Trends and Determinants of Healthy Aging in China 981194153X, 9789811941535

Table of contents :
Preface
Contents
About the Editors
Part I Trends of Healthy Aging in China
1 Twenty Years’ Follow-Up Surveys on Elder People’s Health and Quality of Life
1.1 Introduction
1.2 The Ageing of China’s Population and Policy Changes During the Last Two Decades
1.3 The Beginning of CLHLS and Sampling Design of the Survey
1.4 The CLHLS Questionnaire and Key Modifications Over Time
1.4.1 The Core Components of the CLHLS Questionnaire
1.4.2 Major Modifications to the CLHLS Questionnaire
1.5 Data Quality Assessment
1.6 Data Use and Publications
References
2 Trends in the Impact of Socioeconomic Status on Health with Increases in Ages: Convergence or Divergence?
2.1 Introduction
2.2 Literature Review and Research Design
2.3 Data and Research Methods
2.3.1 Data
2.3.2 Measures
2.3.3 Method
2.4 Results
2.4.1 Effect of Socioeconomic Status on Health
2.4.2 Verification of the “Convergence Effect” and the “Divergence Effect”
2.5 Conclusion
References
3 The Age, Gender, Urban-Rural and Regional Differences in Dynamic Changes of Activity of Daily Living Among the Chinese Oldest-Old
3.1 Introduction
3.2 Basic Information of Respondents
3.3 Results of the Analyses
3.3.1 Trends in the Oldest-Old by Each Basic Activity of Daily Living
3.3.2 Differences and Trends in ADL Performance by Age Groups Among Oldest-Old Respondents
3.3.3 Differences in ADL Performance by Gender Among Oldest-Old Respondents
3.3.4 Differences and Trends in ADL Performance by Urban/Rural Residence Among Oldest-Old Respondents
3.3.5 Differences and Trends in ADL Performance by Region (East/Central/West) Among Oldest-Old Respondents
3.4 Conclusion
Reference
4 Trends of Dynamic Changes in Activities of Daily Living, Physical Performance, Cognitive Function and Mortality Rates Among the Oldest-Old in China
4.1 Introduction
4.2 Methods
4.2.1 Study Design and Participants
4.2.2 Procedures and Variables
4.2.3 Statistical Analyses
4.2.4 Role of Funding Source
4.3 Results
4.4 Discussion
References
5 Trends of Dynamic Changes of Family Support for the Chinese Oldest-Old
5.1 Introduction
5.2 Literature Review of Related Research
5.3 Main Data Resource and Research Methods
5.3.1 Data Resource
5.3.2 The Category of the Family Support for the Oldest-Old and Measurement
5.3.3 Methods
5.4 Changes of Family Support to the Oldest-Old in Different Time Periods
5.4.1 Changes in Family Support of the Oldest-Old Happened in Different Periods
5.4.2 The Perceived Support of the Oldest-Old in Different Periods
5.4.3 The Family Support Resource of the Oldest-Old in Different Periods
5.5 Changes of Family Support for the Oldest-Old from Different Cohorts
5.5.1 Change of Family Care Support for the Oldest-Old of Different Cohorts
5.5.2 Change of Family Support Resource for Oldest-Old of Different Cohorts
5.6 Conclusion
References
6 Analysis of Trends of Future Home-Based Care Needs and Costs for the Elderly in China
6.1 Introduction
6.2 Methods, Data and Projection Scenarios
6.2.1 Methods
6.2.2 Data Sources and Estimates
6.2.3 Projection Scenarios and Parameters
6.3 Results
6.3.1 Age-Specific Trajectories of ADL Statuses Transitions and Home-Based Care Costs
6.3.2 Trends Under the Medium Scenario
6.3.3 Possible Ranges of the Trends
6.3.4 Resources of Care Providers Under Different Fertility Policy Scenarios
6.3.5 Sensitivity Analysis
6.4 Discussions and Policy Considerations
6.5 Concluding Remarks
References
Part II Determinants of Healthy Aging
7 The Impact of Empty-Nested Living on Physical and Psychological Health Among Elderly
7.1 Introduction
7.2 Method, Data and Variables
7.2.1 Method
7.2.2 Data
7.2.3 Measurement
7.3 Empirical Results
7.3.1 The Impact of Empty-Nested Living on ADL Among Older Adults
7.3.2 The Impacts of Empty-Nested Living on Cognitive Function
7.3.3 The Heterogenous Impacts of Empty-Nested Living on Cognition
7.3.4 The Mechanisms Through Which Empty-Nested Living Impacts Cognition
7.4 Conclusion and Policy Implication
References
8 Living Closer to Major Roads May Increase the Risk of Cognitive Decline
8.1 Introduction
8.2 Methods
8.2.1 Study Population
8.2.2 Residential Proximity to Major Roads
8.2.3 Cognitive Function Assessment
8.2.4 Confounders
8.2.5 Statistical Analysis
8.3 Results
8.3.1 Participants’ Characteristics
8.3.2 The Relationship Between Residential Proximity to Major Roadways and Cognitive Function
8.3.3 Analysis of Effect Modification and Interaction
8.4 Discussion
8.5 Conclusion
References
9 Self-assessment of Health and Life Satisfaction Among Older Adults
9.1 Background
9.2 Data and Methods
9.3 Major Findings
9.4 Conclusion and Discussion
References
10 The Analysis on Gender Difference in Self-rated Health Among Elderly in China
10.1 Introduction
10.2 Descriptive Analysis on Gender Difference in Self-rated Health Among Elderly in China
10.3 Multivariate Statistical Analysis on Gender Difference in Self-rated Health Among Elderly in China
10.4 Decomposition Analysis on Gender Difference in Self-rated Health Among Elderly in China
10.5 Conclusion
References
11 A Study on the Intensity of Care Needs Among the Chinese Elderly in Later Life
11.1 Introduction
11.2 Data and Method
11.2.1 Measures
11.2.2 Data Analysis
11.3 Major Findings
11.3.1 Limitation in Daily Life
11.3.2 Intensity of Care by Limitation in Each Item of ADL
11.3.3 Intensity of Care by Total ADL Scores
11.3.4 Intensity of Care by Type of Diseases
11.3.5 Factors Affecting the Care Needs and Care Intensity by Multivariate Analyses
11.4 Discussion
References
12 The Impacts of Improved Urban and Rural Medical Security Level on Health Care Utilization, Financial Burden and Health Status Among the Elderly
12.1 Introduction
12.2 Data, Variables and Descriptive Statistics
12.2.1 Data and Variables
12.2.2 Description on the Outcomes of the Elderly Covered by Medical Insurance
12.3 Trends in the Medical Financial Burden of the Elderly
12.3.1 Both Total and OOP Medical Expenditures Increased Significantly
12.3.2 The Share of OOP Expenses Decreased Significantly, Whereas Catastrophic Medical Expenditures Remained Heavy
12.3.3 Share of OOP Medical Expenditures in Household Income
12.4 Trends in Medical Expenditures, Utilization of Health Services, and Health Outcomes of the Elderly, by Medical Insurance Scheme
12.4.1 Trends in Medical Expenditures of the Elderly, by Medical Insurance Scheme
12.4.2 Trends in Utilization of Health Services Among the Elderly, by Insurance Type
12.4.3 Trends in Health Outcomes Among the Elderly, by Insurance Type
12.5 Empirical Models
12.6 Empirical Results
12.7 Conclusions
Appendix
References
13 The Impacts of the New Rural Pension Scheme on the Elder-Care Pattern in Rural China
13.1 Introduction
13.2 Literature Review
13.3 Theoretical Framework and Empirical Strategies
13.3.1 Theoretical Framework
13.3.2 Empirical Strategy
13.4 Data and Descriptive Statistics
13.4.1 Data Sources and Measures
13.4.2 Descriptive Statistics
13.5 Empirical Results and Analyses
13.5.1 Impact of NRPS on the Elderly Support Patterns
13.5.2 Heterogeneity Analyses
13.5.3 Robustness Test
13.5.4 Analyses and Discussions
13.6 Conclusion
References
14 Effects of the New Rural Society Endowment Insurance Program on Intergenerational Transfer
14.1 Introduction
14.2 Literature Review
14.3 Data and Variables
14.3.1 Data Source
14.3.2 Variable Definition
14.3.3 Descriptive Statistics
14.4 Models
14.4.1 Panel Model (Including Fixed Effect Panel and Tobit Panel)
14.4.2 Instrumental Variable Estimation (IV)
14.4.3 Propensity Score Matching with Difference-In-Difference Estimation (PSMDD)
14.5 Results
14.6 Discussions
Appendix 1: Samples Distribution of NRSEI Pilot in the 2011 CLHLS
Appendix 2: Balance Test of Covariates in PSMDD Analysis
References
15 Impacts of Changes in Living Arrangements on Elderly Mortality Risk
15.1 Introduction
15.2 Study on the Factors Affecting the Death Risk of the Elderly
15.3 Data and Research Methods
15.3.1 Source of Data
15.3.2 Variable Measurement
15.3.3 Research Methodology
15.4 Results of the Analysis
15.5 Conclusion
References
16 Analysis on Influence Factors on Mortality Risk of the Elderly with Mild Cognitive Impairment
16.1 Background
16.2 Literature Review
16.3 Data, Methodology and Variable Measurement
16.3.1 Data
16.3.2 Sample Selection
16.3.3 Variable Measurement
16.3.4 Methodology
16.4 Analysis of Research Results
16.4.1 Overview of the Elderly People with MCI in 2008
16.4.2 Preliminary Analysis of Factors Associated with the Mortality Risk in the Elderly with MCI
16.4.3 Analysis of the Results of COX Regression
16.5 Conclusion and Discussion
References
17 Associations of Environmental Factors with Older Adults’ Health and Mortality in China
17.1 Introduction
17.2 Data Sources, Variable Definitions and Research Methods
17.2.1 Health Outcomes
17.2.2 Community Environmental Variables
17.2.3 Individual-Level Covariates
17.2.4 Statistical Analyses
17.3 Results of the Analysis
17.4 Conclusion
17.4.1 Limitations
17.4.2 Summary
References
18 Effects of Interactions Between Environmental and Genetic Factors on Healthy Aging: A Review on the Relevant Prior Research
18.1 Introduction
18.2 Review on Interdisciplinary Research of Effects of Interactions Between Environmental and Genetic Factors on Healthy Aging
18.2.1 Effects of Interactions Between Environmental (External) and Genetic (Internal) Factors on Healthy Aging
18.2.2 Studies on the Interactions Between Environment and Genetic Factors Can Effectively Improve the Effectiveness of Interventions for Healthy Aging
18.2.3 Latest Progressions of Related Studies in International Scientific Community
18.3 Review on the Relevant Prior Research
18.3.1 Interactions Between Carrying the FOXO Genotypes and Tea Drinking Were Significantly Associated with Cognitive Function and Mortality at Advanced Ages
18.3.2 Interactions Between Carrying the ADRB2 Genotypes and Regular Exercise, Social-Leisure Activities and Negative Emotion Were Significantly Associated with Health at Advanced Ages
18.3.3 Interactions Between Carrying the APOE4 Allele and Life Stress Factors Were Significantly Associated with Self-reported Health of Older Adults
18.4 Conclusion
References
Part III Analyses of Biomarkers of the Older Adults
19 Chronic Diseases, Biomarkers and Their Influencing Factors in the Elderly
19.1 Introduction
19.2 Analysis on Chronic Diseases and Related Biomarkers of the Elderly in Longevity Regions
19.2.1 The Levels of Height, Weight, Waist Circumference and Body Mass Index
19.2.2 Prevalence of Hypertension, Diabetes, and Dyslipidemia
19.2.3 Plasma Levels of Macro and Trace Elements, Hs-CRP, MDA and SOD
19.3 Influencing Factors of the Chronic Diseases
19.4 Relationship Between the Dietary Nutrition and Health in the Old People
19.5 Influencing Factors of ADL in the Elderly
19.6 Conclusion
References
20 Cognitive Impairment and Biomarkers in the Older Adults
20.1 Introduction
20.2 Prevalence of Cognitive Impairment in the Older Adults
20.3 Biomarkers of Cognitive Impairment
20.3.1 Diabetes, Blood Glucose and Cognitive Impairment
20.3.2 Chronic Kidney Disease and Cognitive Impairment in the Elderly
20.3.3 Interaction Between Diabetes Mellitus and Chronic Kidney Disease and Cognitive Impairment in the Elderly
20.3.4 Blood Lipids and Cognitive Impairment in the Elderly
20.3.5 Albumin and Cognitive Impairment in the Elderly
20.3.6 Oxidative Stress Index and Cognitive Impairment in the Elderly
20.3.7 Blood Pressure and Cognitive Impairment in the Elderly
20.3.8 SpO2 and Cognitive Impairment in the Elderly
20.3.9 Vitamin D Levels, Anemia and Cognitive Impairment
20.4 Conclusion
References
21 Mortality Risk and Biomarkers Among the Oldest-Old
21.1 Introduction
21.2 Mortality Risk of the Oldest-Old
21.3 Analysis on Biomarkers of Mortality Risk of the Oldest-Old
21.3.1 Anemia or Low Hemoglobin Levels Increase the Mortality Risk of the Oldest-Old
21.3.2 High LDL-C Reduces the Risk of Mortality in the Oldest Old
21.3.3 The Higher Ratio of LDL-C/HDL-C, TG/HDL-C and AI Was Associated with a Lower Mortality Risk in the Oldest-Old
21.4 Conclusion
References
22 Selenium and Elderly Health
22.1 Introduction
22.2 Selenium Levels in the Environment and in the Elderly Population
22.3 Selenium and Cognitive Function of the Elderly
22.4 Selenium and Hypertension
22.5 Selenium and Dyslipidemia
22.6 Selenium and Diabetes
22.7 Summary
References
Part IV Healthy Aging Related Policy Analyses
23 Long-Term Care Demand and Supply Among the Rural Elderly in China: Evidence from CLHLS Data
23.1 Introduction
23.2 Demand for Long-Term Care Among the Rural Elderly
23.3 Supply of Long-Term Care for the Rural Elderly in China
23.3.1 Long-Term Care for the Elderly Who Live at Home
23.3.2 Institutional Elder Care
23.3.3 Communal Elder Care
23.4 Impact of the New Rural Pension Scheme on Elder Care: Does It Crowd Out or Displace the Care Provided by Children?
23.5 Conclusions and Policy Implications
References
24 Impacts of Childhood Socioeconomic Status on Health of Chinese Elderly: Policy Implications
24.1 Introduction
24.2 Literature Review
24.3 Variables and Research Method
24.4 Results
24.4.1 Effect of Childhood Socioeconomic Status on Middle-Aged and Elderly Health
24.4.2 Pathways of the Impact of Childhood Socioeconomic Status on Health in Middle and Old Age
24.5 Conclusion: Policy Recommendations
References
25 Pattern of Old-Age Care in Rural China—Lessons Learned from Fieldworks in the Villages
25.1 Introduction
25.2 Research Site
25.3 Data and Methods
25.4 Research Results
25.4.1 Male and Female Villagers Have Their Own Definition of “Old”
25.4.2 Elder Houses Secure Living Space for the Elderly
25.4.3 Self-support Becomes the Mainstream of Old-Age Care
25.5 Conclusion
References
26 Welfare Implications of Intergenerational Co-residence: Mutual Benefits for the Elderly and Their Children
26.1 Introduction
26.2 Data and Summary Statistics
26.2.1 Data
26.2.2 Female Labor Supply and Housework Burden
26.3 Model Specification
26.4 Results
26.4.1 Impact of Co-residence with Parents on Female Labor Supply
26.4.2 Why Co-residence with Parents Promotes Female Labor Supply
26.5 Conclusion
References
27 Integrated Strategies to Face the Serious Challenges of Population and Households Aging
27.1 Introduction
27.2 Two Closely Related Major Issues of Population Security in China
27.2.1 Rapidly Aging, Especially in the Rural Areas, and the Imperfect Basic Endowment Insurance System
27.2.2 Imbalance of Sex Ratio at Birth (SRB)
27.3 The Trend of Households Miniaturization and Its Effects on Population Aging and Residential Energy Demands
27.3.1 The Trend of Households Miniaturization
27.3.2 The Effects of Miniaturization of Households on Population Aging and Residential Energy Demands
27.4 Challenges and Opportunities of Retirement Ages
27.5 Challenges and Opportunities of Rural Old Age Insurance Program
27.6 Policy Recommendations
27.6.1 Integrate Administrations of Population Aging and Family Planning
27.6.2 Gradually Increase Age at Retirement
27.6.3 Further Develop the Rural Old Age Insurance Program
References
28 Closing Remarks: Prospects for Future Research
References
Appendix Data Quality Assessment of the 7th and 8th Waves of CLHLS Surveys
A.1 Introduction
A.2 Age Declaration
A.3 Internal Consistency of Major Health Measurements
A.3.1 Reliability
A.3.2 Validity
A.3.3 Logical Error (Inconsistent Response)
A.4 Proxy Use, Nonresponse Rate and Incomplete Data
A.4.1 Proxy Use
A.4.2 Nonresponse Rate and Incomplete Data
A.5 Sample Attrition
A.6 Concluding Remarks
References

