Freidl, W., Roller-Wirnsberger, R., & Gill, T. M. (2022a). Intrinsic Capacity Predicts Negative Health Outcomes in Older Adults. The Journals of Gerontology. Series A, Biological Sciences and Medical Sciences , 77 (1), 101–105. https://doi.org/10.1093/gerona/glab279 Stolz, E., Mayerl, H., Freidl, W., Roller-Wirnsberger, R., & Gill, T. M. (2022b). Intrinsic Capacity Predicts Negative Health Outcomes in Older Adults. The Journals of Gerontology. Series A, Biological Sciences and Medical Sciences , 77 (1), 101–105. https://doi.org/10.1093/gerona/glab279 Sun, J., Amnatsatsue, K., Srisuwan, P., Kerdmongkol, P., & Nityasuddhi, D. (2025). Intrinsic Capacity of Chinese Community-Dwelling Older Adults Using WHO Integrated Care for Older People (ICOPE) Framework: Structural Equation Model Analysis. Journal of Primary Care & Community Health , 16 , 21501319251346433. https://doi.org/10.1177/21501319251346433 Tao, Shengsheng. (2019). Research on Social Network Measurement Index System of OldPeople under Modern Medical Model [Master’s Thesis, Anhui Medical University]. https://kns.cnki.net/KCMS/detail/detail.aspx?dbcode=CMFD&dbname=CMFD201902&filename=1019175215.nh Tian, T., Kong, F., & Li, S. (2021). Effects of Living Conditions, Subjective Integration, and Social Networks on Health-Related Quality of Life among the Migrant Elderly Following Children in Jinan, China. Healthcare (Basel, Switzerland) , 9 (4), 414. https://doi.org/10.3390/healthcare9040414 Tokodi, M., Schwertner, W. R., Kovács, A., Tősér, Z., Staub, L., Sárkány, A., Lakatos, B. K., Behon, A., Boros, A. M., Perge, P., Kutyifa, V., Széplaki, G., Gellér, L., Merkely, B., & Kosztin, A. (2020). Machine learning-based mortality prediction of patients undergoing cardiac resynchronization therapy: The SEMMELWEIS-CRT score. European Heart Journal , 41 (18), 1747–1756. https://doi.org/10.1093/eurheartj/ehz902 Tu, W.-J., Zeng, X., & Liu, Q. (2022). Aging tsunami coming: The main finding from China’s seventh national population census. Aging Clinical and Experimental Research , 34 (5), 1159–1163. https://doi.org/10.1007/s40520-021-02017-4 Willen, S. S., Selim, N., Mendenhall, E., Lopez, M. M., Chowdhury, S. A., Dilger, H., & Migration and Health in Social Context Working Group. (2021). Flourishing: Migration and health in social context. BMJ Global Health , 6 (Suppl 1), e005108. https://doi.org/10.1136/bmjgh-2021-005108 Yang, C., Chen, X., Wan, X., & Cai, Y. (2025). The association between sleep duration trajectories and intrinsic capacity in middle-aged and older adults in China: A longitudinal Chinese study assessing healthy aging. Frontiers in Medicine , 12 , 1595241. https://doi.org/10.3389/fmed.2025.1595241 Yesavage, J. A., Brink, T. L., Rose, T. L., Lum, O., Huang, V., Adey, M., & Leirer, V. O. (1982). Development and validation of a geriatric depression screening scale: A preliminary report. Journal of Psychiatric Research , 17 (1), 37–49. https://doi.org/10.1016/0022-3956(82)90033-4 Yu, J., Si, H., Qiao, X., Jin, Y., Ji, L., Liu, Q., Bian, Y., Wang, W., & Wang, C. (2021). Predictive value of intrinsic capacity on adverse outcomes among community-dwelling older adults. Geriatric Nursing (New York, N.Y.) , 42 (6), 1257–1263. https://doi.org/10.1016/j.gerinurse.2021.08.010 Yu, R., Lai, D., Leung, G., Tam, L.-Y., Cheng, C., Kong, S., Tong, C., & Woo, J. (2024). Transitions in intrinsic capacity among community-dwelling older people and their associated factors: A multistate modelling analysis. The Journal of Nutrition, Health & Aging , 28 (7), 100273. https://doi.org/10.1016/j.jnha.2024.100273 Yu, R., Lai, D., Leung, G., & Woo, J. (2023). Trajectories of Intrinsic Capacity: Determinants and Associations with Disability. The Journal of Nutrition, Health & Aging , 27 (3), 174–181. https://doi.org/10.1007/s12603-023-1881-5 Zhang, C., Zhang, H., Zhao, M., Chen, C., Li, Z., Liu, D., Zhao, Y., & Yao, Y. (2022). Psychometric properties and modification of the 15-item geriatric depression scale among Chinese oldest-old and centenarians: A mixed-methods study. BMC Geriatrics , 22 (1), 144. https://doi.org/10.1186/s12877-022-02833-x Zhang, M., Xu, Y., Xing, Y., & Li, H. (2025). Association between multimorbidity and intrinsic capacity among older Chinese adults: Evidence from the CHARLS 2011-2015. European Geriatric Medicine , 16 (4), 1207–1216. https://doi.org/10.1007/s41999-025-01232-w Zhao, Q., & Hastie, T. (2021). Causal Interpretations of Black-Box Models. Journal of Business & Economic Statistics , 39 (1), 272–281. https://doi.org/10.1080/07350015.2019.1624293 Zhao, Y., Jiang, Y., Tang, P., Wang, X., Guo, Y., & Tang, L. (2024). Adverse health effects of declined intrinsic capacity in middle-aged and older adults: A systematic review and meta-analysis. Age and Ageing , 53 (7), afae162. https://doi.org/10.1093/ageing/afae162 Zhi, K., Chen, Y., & Huang, J. (2021). China’s challenge in promoting older migrants’ health and wellbeing: A productive ageing perspective. BMJ (Clinical Research Ed.) , 375 , n2874. https://doi.org/10.1136/bmj.n2874 Zhou, B., Ma, R., Wang, M., & Wang, Y. (2025a). Dose-response relationship between nighttime sleep duration and intrinsic capacity declines among Chinese elderly: A cross-sectional study from CHARLS. BMC Public Health , 25 (1), 1034. https://doi.org/10.1186/s12889-025-22294-2 Zhou, B., Ma, R., Wang, M., & Wang, Y. (2025b). Dose-response relationship between nighttime sleep duration and intrinsic capacity declines among Chinese elderly: A cross-sectional study from CHARLS. BMC Public Health , 25 (1), 1034. https://doi.org/10.1186/s12889-025-22294-2 Zhou, Y., & Ma, L. (2022). Intrinsic Capacity in Older Adults: Recent Advances. Aging and Disease , 13 (2), 353–359. https://doi.org/10.14336/AD.2021.0818 Additional Declarations No competing interests reported. Supplementary Files AppendixA.docx AppendixB.docx Cite Share Download PDF Status: Under Review Version 1 posted Reviews received at journal 08 Mar, 2026 Reviewers agreed at journal 28 Feb, 2026 Reviewers invited by journal 13 Feb, 2026 Editor assigned by journal 13 Feb, 2026 Editor invited by journal 22 Jan, 2026 Submission checks completed at journal 22 Jan, 2026 First submitted to journal 22 Jan, 2026 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-8610961","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":593339953,"identity":"d6497288-3e09-4b72-888f-69fc9325e764","order_by":0,"name":"Qinghua Zhang","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA0klEQVRIiWNgGAWjYBACNvb+hw8+VDAwG4B4PMRo4ec5w2w44wwpWiRn5LBJ87YxMBCvxeBA7jHJmfPq2M0lEhgfvG1jkDcnrOVcssXHbWzMljMSmA3ntjEY7mwgpOVgg+HNmdt4mA1uJIBdmGBwgJCWwwwG0rxzJEBa2H8TpUWyjcdImrfBAGwLM1Fa+HnYkg1nHEtgNjjzsFlyzjkJww2EtLDJPz744ENNXbLB8eSDH96U2cgTtAUGkhkYGBuAtASR6oHAjnilo2AUjIJRMOIAAEeKPrYxPhS9AAAAAElFTkSuQmCC","orcid":"","institution":"School of Medicine \u0026 Nursing Sciences, Huzhou University, 759-Second Ring East Road","correspondingAuthor":true,"prefix":"","firstName":"Qinghua","middleName":"","lastName":"Zhang","suffix":""},{"id":593339956,"identity":"11bbfc22-b57f-4415-9f54-1d9aba58b7a6","order_by":1,"name":"Yu Wang","email":"","orcid":"","institution":"School of Medicine \u0026 Nursing Sciences, Huzhou University, 759-Second Ring East Road","correspondingAuthor":false,"prefix":"","firstName":"Yu","middleName":"","lastName":"Wang","suffix":""},{"id":593339957,"identity":"f103ee76-417e-4b3a-b8be-ac81305acdbb","order_by":2,"name":"Xiaoxiao Hu","email":"","orcid":"","institution":"School of Medicine \u0026 Nursing Sciences, Huzhou University, 759-Second Ring East Road","correspondingAuthor":false,"prefix":"","firstName":"Xiaoxiao","middleName":"","lastName":"Hu","suffix":""},{"id":593339958,"identity":"824c34cb-3d9d-4504-80e6-37eb95c10aa2","order_by":3,"name":"Xinuo Yao","email":"","orcid":"","institution":"School of Medicine \u0026 Nursing Sciences, Huzhou University, 759-Second Ring East Road","correspondingAuthor":false,"prefix":"","firstName":"Xinuo","middleName":"","lastName":"Yao","suffix":""},{"id":593339961,"identity":"c6d334c7-715e-4c0a-bda6-31a948d7add8","order_by":4,"name":"Yuhan Yang","email":"","orcid":"","institution":"School of Medicine \u0026 Nursing Sciences, Huzhou University, 759-Second Ring East Road","correspondingAuthor":false,"prefix":"","firstName":"Yuhan","middleName":"","lastName":"Yang","suffix":""},{"id":593339965,"identity":"83d23719-e925-444d-a746-487d6711779c","order_by":5,"name":"Danyan Lu","email":"","orcid":"","institution":"School of Medicine \u0026 Nursing Sciences, Huzhou University, 759-Second Ring East Road","correspondingAuthor":false,"prefix":"","firstName":"Danyan","middleName":"","lastName":"Lu","suffix":""},{"id":593339968,"identity":"eaec1d25-2629-4d93-a3fb-db50329091f5","order_by":6,"name":"Shengguang Chen","email":"","orcid":"","institution":"School of Medicine \u0026 Nursing Sciences, Huzhou University, 759-Second Ring East Road","correspondingAuthor":false,"prefix":"","firstName":"Shengguang","middleName":"","lastName":"Chen","suffix":""},{"id":593339972,"identity":"888bdeaa-c210-414d-a4bc-20bda127ad7b","order_by":7,"name":"Xiaoyu Chen","email":"","orcid":"","institution":"School of Medicine \u0026 Nursing Sciences, Huzhou University, 759-Second Ring East Road","correspondingAuthor":false,"prefix":"","firstName":"Xiaoyu","middleName":"","lastName":"Chen","suffix":""}],"badges":[],"createdAt":"2026-01-15 13:38:19","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-8610961/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-8610961/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":102991858,"identity":"852d78f5-7aff-4bbb-ab49-7bcf8088c299","added_by":"auto","created_at":"2026-02-19 11:35:28","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":62020,"visible":true,"origin":"","legend":"\u003cp\u003eDeterminants of decline in IC identified by the importance matrix plot in MOAC.\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-8610961/v1/c75a0613dd8a55b5e6d5448f.png"},{"id":103050139,"identity":"c9c96eb7-beca-4cfe-bf7c-72e09e40b3aa","added_by":"auto","created_at":"2026-02-20 07:48:25","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":130231,"visible":true,"origin":"","legend":"\u003cp\u003eComparison of area under the ROC curve (AUC) of four machine learning models for predicting intrinsic capacity decline in the MOAC.\u003c/p\u003e\n\u003cp\u003eRF: Random Forest; Logit: ANN: Artificial Neural Network; CART: Classification and Regression Tree; Logit: Logistic Regression; AUC: area under the ROC curve.\u003c/p\u003e","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-8610961/v1/96c3ed3cd38989a99c30e842.png"},{"id":102991853,"identity":"699e3118-73b1-42cc-9d8b-4cb2737de835","added_by":"auto","created_at":"2026-02-19 11:35:27","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":68508,"visible":true,"origin":"","legend":"\u003cp\u003ePerformance of random forest (RF) prediction model for predicting intrinsic capacity decline in the MOAC.\u003c/p\u003e\n\u003cp\u003eAUC: area under the ROC curve.\u003c/p\u003e","description":"","filename":"3.png","url":"https://assets-eu.researchsquare.com/files/rs-8610961/v1/6adaebcb61294ae1d42cf3bf.png"},{"id":102991857,"identity":"c5cdda96-fffc-45b3-b74e-46a749d4739d","added_by":"auto","created_at":"2026-02-19 11:35:27","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":245354,"visible":true,"origin":"","legend":"\u003cp\u003ePartial Dependence Plots (PDPs) of each predictor for predicting risk of decline in IC among MOAC.