Interpretable Machine Learning Models for Predicting the Risk of Incident Cardiometabolic Multimorbidity in Chinese Postmenopausal Women

doi:10.12114/j.issn.1007-9572.2025.0358

Chinese General Practice ›› 2026, Vol. 29 ›› Issue (23): 3294-3306.DOI: 10.12114/j.issn.1007-9572.2025.0358

Special Issue: 内分泌代谢性疾病最新文章合辑；心血管最新文章合辑

• Article • Previous Articles Next Articles

Interpretable Machine Learning Models for Predicting the Risk of Incident Cardiometabolic Multimorbidity in Chinese Postmenopausal Women

1. Second College of Clinical Medicine, Beijing University of Chinese Medicine, Beijing 100029, China
2. Affiliated Guangdong Hospital of Integrated Traditional Chinese and Western Medicine of Guangzhou University of Chinese Medicine, Foshan 528200, China
3. Tianjin University of Traditional Chinese Medicine, Tianjin 301617, China
4. Department of Cardiology, Dongfang Hospital Affiliated to Beijing University of Chinese Medicine, Beijing 100078, China
5. Shenzhen Hospital of Southern Medical University, Shenzhen 518000, China

Received:2025-08-12 Revised:2025-12-30 Published:2026-08-15 Online:2026-07-03
Contact: LI Weiwei

预测中国绝经后女性新发心脏代谢性共病风险的可解释机器学习模型研究

1．100029　北京市，北京中医药大学第二临床医学院
2．528200　广东省佛山市，广州中医药大学附属广东中西医结合医院
3．301617　天津市，天津中医药大学
4．100078　北京市，北京中医药大学东方医院心内科
5．518000　广东省深圳市，南方医科大学深圳医院

通讯作者: 李伟炜
作者简介:
作者贡献：
袁杭滔负责研究的思路构思、数据分析、绘制图表与文章撰写；洪妍进行数据的整理与协助撰写；袁沛宏负责文章撰写及语言润色；林博负责统计学计算及提出修改建议；崔晓云负责文章的思路指导与文章修改；李伟炜负责文章的思路指导，质量控制与审查，对文章整体负责，监督管理。
基金资助:
国家自然科学基金面上项目(81774044); 中央高水平中医医院临床科研业务费资助(DFGZRB-2024GJRCO17)

Abstract

Abstract:

Background

Cardiovascular diseases are prevalent in China, with cardiometabolic multimorbidity (CMM) being a common comorbidity pattern. Postmenopausal women represent a high-risk group for cardiovascular diseases, yet there is a lack of predictive models for CMM risk specifically in this population.

Objective

To develop an interpretable machine learning (ML) model to predict the risk of CMM among Chinese postmenopausal women, based on data from the China Health and Retirement Longitudinal Study (CHARLS).

Methods

The study included postmenopausal women aged≥45 years from the CHARLS cohort in 2011 who were free of CMM at baseline. Data on demographic characteristics, family background, health status, and laboratory indicators were collected at baseline and during follow-up in 2013, 2015, 2018, and 2020 to observe CMM incidence. Feature selection was performed using the least absolute shrinkage and selection operator (LASSO) algorithm. Seven ML algorithms were constructed for risk prediction. The optimal model was further optimized on the test set using a combined strategy of "class_weight='balanced' dynamic weighting+optimal threshold selection" and visually interpreted using Shapley Additive Explanations (SHAP). Model performance was evaluated using the area under the receiver operating characteristic curve (AUC), sensitivity, specificity, precision, and F1-score.

Results

A total of 5 575 participants completed the 4 rounds of follow-up and were included, comprising 4 363 in the non-CMM group and 1 212 in the CMM group. Over a median follow-up of 9 years, the cumulative incidence of CMM was 21.74%. LASSO regression identified 22 key features as significant predictors of CMM: self-rated health, mental disorders, arthritis, dyslipidemia, kidney disease, retirement status, systolic blood pressure (SBP), diastolic blood pressure (DBP), mean pulse rate, waist circumference, BMI, headache, lower back pain, serum creatinine (Scr), triglycerides (TG), C-reactive protein (CRP), glycated hemoglobin (HbA_1c), uric acid (UA), age, Center for Epidemiologic Studies Depression Scale (CES-D) score, smoking status, and geographic region. Among the models, the Logistic regression (LR) model demonstrated the best predictive performance (test set AUC=0.758, accuracy =79.2%). The SHAP mean bar plot revealed core predictors: SBP, HbA_1c, geographic region, waist circumference, CES-D score, BMI, DBP, and age. The SHAP summary plot indicated that higher values of SBP, HbA_1c, waist circumference, and others were associated with increased predicted CMM risk.

Conclusion

This study develops a clinically interpretable prediction model for CMM in Chinese postmenopausal women, with the LR algorithm showing favorable performance. Key risk factors include SBP, HbA_1c, and waist circumference. The model provides an evidence-based tool for screening high-risk individuals and guiding personalized interventions.