Citation preview

Yi Zeng Jiehua Lu Xiaoyan Lei Xiaoming Shi Editors

Trends and Determinants of Healthy Aging in China

Trends and Determinants of Healthy Aging in China

Yi Zeng · Jiehua Lu · Xiaoyan Lei · Xiaoming Shi Editors

Trends and Determinants of Healthy Aging in China

Editors Yi Zeng National School of Development at Peking University Peking University Beijing, China School of Medicine at Duke University Duke University Durham, NC, USA Center for Study of Aging and Human Development and Geriatrics Division School of Medicine Duke University Durham, USA

Jiehua Lu Department of Sociology Peking University Beijing, China Xiaoming Shi National Institute of Environmental Health Chinese Center for Disease Control and Prevention Beijing, China

Xiaoyan Lei National School of Development Peking University Beijing, China

ISBN 978-981-19-4153-5 ISBN 978-981-19-4154-2 (eBook) https://doi.org/10.1007/978-981-19-4154-2 Jointly published with Science Press The print edition is not for sale in China mainland. Customers from China mainland please order the print book from Science Press. Translation from the Chinese language edition: “中国健康老龄发展趋势和影响因素研究” by Yi Zeng et al., © Science Press 2018. Published by Science Press. All Rights Reserved. © Science Press 2022 This work is subject to copyright. All rights are solely and exclusively licensed by the Publisher, whether the whole or part of the material is concerned, specifically the rights of reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed. The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. The publishers, the authors, and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publishers nor the authors or the editors give a warranty, expressed or implied, with respect to the material contained herein or for any errors or omissions that may have been made. The publishers remain neutral with regard to jurisdictional claims in published maps and institutional affiliations. This Springer imprint is published by the registered company Springer Nature Singapore Pte Ltd. The registered company address is: 152 Beach Road, #21-01/04 Gateway East, Singapore 189721, Singapore

Preface

Since the beginning of the twenty-first century, the speed of population aging in China has been accelerating. Under the medium or low assumptions, the proportion of older adults aged 65+ among total population size in China is projected to increase dramatically from 13.2% in 2020 to 28.3~30.1% in 2050. Among which, the number of the oldest-old aged 80+ who most likely need daily life assistance was 30.3 millions in 2020, but it will climb extraordinarily quickly to about 125.5 millions in 2050, with an average annual growth rate of twice that of the young-old aged 65–79. Overall, the average annual growth rate of the Chinese older population is more than twice that of the Western countries with a large population. The sustained prolongation of average life expectancy and population aging are inevitable trends, which are the consequences of rapid social and economic development. China’s infant and middle-aged mortality rates have been very low, and there is a small potential for continued decline. However, the average annual mortality rates for senior citizens aged 65+ (especially those oldest-old aged 80+) are still relatively high, with a great potential for continued decline. The current and future increase in average life expectancy in China and the other middle- and high-income countries is mainly due to the decline in the death rate of the older adults, especially the oldest-old, which has led to the continuous and rapid increase in the numbers of surviving older adults. Based on data analyses of the Chinese Longitudinal Healthy Longevity Survey (CLHLS), the paper published in The Lancet in March 2017 (ref. Chap. 4 of this book) shows that the prolonged lifespan has also improved the survival rate of the older adults with poor health. Many older people who may have died if they live under previous medical and living conditions have been now “saved” and survived. This resulted in the reduced average physical and cognitive functions of the Chinese older adults, which has led to the significantly increased disability rate. The increased disability rate of the oldest-old will bring serious challenges to the long-term care of the society and families and may have negative impacts on the quality of life of the old, middle-aged and young, if no effective intervention measurements are taken. Meanwhile, the Chinese economic and social security system lags behind the developed countries, and we are not yet fully prepared for the rapidly aging society. v

vi

Preface

Therefore, further conducting data collections and interdisciplinary in-depth research on healthy aging are extremely important, and we do need to explore how to achieve the healthy survival of older adults and reduce the proportion of sickness and disability of the older population. We do need to analyze what factors and ways can make the older adults live healthier and provide a scientific basis for China to actively respond to the serious challenges of population aging and improve the quality of life of everyone. This is of far-reaching significance to China, which is “getting old before getting rich.” These are not only major scientific and practical issues related to people’s daily life, but also prerequisites for the stability of the country and sustainable development of the economy and society to achieve the goal of healthy China. In order to study and discover the impacts of social, behavioral, environmental, and biomedical factors and their interactions on healthy longevity and family happiness, as well as to provide datasets for academic research and healthy aging policy-making, our Chinese healthy aging research team has been conducting the CLHLS since 1998. The first eight waves of the CLHLS were conducted in randomly selected about half of the counties and city districts of 23 Chinese provinces in 1998–2018. The 9th wave is conducted in 2021 (whose data process is to be completed by the summer of 2022) and extended as “The Chinese Longitudinal Healthy Longevity and Happy Family Study” (its abbreviation remains as CLHLS), through adding several family-relevant questions in addition to CLHLS’ initial questionnaire which contained family-relevant questions that consists of about one-third of the total number of questions. The baseline survey was conducted in 1998, and the follow-up surveys (with new recruitments to replace dead participants) were conducted in 2000, 2002, 2005, 2008–2009, 2011–2012, 2014, and 2017–2018 in randomly selected about half of the counties and city districts of 23 out of 31 Chinese provinces, municipalities and autonomous regions (Liaoning, Jilin, Heilongjiang, Hebei, Beijing, Tianjin, Shanxi, Shaanxi, Shanghai, Jiangsu, Zhejiang, Anhui, Fujian, Jiangxi, Shangdong, Henan, Hubei, Hunan, Guangdong, Guangxi, Sichuan, Chongqing, Hainan). The survey areas covered 985 million persons in the baseline year 1998 and 1156 million persons in the census year 2010, about 84–85% of the total population of China. The 9th wave of CLHLS in 2021 is substantially extended to cover 27 provinces, municipalities, and autonomous regions in China, including four additional provinces of Guizhou, Yunnan, Gansu, and Ningxia. In the 9 waves of the CLHLS conducted in 1998–2021, we have conducted face-to-face home-based 136 thousands participants, including 22.7 thousand centenarians, 30.9 thousands nonagenarians, 34.4 thousands octogenarians, 30.9 thousands younger elders aged 65–79, and 16.7 thousands middle-age adults aged 35– 64. Among the 136 thousands participants, 64.9% were oldest-old aged 80–118. Data on death dates/age and relatively detailed information of health status and care needs/costs, etc., before dying for the 33.2 thousands elders aged 65–118 who died between waves were collected in interviews with a close family member of the deceased. All of the interviews, basic health exams, and biomarkers samples collections are voluntary with standard consensus forms reviewed/signed by the participants (or their direct family members).

Preface

vii

Among the randomly selected survey sites of counties/city districts, CLHLS selected eight longevity areas (counties/cities), where the proportion of centenarians are exceptionally high, to conduct the in-depth study with 3470, 2860, 2639, 2786, and 3129 old participants in the CLHLS 2008–2009, 2011–2012, 2014, 2017– 2018 and 2021 waves. The CLHLS in-depth study in the eight longevity areas conducted face-to-face interviews using the same questionnaire as that used in all other study sites, plus home-based basic health exams by medical doctors/nurses and blood/urine sample collections from the participants, and biomarkers laboratory tests and analyses. Data collections of CLHLS have been coordinated by the Center for Healthy Aging and Development Studies of National School of Development at Peking University, and the survey field works were carried out through the nationwide survey/health service networks by Chinese Center for Scientific Research on Aging for the 1st, 2nd, 3rd, and 4th waves in 1998–2005, and by China Center for Diseases Control and Prevention (China CDC) for the 5th, 6th, 7th, and 8th wave in 2008–2018. The survey field works of the 9th wave in 2021 were carried out by the nationwide survey/health service networks by China Population and Development Research Center for the nationwide surveys and by National Institute of Environmental Health of China CDC for the in-depth study sites of the 8 longevity areas. The CLHLS questionnaire data collected provides information on family structure, living arrangements and proximity to children, activities of daily living (ADL), the capacity of physical performance (picking up a book from the floor; standing up from a chair; turning 360° without help), self-rated health, self-evaluation of life satisfaction, cognitive functioning, chronic disease prevalence, care needs and costs, participation in social activities, diet, smoking and drinking behaviors, psychological characteristics, economic resources, and care giving and family support among older adults respondents and their relatives. Information about the health status of the CLHLS participants who were interviewed in the previous wave(s) but died before the current survey was collected by interviewing a close family member. Information provided consists of cause of death, chronic diseases, ADL before dying, frequency of hospitalization or instances of being bedridden from the last interview until death, whether bedridden before death, length of disability and suffering before death, etc. For the participants from the in-depth study sites of the 8 longevity areas, in addition to the questionnaire interview data, the CLHLS conducted laboratory tests of the following biomarkers using blood/urine samples collected from the voluntary participants. Routine blood tests: white blood cell count, lymphocyte count, percentage of lymphocytes, red blood cell count, hemoglobin, erythrocyte hematocrit, erythrocyte mean corpuscular volume, erythrocyte mean corpuscular hemoglobin, erythrocyte mean corpuscular hemoglobin concentration, platelet count, mean platelet volume, platelet distribution width, plateletocrit; plasma tests: plasma albumin, blood urea nitrogen, total cholesterol, plasma creatine, high-sensitive C-reactive protein, plasma glucose, glycolated plasma protein, high-density lipoprotein cholesterol, low-density lipoprotein cholesterol, triglyceride, urea acid, superoxide dismutase, 25-oh-vitamin D3, malondialdehyde, vitamin B12; urine tests: urine microalbumin, urine creatinine, Ualb/Ucr ratio.

viii

Preface

The baseline, the cross-sectional, and follow-up survey datasets and documents (questionnaires, users’ guides, etc.) of the eight waves of CLHLS conducted in 2000, 2002, 2005, 2008–2009, 2011–2012, 2014, and 2017–2018 are being distributed to the research community. Researchers who are interested in using the CLHLS data need to first sign a Data Use Agreement and then download the datasets free of charge. As of October 15, 2021, according to incomplete statistics, there are 10,327 registered CLHLS data users (excluding their students and other group members), and they produced the following publications using the CLHLS datasets: 17 books (five in English and twelve in Chinese); 431 papers written in English published in peer-reviewed journals outside of China; 731 papers written and published in peerreviewed Chinese journals; 99 Ph.D. dissertations and 678 M.A. theses successfully defended at Universities inside or outside of China. The CLHLS has the world’s largest sample size of the oldest-old aged 80+ with comparable samples of young-old aged 65–79 with rich data information, which has provided great opportunities and potentials for scientific research and policy analyses. Using the CLHLS data, researchers have produced internationally and nationally influential research outcomes. For example, the paper entitled “Associations of Community Environmental Factors with Older adults Health and Mortality” won the Best Paper Award of the American Journal of Public Health (ref. Chap. 17 of this book); papers entitled “Chinese Longitudinal Healthy Longevity Survey and Related Policy Research” and “Research Progress Report on Typical Regions of Healthy and Longevity” won the ‘Excellent Paper Award’ in the 20th World Congress of Gerontology and Geriatrics. Our policy advisory report on the healthy aging of the population and the universal two-child policy and report entitled “Integrate Administrations of Health, Family Planning, and Population Aging, to Promote Well-being of Billions of Families” had been noted in written by President Jinping Xi and Premier Keqiang Li and had been selected twice as the main report of the policy bulletin of Internal Reference for Reforms: High-level Reports sponsored/managed by China National Development and Reforms Commission. Based on in-depth joint research efforts by our team members, we integrated our research results in various aspects and formulated this book entitled Trends and Determinants of Healthy Aging in China. The book consists of four parts (a total of 28 chapters), including: Part I. Trends of Healthy Aging in China; Part II. Determinants of Healthy Aging; Part III. Analyses of Biomarkers of the Older Adults; and Part IV. Healthy Aging related Policy Analyses. We particularly emphasize the interdisciplinary research of social science and biomedical science. The concept of “biomarkers” (analyzed in Part III of this book) has been gradually incorporated into social science from the field of biomedicine since 1990s. It has been proven in various empirical studies that many social and environmental factors (such as socioeconomic status and social relations) have important impacts on health. It is necessary to analyze further how these effects are transmitted to the human body and significantly affect health. However, commonly used health indicators (such as activities of daily living, cognitive function, physical activity, and