\u003c/p\u003e\n\u003cp\u003eNote: Sleep quality (1: Very poor; 2: Poor; 3: Commonly; 4: Good); Migration reason (1: Look after the younger generation; 2: Provide for the aged; 3: Therapy; 4: Family reunions; 5: Others) Economic level (1: Needy; 2: Below the average; 3 Medium; 4 Above the average; 5 Affluent;\u003c/p\u003e\n\u003cp\u003eNumber of chronic diseases (1：0; 2：1; 3: ≥2); Source of finance (1: Pension; 2 Child maintenance; 3 Labor income; 4 Old-age insurance; 5: Others); Previous career (1: Personnel in public institutions; 2: Self-employed workers; 3: Non-agricultural workers; 4: Peasant; 5: Others); Medical insurance: (1: Employee medical insurance; 2: Resident medical insurance; 3: New rural insurance; 4: Other insurance; 5: No insurance); Age (1: 60-69; 2: 70-79; 3: ≥80);\u003c/p\u003e","description":"","filename":"4.png","url":"https://assets-eu.researchsquare.com/files/rs-8610961/v1/4ae1e6777e5a4dd0b3e987d7.png"},{"id":105036860,"identity":"dd20e44b-bc60-48e1-ba3b-adc481380140","added_by":"auto","created_at":"2026-03-20 07:36:23","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1686740,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8610961/v1/a0b1e3d1-5d94-4d5f-8f43-4a0aa8d83bac.pdf"},{"id":102991854,"identity":"6006e07e-c17d-4bec-a4d4-9a1f7ca63775","added_by":"auto","created_at":"2026-02-19 11:35:27","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":665643,"visible":true,"origin":"","legend":"","description":"","filename":"AppendixA.docx","url":"https://assets-eu.researchsquare.com/files/rs-8610961/v1/d6556477ef23fe2b043328a5.docx"},{"id":105032538,"identity":"02f395e1-37ec-49d5-bf2e-63006a8e50a7","added_by":"auto","created_at":"2026-03-20 06:59:58","extension":"docx","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":17721,"visible":true,"origin":"","legend":"","description":"","filename":"AppendixB.docx","url":"https://assets-eu.researchsquare.com/files/rs-8610961/v1/2e4f4d41f82f331ed9cbfebd.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"Machine learning-based predictive modeling of decline in intrinsic capacity among migrant older adults with children","fulltext":[{"header":"1 Introduction","content":"\u003cp\u003eThroughout the world, the number of migrant older adults is increasing dramatically (Holecki et al., \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2020\u003c/span\u003e), and this trend is particularly prominent in China. The Seventh National Population Census reports 375.82\u0026nbsp;million older adults, 18.7% of whom are aged 60 or above (Tu et al., \u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). This aging trend, combined with heightened internal mobility, has led to sustained migration across regions, thereby giving rise to a specific vulnerable population\u0026mdash;migrant older adults with children (MOAC) (Zhi et al., \u003cspan citationid=\"CR52\" class=\"CitationRef\"\u003e2021\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eThe MOAC are those older adults who reside long-term after moving to where their children live to take care of their grandchildren or to receive care from their children (Muneera et al., \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2022\u003c/span\u003e, \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2022\u003c/span\u003e; Stephens et al., \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). After leaving their familiar living environment, the MOAC have to adapt to a new setting and an unfamiliar social network. This adaptation leads to a series of cognitive and psychological challenges, as well as social isolation(H. Li et al., \u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e2022\u003c/span\u003e), which ultimately negatively impacts their intrinsic capacity (Tian et al., \u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e2021\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eAs defined by WHO, intrinsic capacity (IC) is \u0026ldquo;the composite of all the physical and mental capacities that an individual can draw on at any point in time\u0026rdquo; (Bautmans et al., \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e2022\u003c/span\u003e), whose core domains include sensation, locomotion, vitality, cognition and psychology (Y. Zhou \u0026amp; Ma, \u003cspan citationid=\"CR55\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). While existing research on community-dwelling older adults has identified personal traits, health conditions, and lifestyle habits as key correlates of IC (Ma et al., \u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; Muneera et al., \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2022\u003c/span\u003e; Stephens et al., \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e2020\u003c/span\u003e), studies focusing specifically on MOAC remain limited. Moreover, the selection of influencing factors in prior studies has often been insufficiently comprehensive, reflecting certain theoretical gaps. To tackle these limitations, this study employs the health ecology model to explore the factors associated with IC decline among MOAC in a more holistic and multidimensional manner.\u003c/p\u003e \u003cp\u003eThe health ecology model consists of five layers from the inside to the outside, namely demographic factors, health behavior factors, social network factors, living and working conditions factors, and social policy factors (Liang et al., \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). It serves as a theoretical framework that highlights the interactions between health and factors across multiple levels, thereby offering a structured approach to understanding the determinants of individual health. IC is individual and internal in nature, is not static. It evolves dynamically through interaction with the environment and can thus be viewed as the core \u0026ldquo;individual health asset\u0026rdquo; within this framework. Given the complexity of IC, the health ecology model provides a useful structure for characterizing an individual\u0026rsquo;s risk behaviors and helps systematically identify the factors associated with IC decline among MOAC (See Appendix A).\u003c/p\u003e \u003cp\u003eCurrent studies indicate that the rate of IC decline among older adults varies from 19.23% to 89.3% (Ram\u0026iacute;rez-V\u0026eacute;lez et al., \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e2019\u003c/span\u003e; Stolz et al., \u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e2022a\u003c/span\u003e). This decline is linked to adverse health outcomes such as falls, frailty, functional impairment, longer hospital stays, and increased mortality risk (Hernandes et al., \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e2024\u003c/span\u003e; Lee et al., \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; S. Liu et al., \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; Salinas-Rodr\u0026iacute;guez et al., \u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). Therefore, establishing a predictive model for IC decline is essential to accurately identify the MOAC with high-risk of IC decline.\u003c/p\u003e \u003cp\u003eMachine learning (ML) technology has provided advanced analytical methods for developing predictive systems. It enables more accurate classification and prediction by analyzing complex interactions within multi-dimensional datasets, while overcoming the limitations of traditional models that rely on linear assumptions (Tokodi et al., \u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). Additionally, ML can enhance predictive accuracy by assessing feature importance when handling high dimensional data (Stafford et al., \u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e2020\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eThe MOAC population, characterized by the dual vulnerability of aging and mobility, is expanding rapidly. Despite facing complex physiological, psychological, and social challenges, this group remains significantly understudied, and their risks of IC decline have been largely overlooked. To bridge this gap, this study evaluates and compares multiple ML algorithm\u0026mdash;including Random Forest (RF), Artificial Neural Network (ANN), Classification and Regression Tree (CART), and Logistic Regression (Logit)\u0026mdash;to develop an interpretable predictive model for early identification of IC decline among MOAC. The findings offer practical and actionable guidance for community healthcare providers, enabling early identification of high-risk individuals and informing the development of tailored interventions to support IC maintenance and quality of life improvement among MOAC.\u003c/p\u003e"},{"header":"2 Methods","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003e2.1 Study design and participants\u003c/h2\u003e \u003cp\u003eA multi-stage cluster random sampling method was adopted in China from December 2022 to September 2023. Firstly, Zhejiang Province, Jiangsu Province, and Shanghai Municipality were selected as primary sampling units based on their level of economic development. Secondly, three cities/districts were selected from each region as secondary sampling units: the XX Province sample included XX, XX, and XX cities; the XX Province sample included XX, XX, and XX cities; and the XX Municipality sample included XX New Area, XX District, and XX District. Thirdly, two subordinate towns were randomly selected from each city/district as tertiary sampling units. Finally, five communities were selected from each of the tertiary sampling units, totaling 90 communities. Due to the scattered distribution of the MOAC, we adopted a combination of purposive and snowball sampling methods to identify participants in local communities and expand the sample size.\u003c/p\u003e \u003cp\u003eAll participants should meet the following inclusion criteria: (1) Age\u0026thinsp;\u0026ge;\u0026thinsp;60 years; (2) Following their children from original residence to a different area in order to take care of their grandchildren or to receive care from their children; (3) Having lived there at least six months after crossing district/county/province boundaries were included; (4) Having voluntarily signed the informed consent form. The exclusion criteria included serious physical diseases, mental disorders or cognitive disorders. The data were collected by seven trained geriatric care professionals who were assessed by a professor in geriatric psychological care.\u003c/p\u003e \u003cp\u003eAccording to Kendall\u0026rsquo;s sample estimation method, the sample size is at least 10\u0026ndash;20 times the number of variables. Since the study\u0026rsquo;s questionnaire had 23 variables, the benchmark sample size for simple random sampling should be at least 230\u0026ndash;460. Given that an intraclass correlation coefficient (ICC) of approximately 0.0304 was identified in the preliminary small-sample survey conducted by the research team, the required sample size for cluster sampling calculated according to the formula n\u0026thinsp;=\u0026thinsp;n\u003csub\u003e0\u003c/sub\u003e[1+(m-1) \u003cem\u003eρ\u003c/em\u003e] ranges from 461 to 921. Considering the 20% invalid recovery rate (Burmeister \u0026amp; Aitken, \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e2012\u003c/span\u003e), the final sample size ranges from 577 to 1152. Ultimately, approximately 1000 participants were surveyed in each primary sampling unit, with a total of 3055 MOAC enrolled in the study. After excluding samples with missing values, 3016 MOAC remained in the dataset for further analysis.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec4\" class=\"Section2\"\u003e \u003ch2\u003e2.2 Outcome variables\u003c/h2\u003e \u003cp\u003eIC encompasses five domains: locomotion, cognition, psychology, sensory, and vitality. IC was measured by the multidimensional domain evaluation tool which was recommended by the WHO for integrated care of older adults. Evaluation criteria for IC: A score of 0 is assigned if the ability in a certain field decline, and a score of 1 if it remains normal. The total score of the IC composite index ranges from 0 to 5, with a higher score indicating better IC. The evaluation criteria are detailed in Appendix B. The total IC score was obtained by summing the scores from these five domains, categorizing IC into two groups: (a) the maintenance group (total IC score\u0026thinsp;=\u0026thinsp;5) and (b) the decline group (total IC score\u0026thinsp;\u0026lt;\u0026thinsp;5) (B. Zhou et al., \u003cspan citationid=\"CR53\" class=\"CitationRef\"\u003e2025a\u003c/span\u003e). The evaluation tools for each domain are as follows:\u003c/p\u003e \u003cp\u003e(1) Self-reported auditory and visual functions\u003c/p\u003e \u003cp\u003eSelf-reported auditory and visual functions were used to measure the sensory domain. Participants were asked whether they had any visual or hearing impairments and if these impairments affect their daily lives.