Key words: Cardiometabolic multimorbidity, Postmenopausal women, Middle-aged, The elderly, CHARLS, Prediction model

摘要：

背景

心血管疾病是我国人群的常见疾病，心脏代谢性共病（CMM）是其中常见的共病类型，绝经后女性是心血管疾病的高发人群，但针对该人群CMM风险的预测模型尚缺少。

目的

基于中国健康与养老追踪调查（CHARLS）队列，开发可解释机器学习（ML）模型预测中国绝经后女性CMM风险。

方法

本研究纳入2011年参与CHARLS队列且未患有CMM的年龄≥45岁的绝经后女性，收集人口学特征、家庭特征、健康状况及实验室检查指标等特征资料，并于2013年、2015年、2018年、2020年进行随访和数据收集，观察CMM发生情况。通过最小绝对收缩和选择运算符（LASSO）算法方法进行特征选择，构建7种ML算法进行风险预测，对最优模型的测试集采用"class_weight='balanced’动态平衡权重+最优阈值筛选"优化模型并应用Shapley加法解释（SHAP）进行可视化分析；以受试者工作特征曲线下面积（AUC）、灵敏度、特异度、准确率和F1分数评估模型性能。

结果

共有5 575名参与者完成4轮随访并纳入此项研究，其中非CMM者4 363例，CMM者1 212例；中位随访9年，CMM累积发病率为21.74%。LASSO回归确定了22个关键特征作为CMM的重要预测因子：健康自评、精神疾病、关节炎、血脂异常、肾脏疾病、是否退休、收缩压（SBP）、舒张压（DBP）、平均脉搏、腰围、BMI、头痛、腰痛、血肌酐（Scr）、甘油三酯（TG）、C反应蛋白（CRP）、糖化血红蛋白（HbA_1c）、尿酸（UA）、年龄、流调中心抑郁量表（CES-D）、吸烟、所属地区。各模型中，逻辑回归（LR）模型预测性能最优（测试集AUC=0.758，准确率为79.2%）。SHAP平均值条形图显示核心预测因子为：SBP、HbA_1c、所属地区、腰围、CES-D、BMI、DBP、年龄。SHAP汇总图显示，SBP、HbA_1c、腰围等与较高的CMM风险预测值相关。

结论

本研究针对中国绝经后女性的CMM风险开发了一个临床可视化预测模型，LR算法在其中表现出良好性能；SBP、HbA_1c、腰围等特征是关键风险因素；该模型可为高危人群筛查和个体化干预提供循证依据。

关键词: 心脏代谢性共病, 绝经后女性, 中年人, 老年人, CHARLS, 预测模型

YUAN Hangtao,HONG Yan,YUAN Peihong, et al. Interpretable Machine Learning Models for Predicting the Risk of Incident Cardiometabolic Multimorbidity in Chinese Postmenopausal Women[J]. Chinese General Practice, 2026, 29(23): 3294-3306. DOI: 10.12114/j.issn.1007-9572.2025.0358.
袁杭滔,洪妍,袁沛宏等. 预测中国绝经后女性新发心脏代谢性共病风险的可解释机器学习模型研究[J]. 中国全科医学, 2026, 29(23): 3294-3306. DOI: 10.12114/j.issn.1007-9572.2025.0358.