Preface

ix

chronic conditions) cannot provide information of mechanism and process. Therefore, researchers tried to find some intermediary variables to help reveal the intermediate pathways and adjustment mechanisms of external influences such as the social environment on health. Biomarkers (biomedical indicators) are ideal intermediary variables in social medicine research. There are two advantages of biomarkers. First, the medical community has a relatively thorough understanding of the functions of various health-related biomarkers, and their measurement and examination technologies are relatively mature. Second, the cost of biomarkers tests is relatively low and large samples can be obtained, which can be well combined with large-scale sampling surveys in the field of social science research. However, compared with the international frontier research level, China’s current interdisciplinary research on biomarkers in social science and biomedical science is relatively weak. The analyses in Part III of this book play an important role in this field. It is important to note that the concept of “older adults health” (or “health at old ages”) used in some chapters of this book refers to the health status of the older adults aged 65+. The concept of “healthy aging” used in the title of this book and the titles of Parts I, II, and IV refer to population, economics, and society aging healthily. It also means the interdisciplinary research on the social, economic, and biomedical factors at macro-level (national, provincial), middle-level (county, city, community), and micro-level (individual family and molecular genetics) that affect the health of the older adults. The concept of healthy aging includes not only the physical/mental health and social participation statuses of the older adults and the influencing factors, but also the fertility, death, quality of life, and migration factors related to healthy aging. Healthy aging implies the balance of supply and demand of labor resources, the coordination of population and resources, and the environment. Furthermore, it also means building a friendly aging society in which the older adults can obtain enough supports, realize self-worth, receive appropriate medical treatment, learn knowledge or skill they want and enjoy their life. In short, the concept of “healthy aging” is a high-level summary of the statuses, influencing factors, processes, policy research and practices of population, economy, and society aging healthily. There are still limitations in our studies, and we need to further deepen and expand the research to fully explore its huge potential. We outline the limitations of our studies in the last chapter of “Closing Remarks” of this book and discuss how to continue to deepen and expand the CLHLS, a project that benefits the country and the people, to make more concrete contributions to healthy aging in China. Finally, we sincerely thank the Management Science Division of the National Natural Science Foundation of China (NSFC) for the financial support. We sincerely thank China CDC, local CDCs at provincial, prefecture, and county levels, China Population and Development Research Center, aging committees at all levels, and colleagues from various universities for their collaboration and hard work. We sincerely thank all of the interviewed older adults and their family members who participated in the CLHLS surveys. We sincerely thank all members of our team and the students who participated in the project. We sincerely thank the authors of each of the chapters of this book for their contributions and thank the assistances of Xuxi

x

Preface

Zhang, Man Li, Chen Bai, Muqi Guo, Hanmo Yang, Yao Yao, Minhui Liu, Wende Zhang and Xiumei Yu. We will continue to carry forward the scientific spirit of hard working, diligence, rigor, truth-seeking, and innovation and continue to produce high-quality interdisciplinary academic research results and policy advisory reports for healthy aging of population, economy, and society in China and elsewhere. Beijing, China and Durham, USA

Prof. Yi Zeng

Contents

Part I 1

2

3

4

5

6

Twenty Years’ Follow-Up Surveys on Elder People’s Health and Quality of Life . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Zhenzhen Zheng

3

Trends in the Impact of Socioeconomic Status on Health with Increases in Ages: Convergence or Divergence? . . . . . . . . . . . . . Jianxin Li and Cuicui Xia

21

The Age, Gender, Urban-Rural and Regional Differences in Dynamic Changes of Activity of Daily Living Among the Chinese Oldest-Old . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Wenjuan Zhang and Meng Wei Trends of Dynamic Changes in Activities of Daily Living, Physical Performance, Cognitive Function and Mortality Rates Among the Oldest-Old in China . . . . . . . . . . . . . . . . . . . . . . . . . . . Yi Zeng and Qiushi Feng

37

51

Trends of Dynamic Changes of Family Support for the Chinese Oldest-Old . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Xianxin Zeng

77

Analysis of Trends of Future Home-Based Care Needs and Costs for the Elderly in China . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Yi Zeng and Huashuai Chen

95

Part II 7

Trends of Healthy Aging in China

Determinants of Healthy Aging

The Impact of Empty-Nested Living on Physical and Psychological Health Among Elderly . . . . . . . . . . . . . . . . . . . . . . . . 123 Ke Shen

xi

xii

Contents

8

Living Closer to Major Roads May Increase the Risk of Cognitive Decline . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135 Yao Yao, Jin Xurui, Junfeng Zhang, and Yi Zeng

9

Self-assessment of Health and Life Satisfaction Among Older Adults . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151 Zhenzhen Zheng and Ting Feng

10 The Analysis on Gender Difference in Self-rated Health Among Elderly in China . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163 Huachu Zhang, Yanqing Qian, and Jin Ke 11 A Study on the Intensity of Care Needs Among the Chinese Elderly in Later Life . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175 Yun Zhou and Ting Feng 12 The Impacts of Improved Urban and Rural Medical Security Level on Health Care Utilization, Financial Burden and Health Status Among the Elderly . . . . . . . . . . . . . . . . . . . . . . . . . . . 189 Huamin Chai and Xiaoyan Lei 13 The Impacts of the New Rural Pension Scheme on the Elder-Care Pattern in Rural China . . . . . . . . . . . . . . . . . . . . . . . 217 Lingguo Cheng, Ye Zhang, and Zhibiao Liu 14 Effects of the New Rural Society Endowment Insurance Program on Intergenerational Transfer . . . . . . . . . . . . . . . . . . . . . . . . . 241 Huashuai Chen and Yi Zeng 15 Impacts of Changes in Living Arrangements on Elderly Mortality Risk . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 267 Chunhua Li, Jianxin Li, and Wangchun Wu 16 Analysis on Influence Factors on Mortality Risk of the Elderly with Mild Cognitive Impairment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 283 Jiehua Lu and Yue Li 17 Associations of Environmental Factors with Older Adults’ Health and Mortality in China . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 299 Yi Zeng, Danan Gu, Jama Purser, Helen Hoenig, and Nicholas Christakis 18 Effects of Interactions Between Environmental and Genetic Factors on Healthy Aging: A Review on the Relevant Prior Research . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 315 Yi Zeng, Lingguo Cheng, Rongping Ruan, and Huashuai Chen

Contents

xiii

Part III Analyses of Biomarkers of the Older Adults 19 Chronic Diseases, Biomarkers and Their Influencing Factors in the Elderly . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 331 Jiaonan Wang and Xiaoming Shi 20 Cognitive Impairment and Biomarkers in the Older Adults . . . . . . . 345 Zhaoxue Yin, Yuebin Lv, and Xiaoming Shi 21 Mortality Risk and Biomarkers Among the Oldest-Old . . . . . . . . . . . 359 Yuebin Lv, Zhaoxue Yin, and Xiaoming Shi 22 Selenium and Elderly Health . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 371 Liqin Su, Yibin Cheng, Yinlong Jin, and Sujuan Gao Part IV Healthy Aging Related Policy Analyses 23 Long-Term Care Demand and Supply Among the Rural Elderly in China: Evidence from CLHLS Data . . . . . . . . . . . . . . . . . . . 389 Lingguo Cheng 24 Impacts of Childhood Socioeconomic Status on Health of Chinese Elderly: Policy Implications . . . . . . . . . . . . . . . . . . . . . . . . . . 399 Cuicui Xia, Jianxin Li, and Jiehua Lu 25 Pattern of Old-Age Care in Rural China—Lessons Learned from Fieldworks in the Villages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 413 Qing Li, Chuanqi Xu, and Zhou Yun 26 Welfare Implications of Intergenerational Co-residence: Mutual Benefits for the Elderly and Their Children . . . . . . . . . . . . . . 427 Ke Shen and Ping Yan 27 Integrated Strategies to Face the Serious Challenges of Population and Households Aging . . . . . . . . . . . . . . . . . . . . . . . . . . . . 441 Yi Zeng, Angang Hu, and Xuxi Zhang 28 Closing Remarks: Prospects for Future Research . . . . . . . . . . . . . . . . 457 Yi Zeng Appendix: Data Quality Assessment of the 7th and 8th Waves of CLHLS Surveys . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 461 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 485

About the Editors

Yi Zeng is a professor of National School of Development and honorary director of Center for Healthy Aging and Development, Peking University. He is also a professor at the Center for Study of Aging and Human Development and Geriatric Division at School of Medicine, Duke University. He is a member of The World Academy of Sciences for the advancement of science in developing countries and a foreign member of the Royal Netherlands Academy of Arts and Sciences. He received his Ph.D. with Summa Cum Laude from Brussels Free University (1986) and pursued research as Post-doctoral Fellow at Princeton University (1986–87). Up to November, 2021, he has published 388 professional articles in domestic and international journals or as book chapters. He has published thirty-one academic books, including eleven books in English and 20 books in Chinese. He has been awarded four international academic prizes and eleven national academic prizes of China, such as The International Union for Scientific Studies of Populations (IUSSP) 2021 Laureate which is awarded to one scholar annually worldwide, the Dorothy Thomas Prize of the Population Association of America, the Harold D. Lasswell Prize in Policy Science awarded by the journal Policy Sciences and Kluwer Academic Publishers, the Best Paper Award of American Journal of Public Health, the national prizes for advancement of science and technology, the highest academic honor of Peking University: “Prize for Outstanding Contributions in Sciences,” and the national “Chinese Population Prize (Science and Technology).” In 2019, he received “National Medal of Outstanding Contributions” awarded by the Central Government of China. Jiehua Lu is a professor of Department of Sociology and deputy director of Center for Healthy Aging and Development Studies at Peking University. He also serves as the vice president of Chinese Population Association and vice president of Chinese Association of Gerontology and Geriatrics. He received his Ph.D. in Demography from Peking University (1997). His research areas include demography, gerontology, economics of population, and interaction between population and environment. He has been the principal investigator for over ten key projects and published over 70 peer-reviewed journal articles, such as “Chinese Women’s Family Status: Analysis of Chinese Decennial Survey, 1990–2010,” “Determinants Affecting Longevity xv

xvi

About the Editors

at County Level in China,” and “Patterns of living arrangements of the elderly in Mainland China: changes, consequences and policy implications.” He authored/coauthored in over ten book chapters and books, such as “Chap. 3: Changing Patterns of Marriage and Divorce in Today’s China,” Isabelle and Gu, Baochang (eds.), Analyzing China’s Population: Social Change in a New Demographic Era. Berlin: Springer. Xiaoyan Lei is a professor at National School of Development and the director of Center for Healthy Aging and Development at Peking University (PKU) in China. She also serves as the deputy director of PKU Center for Human Capital and National Policy, China. She is the editor of renowned publications such as China Economic Quarterly and Journal of Economics of Aging. She received her Ph.D. in Economics from University of California, Los Angeles, USA (2007). Her areas of research interests include labor economics, health economics, the economics of aging and applied econometrics. She has been awarded many prizes, such as “Chinese Ministry of Education Yangtze River Distinguished Professor, PKU Boya Distinguished Professor” (2019), “China Medical Science and Technology Second Award” (2018), and “Excellent Research in Art and Social Science, Peking University” (2013–2014). She has published over 50 peer-reviewed journal articles and authored/co-authored in many book chapters and books. Xiaoming Shi is a professor and the director of the National Institute of Environmental Health (NIEH) at Chinese Center for Disease Control and Prevention (China CDC). He received his M.D. from Anhui Medical University of China (1999) and Ph.D. in Epidemiology and Health Statistics from China CDC (2005). He joined the branch of Infectious Disease Surveillance at China CDC in 2005 and then worked as the head of the Branch of Monitoring and Evaluation, Division of Chronic Diseases Control and Community Health at China CDC. Since 2015, he was appointed as the director of NIEH at China CDC. His research has focused on the determinants of health and longevity, risk assessments of air pollution, heavy metals and other environmental exposures, etc. He has extensive experiences working with numerous chronic diseases and aging studies in Chinese populations. He has published over 190 peer-reviewed journal articles and authored/co-authored in over ten book chapters and books. He is also the editor of many renowned publications such as Global Health Journal and Journal of Environmental and Occupational Medicine.