\u003c/p\u003e \u003cp\u003e(2) The Short Physical Performance Battery (SPPB)\u003c/p\u003e \u003cp\u003eThe Short Physical Performance Battery (SPPB), originally developed by Guralnik (Guralnik et al., \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e1994\u003c/span\u003e) and subsequently translated into Chinese by Li (W. Li et al., \u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e2023\u003c/span\u003e), was used as the measurement tool for the locomotor domain. This scale assessed lower limb function in older adults and evaluates three key dimensions: balance tests (3 items), chair stand tests (1 item), and gait speed tests (1 item). Each test is scored from 0 to 4, with a total score ranging from 0 to 12. Higher scores indicate better locomotor function. The Cronbach\u0026rsquo;s α coefficient for SPPB in this study was 0.81.\u003c/p\u003e \u003cp\u003e(3) The Mini Nutritional Assessment (MNA)\u003c/p\u003e \u003cp\u003eThe Mini Nutritional Assessment (MNA), originally developed by Guigoz (Guigoz, \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e2006\u003c/span\u003e) and later translated into Chinese by Lei (Lei et al., \u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e2009\u003c/span\u003e), was employed to measure the vitality domain. This scale consists of four dimensions with a total of 18 items: anthropometric measurements (3 items), general health status (6 items), dietary intake (7 items), and subjective health assessment (2 items). Higher scores indicate better nutritional status. The Cronbach\u0026rsquo;s α coefficient for MNA in this study was 0.71.\u003c/p\u003e \u003cp\u003e(4) The Mini-Cog\u003c/p\u003e \u003cp\u003eThe Chinese version of the Mini-Cog developed and translated into Chinese by Borson (Borson et al., \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e2000\u003c/span\u003e) is a brief cognitive test that involves an assessment of an older person\u0026rsquo;s ability to recall three words and draw a clock. In this study, the Cronbach\u0026rsquo;s α coefficient for Mini-cog was 0.89.\u003c/p\u003e \u003cp\u003e(5) 15 - item Geriatric Depression Scale (GDS-15)\u003c/p\u003e \u003cp\u003eThe Chinese version of the 15-item Geriatric Depression Scale (GDS-15), originally developed by Yesavage (Yesavage et al., \u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e1982\u003c/span\u003e) and adapted for Chinese populations by Zhang (C. Zhang et al., \u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e2022\u003c/span\u003e), was applied to measure the psychological domain. This scale evaluates the emotional state of participants over the past week which includes unhappiness (5 items), apathy and anxiety (5 items), loss of hope and morale (3 items), loss of social activity and memory (2 items). The total score ranges from 0 to 15, with higher scores indicating more severe depression. Cronbach\u0026rsquo;s α coefficient for GDS-15 in this study was 0.75.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec5\" class=\"Section2\"\u003e \u003ch2\u003e2.3 Candidate predictors\u003c/h2\u003e \u003cp\u003eTo identify initial predictors associated with IC, we conducted a literature review and applied the health ecology model. This model conceptualizes influences on health across five concentric layers: demographic factors, health behaviors, social networks, living/working conditions, and social policies. To initially identify predictors associated with IC, a thorough review of the relevant literature was performed, and the theory of health ecology model was also applied. This model consists of five layers from the inside to the outside, namely the demographic factors, health behavior factors, social network factors, living and working conditions factors, and social policy factors (Fan Tao et al., \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e2012\u003c/span\u003e). The factors across these five layers interact dynamically, ultimately shaping human health outcomes.\u003c/p\u003e \u003cp\u003eThis theoretical model led to the inclusion of 23 baseline variables in the analytical framework, encompassing five demographic factors (gender, age, household registration, current residence, number of children); five health behavior factors (health literacy, smoking, drinking, sleep quality, number of chronic diseases); two social network factors (marital status, social network); ten living and working condition factors (education, previous career, reemployment, primary source of financial support in the host location, economic level, migration scope, migration reason, migration status, migration duration, stay willingness); and one social policy factor, that is the multi-layered medical insurance coverage (whether through portable benefits from their place of origin, local schemes in the destination, or both) (see Appendix A).\u003c/p\u003e \u003cp\u003eThe health literacy scale used in this study is the 2015 Chinese Resident Health Literacy Questionnaire, further developed by the National Health and Family Planning Commission (Chen, Shuqi, 2017). It covers three dimensions, with the distribution of questions as follows: 24 items assessing basic health knowledge and concepts, 18 items on healthy lifestyles and behaviors, and 14 items measuring basic health skills.\u003c/p\u003e \u003cp\u003eThe social network was assessed by The Elderly Individual Social Network Scale which is developed by Tao Shengsheng and his research team from Anhui Medical University in 2019 (Tao, Shengsheng, 2019) was adopted. It includes 32 items across four dimensions: health guardianship and emotional support, social participation, social interaction, and social support.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec6\" class=\"Section2\"\u003e \u003ch2\u003e2.4 Statistical analyses\u003c/h2\u003e \u003cp\u003eCharacteristics of participants by the maintenance in IC group and the IC decline group were compared by using chi-square test for categorical variables and t-test for continuous variables. Important risk factors were identified by importance matrix plot of the RF model, with importance of each factor ranking in descending order.\u003c/p\u003e \u003cp\u003eAfter features were selected by importance matrix plot, this study adopted RF, ANN, CART and Logit algorithms to construct predictive models for predicting risk of decline in intrinsic capacity. The dataset was randomly split into a training set (2/3 participants) in which models were conducted, and a validation set (1/3 participants) in which model performance was assessed, by using area under the ROC curve (AUC), classification accuracy (ACC), balanced error rate (BER), false positive rate (FPR), positive predictive value (PPV) and negative predictive value (NPV) as the metrics to evaluate of predictive models. Furthermore, PDPs of each predictor were drawn to help explain the relationship between each predictor and risk of IC decline. To further validate this finding, we conducted additional sensitivity analyses using different feature selection methods, confirming the robustness of our results. Data analyses were performed with SPSS (27.0 version) and R (version 4.4.3).\u003c/p\u003e \u003c/div\u003e"},{"header":"3 Results","content":"\u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003e3.1 Sociodemographic characteristics of participants\u003c/h2\u003e \u003cp\u003eA total of 3055 participants completed the questionnaire, with 39 excluded for missing values, resulting in a 98.7% recovery rate and a final sample of 3016 MOAC. Among these MOAC, 1593 (52.8%) were in the IC decline group. Baseline characteristics of the study population are shown in Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e. The mean age of all MOAC was 67.22\u0026thinsp;\u0026plusmn;\u0026thinsp;6.95.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eSociodemographic characteristics of the decline intrinsic capacity among the MOAC\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"5\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCharacteristics\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eTotal\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003ethe decline group\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003ethe maintenance group\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003e\u003cem\u003eP\u003c/em\u003e\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eTotal\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e3016(100)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1423(47.2)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1593(52.8)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eGender\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.326\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMale\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1404(46.6)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e117(47.8)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e100(40.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eFemale\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1612(53.4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e128(52.2)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e150(60.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eAge\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u0026lt;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e60\u0026ndash;69\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2243(74.4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e876(61.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1367(85.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e70\u0026ndash;79\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e627(20.8)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e418(29.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e209(13.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u0026ge;\u0026thinsp;80\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e146(4.8)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e129(9.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e17(1.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eEducation\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u0026lt;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u0026le;Primary school\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1744(57.8)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e901(63.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e843(52.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMiddle school and high school\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1207(40.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e495(34.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e712(44.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u0026ge;College\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e65(2.2)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e27(1.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e38(2.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eMarital status\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u0026lt;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMarried\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2772(91.9)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1242(87.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1530(96%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eUnmarried\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e18(0.6)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e13(0.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e5(0.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDivorced\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e31(1.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e16(1.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e15(0.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eBereaved\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e195(6.5)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e152(10.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e43(2.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eNumber of children\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u0026lt;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e12(0.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e11(0.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1(0.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e759(25.