Figures/Tables 9

References 75

[1]	Emerging Risk Factors Collaboration. Association of cardiometabolic multimorbidity with mortality[J]. JAMA, 2015, 314(1): 52-60. DOI: 10.1001/jama.2015.7008.
[2]	Glynn L G. Multimorbidity: another key issue for cardiovascular medicine[J]. Lancet, 2009, 374(9699): 1421-1422. DOI: 10.1016/S0140-6736(09)61863-8.
[3]	Weiss C O, Boyd C M, Yu Q L, et al. Patterns of prevalent major chronic disease among older adults in the United States[J]. JAMA, 2007, 298(10): 1160-1162. DOI: 10.1001/jama.298.10.1160-b.
[4]	Prados-Torres A, Calderón-Larrañaga A, Hancco-Saavedra J, et al. Multimorbidity patterns: a systematic review[J]. J Clin Epidemiol, 2014, 67(3): 254-266. DOI: 10.1016/j.jclinepi.2013.09.021.
[5]	Zhang D D, Tang X, Shen P, et al. Multimorbidity of cardiometabolic diseases: prevalence and risk for mortality from one million Chinese adults in a longitudinal cohort study[J]. BMJ Open, 2019, 9(3): e024476. DOI: 10.1136/bmjopen-2018-024476.
[6]	Liu J L, Liu M, Chai Z L, et al. Projected rapid growth in diabetes disease burden and economic burden in China: a spatio-temporal study from 2020 to 2030[J]. Lancet Reg Health West Pac, 2023, 33: 100700. DOI: 10.1016/j.lanwpc.2023.100700.
[7]	Zhou M G, Wang H D, Zeng X Y, et al. Mortality, morbidity, and risk factors in China and its provinces, 1990-2017: a systematic analysis for the Global Burden of Disease Study 2017[J]. Lancet, 2019, 394(10204): 1145-1158. DOI: 10.1016/S0140-6736(19)30427-1.
[8]	Moran A, Zhao D, Gu D F, et al. The future impact of population growth and aging on coronary heart disease in China: projections from the Coronary Heart Disease Policy Model-China[J]. BMC Public Health, 2008, 8: 394. DOI: 10.1186/1471-2458-8-394.
[9]	Tu W J, Wang L D, Yan F, et al. China stroke surveillance report 2021[J]. Military Med Res, 2023, 10: 33. DOI: 10.1186/s40779-023-00463-x.
[10]	Salisbury C, Man M S, Bower P, et al. Management of multimorbidity using a patient-centred care model: a pragmatic cluster-randomised trial of the 3D approach[J]. Lancet, 2018, 392(10141): 41-50. DOI: 10.1016/S0140-6736(18)31308-4.
[11]	Celermajer D S, Chow C K, Marijon E, et al. Cardiovascular disease in the developing world: prevalences, patterns, and the potential of early disease detection[J]. J Am Coll Cardiol, 2012, 60(14): 1207-1216. DOI: 10.1016/j.jacc.2012.03.074.
[12]	Mao D H, Mu J F, Li Y J, et al. Network-based machine learning reveals cardiometabolic multimorbidity patterns and modifiable lifestyle factors: a community-focused analysis of NHANES 2015-2018[J]. BMC Public Health, 2025, 25(1): 2297. DOI: 10.1186/s12889-025-23483-9.
[13]	Sakakibara B M, Obembe A O, Eng J J. The prevalence of cardiometabolic multimorbidity and its association with physical activity, diet, and stress in Canada: evidence from a population-based cross-sectional study[J]. BMC Public Health, 2019, 19(1): 1361. DOI: 10.1186/s12889-019-7682-4.
[14]	Gerges S H, El-Kadi A O S. Changes in cardiovascular arachidonic acid metabolism in experimental models of menopause and implications on postmenopausal cardiac hypertrophy[J]. Prostaglandins Other Lipid Mediat, 2024, 173: 106851. DOI: 10.1016/j.prostaglandins.2024.106851.
[15]	Yu Q. Population education of menopausal knowledge in China[J]. Climacteric, 2019, 22(4): 323. DOI: 10.1080/13697137.2019.1609767.
[16]	Esteva A, Kuprel B, Novoa R A, et al. Dermatologist-level classification of skin cancer with deep neural networks[J]. Nature, 2017, 542(7639): 115-118. DOI: 10.1038/nature21056.
[17]	Deo R C. Machine learning in medicine[J]. Circulation, 2015, 132(20): 1920-1930. DOI: 10.1161/CIRCULATIONAHA.115.001593.
[18]	Huang Q, Zou X T, Lian Z H, et al. Predicting cardiovascular outcomes in Chinese patients with type 2 diabetes by combining risk factor trajectories and machine learning algorithm: a cohort study[J]. Cardiovasc Diabetol, 2025, 24(1): 61. DOI: 10.1186/s12933-025-02611-0.
[19]	Qamar T, Bawany N Z. Understanding the black-box: towards interpretable and reliable deep learning models[J]. PeerJ Comput Sci, 2023, 9: e1629. DOI: 10.7717/peerj-cs.1629.
[20]	Nohara Y, Matsumoto K, Soejima H, et al. Explanation of machine learning models using shapley additive explanation and application for real data in hospital[J]. Comput Meth Programs Biomed, 2022, 214: 106584. DOI: 10.1016/j.cmpb.2021.106584.
[21]	Zhao Y H, Hu Y S, Smith J P, et al. Cohort profile: the China health and retirement longitudinal study(CHARLS)[J]. Int J Epidemiol, 2014, 43(1): 61-68. DOI: 10.1093/ije/dys203.
[22]	Si F, Liu Q, Yu J. A prediction study on the occurrence risk of heart disease in older hypertensive patients based on machine learning[J]. BMC Geriatr, 2025, 25(1): 27. DOI: 10.1186/s12877-025-05679-1.
[23]	Xiao Y, Zhao Z J, Su C G, et al. An interpretable machine learning model for predicting depression in middle-aged and elderly cancer patients in China: a study based on the CHARLS cohort[J]. BMC Psychiatry, 2025, 25(1): 610. DOI: 10.1186/s12888-025-07074-x.
[24]	Luo Y X, Yin Z Y, Li X, et al. Cardiometabolic index predicts cardiovascular events in aging population: a machine learning-based risk prediction framework from a large-scale longitudinal study[J]. Front Endocrinol, 2025, 16: 1551779. DOI: 10.3389/fendo.2025.1551779.
[25]	中国成人血脂异常防治指南修订联合委员会. 中国成人血脂异常防治指南(2016年修订版)[J]. 中国循环杂志, 2016, 31(10): 937-950. DOI: 10.3969/j.issn.1000-3614.2016.10.001.
[26]	Gao Y, Chen L, Jia Z H, et al. Social participation and health in middle-aged and older empty nesters: a study on gender differences[J]. SSM Popul Health, 2024, 25: 101641. DOI: 10.1016/j.ssmph.2024.101641.
[27]	American Diabetes Association Professional Practice Committee. 2. Diagnosis and classification of diabetes: standards of care in diabetes-2024[J]. Diabetes Care, 2024, 47(Suppl 1): S20-S42. DOI: 10.2337/dc24-S002.
[28]	Li H B, Zheng D Q, Li Z W, et al. Association of depressive symptoms with incident cardiovascular diseases in middle-aged and older Chinese adults[J]. JAMA Netw Open, 2019, 2(12): e1916591. DOI: 10.1001/jamanetworkopen.2019.16591.
[29]	Zhao Y, Zhang H, Liu X Y, et al. The prevalence of cardiometabolic multimorbidity and its associations with health outcomes among women in China[J]. Front Cardiovasc Med, 2023, 10: 922932. DOI: 10.3389/fcvm.2023.922932.
[30]	Nagel C L, Chen S T, Allore H G, et al. Longitudinal sequencing of cardiometabolic multimorbidity among older adults and association with subsequent dementia onset[J]. PLoS One, 2025, 20(7): e0326309. DOI: 10.1371/journal.pone.0326309.
[31]	Marin A G, Filipescu A, Vladareanu R, et al. Metabolic syndrome and survival outcomes in endometrial cancer[J]. Cureus, 2024. DOI: 10.7759/cureus.60324.
[32]	Sun J Y, Lin J C, Yin D, et al. Androgen receptor inhibitor ameliorates pulmonary arterial hypertension by enhancing the apoptosis level through suppressing the Notch3/Hes5 pathway[J]. Front Pharmacol, 2025, 16: 1572489. DOI: 10.3389/fphar.2025.1572489.
[33]	Kułak-Bejda A, Krajewska-Ferishah K, Szyszko-Perłowska A, et al. Risk assessment of depression amongst women during menopause before and during the COVID-19 pandemic[J]. Int J Environ Res Public Health, 2023, 20(1): 596. DOI: 10.3390/ijerph20010596.
[34]	Siriprapanonkul P, Orprayoon N, Tuntiviriyapun P, et al. Alu hypomethylation in naturally and surgically postmenopausal women; a cross-sectional study[J]. PLoS One, 2022, 17(8): e0273403. DOI: 10.1371/journal.pone.0273403.
[35]	Wu Y J, Sun W J, Liu H, et al. Age at menopause and risk of developing endometrial cancer: a meta-analysis[J]. Biomed Res Int, 2019, 2019: 8584130. DOI: 10.1155/2019/8584130.
[36]	Pu D, Tan R, Yu Q, et al. Metabolic syndrome in menopause and associated factors: a meta-analysis[J]. Climacteric, 2017, 20(6): 583-591. DOI: 10.1080/13697137.2017.1386649.
[37]	Bagheri Lankarani K, Jamalinia M, Zare F, et al. Liver-kidney-metabolic health, sex, and menopause impact total scores and monovessel vs. multivessel coronary artery calcification[J]. Adv Ther, 2025, 42(4): 1729-1744. DOI: 10.1007/s12325-025-03121-x.
[38]	Lee D H, Park J E, Kim S Y, et al. Association between the triglyceride-glucose(TyG) index and increased blood pressure in normotensive subjects: a population-based study[J]. Diabetol Metab Syndr, 2022, 14(1): 161. DOI: 10.1186/s13098-022-00927-5.
[39]	Hotamisligil G S. Inflammation, metaflammation and immunometabolic disorders[J]. Nature, 2017, 542(7640): 177-185. DOI: 10.1038/nature21363.