Part I

Trends of Healthy Aging in China

Chapter 1

Twenty Years’ Follow-Up Surveys on Elder People’s Health and Quality of Life Zhenzhen Zheng

1.1 Introduction Population of China, which is the largest developing country with 1.41 billion people in total in 2020, is aging rapidly due to rapid declines in both fertility and mortality. Under the medium or low mortality assumptions, the total number of elderly aged 65+ in China is estimated to increase dramatically from 176 million in 2019 (12.6% of the total population) to 337–400 million in 2050 (23.9–26.9% of the total population); the number of oldest-old aged 80+ who most likely need daily life assistance will climb extraordinarily to about 107–150 million in 2050. The average annual rate of increase of the oldest-old from 2000 to 2050 is about 4.4–5.1% in China, more than twice that of the industrialized countries. The unavoidable worldwide rapid aging trends raised fundamental research questions: whether the rapid population aging will be accompanied by disability expansion, or morbidity compression, or dynamic equilibrium? Why do some people survive to advanced ages with good health while others suffer severe disability and diseases? What and how to keep the aging societies healthy? More specifically, what are the factors which may affect old adults’ health status and the process of population healthy aging? What kind of policy reactions and program interventions need to be taken to respond to the serious challenges of population aging? Adequately addressing these questions will improve quality of life not only for the elderly but for all members of the society, but so far, there are limited answers to these important questions.

Z. Zheng (B) Institute of Population and Labor Economics, Chinese Academy of Social Sciences, Beijing, China e-mail: [email protected] Center for Healthy Aging and Development Studies, National School of Development, Peking University, Beijing, China © Science Press 2022 Y. Zeng et al. (eds.), Trends and Determinants of Healthy Aging in China, https://doi.org/10.1007/978-981-19-4154-2_1

3

4

Z. Zheng

To address the important questions outlined above through discovering the social, behavioral, environmental, and genetic factors and their interactions that may influence healthy longevity, as well as providing data for academic research, and information for healthy aging policy analyses, the baseline survey of CLHLS was conducted in 1998 and the follow-up surveys were conducted in 2000, 2002, 2005, 2008– 2009, 2011–2012, 2014 and 2017–2018 in randomly selected about half of the counties and city districts of 23 out of 31 Chinese provinces (Liaoning, Jilin, Heilongjiang, Hebei, Beijing, Tianjin, Shanxi, Shaanxi, Shanghai, Jiangsu, Zhejiang, Anhui, Fujian, Jiangxi, Shandong, Henan, Hubei, Hunan, Guangdong, Guangxi, Sichuan, Chongqing, Hainan). The survey areas covered about 85% of the total population of China. Data collections have been coordinated by the Center for Healthy Aging and Development Studies of National School of Development at Peking University, and the survey field works were carried out by the nationwide survey/health service networks by Chinese Center for Scientific Research on Aging for the 1st, 2nd, 3rd and 4th waves in 1998–2005, and by China Center for Diseases Control and Prevention for the 5th, 6th, 7th and 8th waves in 2008–2018. The survey field works of the 9th wave in 2021 has implemented by China Population and Development Research Center for the nationwide surveys and by National Institute of Environmental Health of China CDC for the in-depth study sites of the 8 longevity areas. In the 8 waves of the CLHLS conducted in 1998–2018, we have conducted face to-face home-based 113 thousands interviews, including 19.5 thousands centenarians, 26.8 thousands nonagenarians, 29.7 thousands octogenarians, 25.5 thousands younger elders aged 65–79, and 11.3 thousands middle-age adults aged 35–64. Data on mortality and relatively detailed information of health status and care needs/costs before dying for the 28.9 thousands elders aged 65–110 who died between waves were collected in interviews with a close family member of the deceased. Note that slightly more than two-thirds of CLHLS interviewees were oldest-old aged 80+, because much more dramatic increase of the oldest-old who are more likely need services and medical care than the young-old aged 65–79, as discussed earlier. In the 8 waves of the CLHLS, we have collected DNA samples (blood or saliva) from 40.8 thousands participants for interdisciplinary research between natural and social scientists. All of the interviews, basic health exams and DNA sample collections are voluntary with standard consensus forms reviewed/signed by the participants or their direct family members. The sample of CLHLS is national representative with longitudinal data of a large sample size and satisfactory data quality. The datasets of CLHLS are accessible to academia, and have been widely used in the field of ageing studies. After introducing elder population changes and relevant policy during the last two decades as the background of CLHLS, this chapter introduces sampling design used by CLHLS, lists core information collected by the survey, and identifies major items added to follow-up surveys. As the dataset from the 6th and 7th waves is major source for analysis of other chapters in this book, data quality of the two waves is introduced. Finally, the chapter summarizes how CLHLS data is used and by what types of publications.

1 Twenty Years’ Follow-Up Surveys on Elder People’s Health …

5

1.2 The Ageing of China’s Population and Policy Changes During the Last Two Decades China’s population has aged rapidly during the last two decades. The share of the population aged 65 and above increased from 7.0% in 2000 to 8.9% in 2010, and to 13.5% in 2020. The number of people aged 65 and above increased from 88.2 million in 2000 to 118.9 million in 2010, while the number of centennials doubled during this decade. The number of people aged 65 and above is expected to approach 400 million by the middle of twenty-first century. Because the Chinese “baby boomers,” who were born in the 1950s and 1960s, will fall into the category of the “oldest-old” after 2030, the number of people aged 80 and above and the proportion of this group in the total population is continuing to rise, with the total number expected to exceed 100 million by the middle of the century and reach a peak of 160 million around 2070 (Wang 2019). China has also undergone rapid socio-economic changes over the last half a century, and there have been numerous changes to public policies affecting urban and rural residents. For the elderly, in addition to increased numbers and a larger proportion of the population, there have been changes in the structure of the elderly population, such as in the educational structure and living arrangements of the elderly. The level of educational attainment plays an important role in determining the health and socio-economic status of an individual, and the elderly are no exception. The overall level of educational attainment of China’s elderly population has changed greatly in recent decades. Before the 1950s, few Chinese residents could go to school and learn to read; rural women in particular had almost no opportunities to go to school and were largely illiterate. Literacy campaigns carried out in the 1950s and 1960s allowed large numbers of urban and rural residents to learn to read, and primary schools gradually became more common in rural areas during the 1960s and 1970s. The development of education during these years has had an important impact on the educational composition of today’s elderly population. The educational composition of the elderly aged 65 and above in 2000 and 2015 shows the changes brought about by the availability of primary education in both urban and rural areas. Elderly women, especially, benefited from this change. In 2000, nearly three quarters of women aged 65 and above had never attended school, while only 21.6% had attended primary school, and even fewer had junior high school or above education. By 2015, 43.1% of elderly women had primary school education, 11.6% had attended junior high school, and 6.0% had high school or higher education. Changes in the composition of education attainments of the elderly imply changes in employment histories, income, pension and related social welfare benefits for the elderly, as well as changes in intergenerational relationship and living arrangement within a family. These changes undoubtedly affect the lives and health of the elderly people as they age. The changes of family structure and to the living arrangements of the elderly are directly related to the nature of intergenerational relationships and how elderly people are cared for. Since China’s reforms and opening up began some 40 years ago, Chinese households have changed in a number of ways. Average family size has become

6

Z. Zheng

smaller and the structure of families simpler. The average age of family members has increased, living arrangement patterns have changed, and a large number of nontraditional families has emerged. The proportion of one to three-person households rose rapidly after 1990, reaching 65% in 2010, and the average size of households shrank to 3.09,1 more than 30% smaller than in 1982. The proportion of families with at least one elderly member and the proportion of elderly family members in families is on the rise. In 2010, 88.036 million households had elderly family members aged 65 and above, accounting for 21.9% of all households. Almost 30 million households were made up entirely of people aged 65 and above. Among these were two generation elderly households with young old and the oldest old living together, and single generation elderly households in which elderly brothers and sisters lived together. In terms of living arrangements, compared with 1990, the proportion of elderly people living with their children had decreased by 10 percentage points in 2000. In 2010, 63.8% of people aged 80 and above lived with their children, a further drop of 12 percentage points from 2000 (Peng and Hu 2015). In addition to the changes in the characteristics of the elderly population and their families, the continuous improvement of social and economic conditions as a result of development and improved social welfare programs have also changed the lives and living conditions of the elderly. Throughout the 1990s and the first years of the twenty-first century, China experienced rapid economic growth, with GDP per capita maintaining double-digit growth for a number of years, rising from less than 2000 RMB in the early 1990s to nearly 50,000 RMB in 2015. China has moved from being a low-income country to a middle-income country. At the same time, China’s social security system, which is closely related to the welfare of the elderly, has also improved. In the 1990s, the State Council issued decisions on basic old age insurance for enterprise workers and basic medical insurance for urban workers. Local governments have successively established a basic old-age insurance system for employees and a basic medical insurance system for workers, and the income and health care of urban workers after retirement have been basically guaranteed. In 1999, the State Council promulgated regulations on minimum living security for urban residents, and it was implemented in all the cities, ensuring the basic livelihood of the elderly urban poor. In the twenty-first century, social welfare coverage for elderly people has been extended to include rural areas as well as urban areas, and from including salaried workers to including almost everyone. In 2012, the State Council introduced a new type of social old-age insurance in rural areas and social old-age insurance for urban residents, extending basic old-age insurance coverage to the entire country. The new rural cooperative health insurance system began trials in 2003 and scaled up rapidly after 2005. In 2007, pilot programs for the basic health insurance system for urban residents got underway. In 2016, the health insurance system in rural and urban areas was integrated into a basic health insurance system for both urban and rural residents, achieving full coverage nationally. At the same time, there has been a steady improvement to the medical assistance system in both 1

The average household size further decreased to 2.62 in 2020, according to the latest population census.

1 Twenty Years’ Follow-Up Surveys on Elder People’s Health …

7

urban and rural areas, giving full play to the supporting function of social assistance (He et al. 2018). In addition to changes in the elderly population and development of the social welfare system, China’s policies for the elderly have also undergone important changes, and the policy system for the elderly has gradually developed and been improved. In 1999, the State Council established the National Working Commission on Ageing, and in 2001, introduced the first development plan on ageing. The State Council also included sections on the Development of Undertakings on Ageing in the 11th and 12th Five-year Plans, providing top-level guidance for the formulation of systematic policies on ageing. By 2013, China’s ageing policy system had taken shape, covering many aspects of life for the elderly population, such as old-age security, health care for the elderly, services for the elderly, education for the elderly, social participation of the elderly, and efforts designed to ensure a friendly social environment for the elderly. Since 2013, many major policies and systems related to ageing have made breakthroughs. The work on elderly policy is regarded as a systematic project related to the overall social situation. Since 2018, the State Council and various government departments have issued nearly 300 policies related to elderly people and ageing issues, and there are more than 20 special plans for the elderly in the national-level 13th Five-year Plan. The areas covered by the policy system have steadily expanded (Zhu et al. 2018). The establishment and improvement of the policy system for the elderly benefit elderly people, both directly and indirectly, and we can look forward to a corresponding improvement in the quality of life for all segments of the elderly population.

1.3 The Beginning of CLHLS and Sampling Design of the Survey Initiated in 1997, the Chinese Longitudinal Healthy Longevity Survey (CLHLS) was aimed at discovering the social, behavioral, environmental, and genetic factors and their interactions that may influence healthy longevity. It was also intended to provide data for academic research and policy dialogue, and especially to provide information about the oldest old people (aged 80 and above), a sub-group of elderly which had not been adequately represented in previous surveys because of small sample size. The 1998 baseline survey provided a clear picture about China’s oldest old with respect to their living arrangements, socioeconomic status, and health status (Zeng et al. 2002). The sampling design of CLHLS adopted a multi-stage disproportionate and targeted random sampling method, taking into account the needs for a sample that was representative, methods that allowed for the collection of reliable data, and a fieldwork program that was feasible. As age is a key variable in data analysis, site selection took into consideration the need for reliable and accurate age reporting of the population at the site. Some areas were not included in the survey because of proven

8

Z. Zheng

significant misreporting of age in population censuses. The first CLHLS conducted in 1998 covered 22 provinces: Liaoning, Jilin, Heilongjiang, Hebei, Beijing, Tianjin, Shanxi, Shaanxi, Shanghai, Jiangsu, Zhejiang, Anhui, Fujian, Jiangxi, Shandong, Henan, Hubei, Hunan, Guangdong, Guangxi, Sichuan and Chongqing. The population of the 22 provinces was 985 million in 1990, about 85% of the total population of China at that time. Although other social surveys of the elderly population existed at that time, information pertaining to the oldest old group was very limited because the sample size of this group in other surveys was relatively small. This was unfortunate, given the numerous health issues and care needs of this group. However, any sample selection that was proportionate, based on the actual age structure of the population, would be highly concentrated towards relatively younger elderly people and female elderly; the sample of oldest old would be too small for in-depth analysis. To obtain a sample with enough of the oldest old, especially of those over 100 years of age, for meaningful analysis, a targeted and disproportionate sampling method had to be adopted. The sampling method for the survey took the following steps: (1) Roughly 50% of the counties (or county-level cities or districts) with a total of 631 county level administrative units in the 22 provinces were randomly selected. (2) All centenarians (people aged 100 or above) living in the selected sites who voluntarily participate in the survey were visited; and for each centenarian interviewed, one 80–89 year-old, one 90–99 year-old, one 70–79 year-old and 0.5 65–69 year-old resident nearby were interviewed. To be more specifically, interview one elderly of age 80–84, one of age 90–94, and one of age 70–74 for the centenarian interviewed with the last digit of id number 0–4; and interview one elderly of age 85–89, one of age 95–99, one of age 75–79 and one of age 65–69 for the centenarian interviewed with the last digit of id number 5–9. (3) The gender of interviewees age 65–99 is defined by birth date of the centenarian interviewed, male elderly is interviewed if birth month of the centenarian is between January and June, and female elderly is interviewed if the birth month of the centenarian is between July and December. The basic idea of selecting the sample in this way was to ensure that roughly equal numbers of centenarians and men and women in their 80s and 90s were interviewed. While the number of male and female interviewees in age 80s and 90s are also roughly the same by single age. The purpose and advantage of the targeted random sampling is to ensure adequate sample size for oldest old especially male oldest old sub-groups, to eliminate problems often encountered by proportionate sampling, such as difficult to analyze determinants of health and gender differences among oldest old because of small sample size. Meanwhile by apply a weighting by age and gender structure of population census to the sample, we are able to estimate the weighted average of the population by age and gender. The validity of the targeted random sampling is tested and accepted by scholars who use seven waves of CLHLS datasets.