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e322(22.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e437(27.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1401(46.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e608(42.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e793(49.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u0026ge;\u0026thinsp;3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e844(28%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e482(33.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e362(22.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eHousehold registration\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.006\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRural\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2400(79.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1102(77.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1298(81.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eUrban\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e616(20.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e321(22.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e295(18.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eCurrent residence\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.026\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRural\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e53(1.8)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e33(2.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e20(1.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eUrban\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2963(98.2)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1390(97.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1573(98.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eMigration scope\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.055\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCross the district\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e315(10.4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e158(11.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e157(9.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCross the county\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e556(18.4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e238(16.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e318(20%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCross the province\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2145(71.1)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1027(72.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1118(70.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eMigration reason\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u0026lt;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLook after the younger generation\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1345(44.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e477(33.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e868(54.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eProvide for the aged\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e426(14.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e277(19.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e149(9.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTherapy\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e126(4.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e108(7.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e18(1.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eFamily reunions\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e853(28.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e446(31.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e407(25.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eOthers\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e266(8.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e115(8.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e151(9.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eMigration status\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.004\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePersonal migration\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2327(77.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e358(25.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e331(20.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMigrate with spouse\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e689(22.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1065(74.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1262(79.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eMigration duration\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.011\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e0.5 to 1 (including 1 year)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e424(14.1)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e212(14.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e212(13.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e1 to 3(including 3 years)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e667(22.1)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e342(24%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e325(20.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e3 to 6 (including 6 years)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e659(21.9)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e316(22.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e343(21.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e6 to 9 (including 9 years)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e342(11.3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e142(10%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e200(12.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u0026gt;9\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e924(30.6)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e411(28.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e513(32.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eStay willingness\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.838\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1923(63.8)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e910(63.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1013(63.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1093(36.2)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e513(36.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e580(36.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003ePrevious career\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u0026lt;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePersonnel in public institutions\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e419(13.9)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e261(18.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e158(9.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSelf-employed workers\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e248(8.2)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e96(6.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e152(9.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNon-agricultural workers\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1215(40.3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e412(29%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e803(50.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePeasant\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1131(37.5)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e653(45.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e478(30%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eOthers\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e3(0.1)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1(0.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e2(0.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eReemployment\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u0026lt;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e666(22.1)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e215(15.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e451(28.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2350(77.9)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1208(84.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1142(71.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003ePrimary source of financial support in the host location\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u0026lt;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePension\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e635(21.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e350(24.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e285(17.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eChild maintenance\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e969(32.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e534(37.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e435(27.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLabor income\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1061(35.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e387(27.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e674(42.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eOld-age insurance\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e315(10.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e132(9.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e183(11.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eOthers\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e36(1.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e20(1.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e16(1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eEconomic level\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u0026lt;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNeedy\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e226(7.5)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e189(13.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e37(2.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eBelow the average\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e685(22.7)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e419(29.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e266(16.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMedium\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1717(56.9)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e636(44.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1081(67.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAbove the average\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e334(11.1)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e149(10.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e185(11.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAffluent\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e54(1.8)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e30(2.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e24(1.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eHealth literacy\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e32.78\u0026thinsp;\u0026plusmn;\u0026thinsp;12.05\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e40.75\u0026thinsp;\u0026plusmn;\u0026thinsp;11.87\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u0026lt;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eSocial participation\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e6.47\u0026thinsp;\u0026plusmn;\u0026thinsp;3.99\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e8.65\u0026thinsp;\u0026plusmn;\u0026thinsp;3.57\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u0026lt;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eSocial network\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e91.58\u0026thinsp;\u0026plusmn;\u0026thinsp;16.80\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e110.39\u0026thinsp;\u0026plusmn;\u0026thinsp;23.84\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u0026lt;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eSmoking\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.969\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e764(25.3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1063(74.