[40]	Suren Garg S, Kushwaha K, Dubey R, et al. Association between obesity, inflammation and insulin resistance: insights into signaling pathways and therapeutic interventions[J]. Diabetes Res Clin Pract, 2023, 200: 110691. DOI: 10.1016/j.diabres.2023.110691.
[41]	Liu F H, Meng Y, Wang R R, et al. Combined influence of depression symptoms and ratio of triglyceride to high-density lipoprotein cholesterol on cardiometabolic multimorbidity: findings from the China Health and Retirement Longitudinal Study 2011-2018[J]. J Affect Disord, 2024, 360: 242-248. DOI: 10.1016/j.jad.2024.05.159.
[42]	Qin K, Bai S G, Chen W Q, et al. Association of comorbid depression and obesity with cardiometabolic multimorbidity among middle-aged and older Chinese adults: a cohort study[J]. Arch Gerontol Geriatr, 2023, 107: 104912. DOI: 10.1016/j.archger.2022.104912.
[43]	He Q L, Yu W, Long C B, et al. Obesity-depression-metabolism triad: amplification of cardiometabolic multimorbidity risk in Chinese adults(CHARLS 2011-2018)[J]. Lipids Health Dis, 2025, 24(1): 295. DOI: 10.1186/s12944-025-02724-1.
[44]	Mills K T, Stefanescu A, He J. The global epidemiology of hypertension[J]. Nat Rev Nephrol, 2020, 16(4): 223-237. DOI: 10.1038/s41581-019-0244-2.
[45]	Wang Z W, Chen Z, Zhang L F, et al. Status of hypertension in China: results from the China hypertension survey, 2012-2015[J]. Circulation, 2018, 137(22): 2344-2356. DOI: 10.1161/CIRCULATIONAHA.117.032380.
[46]	Stanciu S, Rusu E, Miricescu D, et al. Links between metabolic syndrome and hypertension: the relationship with the current antidiabetic drugs[J]. Metabolites, 2023, 13(1): 87. DOI: 10.3390/metabo13010087.
[47]	Singleton C M H, Brar S, Robertson N, et al. Cardiometabolic risk factors in South American children: a systematic review and meta-analysis[J]. PLoS One, 2023, 18(11): e0293865. DOI: 10.1371/journal.pone.0293865.
[48]	Poulsen M K, Henriksen J E, Dahl J, et al. Left ventricular diastolic function in type 2 diabetes mellitus: prevalence and association with myocardial and vascular disease[J]. Circ Cardiovasc Imaging, 2010, 3(1): 24-31. DOI: 10.1161/CIRCIMAGING.109.855510.
[49]	Liu M H, Chen R M, Zheng Z W, et al. Mechanisms of inflammatory microenvironment formation in cardiometabolic diseases: molecular and cellular perspectives[J]. Front Cardiovasc Med, 2025, 11: 1529903. DOI: 10.3389/fcvm.2024.1529903.
[50]	Guardiola M, Rehues P, Amigó N, et al. Increasing the complexity of lipoprotein characterization for cardiovascular risk in type 2 diabetes[J]. Eur J Clin Invest, 2024, 54(8): e14214. DOI: 10.1111/eci.14214.
[51]	Ross R, Neeland I J, Yamashita S, et al. Waist circumference as a vital sign in clinical practice: a Consensus Statement from the IAS and ICCR Working Group on Visceral Obesity[J]. Nat Rev Endocrinol, 2020, 16(3): 177-189. DOI: 10.1038/s41574-019-0310-7.
[52]	Shikuma J, Nagai Y, Otaki M, et al. Anterior-posterior abdominal diameter and waist circumference have comparable predictive ability for visceral fat area[J]. Obes Sci Pract, 2025, 11(1): e70047. DOI: 10.1002/osp4.70047.
[53]	Su Y, Sun J Y, Zhou Y, et al. The relationship of waist circumference with the morbidity of cardiovascular diseases and all-cause mortality in metabolically healthy individuals: a population-based cohort study[J]. Rev Cardiovasc Med, 2024, 25(6): 212. DOI: 10.31083/j.rcm²506212.
[54]	Chen Q C, Li L L, Yi J Z, et al. Waist circumference increases risk of coronary heart disease: evidence from a Mendelian randomization study[J]. Mol Genet Genomic Med, 2020, 8(4): e1186. DOI: 10.1002/mgg3.1186.
[55]	Levine D A, Calhoun D A, Prineas R J, et al. Moderate waist circumference and hypertension prevalence: the REGARDS Study[J]. Am J Hypertens, 2011, 24(4): 482-488. DOI: 10.1038/ajh.2010.258.
[56]	Ai F Z, Li E G, Ji Q Q, et al. Construction of a machine learning-based risk prediction model for depression in middle-aged and elderly hypertensive people in China: a longitudinal study[J]. Front Psychiatry, 2024, 15: 1398596. DOI: 10.3389/fpsyt.2024.1398596.
[57]	Haam J H, Kim B T, Kim E M, et al. Diagnosis of obesity: 2022 update of clinical practice guidelines for obesity by the Korean society for the study of obesity[J]. J Obes Metab Syndr, 2023. DOI: 10.7570/jome23031 DOI: 10.7570/jome23031.
[58]	Yin X, Chen Y C, Zhou L H, et al. Association between chronic pain and risk of cardiometabolic multimorbidity: a prospective cohort study[J]. Reg Anesth Pain Med, 2025, 50(11): 858-864. DOI: 10.1136/rapm-2024-105486.
[59]	Pimenta S, Gonçalves H, Pimenta M, et al. Central obesity and radiographic severity are associated with symptomatic hand osteoarthritis: a population-based cross-sectional study[J]. Rheumatol Int, 2025, 45(6): 143. DOI: 10.1007/s00296-025-05891-7.
[60]	Hsueh T Y, Chen Y P. The association between dyslipidemia and hand osteoarthritis: a systematic review and meta-analysis[J]. Clin Rheumatol, 2025, 44(4): 1439-1448. DOI: 10.1007/s10067-025-07384-1.
[61]	Pavão A L B, Werneck G L, Campos M R. Self-rated health and the association with social and demographic factors, health behavior, and morbidity: a national health survey[J]. Cad Saude Publica, 2013, 29(4): 723-734.
[62]	Wang L D, Dong W Z, Ou Y Q, et al. Regional differences and determinants of self-rated health in a lower middle income rural Society of China[J]. Int J Equity Health, 2018, 17(1): 162. DOI: 10.1186/s12939-018-0875-0.
[63]	Song D X, Liu D S, Ning W H, et al. Incidence, prevalence and characteristics of multimorbidity in different age groups among urban hospitalized patients in China[J]. Sci Rep, 2023, 13(1): 18798. DOI: 10.1038/s41598-023-46227-4.
[64]	Huang Q, Jiang Z H, Shi B, et al. Characterisation of cardiovascular disease(CVD) incidence and machine learning risk prediction in middle-aged and elderly populations: data from the China health and retirement longitudinal study(CHARLS)[J]. BMC Public Health, 2025, 25(1): 518. DOI: 10.1186/s12889-025-21609-7.
[65]	Choi S, Park T, Je Y. Long-term alcohol consumption and incident health risk conditions related to cardiometabolic risk markers: a 20-year prospective cohort study[J]. Addiction, 2025, 120(9): 1840-1852. DOI: 10.1111/add.70092.
[66]	Liu J L, Shen M W, Zhuang G H, et al. Investigating the temporal trends of diabetes disease burden in China during 1990-2019 from a global perspective[J]. Front Endocrinol, 2024, 15: 1324318. DOI: 10.3389/fendo.2024.1324318.
[67]	Zhang H H, Zhao Z Y, Wu Z, et al. Identifying interactions among air pollutant emissions on diabetes prevalence in Northeast China using a complex network[J]. Int J Biometeorol, 2024, 68(2): 393-400. DOI: 10.1007/s00484-023-02597-y.
[68]	Turner J B, Kumar A, Koch C A. The effects of indoor and outdoor temperature on metabolic rate and adipose tissue - the Mississippi perspective on the obesity epidemic[J]. Rev Endocr Metab Disord, 2016, 17(1): 61-71. DOI: 10.1007/s11154-016-9358-z.
[69]	Du W Q, Liu L X, Ma Y, et al. Analysis of the gut microbiome in obese native Tibetan children living at different altitudes: a case-control study[J]. Front Public Health, 2022, 10: 963202. DOI: 10.3389/fpubh.2022.963202.
[70]	Chan C B, Ryan D A J, Tudor-Locke C. Relationship between objective measures of physical activity and weather: a longitudinal study[J]. Int J Behav Nutr Phys Act, 2006, 3: 21. DOI: 10.1186/1479-5868-3-21.
[71]	Zhang G G, Dong S J, Wang L. Construction of a machine learning-based risk prediction model for depression in middle-aged and elderly patients with cardiovascular metabolic diseases in China: a longitudinal study[J]. BMC Public Health, 2025, 25(1): 1904. DOI: 10.1186/s12889-025-23075-7.
[72]	Wu Y X, Xin B, Wan Q Y, et al. Risk factors and prediction models for cardiovascular complications of hypertension in older adults with machine learning: a cross-sectional study[J]. Heliyon, 2024, 10(6): e27941. DOI: 10.1016/j.heliyon.2024.e27941.
[73]	Griffith L E, Gilsing A, Mangin D, et al. Multimorbidity frameworks impact prevalence and relationships with patient-important outcomes[J]. J Am Geriatr Soc, 2019, 67(8): 1632-1640. DOI: 10.1111/jgs.15921.
[74]	Mao D H, Li G K, Liang M X, et al. Dietary patterns and multiple chronic diseases in older adults[J]. Nutr Metab, 2024, 21(1): 36. DOI: 10.1186/s12986-024-00814-y.
[75]	Bao Y N, Lu P Y, Wang M J, et al. Exploring multimorbidity profiles in middle-aged inpatients: a network-based comparative study of China and the United Kingdom[J]. BMC Med, 2023, 21(1): 495. DOI: 10.1186/s12916-023-03204-y.