1 Twenty Years’ Follow-Up Surveys on Elder People’s Health …

9

Additional 4478 adult children aged 35–65 of elder interviewees were interviewed in 2002 and 2005 wave in eight provinces (Guangxi, Guangdong, Fujian, Jiangsu, Zhejiang, Shandong, Beijing, and Shanghai). The sampling strategy for adult children is to select the child by birth month if the elder interviewee has two or more children. More specifically, if the elder interviewee has two children age 35–65, the first child is selected for interview if the birth month of the elder interviewee is between January and June, or the second child is selected if the birth month is between July and December, and so on. The method of sample selection is easy to follow and the result is equivalent to random sampling. Seven longevity areas were included in the CLHLS after 2009, they are: Sanshui of Guangdong, Yongfu of Guangxi, Chengmai of Hainan, Xiayi of Henan, Zhongxiang of Hubei, Mayang of Hunan, and Laizhou of Shandong; and another site, Rudong of Jiangsu, was included since 2012. A control group was added in 2008 survey, included a sample of aged 40–59 who have no blood ties with the elderly interviewees (could be family members of the interviewee such as daughter-in-law or son-in-law, selected by month of birth and the last digit of id number of the interviewee). Children aged 60 and above of elderly interviewees in longevity areas were interviewed in 2009. The follow-up surveys revisited elder people who participated in the previous survey or visited family members of the deceased previous respondents. In order to ensure that the total sample size remained roughly the same and that the samples were comparable between the waves, elder people who died or lost in follow-up were replaced by samples nearby from the same region, and of the same sex and same age in 2000, 2002, 2005, and 2008 waves. The 2011 and 2014 CLHLS waves only visited survived interviewees and family members of elderly who died after the last interview, except for the longevity areas. The follow-up in eight longevity areas conducted health and longevity survey in 2012 and 2014, and replaced cases of died or lost in follow-up. CLHLS continued for 20 years; the eighth follow-up survey was completed in 2018. The database now contains approximately 130,000 records of surviving and deceased elderly people aged 65 and above (see Table 1.1). In addition, 40,800 DNA samples from blood or saliva were collected during the eight waves of CLHLS. The over-proportionate sampling of the oldest old and male elderly resulting from the sample design provides a satisfactory sample size for estimating and analyzing the relevant indicators of some sub-groups. However, if users of the data want to calculate the mean or the distribution of variables to reflect the overall situation of the elder population, or when making comparisons, the use of weight is recommended (and weights were attached to the datasets provided by the research team2 ). Weighting is not necessary for sample description without comparison between different subgroups. Multivariate analysis could be un-weighted if variables such as age, gender, and place of residence were controlled. It is found that un-weighted results could be biased if weight is a function of dependent variable, so weighted analysis is required in this case. There are different opinions about whether weighting is necessary when 2

For details about weighting design, refer to Appendix A of Zeng et al. (2001), or Appendix of Zeng (2008, pp. 34–36).

10

Z. Zheng

Table 1.1 Sample size of CLHLS (person) Year of the survey (age range)

Persons interviewed Elderly interviewed

Family members of deceased elderly interviewed

1998 (80+)

9093

–

2000 (80+)

11,199

3368

2002 (65+)

16,064

3343

2005 (65+)

15,638

5874

2008 (65+)

16,954

5228

2011/12 (65+)

9765

5642

2014 (65+)

7192

2879

2018 (65+)

15,874

2226

Records in total (65+)

101,779

28,560

weight is only function of independent variables. Weighted and un-weighted results are often found similar when weight is not a function of dependent variable.

1.4 The CLHLS Questionnaire and Key Modifications Over Time 1.4.1 The Core Components of the CLHLS Questionnaire The main contents of the questionnaire used in the first survey remained unchanged in the follow-up surveys, but some items were adjusted or revised to keep up with rapidly changing socio-economic conditions or to satisfy new research demands. The core content of CLHLS constituted the main body of all of the surveys; there were no major changes and adjustments in the follow-ups. The core parts to the questionnaire for people aged 65 and above are listed below. 1. Part A—Basic Information. Interviewees were asked their age, sex, ethnic group, place of birth and current living arrangements. If the interviewees lived with family members or other people, information about the gender and age of those they lived with were collected, as well as information about their relationship with the elderly interviewee. Interviewees who lived alone or in a nursing home were asked when they began living alone or in a nursing home. Interviewees who lived in a nursing home were asked about the reason and cost. 2. Part B—Self-evaluation. Interviewees assessed their current situation and their personality and emotional characteristics. This part of the interview had to be completed by the elderly person being interviewed. For those who were unable to answer by themselves, there was an “unable to answer” option. In this part of

1 Twenty Years’ Follow-Up Surveys on Elder People’s Health …

11

the interview, elderly respondents self-evaluated their quality of life (What do you think of your life now?) and health (What do you think of your own health?), and whether their health had changed in the past year. A series of seven items were used to measure the personality characteristics of the interviewees. 3. Part C—Cognitive Ability. Questions in this part of the interview also had to be completed by the elderly interviewees. Among the questions in this section, respondents were asked what time the interview was taking place, and what month and season it was, the date of the lunar calendar Mid-Autumn Festival, and the name of the district or township where they lived. The ability to react, attention ability and computing ability of the respondents were tested, with interviewees required to retell words (nouns), make calculations, and draw figures shown in Appendix of the questionnaire. Interviewees were asked to repeat the three words mentioned above to test their short-term recall ability. Three questions tested speaking and understanding ability of the elderly interviewees. To test coordination, respondents were asked to fold a piece of paper in half and put it on the ground. Finally, at the end of Part C, interviewers noted whether interviewees had been able to answer all of the questions in Parts B and C, and if not, to note what difficulties the elderly respondents had had. 4. Part D—Lifestyle. In this part, respondents were asked questions about diet, smoking, drinking, exercise, housework and their participation in social activities. There were questions about the types and quantities of staple foods interviewees ate, whether they often ate fresh fruits and vegetables, and whether they often ate meat, aquatic products, eggs, bean products, pickles, sugar, and garlic, and whether they drank tea. Respondents were also asked about the sources of their drinking water when they were children, when they were 60 years old and at the time of the interview. Smoking and drinking questions asked not only about current behavior, but also queried the age at which respondents had started and, if applicable, had stopped smoking or drinking, the amounts they smoked and drank, and the types of alcohol they consumed. Physical exercise questions asked about purposeful fitness activities, and included question about the ages at which physical activity began and, if applicable, stopped. Interviewees were also asked the ages at which they began and stopped being engaged in physical labor. In addition, a series of questions were asked about participation in housework, outdoor activities, gardening and the keeping of pets, reading books and newspapers, raising poultry or livestock, playing cards or mahjong, watching TV and listening to the radio, and participating in organized social activities. Respondents were also asked about the number of tourist trips they had taken in the previous two years. 5. Part E—Ability to Perform Daily Activities. The questions included in all surveys are the six items on the ability to take care of daily life (that is, activity of daily living, ADL). Beginning with the 4th wave in 2005, if the elderly respondents reported needing help from others, additional questions were asked about when they began needing assistance. After 2002, questions were added to the survey about the 8 items of instrumental activity of daily living (IADL).

12

Z. Zheng

6. Part F—Personal Background and Family Structure. Interviewees were asked about number of years of schooling, the main types of work they did before age 60, and their main source and other sources of income at the time of the survey. Interviewees were asked questions about their current marital status and their marital history, and there were questions about the number of years of school attended and the main job before age 60 of the current (or most recent) spouse. Interviewees were asked to identify their primary source of care provider when they were ill. They were asked if they could be treated for serious illness in a timely manner, and why, if treatment was not timely, they could not. There were also questions about whether interviewees had access to timely treatment of illness at age 60. And there was question about who paid for their medical expenses. Interviewees were also asked if they often went hungry in childhood. In addition, information about the gender, age and whether they were still alive was collected about the parents, siblings and children of the elderly interviewees. The final questions in this part asked who, if anyone, the interviewee confided in or could ask for help, and if there were monetary or material exchanges between the interviewees and their children. 7. Part G—Physical Health. Information in this part was provided by exam or observation of interviewer. There was data for each elderly respondent on eyesight and teeth as well as oral hygiene, hand use and dominance, blood pressure and heart rate, height and weight, and the ability to move upper limbs and legs. Finally, there was information about any illnesses or symptoms currently afflicting interviewees, and whether they were hospitalized or bedridden. 8. Questions for Interviewer. In the final part of the questionnaire, interviewers recorded their observation of the interview. They indicated whether interviewees could hear the questions clearly, whether they accepted physical examination, the health status of the interviewee by observation, and whether the age information reported was accurate. They also noted if someone answered on behalf of the interviewee during the visit, and recorded the relationship of proxy reporter to the interviewee, as well as other issues have to make a record or to explain. Biomarkers samples were collected in several waves. In the baseline survey of CLHLS in 1998, 4116 fingertip blood samples were collected from interviewees aged 80 and above; 14,000 saliva samples were collected among interviewees aged 40–110 in 2008; about 4800 blood samples and urine samples were collected in the eight longevity areas in 2009 and 2012; 2561 blood samples and 2448 urine samples were collected in longevity areas in 2014; blood samples or saliva samples were also collected in 2018. For those elderly who died during intervals between two surveys, their situation before death was recorded based on the memories of family members or informants close to the deceased. Information was collected on the marital status and living arrangements at the time of death, the number of family members living together and the number of generations living together, the place and cause of death, the main caregivers who assisted with daily life activities before death and the number of days of care provided during the month before death. There was information on whether

1 Twenty Years’ Follow-Up Surveys on Elder People’s Health …

13

the deceased was bedridden at the time of death and the number of days before death that he or she was confined to a bed; and whether the elderly can be treated in a timely manner, and the number of serious illnesses and days of illness reported after the previous survey until the time of death. The main sources of income and the average annual income of the family during the year before the elderly person’s death were recorded, as were the actual medical expenses incurred in the year before the death. Information was also collected about the facilities available in the home to the elderly person before death and, if the deceased was a rural resident, whether there was a doctor in the village. The ADL of the deceased at the time of death was recorded, and if ADL limited the number of days of ADL limitation. Information was also collected on the number of days that the deceased was completely dependent on others for care before death, the total cost of care, the cost of direct care in the month before death, and the main payer of the cost of care before death. Finally, information on the activities, and the smoking and drinking habits of the deceased. County level information of the sample sites (county, city, or district) were collected in both baseline and follow-ups of CLHLS, including more than 30 items related to geographical setting, demographic and socioeconomic indicators, and environmental information (such as pollution and disasters).

1.4.2 Major Modifications to the CLHLS Questionnaire Since the launch of the first CLHLS survey in 1998, numerous changes to the characteristics, family structure and living arrangements of the elderly in China have taken place. These changes have occurred against a background of rapid social and economic development and improvements to public policies. There is no doubt that the lives and health conditions of China’s elderly have changed, and comparisons of the results of the different waves of CLHLS illustrate these changes. Over the years, changing conditions have created new needs for more information from the survey. In response, CLHLS has not only retained the core components of the questionnaire, but the contents of questionnaire have also been expanded to meet the needs for more information and to satisfy the demands of emerging research initiatives. When the first CLHLS was completed in 1998, much about the health and family situations, and the socio-economic status of the elderly population, and especially of the oldest old, were unknown or not clear. The results of the survey in 1998 provided a demographic profile of the elderly population that filled many of the gaps in knowledge. As the follow-up surveys developed, more information became available, and in-depth analysis of this information revealed new issues that required exploration. The CLHLS research team organized symposiums after most of the surveys. These were seen as platforms for scholars to exchange information, increase awareness of limitations to the questionnaire, and identify areas of research interest that required more information. Therefore, in addition to retaining the core content, each follow-up survey added and adjusted items in the questionnaire to address

14

Z. Zheng

ongoing changes to the elderly population, socio-economic conditions and the public policy environment, and to meet the needs of research and decision-making. In the 2002 and 2005 follow-up surveys, 20 items were added to the questionnaire for the elderly interviewees and 5 items were added to the questionnaire for deceased elderly. The 2008 follow-up survey added two questions to the part about speaking, understanding and self-coordination ability, and the mini-mental state assessment scale (MMS) was modified. In 2008, the survey in longevity areas applied additional health exam form in eight longevity sites and conducted lab tests of some biomarkers using blood or urine samples. Data collection for the 2011–2012 follow-up survey added 13 new PhenX indicators, and there was a total of 32 items in the questionnaire.3 The PhenX indicators are selected from items related to elderly health in the PhenX toolkit, which also applicable to the context of Chinese culture and contemporary social life. The PhenX toolkit provides standardized measurement on complex features, complex diseases, phenotypic characteristics, and environmental exposures. The standardized measurement makes possible comparative study between different countries and regions. The 2018 survey added data collection on mental health and cognitive ability. In addition, the 2018 survey added questions about the ventilation of indoor living environments and the use of insecticides, air fresheners, and household cleaners in the homes of interviewees. With regard to data gathered for deceased elderly, beginning in 2011, additional information was collected about the retirement situation of the deceased before death and what pension or old-age insurance benefits they received. Information about family members who had lived with the deceased and housing conditions was collected, as well as information about oral hygiene, dental problems, and hearing loss of the elderly before death. These questions were similar to those added to the questionnaire in 2011 for surviving elderly people aged 65 and over. Because hospice care for the elderly is labor intensive, elderly people may have more than one caregiver, so the 2018 questionnaire added information about the relationships of second and third caregivers to the elderly person and the number of days of care provided. The 2018 questionnaire also added 16 questions (IQCODE) about the cognitive function of the elderly, asking relatives (or other informants) to recall the state of cognitive functions of elderly individuals during the six months immediately prior to their deaths.

1.5 Data Quality Assessment Based on international comparisons of surveys among elder people, CLHLS data quality is quite satisfactory after a comprehensive assessment. The assessments focus 3

For information about PhenX indicators, see https://www.phenxtoolkit.org/, and for information about the construction of the indicators, see Hamilton et al. (2011).