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1189(74.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2252(74.7)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e360(25.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e404(25.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eDrinking\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.171\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e869(28.8)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1030(72.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1117(70.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2147(71.2)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e393(27.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e476(29.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eSleep quality\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u0026lt;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGood\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1502(49.8)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e420(29.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1082(67.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCommonly\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1174(38.9)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e731(51.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e443(27.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePoor\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e316(10.5)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e253(17.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e63(4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eVery poor\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e24(0.8)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e19(1.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e5(0.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eMulti-layered medical insurance coverage\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u0026lt;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eEmployee medical insurance\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e541(17.9)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e302(21.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e239(15%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eResident medical insurance\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e399(13.2)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e155(10.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e244(15.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNew rural insurance\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1774(58.8)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e835(58.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e939(58.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eOther insurance\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e83(2.8)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e26(1.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e57(3.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNo insurance\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e219(7.3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e105(7.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e114(7.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eNumber of chronic diseases\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u0026lt;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1391(41.6)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e464(32.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e927(58.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1035(34.3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e528(37.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e507(31.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u0026ge;\u0026thinsp;2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e590(19.6)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e431(30.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e159(10%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec9\" class=\"Section2\"\u003e \u003ch2\u003e3.2 Feature selection\u003c/h2\u003e \u003cp\u003eThis study employed the importance matrix plot, a wrapper algorithm that leveraged the random forest algorithm to rank variable importance, for feature selection. The importance matrix plot employs the Gini index as the metric for split point identification\u0026mdash;an indicator representing the probability of misclassifying a randomly selected sample in the dataset. Factors that have a significant impact on the decline in intrinsic abilities among the MOAC were screened out using the importance matrix plot, including social network, health literacy, sleep quality, migration reason, economic level, number of chronic diseases, age, primary source of financial support in the host location, previous career, multi-layered medical insurance coverage (see Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec10\" class=\"Section2\"\u003e \u003ch2\u003e3.3 Model training and evaluation\u003c/h2\u003e \u003cp\u003eBy using factors identified important by importance matrix plot, predictive models were developed by using RF, ANN, CART and Logit method in training dataset, and predictive accuracy of those models in test dataset is presented in Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e and Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e. As RF model performed the best, we developed the predictive model for predicting risk of IC decline among the MOAC by RF model, with mtree of 5 and ntree of 500. For model performance, the AUC was 0.843 (95% \u003cem\u003eCI\u003c/em\u003e: 0.820\u0026ndash;0.867), the sensitivity was 0.772, and the specificity was 0.772, suggesting that the predictive model had a good prediction effect (see Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eComparison of predictive accuracy of four machine learning models in test dataset.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"8\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c8\" colnum=\"8\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMODEL\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eSENSITIVITY\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eSPECIFICITY\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eACC\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eBER\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003eFRP\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c7\"\u003e \u003cp\u003ePPV\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c8\"\u003e \u003cp\u003eNPV\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRF\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e74.47\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e77.97\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e76.32\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e23.78\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e22.03\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e75.11\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e77.38\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLogistic\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e71.94\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e76.65\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e74.43\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e25.71\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e23.35\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e73.33\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e75.37\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCART\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e73.84\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e72.50\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e73.13\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e26.83\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e27.50\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e70.56\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e75.64\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eANN\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e73.00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e76.84\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e75.02\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e25.08\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e23.16\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e73.77\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e76.12\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003e3.4 Elucidating Feature-Prediction Associations\u003c/h2\u003e \u003cp\u003ePDPs, an agnostic method, reveal how changes in a single feature influence prediction, aiding in understanding the marginal effect of a variable on the predicted outcome of an ML model. The result presented the non-linear effect of each predictor on the risk of IC decline after adjusting for all other factors. Social network and economic level roughly showed a V-shaped curve with the risk of IC decline (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e). As health literacy increases, the risk of IC decline also decreases. Both the increase in the number of chronic diseases and advancing age elevate the risk of IC decline, while the deterioration of sleep quality also contributes to a higher risk of IC decline. Differences in migration reasons, primary source of financial support in the host location, previous career and multi-layered medical insurance coverage are associated with variations of the IC decline among the MOAC. Among them, MOAC who migrated to care for the younger generation, labor income, non-agricultural workers, and those with employee multi-layered medical insurance coverage exhibited the mildest IC decline. While those who migrated for therapeutic purposes, pension, personnel in public institutions, no insurance faced the highest risk of IC decline. Among them, MOAC who migrated to care for the younger generation, individuals with labor income, non-agricultural workers, and those covered by employee multi-layered medical insurance coverage exhibited the mildest IC decline. In contrast, those who migrated for therapeutic purposes, retirees relying on pensions, personnel in public institutions, and the uninsured faced the highest risk of IC decline.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e"},{"header":"4 Discussion","content":"\u003cp\u003eBased on multi-center, large-sample data (N\u0026thinsp;=\u0026thinsp;3,016), this study found that over half of the MOAC have experienced a IC decline, which leads to worsening function, lower quality of life, and poor health outcomes (Stolz et al., \u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e2022b\u003c/span\u003e; J. Yu et al., \u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; Y. Zhao et al., \u003cspan citationid=\"CR51\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). Therefore, this study conducted further analysis to identify key factors influencing IC decline.\u003c/p\u003e \u003cp\u003eThis was the first time that ML methods had been applied to predict the risk of IC decline among the MOAC. Based on the five dimensions of the health ecology model, namely the demographic factors, health behavior factors, social network factors, living and working conditions factors, and social policy factors, this study incorporated 23 variables. Meanwhile, the importance matrix plot was employed to identify the top 10 predictors with the most significant impact on the risk of IC decline. Four predictive models were developed in this study, and among them, the RF model showed the best predictive performance and clinical utility. Furthermore, after adjusting for all other factors, PDPs were employed to analyze the non-linear influence of these top predictors on the risk of IC decline, revealing the marginal effects of changes.