组别	例数	年龄[M（P₂₅，P₇₅），岁]	婚姻状况[例（%）]		居住地[例（%）]		血脂异常[例（%）]
组别	例数	年龄[M（P₂₅，P₇₅），岁]	其他	已婚/同居	城市	农村	无	有
非CMM者	4 363	56.0（49.0，62.0）	492（11.28）	3 871（88.72）	1 552（35.57）	2 811（64.43）	4 100（93.97）	263（6.03）
CMM者	1 212	59.0（53.0，65.0）	182（15.02）	1 030（84.98）	455（37.54）	757（62.46）	1 052（86.80）	160（13.20）
χ²（Z）值		-10.54^a	12.48		1.60		69.61
P值		<0.001	<0.001		0.206		<0.001
组别	肾脏疾病[例（%）]		受教育水平[例（%）]				饮酒史[例（%）]
组别	无	有	小学以下	小学	中学	大学及以上	无	有
非CMM者	4 187（95.97）	176（4.03）	2 529（57.96）	749（17.17）	712（16.32）	373（8.55）	3 717（85.19）	646（14.81）
CMM者	1 131（93.32）	81（6.68）	736（60.73）	207（17.08）	184（15.18）	85（7.01）	1 022（84.32）	190（15.68）
χ²（Z）值	15.14		4.73				0.56
P值	<0.001		0.193				0.453
组别	吸烟史[例（%）]		所属地区[例（%）]
组别	无	有	东北地区	华北地区	华中地区	东南沿海地区	华东地区	西北地区	西南地区
非CMM者	4 124（94.52）	239（5.48）	242（5.55）	445（10.20）	702（16.09）	570（13.06）	1 262（28.93）	289（6.62）	853（19.55）
CMM者	1 146（94.55）	66（5.45）	90（7.43）	201（16.58）	226（18.65）	100（8.25）	330（27.23）	95（7.84）	170（14.03）
χ²（Z）值	<0.01		79.77
P值	0.965		<0.001