1 Twenty Years’ Follow-Up Surveys on Elder People’s Health …

15

on major indicators such as: accuracy of age reporting, reliability and validity of main health measurements, proxy use and nonresponse, missing data, logical error, and reliability of mortality. The assessments support that the data quality of seven waves (1998–2014) of CLHLS is generally good (Gu 2008, pp. 39–60; Bongaarts 2009; Goodkind 2009; and Appendix in this book). As Goodkind (2009) commented, “the quality of reporting, despite a few flaws endemic to this kind of survey, was fairly good.” The technical report of CLHLS datasets provides explanations to data limitations need to pay attention when use the datasets. The research team of CLHLS assessed data quality of three waves (1998, 2000, and 2002) in 2004, focused on reliability and validity of age reporting of oldest old and major health measurements, as well as analysis of proxy, nonresponse, and lost in follow-up. The same data quality assessment strategy and methodology has been applied to other follow-up waves. Based on comparisons with age reporting in China and other countries, analysis in 2004 found that the age reporting quality of Han oldest old interviewees aged 105 and younger is close to that in developed countries. The bias in age reporting increases by increase of age, quality of age reporting is not as good among elderly aged 106 and above. Age reporting of minority ethnic groups (about 7% of total sample) is fairly good. The data quality of main health measurements is as good as other similar surveys and has good convergent and discriminant validity (Zeng and Gu 2008; Gu 2008). The results of analysis show that the data quality of main indicators of 2011 and 2014 CLHLS datasets are fairly good (see Appendix in this book). We briefly report major findings from data quality assessment on age reporting, major health measurements, logical error, proxy, nonresponse and missing data, lost in follow-up, and mortality data of 2011 and 2014. For more detailed report of the analysis, please refer to the data quality report in Appendix. 1. Accuracy of age reporting Han ethnicity makes 94.13% in 2011 CLHLS, and minority ethnic groups mainly are Zhuang (3.26%), Hui (0.76%), Yao (0.36%), Manchu (0.45%), Mongolian (0.06%), and Korean (0.02%), they make 99.04% of the sample. In 2014 CLHLS follow-up survey, Han ethnicity makes 92.54%, and other minority ethnic groups make 2.94, 0.62, 0.33, 0.32, 0.03 and 0.02% of the sample, about 96.80% in total. The age reporting of Han and six minority ethnic groups are proved “very good.” The sex ratio of Han elderly aged 100–104 and 105–109 is quite similar to that of centenarians from Sweden, but sex ratios of centenarians of the six minority ethnic groups are slightly different, it is possibly because that the sample size of minority centenarians is too small. 2. Reliability and validity of major health measurements The major variables of health in CLHLS are health measurements often applied to assess health status of elder people, such as ADL, IADL, and MMSE. As the measurements have been well developed and widely used, it is possible to make international comparison on quality of data with other similar surveys. The results of quality assessment show that internal consistency of ADL, IADL and MMSE is

16

Z. Zheng

quite close to other similar surveys, the major health indicators of 2011 and 2014 surveys are reliable. Construct validity (convergent and discriminant validity) is the key validity, often tested by correlation analysis. The results of 2011 and 2014 health indicator analysis show that indicators of the same or similar dimension have stronger correlation than that of other different dimensions, imply the construct validity is good. Another approach to assess validity is factor analysis, which checks agreement on response from items of same dimension. The responses from the items of same dimension should be included in the same factor and with similar loading for valid measurements. The results of factor analysis for daily life activities and cognition ability support a good validity of the indicators. 3. Logical error Logical error is often caused by inadvertent error during interview or data entry. The rate of inconsistent response is slightly higher than 1% in 2011 and 2014, mainly found as inconsistency between daily life activity reported and physical ability observed, and some cases marked proxy for the response but the interviewer did not report proxy, however the rates of inconsistency are not high. 4. Proxy, nonresponse and missing data Previous studies have shown that proxies have been used as an alternative method to reduce nonresponse or missing data in the survey of the health status of elder people, especially for the oldest old interviewees due to impaired hearing/vision, frail health or recall problems. Proxy has a better effect on factual questions, and deviation is smaller when the information is more specific and observable. However proxy should not be used for questions such as personal feelings, satisfaction, depression and cognition, this is the principle of CLHLS as well as other surveys. The proportion of all questions answered with help is zero in the CLHLS surveys, and about 90% of the proxy reporter is spouse or children/grandchildren of the interviewee, indicating that the deviation of the answer is not very large. Note that the proportion of proxy use increases obviously with the increase of age. Some questions have a high proportion of “Don’t know” or “Missing”, such as parents’ death age and age of the elderly when their parents died, and the proportion of missing is close to 20% in 2014 wave. Special attention should be taken in the application of these variables. The proportion of nonresponse has a direct impact on estimation. Compared with the first five waves from 1998 to 2008, the proportion of nonresponse in the 2011 wave was similar to that in 2008, at 6.92%, and decreased to 2.92% in 2014. The proportion of nonresponse is in an acceptable range. Women’s proportion of nonresponse in the two surveys was slightly higher than that of men, and the proportion of nonresponse increased with the increase of age. Although the proportion of nonresponses in CLHLS is relatively low, if every respondent has a nonresponse in some questions would lead to the incompleteness of the data. Nonresponse can be further classified into “Don’t know” and “Missing”. When it comes to attitudes, feelings

1 Twenty Years’ Follow-Up Surveys on Elder People’s Health …

17

and expectations, the proportion of oldest old people who answer “Don’t know” is relatively high. The proportion of incomplete item in CLHLS is less than 10%, which is relatively lower based on international comparison. The proportion of nonresponse of oldest old is higher than that of the younger elder people. The questions with higher proportion of nonresponse are the cognitive function of oldest old people, the frequency of smoking and passive smoking, the type of serious illness and the length of stay in hospital in the past two years. Therefore, special attention should be taken when using these variables for analysis. If the missing is completely random, the deviation will not be very large. However, if the missing is not random, then ignoring them in the analysis will lead to a certain deviation. In this case, it should be tested to identify related factors. Multivariate logistic regression analysis found that nonresponse was related to ethnic groups, marital status, urban and rural residence, cognitive function, and self-rated health. The proportion of incomplete items was higher among the elderly who were older, female, urban residence, ethnic minority, without spouse, and in poor health. 5. Lost in follow-up Respondent lost in follow-up is one of the serious problems in longitudinal survey. Similar to nonresponse situation, if the lost in follow-up is completely random, it will not have a significant impact. However, if the lost in follow-up is not random, it will affect the results. There were 1690 elderly surveyed in 2008 lost in follow-up in the 2011 wave, accounting for 11.53% of those interviewed in 2008, and the proportion decreased to 8.38% in 2014. The proportion of lost in follow-up is in a lower range with an international comparison. The rejection rates in 2011 and 2014 were 2.40% and 1.54%, respectively, both of which were very low. Multivariate logistic regression analysis shows that women, elder people with poor physical health and cognitive function, and had less social communication were more likely to be lost in follow-up. Meanwhile non-illiterate elderly are more likely to be lost in follow-up than illiterate elderly, and urban elderly are more likely to be lost in follow-up than rural elderly, partly because urban residents are more likely to be relocated than that in rural areas. In addition, the Han elderly are more likely to be lost in follow-up than those of ethnic minorities. After making comparison with other similar surveys, we believe that the proportion and pattern of lost in follow-up in CLHLS including these two surveys will not cause a significant bias. 6. Data quality of mortality According to the comparison of the 2008–2011 and 2011–2014 age-sex-specific mortality from the CLHLS with those from the 2010 census, the age-sex-specific mortality of the CLHLS survey is reliable. The results of internal consistency analysis and factor analysis on the main health variables of elderly before death show that the reliability and validity of the main health variables are good. The overall quality of the data is relatively high, the same as or better than the quality of similar surveys in developed countries.

18

Z. Zheng

The main problems in the data of the deceased elderly are missing data and nonresponse. For example, the proportion of incomplete items in the health status before death of the deceased elderly in 2014 survey was 8.8%. The questions with a relatively high proportion of incomplete items include suffer from any disease before death, whether there are dreams of death, whether there is a doctor in the village, the cost of end-of-life care, which should be paid attention to when analyzing these variables.

1.6 Data Use and Publications Data cleaning and evaluations that included validity and reliability testing and analysis of non-responses and loses to follow-up surveys were performed after each survey. The data proved to be reliable and the quality satisfactory. During the last 20 years, a number of interdisciplinary and multi-institute cooperative studies have been carried out using the survey data. Nearly eight hundreds academic articles have been published in Chinese and English, more than one hundred master theses or doctoral dissertations have used the data, and eleven books with findings from the CLHLS have been published in Chinese or English. Several policy briefs and advisory reports have been produced as well. The use of CLHLS data has increased over time as more waves of survey data have become available. Before 2011, an average of 17 papers per year were using CLHLS data, but during the years 2012–2019 that number jumped to an average of 66 papers per year. Several book-length collections of research findings presented at international seminars and conferences focused on CLHLS data analyses (e.g. Zeng et al. 2008). The significant increase in Chinese language publications suggests there is more research interest and policy discussions related to ageing issues in China. The publications cover multiple research topics. Figure 1.1 breaks down by research topics papers published in Chinese and in English in peer reviewed academic journals from 1998 to 2019. Health status and the determinants, measurements of health, longevity and the risk of death, and health related biologic, genetic and environmental factors are the most frequently studied topics. Issues related to families, such as living arrangement, intergenerational relationships, and the relationship the family situation has to the health status of elderly people are also frequently studied topics. Studies related to the care of the elderly are, for the most part, published in Chinese language journals, indicating that both the society and academia are paying more attention to this issue in response to the rapid ageing of China’s population. Topics related to health care, medical insurance, and social support also appear mainly in Chinese language publications, a response to the changing situation for health insurance and social welfare programs mentioned earlier. English language studies are often related to psychological and cognitive health, as well as to lifestyle and health. Earlier publications were concerned primarily with the general status of the oldest old, and provided demographic and socioeconomic profiles of this subgroup that

1 Twenty Years’ Follow-Up Surveys on Elder People’s Health …

19

Survey/methodology Old-age care Healthcare/social support Social-economic status Family/living arrangement Lifestyle and health Mental health Health and measurement Cognitive function ADL/IADL and related factors Self assessment Demographic/socioeconomic profile 0 10 20 30 40 50 60 70 80 90 100110

In Chinese

In English

Fig. 1.1 Number of papers using CLHLS data published by research topics, 1998–2019

helped to fill gaps in the knowledge of the oldest old. Health status and the determinants, especially those factors related to healthy longevity, have always been topics of research interest. Because more CLHLS data became available with each followup study and the types of data collected expanded over the years, studies using the data have been able to address virtually every dimension of health status and the determinants of that status for elderly people. Over time increasingly sophisticated methodologies have been applied to explore complex relationships in-depth. In the last decade numerous research findings about the socio-economic status and social welfare of elderly people have been published, and more papers than before have been published about psychological health, lifestyle and health, and the relationship of bio-indicators and environmental factors to the health of elderly people. In recent years, a number of papers have focused on the interrelationships of biomarkers, socio-economic and environmental factors as they affect the mortality, disability, and cognition of elderly people. Several papers have examined issues related to the quality of life before death. Longitudinal studies utilizing information from multiple CLHLS waves have also become more common in recent years. The CLHLS dataset (1998–2014) and all related documents is available in the Peking University Open Access Research Database (http://opendata.pku.edu.cn) and the National Archive of Computerized Data on Ageing, Inter-university Consortium for Political and Social Research (https://www.icpsr.umich.edu/icpsrweb/NACDA/ series/487). When using CLHLS data, please note accompanying explanations describing the sample for each wave and adjustments made to each of the survey

20

Z. Zheng

questionnaires in order to find the data content corresponding to the appropriate survey year and to meet the needs of specific research topics. Acknowledgements This study is supported by National Science Foundation of China (71233001, 71490732).

References Bongaarts J (2009) Book review on Zeng Yi, D. L. Poston, D. A. Vlosky, and Danan Gu (eds.), Healthy longevity in China: demographic, socioeconomic, and psychological dimensions. Popul Dev Rev 35(2):434–435 Goodkind D (2009) Review on the book “Healthy longevity in China: demographic, socioeconomic, and psychological dimensions”, Edited by ZENG YI, DUDLEY L. POSTON JR., DENESE ASHBAUGH VLOSKY, and DANAN GU. Dordrecht: Springer. Pp. xv_435. t134.95. ISBN: 978-1-4020-6751-8. Popul Stud 63:312–313 Gu D (2008) General data quality assessment of the CLHLS. In: Zeng Y, Poston D, Vlosky DA, Gu D (eds) Healthy longevity in China: demographic, socioeconomic, and psychological dimensions. Springer, Dordrecht Hamilton et al (2011) The PhenX toolkit: get the most from your measures. Am J Epidemiol 174(3):253–260 He W, Hu X, Hou Y et al (2018) The development of economic security system for the elderly people in China. In: Du P, Lu J, He W (eds) Annual report on development of actively coping with population ageing in the new era. Hualing Press, Beijing (in Chinese) Peng X, Hu Z (2015) The contemporary transition of the Chinese family and the reconstruction of family policy. Soc Sci China 12:113–132 Wang G (2019) Seventy years of China: the changes of population age structure and the trend of population ageing. Chin J Popul Sci 5:1–15 (in Chinese) Zeng Y (2008) Introduction to the Chinese longitudinal healthy longevity survey (CLHLS). In: Zeng Y, Poston D, Vlosky DA, Gu D (eds) Healthy longevity in China: demographic, socioeconomic, and psychological dimensions. Springer, Dordrecht Zeng Y, Gu D (2008) Reliability of age reporting among the Chinese oldest-old in the CLHLS datasets. In: Zeng Y, Poston D, Vlosky DA, Gu D (eds) Healthy longevity in China: demographic, socioeconomic, and psychological dimensions. Springer, Dordrecht Zeng Y, Vaupel JW, Xiao Z et al (2001) The healthy longevity survey and the active life expectancy of the oldest old in China. Popul Engl Select 13(1):95–116 Zeng Y, Vaupel JW, Xiao Z, Zhang C, Liu Y (2002) Sociodemographic and health profiles of the oldest old in China. Popul Dev Rev 28(2):251–273 Zeng Y, Poston D, Vlosky DA, Gu D (eds) (2008) Healthy longevity in China: demographic, socioeconomic, and psychological dimensions. Springer, Dordrecht Zhu H, Lu J, Zhang Y, Cui B (2018) Responses to population ageing in the new era: a national condition report from China. China Popul Dev Stud 2:272–283