\u003c/p\u003e \u003cp\u003eThis study employed an importance matrix plot to identify key influencing factors. The analysis revealed that social network, health literacy, sleep quality, migration reason, economic level, number of chronic diseases, primary source of financial support in the host location, previous career, multi-layered medical insurance coverage and age are the top ten most important risk factors for IC decline, which indicates that favorable individual health and social characteristics are beneficial for preventing IC decline. The selection of these influencing factors contributes to enhancing the credibility of practical ML applications (Degenhardt et al., \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e2019\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eIn a comparison of four ML algorithms, the RF model achieved the highest accuracy in predicting IC decline, supporting its potential as a community-level screening tool for healthcare providers to identify who may be at risk of IC decline among the MOAC.\u003c/p\u003e \u003cp\u003eRF models are like \u0026ldquo;black boxes\u0026rdquo; where only the inputs and outputs are visible, while the internal processes are obscured by complex calculations. PDPs clarify the relationship between a specific feature and the prediction outcome by keeping all other features constant. This method extends the interpretation of feature effects beyond simple linear regression and is applicable to more complex models (Q. Zhao \u0026amp; Hastie, \u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). PDPs effectively depict the relationship between the response variable and independent variables, regardless of whether the relationship is linear, monotonic, or complex. They show how the response variable changes as the value of an independent variable changes, while accounting for the average effects of all other independent variables in the model (Boehmke \u0026amp; Greenwell, \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e2019\u003c/span\u003e). Therefore, the PDPs method provides a clear and robust way to interpret complex models because they are easy to implement and simple to explain.\u003c/p\u003e \u003cp\u003eThe study visualized the results via PDPs, revealing that social network and economic level exhibit a V-shaped curve in relation to the risk of IC decline. The risk of IC decline decreases with the increase in social network (Guo et al., \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e2024\u003c/span\u003e) (Sun et al., \u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e2025\u003c/span\u003e); however, 175 points is a turning point, beyond which a larger social network leads to an increase in the risk of IC decline. This may be attributed to either excessive reliance on others within the social network or excessive social engagement that leads to mental and physical exhaustion, both of which contribute to a IC decline (Arafat \u0026amp; Thoma, \u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e2024\u003c/span\u003e; Qin et al., \u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). Economic level showed a similar pattern: risk declined until a medium level was reached, then rose among affluent groups When the economic level was below the medium level, the higher the economic level, the lower the risk of IC decline. When the economic level was above average and affluent, the risk of IC decline increased instead (Salinas-Rodr\u0026iacute;guez et al., \u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e2024\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eWithin the non-modifiable dimensions of the health ecology model, age, previous career, number of chronic diseases and migration reasons were identified as key factors influencing the risk of IC decline among the MOAC. Specifically, MOAC who are elder (R. Yu et al., \u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e2023\u003c/span\u003e, \u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e2024\u003c/span\u003e), non-agricultural workers, (C.-Y. Li et al., \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e2002\u003c/span\u003e), have multiple chronic conditions (M. Zhang et al., \u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e2025\u003c/span\u003e), or migrated for medical therapy (Willen et al., \u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e2021\u003c/span\u003e) face a higher risk of IC decline. Therefore, community primary healthcare providers should prioritize these populations for early screening to enable timely identification and intervention of related risks.\u003c/p\u003e \u003cp\u003eWithin the health ecology model, health literacy stands out as the foremost modifiable factor protective against IC decline, the finding consistent with prior research (Q. Liu et al., \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). Given this primary status, community healthcare providers should adopt a proactive strategy that integrates health literacy promotion into social activities and health education programs, aiming to improve health outcomes for the MOAC.\u003c/p\u003e \u003cp\u003eThis study also analyzed other modifiable factors, specifically the primary source of financial support in the host location, access to multi-layered medical insurance coverage, and sleep quality. The results indicate that MOAC with poorer sleep quality, reliance on self-earned income, or lack of adequate multi-layered medical insurance coverage face a heightened risk of IC decline, which aligns with existing evidence (Kang et al., \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e2025\u003c/span\u003e; Rodriguez \u0026amp; Saenz, \u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e2022\u003c/span\u003e; Yang et al., \u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e2025\u003c/span\u003e; B. Zhou et al., \u003cspan citationid=\"CR54\" class=\"CitationRef\"\u003e2025b\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eThis elevated risk likely stems from an interconnected decline: age and multimorbidity impair sleep and physiological resilience, while insufficient financial support and inadequate insurance coverage exacerbate psychological vulnerability (Arokiasamy et al., \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2021\u003c/span\u003e), collectively driving the risk of IC decline.\u003c/p\u003e \u003cp\u003eTo summarize, this study compared several ML algorithms and identified the RF model as having the optimal predictive performance. After screening out the top ten key predictors of IC decline among the MOAC population, the study further used PDPs to make up for the inherent \u0026ldquo;black-box\u0026rdquo; limitation of the RF model. This approach clearly delineated the core impact trends of these ten factors on IC decline, thereby providing a robust scientific foundation for formulating precise intervention strategies to improve the IC among the MOAC.\u003c/p\u003e \u003cp\u003eRegarding non-modifiable risk factors, this study identifies five high-risk subgroups of MOAC requiring prioritization by community primary healthcare providers: older individuals, those with multimorbidity, migrants for medical therapy, have a previous career as non-agricultural workers, and those without adequate multi-layered medical insurance coverage in the host location.\u003c/p\u003e \u003cp\u003eFurthermore, guided by the health ecology model, a multilevel intervention pathway is proposed: a) At the individual level, healthcare providers can strengthen health literacy and improve sleep quality. b) At the social/community level, tailored activities should be designed to expand meaningful social networks and enhance social engagement. c) At the family level, family members are encouraged to support MOAC\u0026rsquo;s living and working conditions, such as stabilizing economic status and diversifying income sources. Through such coordinated, multidimensional interventions, IC decline among the MOAC population can be effectively delayed, supporting sustained improvements in their quality of life.\u003c/p\u003e"},{"header":"5 Limitations","content":"\u003cp\u003eThis study has several limitations. First, the cross-sectional design limits the ability to establish causal relationships among the identified risk factors for IC decline. Future longitudinal studies are needed to further explore these causal links. Second, while recent research has identified promising biomarkers associated with IC (Fang et al., \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e2025\u003c/span\u003e; Guyonnet et al., \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2025\u003c/span\u003e), this study did not incorporate such biological measures. Subsequent research should aim to identify accessible biomarkers for comprehensively assessing IC decline and integrate them to improve the accuracy of predictive models.\u003c/p\u003e"},{"header":"6 Conclusion","content":"\u003cp\u003eThis study demonstrates that the RF model provides superior accuracy in predicting IC decline among the MOAC. Guided by the health ecology model, communities healthcare providers should pay more attention to screen the high-risk individuals based on key non-modifiable factors, including advanced age, multimorbidity, a non‑agricultural work history, and migrant for medical reasons. Furthermore, tailored multilevel interventions targeting modifiable determinants should be designed to improve IC, with a focusing on behavioral and interpersonal factors such as health literacy, social networks, sleep quality, diversified income sources, and adequate multi-layered medical insurance coverage in the host location for MOAC.\u0026nbsp;\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eDeclaration of Conflicting Interests\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll authors certify that they have no affiliations with or involvement in any organization or entity with any financial interest or non-financial interest in the subject matter or materials discussed in this manuscript.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding information\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis work was supported by the National Social Science Fund of China (22BGL251).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eInformed consent was obtained from all individual participants included in the study participants signed informed consent regarding publishing their data. This study did not involve participant identity images or other personal or clinical details that would compromise anonymity.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eEthics approval\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eApproval was obtained from the Institutional Review Board of Huzhou University (IRB No. 2022-03-05). The procedures used in this study adhere to the tenets of the Declaration of Helsinki.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAvailability of Data and Materials\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll data generated or analyzed during this study are included in this manuscript. The data that support the findings of this study are available from the corresponding author upon reasonable request.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eClinical trial number\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent for publication\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot Applicable.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n \u003cli\u003eArafat, D., \u0026amp; Thoma, P. (2024). Impairments of Sociocognitive Functions in Individuals with Behavioral Addictions: A Review Article. \u003cem\u003eJournal of Gambling Studies\u003c/em\u003e, \u003cem\u003e40\u003c/em\u003e(2), 429\u0026ndash;451. https://doi.org/10.1007/s10899-023-10227-w\u003c/li\u003e\n \u003cli\u003eArokiasamy, P., Selvamani, Y., Jotheeswaran, A. T., \u0026amp; Sadana, R. (2021). Socioeconomic differences in handgrip strength and its association with measures of intrinsic capacity among older adults in six middle-income countries. \u003cem\u003eScientific Reports\u003c/em\u003e, \u003cem\u003e11\u003c/em\u003e(1), 19494. https://doi.org/10.1038/s41598-021-99047-9\u003c/li\u003e\n \u003cli\u003eBautmans, I., Knoop, V., Amuthavalli Thiyagarajan, J., Maier, A. B., Beard, J. R., Freiberger, E., Belsky, D., Aubertin-Leheudre, M., Mikton, C., Cesari, M., Sumi, Y., Diaz, T., Banerjee, A., \u0026amp; WHO Working Group on Vitality Capacity. (2022). WHO working definition of vitality capacity for healthy longevity monitoring. \u003cem\u003eThe Lancet. Healthy Longevity\u003c/em\u003e, \u003cem\u003e3\u003c/em\u003e(11), e789\u0026ndash;e796. https://doi.org/10.1016/S2666-7568(22)00200-8\u003c/li\u003e\n \u003cli\u003eBoehmke, B., \u0026amp; Greenwell, B. (2019). Interpretable Machine Learning. In \u003cem\u003eHands-On Machine Learning with R\u003c/em\u003e. Chapman and Hall/CRC.