组别	例数	年龄[M（P₂₅，P₇₅），岁]	婚姻状况[例（%）]		居住地[例（%）]		血脂异常[例（%）]
组别	例数	年龄[M（P₂₅，P₇₅），岁]	其他	已婚/同居	城市	农村	无	有
非CMM者	4 363	56.0（49.0，62.0）	492（11.28）	3 871（88.72）	1 552（35.57）	2 811（64.43）	4 100（93.97）	263（6.03）
CMM者	1 212	59.0（53.0，65.0）	182（15.02）	1 030（84.98）	455（37.54）	757（62.46）	1 052（86.80）	160（13.20）
χ²（Z）值		-10.54^a	12.48		1.60		69.61
P值		<0.001	<0.001		0.206		<0.001
组别	肾脏疾病[例（%）]		受教育水平[例（%）]				饮酒史[例（%）]
组别	无	有	小学以下	小学	中学	大学及以上	无	有
非CMM者	4 187（95.97）	176（4.03）	2 529（57.96）	749（17.17）	712（16.32）	373（8.55）	3 717（85.19）	646（14.81）
CMM者	1 131（93.32）	81（6.68）	736（60.73）	207（17.08）	184（15.18）	85（7.01）	1 022（84.32）	190（15.68）
χ²（Z）值	15.14		4.73				0.56
P值	<0.001		0.193				0.453
组别	吸烟史[例（%）]		所属地区[例（%）]
组别	无	有	东北地区	华北地区	华中地区	东南沿海地区	华东地区	西北地区	西南地区
非CMM者	4 124（94.52）	239（5.48）	242（5.55）	445（10.20）	702（16.09）	570（13.06）	1 262（28.93）	289（6.62）	853（19.55）
CMM者	1 146（94.55）	66（5.45）	90（7.43）	201（16.58）	226（18.65）	100（8.25）	330（27.23）	95（7.84）	170（14.03）
χ²（Z）值	<0.01		79.77
P值	0.965		<0.001