Chapter 2

Trends in the Impact of Socioeconomic Status on Health with Increases in Ages: Convergence or Divergence? Jianxin Li and Cuicui Xia

2.1 Introduction The relationship between socioeconomic status and health has been an ongoing research topic in many disciplines. In this field, the positive effect of socioeconomic status on individual health has been confirmed by many studies (Williams 1990; Link and Phelan 1995; Zhu and Xie 2007; Zeng et al. 2007; Wang 2012). However, there were two different outcomes on the effect of socioeconomic status on health among different age groups. One is the “convergence effect” (also known as the divergence-convergence hypothesis), which suggests that the health divergence of different socioeconomic status groups firstly increases with age, but then becomes smaller or even disappear in advanced ages (House et al. 1990), eventually “converging” to no differences. The other is the “divergence effect” (also known as the accumulative hypothesis), which suggests that the effect of socioeconomic status on people’s health increases with age (Ross and Wu 1996) and eventually “diverges” even greater disparities. In fact, by using data from different regions and time periods, previous empirical studies have found the existence of the “convergence effect” and “divergence effect”. However, there is little study has examined the effect of socioeconomic status on health, since variations in times, regions, populations and socioeconomic environments will lead to different patterns of health effects. Based on the analyses of the China Family Panel Studies (CFPS) data, this chapter examined the effect of socioeconomic status on individual health, and discuss whether this effect changes with age. J. Li (B) Department of Sociology, Peking University, Beijing, China e-mail: [email protected] C. Xia Institute of Population and Labor Economics, Chinese Academy of Social Sciences, Beijing, China e-mail: [email protected] © Science Press 2022 Y. Zeng et al. (eds.), Trends and Determinants of Healthy Aging in China, https://doi.org/10.1007/978-981-19-4154-2_2

21

22

J. Li and C. Xia

2.2 Literature Review and Research Design Currently, there are two different views of causality in the study of the relationship between socioeconomic status and health. One is that socioeconomic status determines health, with different socioeconomic status having different levels of health. Those with higher socioeconomic status are healthier (Dahl 1996). The other is health selection, which suggests that health differences will result in socioeconomic status differences. For example, people with poorer health are less likely to achieve higher socioeconomic status and higher income (West 1991). It can be argued that both views are reasonable and have been verified. In this chapter, we will focus on the socioeconomic status as determinants of health and explore the impact of status on health and the possible age pattern. Socioeconomic status is the position of people in the socioeconomic stratification, which influences their socioeconomic behavior, psychological state, knowledge and access to resources, etc. The British scholar Marmot has systematically discussed the aspect of socioeconomic status in determining people’s health. His conclusion goes that the higher status of a person, the higher level of his/her health. Marmot also outlines how socioeconomic status affects health in several ways, namely behavioral styles, welfare levels, psychological stress, loneliness and social relationships, and parental status inheritance (Michael et al. 2008). The perspective that status determines behaviors suggests better education will enhance people’s ability to solve problems, regulate their health, promote mental maturity, and bring a sense of fulfillment (Winkleby et al. 1990; Ross and Mirowsky 2010). However, people with lower levels of education are less likely to recognize the dangers of adverse behaviors, and those with lower income are more likely to adopt behavioral habits that are detrimental to health, such as smoking and alcohol abuse (Mheen et al. 1999). Additionally, people with lower socioeconomic status experience more crises and uncontrollable life events. For example, those with lower living conditions are more vulnerable to dramatic life transitions like crime, violence, discrimination, illness, and death of a child. They are more susceptible to the burdens of life and the uncertainty of income as well. The welfare level perspective argues that those with higher socioeconomic status are more likely to have access to a good residential living environment, better nutritional status, and health services (Dahl 1996). The psychological stress perspective suggests that people with lower income tend to face more acute and chronic stress from life transitions (Belle 1990; House and Robbins 1983), which are the key factors contributing to poorer personal health. Loneliness and social relationships will also affect health. Marmot argues that opportunities to participate in life and the sense of control over one’s life will become determinants of health when people reach a certain degree of material living conditions. And people from higher socioeconomic classes possess more health knowledge and health awareness. There are also two main views on the trend of socioeconomic status on health with increases in ages: the “convergence effect” and the “divergence effect”. The “convergence effect” suggests that health differences among socioeconomic status groups are smaller in the early stage of life, more divergent in middle and old age,

2 Trends in the Impact of Socioeconomic Status on Health …

23

and become smaller or even disappear in advanced old age (House et al. 1990). This view has established that biological and physiological factors play a major role in the level of health at young-age and advanced-age, when people’s physical function are in periods of peak and rapid decline. Thus, during these periods socioeconomic status is of less importance in shaping the individual’s body (Baum et al. 1999). Related empirical studies have also demonstrated that differences in health indicators such as self-rated health, chronic diseases, and physical functioning are smaller in younger age groups, larger in adulthood, and then narrow or disappear at old age (Robert and House 1996). In addition to physiological factors, the survival selectivity also matters in the convergence effect, with those in poorer health having a lower probability of surviving to older ages. Thus, older adults from different socioeconomic status groups all have undergone survival selection. The theory of “divergence effect” is derived from the exploration of cumulative advantage and cumulative disadvantage. In economics, there is a “Matthew effect” of cumulative disadvantage, which suggested that income inequality is greatest in the elderly group and the Gini coefficient is higher in the elderly group than in other age groups (Easterlin et al. 1993). The cumulative advantage theory proposes that the accumulation of experiences over the life course will influence later life patterns. In this regard, a number of studies have pointed out that the life circumstances and events that people experience at different ages will play a role in the aging process (Hu 2009). And socioeconomic status in early life will have a long- lasting impact on an individual’s health (Shen 2008). Some studies have also corroborated the divergent effects in terms of the rate of health decline with age. What they find is that compared with people from low socioeconomic status, those with high socioeconomic status show greater advantage in health as age grows, thus producing a divergence effect of health disparities at older ages (Prus 2007; Lowry and Xie 2009). Although many studies have focused on the relationship between socioeconomic status, age and health, it is necessary to study the trends in health disparities with increase in ages in different regions and over different time spans. This is because socioeconomic context—such as political background, family structure and cultural structure—may lead to differences shapes of impact pattern curve (House et al. 1990; Lowry and Xie 2009). Based on previous research, this chapter builds a framework for the impact of socioeconomic status on health and attempts to make improvements in the following two areas. Firstly, the relationship between socioeconomic status and health is examined using multidimensional health indicators from the WHO definition of health. The definition goes that health is a state of physical, mental and socioeconomic well-being (WHO 1947). According to this definition, three indicators will be selected to measure health, which is illness, mental health and self-reported health. Secondly, the CFPS is a comprehensive, longitudinal socioeconomic survey in China. Since it is well represented, it can provide a more profound understanding of the relationship between socioeconomic status and health. In this chapter, two main questions will be answered. The first is whether one’s socioeconomic status will influence health conditions. The second is whether this effect on health changes with increase in ages and whether the trend is “convergence effect” or “divergence effect” or other possible effects.

24

J. Li and C. Xia

2.3 Data and Research Methods 2.3.1 Data We use data from the China Family Panel Studies (CFPS) 2012 wave. Through a multistage probability sampling strategy, the baseline survey in 2010 drew a sample of 14,798 households from 25 out of mainland China’s 31 provinces excluding Xinjiang, Tibet, Qinghai, Inner Mongolia, Ningxia and Hainan. Since the survey covers over 95% of the whole population, it can be considered as a nationally representative sample (Xie et al. 2013). In 2012, a follow-up survey was conducted on households interviewed in 2010, with 13,459 households and 44,693 individuals (36,063 adults) being interviewed. In our study, 19,841 respondents were included, and the data variables covered individual education, income, hukou, health status, and household incomes per capita.

2.3.2 Measures The dependent variable is health condition, with three indicators of health measurement. The first one is the measure of illness, using the indicator “whether you are sick in last 2 weeks”. Result shows that 29.28% of the respondents claimed to be sick. The second one is mental health, which is measured by converting the mental health scale of the CFPS questionnaire into a score between 0 and 100, with higher scores representing better mental health. The third one is self-reported health. Five options including “extremely healthy”, “very healthy”, “somewhat healthy”, “average” and “unhealthy”. In our analysis, we pay more attention to those who rated themselves as “unhealthy”, which accounts for 17.56% of the respondents in the sample. We then combine the other four choices into one category called “healthy”, which accounts of 82.4%. As for the main independent variables, which is the socioeconomic status, two indicators were selected in this chapter: individual education level and per capita household income. Generally speaking, the common measures of socioeconomic status are income, occupation, and education. Some studies include housing and assets in the measurement as well (Zhu and Xie 2007; House et al. 1990). On the one hand, education and household income per capita can adequately reflect the socioeconomic status of an individual; on the other hand, it has also been found that education is more sensitive to health among the indicators of socioeconomic status (Winkleby et al. 1990). What’s more, in Chinese society, the concept of “family” is strong, family income per capita is a more realistic indicator of an individual’s economic ability. Therefore, it is reasonable to use both education and household income per capita to reflect socioeconomic status. We reclassified the type of education in the original questionnaire into three categories, illiterate and elementary school, middle and high school, and college and above, which accounts for 41.03%, 49.38%, and

2 Trends in the Impact of Socioeconomic Status on Health … Table 2.1 Descriptive statistics of all variables (%)

Variable

25

% or number Variable

%

Age

Dependent variables Illness: sick in last 2 weeks

29.28

16–29

15.43

Average mental health score

79.73

30–39

17.21

Self-reported healthy

82.44

40–49

19.87

50–59

18.28

Education

60–69

16.74

Illiterate and elementary 41.03 school

70+

Independent variables

Middle and high school College and above

49.38

Gender

9.59

Female

Income

8.62

50.36

Marital status

Low income

24.98

Middle income

50.00

With spouse 79.76 Hukou

High income

25.02

Urban

26.78

N = 19,841 Data sources CFPS 2012

9.59% of the population respectively. The variable of household income per capita was classified into low income, middle income and high income according to the cut-off points of 25 and 75%. Age was also transferred into categorical variable. Specifically, it was divided into six groups: 16–29, 30–39, 40–49, 50–59, 60–69, and 70 and above, with the percentage of 15.43%, 17.21%, 19.87%, 18.28%, 16.74%, and 8.62% of the total sample respectively. The main control variables are hukou and other personal information. Because of the uneven distribution of health care resources between urban and rural areas (Hu and Hu 2003), differences also lay in health status between urban and rural areas. So, it is necessary to control urban and rural residence in the analysis, and the indicator “hukou” is used to roughly reflect the differences, with those who have urban hukou accounting for 26.78% of the total sample. In terms of the measurement of marital status, it is recoded into two categories, “with spouse” and “without spouse”, and the proportion of those with spouse is 79.76%. In addition, the percentage of females is 50.36%. The variables are described in Table 2.1.

2.3.3 Method In this chapter, OLS regression and logistic regression models are applied. For the three dependent variables, we first define a base model with three core independent

26

J. Li and C. Xia

Table 2.2 Overall significance of different model Model

Model description

Model 1

No interaction term

Illness

Mental health

Self-reported health

Model 2 Model 3

Education × age

Not significant

Not significant

Significant ***

Income × age

Significant***

Significant **

Model 4

Not significant

Income × age + education × age

Not significant

Not significant

Not significant

**p < 0.01, ***p < 0.001

variables: education, household income per capita, and age; and three control variables: gender, marriage, and hukou. Second, we assume that the effect of socioeconomic status on health status will change as age increases and produce “convergence effect” or “divergence effect”. In order to test these hypotheses, the interaction terms of education and age, income and age were added to the base model. The models are shown in Table 2.2. In details, Models 2 and 3 were compared with the model without interaction terms, and Model 4 was compared with Models 2 or 3. The likelihood ratio is tested by using chi-square. Since mental health score is a continuous variable, R-square incremental test is used here.

2.4 Results 2.4.1 Effect of Socioeconomic Status on Health To test the effect of socioeconomic status on different health indicators, we examined the health differences among groups with different levels of education and income, controlling for basic demographic information. The results of the underlying model are shown in Table 2.3. Compared with those who are illiterate and only finish elementary school, the middle and high school groups are less likely to be sick, and the odds ratio of illness is 0.87. However, other education and income coefficients are not significant. This may due to the measurement of the dependent variable, which could be verified using a more valid measure of physical health. In terms of mental health, the higher the education and income, the higher the mental health score. The mental health score of the middle and high school group is about 3.3 points higher than the reference group, while the college and higher group is 4.1 points higher. The score of middle-income earners is 1.2 points higher than low-income earners on average, while the score of high-income earners is about 2.5 points higher than low-income earners, and all the differences are significant. In terms of self-reported health, both education and income exert a positive and significant effect on selfreported health. The odds ratio of self-reported health is 1.64 times higher for those whose education level is middle and high school than the reference group, and 3.16

2 Trends in the Impact of Socioeconomic Status on Health … Table 2.3 Effects (odds ratios) of ages and socioeconomic status on health

Variables

Illness

27 Mental health

Self-reported health

Age (16–29 = 0) 30–39

0.460***

−1.802***

−1.172***

40–49

0.719***

−2.118***

−1.821***

50–59

1.074***

−3.256***

−2.371***

60–69

1.174***

−2.263***

−2.428***

70 and above

1.194***

−2.727***

−2.499***

−3.150***

−0.405***

Gender (Male = 0) Female

0.365***

Marriage (Without spouse = 0) With spouse

−0.065

3.190***

0.124*

1.883***

0.077

Hukou (Rural = 0) Urban

0.003

Education (Illiterate/elementary school = 0) Middle and high school

−0.134***

3.320***

0.496***

College and above

−0.009

4.125***

1.150***

Income (Low income = 0) Middle income

0.007

1.201***

0.187***

High income

0.009

2.459***

0.333***

N = 19,841; *p < 0.05, ***p < 0.001

times higher for the college and above group. The odds ratio of middle-income and high-income earners are 1.21 times and 1.40 times higher than the reference group. Indeed, age is also the main variable of great importance. No matter which indicator we adopt, there is a tendency for health status to become worse as age increases. In summary, the effects of education and income on different health indicators are different. Mental health and self-reported health are more likely to be influenced by education and income, while illness is less or even not influenced by socioeconomic status. This result may also be due to the inconsistent direction of the effect of education and income on illness at different ages, leading to an overall insignificant or small effect, which we will discuss in following section.