\u003c/li\u003e\n \u003cli\u003eBorson, S., Scanlan, J., Brush, M., Vitaliano, P., \u0026amp; Dokmak, A. (2000). The mini-cog: A cognitive \u0026ldquo;vital signs\u0026rdquo; measure for dementia screening in multi-lingual elderly. \u003cem\u003eInternational Journal of Geriatric Psychiatry\u003c/em\u003e, \u003cem\u003e15\u003c/em\u003e(11), 1021\u0026ndash;1027. https://doi.org/10.1002/1099-1166(200011)15:11%253C1021::aid-gps234%253E3.0.co;2-6\u003c/li\u003e\n \u003cli\u003eBurmeister, E., \u0026amp; Aitken, L. M. (2012). Sample size: How many is enough? \u003cem\u003eAustralian Critical Care: Official Journal of the Confederation of Australian Critical Care Nurses\u003c/em\u003e, \u003cem\u003e25\u003c/em\u003e(4), 271\u0026ndash;274. https://doi.org/10.1016/j.aucc.2012.07.002\u003c/li\u003e\n \u003cli\u003eChen, Shuqi. (2017). \u003cem\u003eStudy on Change Tendency of Health Literacy and ItsInfluencing Factors among Rural Residents in KunshanCity\u003c/em\u003e [Master\u0026rsquo;s Thesis, Soochow University]. https://kns.cnki.net/KCMS/detail/detail.aspx?dbcode=CMFD\u0026amp;dbname=CMFD201701\u0026amp;filename=1016212474.nh\u003c/li\u003e\n \u003cli\u003eDegenhardt, F., Seifert, S., \u0026amp; Szymczak, S. (2019). Evaluation of variable selection methods for random forests and omics data sets. \u003cem\u003eBriefings in Bioinformatics\u003c/em\u003e, \u003cem\u003e20\u003c/em\u003e(2), 492\u0026ndash;503. https://doi.org/10.1093/bib/bbx124\u003c/li\u003e\n \u003cli\u003eFan Tao, Cao Qian, Jiang Lulu, \u0026amp; Xu Zhengzi. (2012). Explaining the Influencing Factors on Chronic Diseases of the Elderly Using Health Ecological Model. \u003cem\u003eChinese General Practice\u003c/em\u003e, \u003cem\u003e15\u003c/em\u003e(1), 33\u0026ndash;36, 40.\u003c/li\u003e\n \u003cli\u003eFang, J., Tan, J., Chen, X., Hua, K., Ding, Y., Wu, Y., Tan, L., Hua, Q., Shao, S., Liu, T., \u0026amp; Hua, Y. (2025). Association between homocysteine levels and decreased intrinsic capacity in Chinese community-dwelling older adults: A cross-sectional study. \u003cem\u003eMedicine\u003c/em\u003e, \u003cem\u003e104\u003c/em\u003e(32), e43771. https://doi.org/10.1097/MD.0000000000043771\u003c/li\u003e\n \u003cli\u003eGuigoz, Y. (2006). The Mini Nutritional Assessment (MNA) review of the literature\u0026mdash;What does it tell us? \u003cem\u003eThe Journal of Nutrition, Health \u0026amp; Aging\u003c/em\u003e, \u003cem\u003e10\u003c/em\u003e(6), 466\u0026ndash;485; discussion 485-487.\u003c/li\u003e\n \u003cli\u003eGuo, W., Meng, L., Han, J., Yang, B., Sun, J., Guo, Y., Wu, J., \u0026amp; Liu, Y. (2024). Intrinsic capacity and its association with predictors among Chinese empty nest older adults in communities: A latent class analysis. \u003cem\u003eBMC Geriatrics\u003c/em\u003e, \u003cem\u003e24\u003c/em\u003e(1), 1008. https://doi.org/10.1186/s12877-024-05583-0\u003c/li\u003e\n \u003cli\u003eGuralnik, J. M., Simonsick, E. M., Ferrucci, L., Glynn, R. J., Berkman, L. F., Blazer, D. G., Scherr, P. A., \u0026amp; Wallace, R. B. (1994). A short physical performance battery assessing lower extremity function: Association with self-reported disability and prediction of mortality and nursing home admission. \u003cem\u003eJournal of Gerontology\u003c/em\u003e, \u003cem\u003e49\u003c/em\u003e(2), M85-94. https://doi.org/10.1093/geronj/49.2.m85\u003c/li\u003e\n \u003cli\u003eGuyonnet, S., Hooper, C., Cantet, C., Coley, N., Andrieu, S., Vellas, B., \u0026amp; IHU HealthAge INSPIRE/Open Science study group. (2025). The INSPIRE-T longitudinal observational study centred on intrinsic capacity: Baseline data. \u003cem\u003eThe Journals of Gerontology. Series A, Biological Sciences and Medical Sciences\u003c/em\u003e, glaf181. https://doi.org/10.1093/gerona/glaf181\u003c/li\u003e\n \u003cli\u003eHernandes, E. C. R., Aliberti, M. J. R., Guerra, R. O., Ferriolli, E., \u0026amp; Perracini, M. R. (2024). Intrinsic capacity and hospitalization among older adults: A nationally representative cross-sectional study. \u003cem\u003eEuropean Geriatric Medicine\u003c/em\u003e, \u003cem\u003e15\u003c/em\u003e(3), 843\u0026ndash;852. https://doi.org/10.1007/s41999-024-00933-y\u003c/li\u003e\n \u003cli\u003eHolecki, T., Rogalska, A., Sobczyk, K., Woźniak-Holecka, J., \u0026amp; Romaniuk, P. (2020). Global Elderly Migrations and Their Impact on Health Care Systems. \u003cem\u003eFrontiers in Public Health\u003c/em\u003e, \u003cem\u003e8\u003c/em\u003e, 386. https://doi.org/10.3389/fpubh.2020.00386\u003c/li\u003e\n \u003cli\u003eKang, Y., Shi, H., Zhang, J., Meng, X., Zhang, C., Shen, J., \u0026amp; Zhang, P. (2025). The Bidirectional Relationship Between Intrinsic Capacity and Catastrophic Health Expenditure in China: A Longitudinal Study. \u003cem\u003eThe Journals of Gerontology. Series A, Biological Sciences and Medical Sciences\u003c/em\u003e, \u003cem\u003e80\u003c/em\u003e(3), glae301. https://doi.org/10.1093/gerona/glae301\u003c/li\u003e\n \u003cli\u003eLee, W.-J., Peng, L.-N., Lin, M.-H., Loh, C.-H., Hsiao, F.-Y., \u0026amp; Chen, L.-K. (2023). Intrinsic capacity differs from functional ability in predicting 10-year mortality and biological features in healthy aging: Results from the I-Lan longitudinal aging study. \u003cem\u003eAging\u003c/em\u003e, \u003cem\u003e15\u003c/em\u003e(3), 748\u0026ndash;764. https://doi.org/10.18632/aging.204508\u003c/li\u003e\n \u003cli\u003eLei, Z., Qingyi, D., Feng, G., Chen, W., Hock, R. S., \u0026amp; Changli, W. (2009). Clinical study of mini-nutritional assessment for older Chinese inpatients. \u003cem\u003eThe Journal of Nutrition, Health \u0026amp; Aging\u003c/em\u003e, \u003cem\u003e13\u003c/em\u003e(10), 871\u0026ndash;875. https://doi.org/10.1007/s12603-009-0244-1\u003c/li\u003e\n \u003cli\u003eLi, C.-Y., Wu, S. C., \u0026amp; Sung, F.-C. (2002). Lifetime principal occupation and risk of cognitive impairment among the elderly. \u003cem\u003eIndustrial Health\u003c/em\u003e, \u003cem\u003e40\u003c/em\u003e(1), 7\u0026ndash;13. https://doi.org/10.2486/indhealth.40.7\u003c/li\u003e\n \u003cli\u003eLi, H., Pang, M., Wang, J., Xu, J., \u0026amp; Kong, F. (2022). Effects of Health Service Utilization and Informal Social Support on Depression, Anxiety, and Stress among the Internal Migrant Elderly following Children in Weifang, China. \u003cem\u003eInternational Journal of Environmental Research and Public Health\u003c/em\u003e, \u003cem\u003e19\u003c/em\u003e(22), 14640. https://doi.org/10.3390/ijerph192214640\u003c/li\u003e\n \u003cli\u003eLi, W., Rao, Z., Fu, Y., Schwebel, D. C., Li, L., Ning, P., Huang, J., \u0026amp; Hu, G. (2023). Value of the short physical performance battery (SPPB) in predicting fall and fall-induced injury among old Chinese adults. \u003cem\u003eBMC Geriatrics\u003c/em\u003e, \u003cem\u003e23\u003c/em\u003e(1), 574. https://doi.org/10.1186/s12877-023-04290-6\u003c/li\u003e\n \u003cli\u003eLiang, H., Zheng, J., \u0026amp; Sun, Y. (2023). Prevalence and predictive risk factors of self-stigmatization in patients with rare diseases: Based on health ecology model. \u003cem\u003eThe International Journal of Social Psychiatry\u003c/em\u003e, \u003cem\u003e69\u003c/em\u003e(6), 1541\u0026ndash;1545. https://doi.org/10.1177/00207640231168023\u003c/li\u003e\n \u003cli\u003eLiu, Q., Li, X., Hu, M., Zhao, Y., Wu, S., \u0026amp; Feng, H. (2024). Factors influencing the self-management ability among older adults experiencing intrinsic capacity decline: A cross-sectional study. \u003cem\u003eFrontiers in Aging Neuroscience\u003c/em\u003e, \u003cem\u003e16\u003c/em\u003e, 1456167. https://doi.org/10.3389/fnagi.2024.1456167\u003c/li\u003e\n \u003cli\u003eLiu, S., Kang, L., Liu, X., Zhao, S., Wang, X., Li, J., \u0026amp; Jiang, S. (2021). Trajectory and Correlation of Intrinsic Capacity and Frailty in a Beijing Elderly Community. \u003cem\u003eFrontiers in Medicine\u003c/em\u003e, \u003cem\u003e8\u003c/em\u003e, 751586. https://doi.org/10.3389/fmed.2021.751586\u003c/li\u003e\n \u003cli\u003eMa, L., Chhetri, J. K., Zhang, L., Sun, F., Li, Y., \u0026amp; Tang, Z. (2021). Cross-sectional study examining the status of intrinsic capacity decline in community-dwelling older adults in China: Prevalence, associated factors and implications for clinical care. \u003cem\u003eBMJ Open\u003c/em\u003e, \u003cem\u003e11\u003c/em\u003e(1), e043062. https://doi.org/10.1136/bmjopen-2020-043062\u003c/li\u003e\n \u003cli\u003eMuneera, K., Muhammad, T., \u0026amp; Althaf, S. (2022). Socio-demographic and lifestyle factors associated with intrinsic capacity among older adults: Evidence from India. \u003cem\u003eBMC Geriatrics\u003c/em\u003e, \u003cem\u003e22\u003c/em\u003e(1), 851. https://doi.org/10.1186/s12877-022-03558-7\u003c/li\u003e\n \u003cli\u003eQin, C., Li, Y., Wang, T., Zhao, J., Tong, L., Yang, J., \u0026amp; Liu, Y. (2024). Too much social media? Unveiling the effects of determinants in social media fatigue. \u003cem\u003eFrontiers in Psychology\u003c/em\u003e, \u003cem\u003e15\u003c/em\u003e, 1277846. https://doi.org/10.3389/fpsyg.2024.1277846\u003c/li\u003e\n \u003cli\u003eRam\u0026iacute;rez-V\u0026eacute;lez, R., Correa-Bautista, J. E., Garc\u0026iacute;a-Hermoso, A., Cano, C. A., \u0026amp; Izquierdo, M. (2019). Reference values for handgrip strength and their association with intrinsic capacity domains among older adults. \u003cem\u003eJournal of Cachexia, Sarcopenia and Muscle\u003c/em\u003e, \u003cem\u003e10\u003c/em\u003e(2), 278\u0026ndash;286. https://doi.org/10.1002/jcsm.12373\u003c/li\u003e\n \u003cli\u003eRodriguez, F. S., \u0026amp; Saenz, J. (2022). Working in old age in Mexico: Implications for Cognitive Functioning. \u003cem\u003eAgeing and Society\u003c/em\u003e, \u003cem\u003e42\u003c/em\u003e(11), 2489\u0026ndash;2509. https://doi.org/10.1017/s0144686x2100012x\u003c/li\u003e\n \u003cli\u003eSalinas-Rodr\u0026iacute;guez, A., Fern\u0026aacute;ndez-Ni\u0026ntilde;o, J. A., Rivera-Almaraz, A., \u0026amp; Manrique-Espinoza, B. (2024). Intrinsic capacity trajectories and socioeconomic inequalities in health: The contributions of wealth, education, gender, and ethnicity. \u003cem\u003eInternational Journal for Equity in Health\u003c/em\u003e, \u003cem\u003e23\u003c/em\u003e(1), 48. https://doi.org/10.1186/s12939-024-02136-0\u003c/li\u003e\n \u003cli\u003eSalinas-Rodr\u0026iacute;guez, A., Gonz\u0026aacute;lez-Bautista, E., Rivera-Almaraz, A., \u0026amp; Manrique-Espinoza, B. (2022). Longitudinal trajectories of intrinsic capacity and their association with quality of life and disability. \u003cem\u003eMaturitas\u003c/em\u003e, \u003cem\u003e161\u003c/em\u003e, 49\u0026ndash;54. https://doi.org/10.1016/j.maturitas.2022.02.005\u003c/li\u003e\n \u003cli\u003eStafford, I. S., Kellermann, M., Mossotto, E., Beattie, R. M., MacArthur, B. D., \u0026amp; Ennis, S. (2020). A systematic review of the applications of artificial intelligence and machine learning in autoimmune diseases. \u003cem\u003eNPJ Digital Medicine\u003c/em\u003e, \u003cem\u003e3\u003c/em\u003e, 30. https://doi.org/10.1038/s41746-020-0229-3\u003c/li\u003e\n \u003cli\u003eStephens, C., Allen, J., Keating, N., Szab\u0026oacute;, \u0026Aacute;., \u0026amp; Alpass, F. (2020). Neighborhood environments and intrinsic capacity interact to affect the health-related quality of life of older people in New Zealand. \u003cem\u003eMaturitas\u003c/em\u003e, \u003cem\u003e139\u003c/em\u003e, 1\u0026ndash;5. https://doi.org/10.1016/j.maturitas.2020.05.008\u003c/li\u003e\n \u003cli\u003eStolz, E., Mayerl, H., Freidl, W., Roller-Wirnsberger, R., \u0026amp; Gill, T. M. (2022a). Intrinsic Capacity Predicts Negative Health Outcomes in Older Adults. \u003cem\u003eThe Journals of Gerontology. Series A, Biological Sciences and Medical Sciences\u003c/em\u003e, \u003cem\u003e77\u003c/em\u003e(1), 