变量	2013年		2015年		2018年		2020年
变量	例数	CMM	例数	CMM	例数	CMM	例数	CMM
年龄
45~60岁	3 842	93（2.42）	3 707	261（7.04）	3 765	575（15.27）	3 760	701（18.64）
>60岁	2 211	91（4.12）	2 040	227（11.13）	1 934	469（24.25）	1 815	511（28.15）
χ²值		13.68		28.28		68.83		65.08
P值		<0.001		<0.001		<0.001		<0.001
婚姻状况
已婚/同居	5 172	145（2.80）	4 947	392（7.92）	4 957	866（16.74）	4 901	1 030（21.02）
其他	881	39（4.43）	800	96（12.00）	742	178（23.99）	674	182（27.00）
χ²值		6.73		14.72		18.33		12.48
P值		0.013		<0.001		<0.001		<0.001
受教育水平
小学以下	3 612	113（3.13）	3 465	315（9.09）	3 371	645（19.13）	3 265	738（22.60）
小学	1 011	26（2.57）	954	73（7.65）	968	175（18.08）	958	207（21.61）
中学	934	24（2.57）	901	69（7.66）	885	149（16.84）	894	182（20.36）
大学及以上	496	21（4.23）	427	31（7.26）	475	75（15.79）	458	85（18.56）
χ²值		3.95		4.11		4.86		5.17
P值		0.267		0.250		0.182		0.160
饮酒史
无	5 294	165（3.12）	5 038	447（8.87）	4 978	918（18.44）	4 872	1 069（21.94）
有	759	19（2.50）	709	41（5.78）	721	126（17.48）	703	143（20.34）
χ²值		0.85		7.64		0.39		0.92
P值		0.419		0.007		0.565		0.361
吸烟史
无	5 688	171（3.01）	5 408	465（8.60）	5 375	982（18.27）	5 270	1 146（21.75）
有	365	13（3.56）	339	23（6.78）	324	62（19.14）	305	66（21.64）
χ²值		0.36		1.35		0.15		<0.01
P值		0.659		0.288		0.751		1.000
所属地区
东北地区	419	22（5.25）	386	50（12.95）	378	92（24.34）	342	94（27.49）
华北地区	714	39（5.46）	662	102（15.41）	684	185（27.05）	651	195（29.95）
华中地区	997	31（3.11）	922	74（8.03）	938	185（19.72）	930	225（24.19）
东南沿海地区	723	14（1.94）	705	29（4.11）	667	73（10.94）	679	101（14.87）
华东地区	1 678	43（2.56）	1 592	118（7.41）	1 583	280（17.69）	1 579	333（21.09）
西北地区	436	19（4.36）	418	52（12.44）	405	83（20.49）	379	95（25.07）
西南地区	1 086	16（1.47）	1 062	63（5.93）	1 044	146（13.98）	1 015	169（16.65）
χ²值		37.07		88.02		84.26		72.86
P值		<0.001		<0.001		<0.001		<0.001

变量	2013年		2015年		2018年		2020年
变量	例数	CMM	例数	CMM	例数	CMM	例数	CMM
年龄
45~60岁	3 842	93（2.42）	3 707	261（7.04）	3 765	575（15.27）	3 760	701（18.64）
>60岁	2 211	91（4.12）	2 040	227（11.13）	1 934	469（24.25）	1 815	511（28.15）
χ²值		13.68		28.28		68.83		65.08
P值		<0.001		<0.001		<0.001		<0.001
婚姻状况
已婚/同居	5 172	145（2.80）	4 947	392（7.92）	4 957	866（16.74）	4 901	1 030（21.02）
其他	881	39（4.43）	800	96（12.00）	742	178（23.99）	674	182（27.00）
χ²值		6.73		14.72		18.33		12.48
P值		0.013		<0.001		<0.001		<0.001
受教育水平
小学以下	3 612	113（3.13）	3 465	315（9.09）	3 371	645（19.13）	3 265	738（22.60）
小学	1 011	26（2.57）	954	73（7.65）	968	175（18.08）	958	207（21.61）
中学	934	24（2.57）	901	69（7.66）	885	149（16.84）	894	182（20.36）
大学及以上	496	21（4.23）	427	31（7.26）	475	75（15.79）	458	85（18.56）
χ²值		3.95		4.11		4.86		5.17
P值		0.267		0.250		0.182		0.160
饮酒史
无	5 294	165（3.12）	5 038	447（8.87）	4 978	918（18.44）	4 872	1 069（21.94）
有	759	19（2.50）	709	41（5.78）	721	126（17.48）	703	143（20.34）
χ²值		0.85		7.64		0.39		0.92
P值		0.419		0.007		0.565		0.361
吸烟史
无	5 688	171（3.01）	5 408	465（8.60）	5 375	982（18.27）	5 270	1 146（21.75）
有	365	13（3.56）	339	23（6.78）	324	62（19.14）	305	66（21.64）
χ²值		0.36		1.35		0.15		<0.01
P值		0.659		0.288		0.751		1.000
所属地区
东北地区	419	22（5.25）	386	50（12.95）	378	92（24.34）	342	94（27.49）
华北地区	714	39（5.46）	662	102（15.41）	684	185（27.05）	651	195（29.95）
华中地区	997	31（3.11）	922	74（8.03）	938	185（19.72）	930	225（24.19）
东南沿海地区	723	14（1.94）	705	29（4.11）	667	73（10.94）	679	101（14.87）
华东地区	1 678	43（2.56）	1 592	118（7.41）	1 583	280（17.69）	1 579	333（21.09）
西北地区	436	19（4.36）	418	52（12.44）	405	83（20.49）	379	95（25.07）
西南地区	1 086	16（1.47）	1 062	63（5.93）	1 044	146（13.98）	1 015	169（16.65）
χ²值		37.07		88.02		84.26		72.86
P值		<0.001		<0.001		<0.001		<0.001

模型	准确率（%）	灵敏度（%）	特异度（%）	F1分数	AUC
LR	79.2	22.5	95.0	0.320	0.758
RF	79.1	15.1	96.9	0.240	0.728
XGB	77.3	26.4	91.5	0.336	0.721
NB	74.1	36.5	84.6	0.381	0.703
Radial Svm	78.7	8.5	98.2	0.148	0.688
KNN	76.3	24.2	90.8	0.308	0.618
DT	71.4	38.7	80.4	0.371	0.596

Interpretable Machine Learning Models for Predicting the Risk of Incident Cardiometabolic Multimorbidity in Chinese Postmenopausal Women

预测中国绝经后女性新发心脏代谢性共病风险的可解释机器学习模型研究

RichHTML

PDF

Knowledge

Abstract

Cite this article

share this article

Figures/Tables 9

References 75

Related Articles 15

Recommended Articles

Metrics

Comments

[1]	TANG Haoqing, LI Mingyue, LONG Ren, ZHU Yinbo, LIANG Weihuan, ZHU Ping, ZHAO Wei, ZHAO Yang, LIU Xiaoyun. Management Burden of Diabetic Mellitus Patients with Multimorbidity among Middle Aged and Older Adults: a Cross-sectional Survey in 6 Community Medical Institutes of China [J]. Chinese General Practice, 2026, 29(32): 4786-4797.
[2]	LI Jiao, XIANG Qingwei, LIU Jinjin, PENG Lang, ZHOU Jianjie, YANG Qiong, MEI Yingbing, LIU Hui, FANG Yingsong, TAN Zihu. Development and Validation of Risk Prediction Model for Fall-related Hospitalization in Patients with Alzheimer's Disease [J]. Chinese General Practice, 2026, 29(32): 4721-4729.
[3]	WANG Zhongkai, WU Changyong, LI Menghan, LUO Yihua, ZHANG Bingqing, PENG Yunzhu. Association between Remnant Cholesterol Inflammatory Index and Risk of Cardiometabolic Multimorbidity in Middle-aged and Older Adults [J]. Chinese General Practice, 2026, 29(31): 4573-4581.
[4]	ZHANG Shuaishuai, CHEN Yue, QIU Jiaojiao, ZHU Ping, WANG Shuxia. Association between Remnant Cholesterol Inflammatory Index and Biological Age Acceleration in Middle-aged and Older Chinese Adults: a Nationwide Cohort Study [J]. Chinese General Practice, 2026, 29(30): 4397-4405.
[5]	LI Jiao, ZHOU Jianjie, PENG Lang, MEI Yingbing, TAN Zihu. Development and Validation of a Risk Prediction Model for Poor Prognosis in Alzheimer's Disease Patients with Community-acquired Pneumonia [J]. Chinese General Practice, 2026, 29(29): 4248-4257.
[6]	SHI Li, ZHOU Zhenkun, LIU Jie, PAN Guofeng. A Randomized Controlled Trial of Sinew Meridian Theory-based Combined Acupuncture and Cupping Therapy for Postmenopausal Women with Stress Urinary Incontinence [J]. Chinese General Practice, 2026, 29(28): 4141-4146.
[7]	ZHAO Yuanyuan, LIU Pulin, DAI Juan, ZHANG Gang, ZHANG Wei, ZHANG Xiaoxia, DENG Qing, YAN Yaqiong. Association Study between Chinese Visceral Adiposity Index and Risk of All-cause and Cause-specific Mortality in the Elderly [J]. Chinese General Practice, 2026, 29(26): 3781-3787.
[8]	ZHANG Wenzheng, ZHANG Hongling. Association between Cardiometabolic Index and Hypertension in Elderly Patients with Metabolic Associated Fatty Liver Disease in the Community Population: a Prospective Cohort Study [J]. Chinese General Practice, 2026, 29(23): 3261-3267.
[9]	AN Yanhong, WANG Shidong, LI Xiaoran, GUO Jiayang, WANG Zhe, SHA Peilin, MENG Yijun, LI Xiaoxuan, SHI Xue, YU Zexing, XIAO Yonghua. Study on Risk Factors and Nomogram Prediction Model for Diabetic Kidney Disease: Based on Contrast-enhanced Ultrasound Technology [J]. Chinese General Practice, 2026, 29(21): 2995-3003.
[10]	CHAI Zhihao, ZHANG Weisen, LUO Jiaoling, ZHU Feng, ZHU Tong, JIN Yali, PAN Jing, LU Yingjun, JIANG Chaoqiang. The Mediation Effect of Socioeconomic Status in Childhood Parental Literacy and Cognitive Function in Middle-aged and Older People: Based on Guangzhou Biobank Cohort Study [J]. Chinese General Practice, 2026, 29(21): 2959-2966.
[11]	YANG Tingting, ZHENG Li, LIANG Daqiang, WU fan, YUAN Xianxian, WEI lai, HU Jin, WANG Junhua, WANG Ziyun. Association between Sleep Fragmentation and Arteriosclerosis among Population Aged 40-65 Years [J]. Chinese General Practice, 2026, 29(16): 2190-2197.
[12]	LIU Fan, CHEN Qiuyu, LI Jing. Development and Validation of a Nighttime Hypertension Screening Model for Subclinical High-risk Populations: a Single-center Cohort Study [J]. Chinese General Practice, 2026, 29(15): 1998-2005.
[13]	WANG Siman, ZHANG Mengchu, LI Wen, XU Ai, XU Jin, GUO Rui, YAN Haixia. Research on the Construction of a Risk Prediction Model for Primary Hypertension with Left Ventricular Hypertrophy Based on TCM Syndromes and Pulse Graph Parameters [J]. Chinese General Practice, 2026, 29(14): 1840-1848.
[14]	YUAN Weixia, AO Panpan, MA Yun, WU Fengfu, XU Facui, YU Juan, ZENG Qun, WEI Shaofeng, YUAN Lijia. The Predictive Value of Age-adjusted Charlson Comorbidity Index for Sarcopenia in Older Adults [J]. Chinese General Practice, 2026, 29(13): 1673-1681.
[15]	YU Xinyan, MA Zhong, CAO Fan, SU Peng, LIN Ying, ZHANG Haicheng. Predicting the Risk of Depression in Elderly Patients with Cardiovascular Metabolic Diseases Using Single-lead Wearable Electrocardiography at the Community Level [J]. Chinese General Practice, 2026, 29(10): 1300-1310.