101\u0026ndash;105. https://doi.org/10.1093/gerona/glab279\u003c/li\u003e\n \u003cli\u003eStolz, E., Mayerl, H., Freidl, W., Roller-Wirnsberger, R., \u0026amp; Gill, T. M. (2022b). Intrinsic Capacity Predicts Negative Health Outcomes in Older Adults. \u003cem\u003eThe Journals of Gerontology. Series A, Biological Sciences and Medical Sciences\u003c/em\u003e, \u003cem\u003e77\u003c/em\u003e(1), 101\u0026ndash;105. https://doi.org/10.1093/gerona/glab279\u003c/li\u003e\n \u003cli\u003eSun, J., Amnatsatsue, K., Srisuwan, P., Kerdmongkol, P., \u0026amp; Nityasuddhi, D. (2025). Intrinsic Capacity of Chinese Community-Dwelling Older Adults Using WHO Integrated Care for Older People (ICOPE) Framework: Structural Equation Model Analysis. \u003cem\u003eJournal of Primary Care \u0026amp; Community Health\u003c/em\u003e, \u003cem\u003e16\u003c/em\u003e, 21501319251346433. https://doi.org/10.1177/21501319251346433\u003c/li\u003e\n \u003cli\u003eTao, Shengsheng. (2019). \u003cem\u003eResearch on Social Network Measurement Index System of OldPeople under Modern Medical Model\u003c/em\u003e [Master\u0026rsquo;s Thesis, Anhui Medical University]. https://kns.cnki.net/KCMS/detail/detail.aspx?dbcode=CMFD\u0026amp;dbname=CMFD201902\u0026amp;filename=1019175215.nh\u003c/li\u003e\n \u003cli\u003eTian, T., Kong, F., \u0026amp; Li, S. (2021). Effects of Living Conditions, Subjective Integration, and Social Networks on Health-Related Quality of Life among the Migrant Elderly Following Children in Jinan, China. \u003cem\u003eHealthcare (Basel, Switzerland)\u003c/em\u003e, \u003cem\u003e9\u003c/em\u003e(4), 414. https://doi.org/10.3390/healthcare9040414\u003c/li\u003e\n \u003cli\u003eTokodi, M., Schwertner, W. R., Kov\u0026aacute;cs, A., Tős\u0026eacute;r, Z., Staub, L., S\u0026aacute;rk\u0026aacute;ny, A., Lakatos, B. K., Behon, A., Boros, A. M., Perge, P., Kutyifa, V., Sz\u0026eacute;plaki, G., Gell\u0026eacute;r, L., Merkely, B., \u0026amp; Kosztin, A. (2020). Machine learning-based mortality prediction of patients undergoing cardiac resynchronization therapy: The SEMMELWEIS-CRT score. \u003cem\u003eEuropean Heart Journal\u003c/em\u003e, \u003cem\u003e41\u003c/em\u003e(18), 1747\u0026ndash;1756. https://doi.org/10.1093/eurheartj/ehz902\u003c/li\u003e\n \u003cli\u003eTu, W.-J., Zeng, X., \u0026amp; Liu, Q. (2022). Aging tsunami coming: The main finding from China\u0026rsquo;s seventh national population census. \u003cem\u003eAging Clinical and Experimental Research\u003c/em\u003e, \u003cem\u003e34\u003c/em\u003e(5), 1159\u0026ndash;1163. https://doi.org/10.1007/s40520-021-02017-4\u003c/li\u003e\n \u003cli\u003eWillen, S. S., Selim, N., Mendenhall, E., Lopez, M. M., Chowdhury, S. A., Dilger, H., \u0026amp; Migration and Health in Social Context Working Group. (2021). Flourishing: Migration and health in social context. \u003cem\u003eBMJ Global Health\u003c/em\u003e, \u003cem\u003e6\u003c/em\u003e(Suppl 1), e005108. https://doi.org/10.1136/bmjgh-2021-005108\u003c/li\u003e\n \u003cli\u003eYang, C., Chen, X., Wan, X., \u0026amp; Cai, Y. (2025). The association between sleep duration trajectories and intrinsic capacity in middle-aged and older adults in China: A longitudinal Chinese study assessing healthy aging. \u003cem\u003eFrontiers in Medicine\u003c/em\u003e, \u003cem\u003e12\u003c/em\u003e, 1595241. https://doi.org/10.3389/fmed.2025.1595241\u003c/li\u003e\n \u003cli\u003eYesavage, J. A., Brink, T. L., Rose, T. L., Lum, O., Huang, V., Adey, M., \u0026amp; Leirer, V. O. (1982). Development and validation of a geriatric depression screening scale: A preliminary report. \u003cem\u003eJournal of Psychiatric Research\u003c/em\u003e, \u003cem\u003e17\u003c/em\u003e(1), 37\u0026ndash;49. https://doi.org/10.1016/0022-3956(82)90033-4\u003c/li\u003e\n \u003cli\u003eYu, J., Si, H., Qiao, X., Jin, Y., Ji, L., Liu, Q., Bian, Y., Wang, W., \u0026amp; Wang, C. (2021). Predictive value of intrinsic capacity on adverse outcomes among community-dwelling older adults. \u003cem\u003eGeriatric Nursing (New York, N.Y.)\u003c/em\u003e, \u003cem\u003e42\u003c/em\u003e(6), 1257\u0026ndash;1263. https://doi.org/10.1016/j.gerinurse.2021.08.010\u003c/li\u003e\n \u003cli\u003eYu, R., Lai, D., Leung, G., Tam, L.-Y., Cheng, C., Kong, S., Tong, C., \u0026amp; Woo, J. (2024). Transitions in intrinsic capacity among community-dwelling older people and their associated factors: A multistate modelling analysis. \u003cem\u003eThe Journal of Nutrition, Health \u0026amp; Aging\u003c/em\u003e, \u003cem\u003e28\u003c/em\u003e(7), 100273. https://doi.org/10.1016/j.jnha.2024.100273\u003c/li\u003e\n \u003cli\u003eYu, R., Lai, D., Leung, G., \u0026amp; Woo, J. (2023). Trajectories of Intrinsic Capacity: Determinants and Associations with Disability. \u003cem\u003eThe Journal of Nutrition, Health \u0026amp; Aging\u003c/em\u003e, \u003cem\u003e27\u003c/em\u003e(3), 174\u0026ndash;181. https://doi.org/10.1007/s12603-023-1881-5\u003c/li\u003e\n \u003cli\u003eZhang, C., Zhang, H., Zhao, M., Chen, C., Li, Z., Liu, D., Zhao, Y., \u0026amp; Yao, Y. (2022). Psychometric properties and modification of the 15-item geriatric depression scale among Chinese oldest-old and centenarians: A mixed-methods study. \u003cem\u003eBMC Geriatrics\u003c/em\u003e, \u003cem\u003e22\u003c/em\u003e(1), 144. https://doi.org/10.1186/s12877-022-02833-x\u003c/li\u003e\n \u003cli\u003eZhang, M., Xu, Y., Xing, Y., \u0026amp; Li, H. (2025). Association between multimorbidity and intrinsic capacity among older Chinese adults: Evidence from the CHARLS 2011-2015. \u003cem\u003eEuropean Geriatric Medicine\u003c/em\u003e, \u003cem\u003e16\u003c/em\u003e(4), 1207\u0026ndash;1216. https://doi.org/10.1007/s41999-025-01232-w\u003c/li\u003e\n \u003cli\u003eZhao, Q., \u0026amp; Hastie, T. (2021). Causal Interpretations of Black-Box Models. \u003cem\u003eJournal of Business \u0026amp; Economic Statistics\u003c/em\u003e, \u003cem\u003e39\u003c/em\u003e(1), 272\u0026ndash;281. https://doi.org/10.1080/07350015.2019.1624293\u003c/li\u003e\n \u003cli\u003eZhao, Y., Jiang, Y., Tang, P., Wang, X., Guo, Y., \u0026amp; Tang, L. (2024). Adverse health effects of declined intrinsic capacity in middle-aged and older adults: A systematic review and meta-analysis. \u003cem\u003eAge and Ageing\u003c/em\u003e, \u003cem\u003e53\u003c/em\u003e(7), afae162. https://doi.org/10.1093/ageing/afae162\u003c/li\u003e\n \u003cli\u003eZhi, K., Chen, Y., \u0026amp; Huang, J. (2021). China\u0026rsquo;s challenge in promoting older migrants\u0026rsquo; health and wellbeing: A productive ageing perspective. \u003cem\u003eBMJ (Clinical Research Ed.)\u003c/em\u003e, \u003cem\u003e375\u003c/em\u003e, n2874. https://doi.org/10.1136/bmj.n2874\u003c/li\u003e\n \u003cli\u003eZhou, B., Ma, R., Wang, M., \u0026amp; Wang, Y. (2025a). Dose-response relationship between nighttime sleep duration and intrinsic capacity declines among Chinese elderly: A cross-sectional study from CHARLS. \u003cem\u003eBMC Public Health\u003c/em\u003e, \u003cem\u003e25\u003c/em\u003e(1), 1034. https://doi.org/10.1186/s12889-025-22294-2\u003c/li\u003e\n \u003cli\u003eZhou, B., Ma, R., Wang, M., \u0026amp; Wang, Y. (2025b). Dose-response relationship between nighttime sleep duration and intrinsic capacity declines among Chinese elderly: A cross-sectional study from CHARLS. \u003cem\u003eBMC Public Health\u003c/em\u003e, \u003cem\u003e25\u003c/em\u003e(1), 1034. https://doi.org/10.1186/s12889-025-22294-2\u003c/li\u003e\n \u003cli\u003eZhou, Y., \u0026amp; Ma, L. (2022). Intrinsic Capacity in Older Adults: Recent Advances. \u003cem\u003eAging and Disease\u003c/em\u003e, \u003cem\u003e13\u003c/em\u003e(2), 353\u0026ndash;359. https://doi.org/10.14336/AD.2021.0818\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"bmc-geriatrics","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"bgtc","sideBox":"Learn more about [BMC Geriatrics](http://bmcgeriatr.biomedcentral.com/)","snPcode":"","submissionUrl":"https://www.editorialmanager.com/bgtc/default.aspx","title":"BMC Geriatrics","twitterHandle":"BMC_series","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"em","reportingPortfolio":"BMC Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"migrant older adults, intrinsic capacity, machine learning","lastPublishedDoi":"10.21203/rs.3.rs-8610961/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-8610961/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003ch2\u003eBackground\u003c/h2\u003e \u003cp\u003eWhile migrating to join adult children can provide familial support, it may also pose risks to the intrinsic capacity (IC) of older adults due to disrupted social networks and environmental stressors. To address the paradox, we applied machine learning within a health ecology framework to develop risk predictive models and to rank key drivers of IC decline among migrant older adults with children (MOAC).\u003c/p\u003e\u003ch2\u003eMethods\u003c/h2\u003e \u003cp\u003eThis multi-center, large-sample cross-sectional study was conducted from December 2022 to September 2023. A total of 3016 MOAC were randomly recruited across three Chinese regions (XX, XX, XX). The health ecology model was operationalized into a structured questionnaire to assess intrinsic capacity and its potential risk factors across five domains. Data were analyzed using interpretable machine learning techniques, including predictor importance ranking and partial dependence plots (PDPs).\u003c/p\u003e\u003ch2\u003eResults\u003c/h2\u003e \u003cp\u003eThe top ten predictors of IC decline were identified as social network, health literacy, sleep quality, migration reason, economic level, number of chronic diseases, primary source of financial support, prior career, multi-layered medical insurance coverage, and age. The Random Forest model provided the most robust prediction of IC decline (AUC: 0.843; 95% CI: 0.820\u0026ndash;0.867). PDPs further delineate the non-linear patterns through which these predictors influence IC decline.\u003c/p\u003e\u003ch2\u003eConclusions\u003c/h2\u003e \u003cp\u003eThe findings provide practical and actionable guidance for community healthcare providers. Screening based on non-modifiable risks (advanced age, chronic diseases, etc) enable the early identification of high-risk MOAC individuals. Building on this, tailored interventions can be developed to target modifiable factors such as health literacy and social networks, thereby promoting IC and enhancing quality of life.\u003c/p\u003e","manuscriptTitle":"Machine learning-based predictive modeling of decline in intrinsic capacity among migrant older adults with children","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-02-19 11:35:22","doi":"10.21203/rs.3.rs-8610961/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"editorInvitedReview","content":"","date":"2026-03-08T17:16:50+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"45397348089165533172544570545693224460","date":"2026-02-28T22:02:37+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2026-02-13T12:45:09+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2026-02-13T08:16:37+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"","date":"2026-01-22T11:20:14+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2026-01-22T06:33:10+00:00","index":"","fulltext":""},{"type":"submitted","content":"BMC Geriatrics","date":"2026-01-22T06:23:08+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"bmc-geriatrics","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"bgtc","sideBox":"Learn more about [BMC Geriatrics](http://bmcgeriatr.biomedcentral.com/)","snPcode":"","submissionUrl":"https://www.editorialmanager.com/bgtc/default.aspx","title":"BMC Geriatrics","twitterHandle":"BMC_series","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"em","reportingPortfolio":"BMC Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"f50cb28f-a08b-4fd0-8f10-3ea2a0a8ba39","owner":[],"postedDate":"February 19th, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[],"tags":[],"updatedAt":"2026-02-19T11:35:23+00:00","versionOfRecord":[],"versionCreatedAt":"2026-02-19 11:35:22","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-8610961","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-8610961","identity":"rs-8610961","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}