28

J. Li and C. Xia

2.4.2 Verification of the “Convergence Effect” and the “Divergence Effect” To verify whether the effect of socioeconomic status on health changes with age, this section focuses on the results of the model with the interaction term. As shows in Table 2.2, the explanatory power of the logistic regression of illness with the interaction term between income and age is significantly better than that of the model without the interaction term. As for the linear regression of mental health, the model with the interaction term between income and age is more significant than that of the model without the interaction term. And when it comes to the regression of self-reported health, the model with the interaction term between education and age is more significant than that of the model without the interaction term. For each dependent variable, based on the selection of a suitable model, we evaluate the probability of becoming healthier among different socioeconomic status groups in three dimensions that mentioned above. Line graphs are drawn to reflect the trend of health status among different status groups with increase in ages. Figures 2.1 and 2.2 reflect the trends in the variation of the possibility of falling ill with age among different socioeconomic status groups. Although Table 2.3 reflects that the effect of education and income on illness is small, the interaction term model of income and age is significant, indicating that there are differences in illness among income groups. And the differences may not be the same during diverse age periods. As has been depicted in Figs. 2.1 and 2.2, the possibility of falling ill among education groups is not significant at all ages, which is consistent with the finding in Table 2.3. Difference between income groups is only significant in the 16–29 age group and not in other age groups. In general, the two socioeconomic status variables—education and income—have little effect on the possibility of falling ill. There is a certain tendency for the effect of income on illness to vary with age, but it is non-significant in most age groups. In addition, the shape of the curves in Figs. 2.1 and 2.2 shows that, the probability of people being sick can be divided into two sections which is bounded by the 50–59 age group. It is obvious that there is a sharp increase with age before the age of 50–59, but the curve becomes flatter and the probability of being sick increases at a slower rate after age group 50–59. Next, we move on to examine the mental health dimensions. Figures 2.3 and 2.4 reflect how the effect of socioeconomic status on mental health changes with age. The effect of education on mental health is significant among all age groups. It suggests higher mental health scores for those with college and above, and lower mental health scores for those with illiterate or primary education. However, the mental health differences among the different education groups are approximately the same at each age, and there is no trend of health differences with age, showing a “parallel effect” trend. Figure 2.4 depicts the pattern of people from different income groups. It can be seen that the variation in mental health of different income groups is significant in all age groups except the groups whose age is between 16 and 29. People with high incomes have the highest mental health scores, followed by those with middle incomes, and those with low incomes have the lowest mental health

2 Trends in the Impact of Socioeconomic Status on Health …

29

0.45

Probability

0.40 0.35 0.30 0.25

Illiterate and elementary school Middle and high school College and above

0.20 0.15 16-29

30-39

40-49

50-59

60-69

70+

Age Fig. 2.1 Trends of illness with age in different educational groups

0.50 0.45 0.40 Probability

0.35 0.30 0.25 0.20 0.15

Low income Middle income High income

0.10 0.05 0.00 16-29**

30-39

40-49 Age

50-59

Fig. 2.2 Trends of illness with age in different income groups

60-69

70+

30

J. Li and C. Xia

scores. What’s more, the effect is significantly diverse in different age groups. As depicted in Fig. 2.4, the line of mental health differences among different income groups shows a “divergence effect”, which means that the health differences are significantly greater among older people. Specifically, in the age group of 16–29, there is no significant difference in mental health scores among different income groups. In the age group of 30–39, the difference begins to appear, which is as large as 1.3 between high-income earners and low-income earners. Then the difference is further widened to 2.8 in the age group of 40–49, and increases to 3.3 in the age group of 50–59 and 60–69. In the age group of 70 and above, the difference in mental health scores among different income groups is the largest, which is 4.3. This “divergence effect” reflects the impact of socioeconomic status on people’s mental health, which means that the difference in people’s mental health will gradually widen when they live in poor conditions for a long time. Even until the old age, the trend still exists. In addition, the curves of mental health change with age present a “U” shape. In the age groups of 16–29, 30–39, 40–49 and 50–59, there is a clear tendency for mental health scores to decline, and then begin to rise after ages 50–59. This “U” shaped curve shows that people’s mental health level becomes better after retirement. As the age increases, the age maturation effect and the historical age effect have a positive effect on people’s mental health. In other words, older people are more experienced and have stronger psychological adaptation ability, and they can also avoid the stress caused by work after retirement. Therefore, the mental health score curves in Figs. 2.3 and 2.4 show a “U” shape as age increases. Figures 2.5 and 2.6 show how the effect of socioeconomic status on self-reported health changes with age, controlling for the other variables. The effect of educational 86 84

Score

82 80 78 76

Illiterate and elementary school Middle and high school College and above

74 72 16-29***

30-39***

40-49*** 50-59*** Age

60-69***

Fig. 2.3 Trends of mental health with age in different educational groups

70+**

2 Trends in the Impact of Socioeconomic Status on Health …

31

83 82 81

Score

80 79 78 77

Low income Middle income High income

76 75 74 16-29

30-39**

40-49*** 50-59*** Age

60-69**

70+*

Fig. 2.4 Trends of mental health with age in different income groups

attainment on self-reported health is significant in all age groups except the 70-yearold group, showing a trend toward better self-reported health in the high educated group. The difference in self-reported health of those from varied educational groups is significantly different among all the ages. In the age group of 16–19, the probability of self-reported health in the college and above group is only about 0.06 higher than that of the illiterate or elementary school group. When age rises to 30–39 and 40–49, the difference tends to widen to 0.1 and 0.13 respectively. And at the age of 50–59, the difference in self-reported health among people with different levels of education reaches its highest point, which is about 0.2. However, this difference begins to narrow to 0.16 in the age group of 60–69, and is no longer significant in the age group of 70 years and above. This trend shows that the differences become smaller at first and then shrink or disappear, which is the “convergence effect”. Previous studies have suggested that there is a tendency for the differences of self-reported health to increase with age among different socioeconomic status groups, and the differences will still be significant in the older age groups (Zeng et al. 2007; Lowry and Xie 2009). Since the proportion of elderly people who are more than 80 years old in the data is relatively low, which is quite different from that in earlier studies (Zeng et al. 2007). Plus, the year in which the data was collect is not the same. In fact, the convergence effect of self-reported health can be explained in two ways. Firstly, this trend reflects the predominant role of biological factors on human health as age increases. Secondly, the survival effect also reduces health differences among socioeconomic status groups, which means that older adults who survive to age 70 or above have gone through health selectivity. In terms of the effect of income on

32

J. Li and C. Xia

1.00 0.95 0.90 0.85 Probability

0.80 0.75 0.70 0.65

Illiterate and elementary school Middle and high school College and above

0.60 0.55 0.50 16-29***

30-39***

40-49*** 50-59*** Age

60-69**

70+

Fig. 2.5 Trends of self-reported health with age in different educational groups

self-reported health, it is significant in all age groups except the 16–29 years group. However, the interaction term between income and age is not significant, indicating that there is no substantial difference among the effect at all ages. Or in other words, the effect of income on self-reported health will not increase with age. In addition, the shape of the curves in Figs. 2.5 and 2.6 show that self-reported health deteriorates very rapidly with age up to 50–59, but after this point it flattens out and the rate slows down. In fact, this trend is similar to the curve of changes in illness as age increases.

2.5 Conclusion Using the data from the CFPS 2012, this chapter adopts multidimensional health indicators to explore and analyze the impact of socioeconomic status on health in China and test whether this impact varies with ages. Compared with the results of related studies in China and abroad, this study has some similarities and differences, as well as some valuable findings. Our study shows that the effect of socioeconomic status on health varies across health indicators. For example, socioeconomic status has a large and significant effect on mental health and self-reported health, but not so on the illness. This result corroborates the results of previous studies (Qi and Wang 2010) and shows that there are differences in the sensitivity of different health indicators to socioeconomic stratification, which is why we use multidimensional health indicators in research. Specifically, our study found that the 50–59 age group is a very critical age group for health transitions. This is the age at which the probability

2 Trends in the Impact of Socioeconomic Status on Health …

33

1.00 0.95

Probability

0.90 0.85 0.80 0.75

Low income Middle income High income

0.70 0.65 0.60 16-29

30-39*

40-49** 50-59*** Age

60-69**

70+*

Fig. 2.6 Trends of self-reported health with age in different income groups

of illness and self-rated health decline at a significantly slower rate, and it is also the age at which the “U” curve of mental health reaches its lowest point. In addition, studies on the trend of socioeconomic status effects on health with increase of ages have yielded the same “convergence effect” and “divergence effect” findings as in previous studies (House et al. 1990; Lowry and Xie 2009). But this chapter also finds another pattern called “parallel effect”, which means that the differences in the impact of socioeconomic status on health do not vary with age. Of course, the finding of different effects of socioeconomic status on different health dimensions may be related to the selection of health indicators and socioeconomic status indicators. Since we have used physical, psychological and self-reported health indicators, this study is closer to the Chinese reality. Our findings suggest that socioeconomic status has different effects on different dimensions of people’s health. What’s more, the effect is long-lasting, across all ages, and is cumulative for some health indicators. At present, China is in the midst of accelerated socioeconomic development and transformation, and socioeconomic differentiation will continue to exist. China has entered an aging society, and the size of the elderly population will continue to increase. Since the elderly are a special group with greater differentiation and changes in socioeconomic status and health, it can be predicted that the problem of socioeconomic status and health inequality in China will continue to exist in the future. Therefore, adjusting socioeconomic public health policies, focusing on socioeconomic equity, narrowing socioeconomic status such as income and education gaps, actively expanding the positive effects of socioeconomic status on health, and exploring the accumulation or abatement

34

J. Li and C. Xia

factors of socioeconomic the status’s effects on old people’s health are of great practical significance for improving the overall health of the population and building a healthy aging society. Needless to say, there are shortcomings in this study. Due to the extremely small sample size of senior citizens aged more than 80 years in the CFPS data, we did not further divide subgroups for those aged 70 years and above to avoid unstable regression results. At the same time, since data are cross-sectional, it is not possible to distinguish between age effects and cohort effects. What we did here is suboptimal. Also, the mechanisms of the “divergence effect”, “convergence effect” and “parallel effect” need to be further explored. Acknowledgements This study was supported by the National Natural Science Foundation of China (Grant No: 71233001, 71490732).

References Baum A, Garofalo JP, Yali AM (1999) Socioeconomic status and chronic stress: does stress account for SES effects on health? Ann N Y Acad Sci 896(1):131–144 Belle D (1990) Poverty and women’s mental health. Am Psychol 45(3):385 Dahl E (1996) Social mobility and health: cause or effect? BMJ Br Med J 313(7055):435 Easterlin RA, Macunovich DJ, Crimmins EM (1993) Economic status of the young and old in the working age population, 1964 and 1987. In: The changing contract across generations, pp 67–86 House JS, Robbins C (1983) Age, psychosocial stress, and health. In: Aging in society: selected reviews of recent research, pp 175–197 House JS, Kessler RC, Herzog AR (1990) Age, socioeconomic status, and health. Milbank Q 383–411 Hu W (2009) Cumulative heterogeneity: differentiation of older adults from the life-course perspective. Chin J Sociol 29(02):112–130+225–226 (in Chinese) Hu L, Hu A (2003) From unfair to a fairer health development: analysis and suggestions of disease patterns in urban and rural China. Manage World 01:78–87 (in Chinese) Link BG, Phelan JC (1995) Social conditions as fundamental causes of disease. J Health Soc Behav 35:80–94 Lowry D, Xie Y (2009) Socioeconomic status and health differentials in China: convergence or divergence at older ages? Research report 09-690. Population Studies Center, University of Michigan Marang-van de Mheen PJ, Smith GD, Hart CL (1999) The health impact of smoking in manual and non-manual social class men and women: a test of the Blaxter hypothesis. Soc Sci Med 48(12):1851–1856 Michael M, Feng X, Wang Q (2008) The status decide your health: the states syndrome: how social standing affects our health and longevity. Renmin University of China Press (in Chinese) Prus SG (2007) Age, SES, and health: a population level analysis of health inequalities over the life course. Sociol Health Illn 29:2 Qi L, Wang C (2010) Health and socioeconomic status: a research based on multiple indicators. Chin Health Econ 29(8):47–50 (in Chinese) Robert S, House JS (1996) SES differentials in health by age and alternative indicators of SES. J Aging Health 8(3):359–388 Ross CE, Mirowsky J (2010) Why education is the key to socioeconomic differentials in health. In: Handbook of medical sociology, pp 33–51

2 Trends in the Impact of Socioeconomic Status on Health …

35

Ross CE, Wu CL (1996) Education, age, and the cumulative advantage in health. J Health Soc Behav 104–120 Shen K (2008) The impact of childhood SES on the mortality risk of China’s oldest old. Chin J Popul Sci 03:56–63+96 (in Chinese) Wang F (2012) Socioeconomic status and health inequity. Chin J Sociol 32(02):125–143 (in Chinese) West P (1991) Rethinking the health selection explanation for health inequalities. Soc Sci Med 32(4):373–384 Williams DR (1990) Socioeconomic differentials in health: a review and redirection. Soc Psychol Q 53(2):81–99 Winkleby MA, Fortmann SP, Barrett DC (1990) Social class disparities in risk factors for disease: eight-year prevalence patterns by level of education. Prev Med 19(1):1–12 World Health Organization (1947) Constitution of the World Health Organization: signed at the international health conference, New York, 22 July 1946. World Health Organization, Interim Commission Xie Y et al (2013) China family panel studies 2013. Peking University Press (in Chinese) Zeng Y, Gu D, Land KC (2007) The association of childhood socioeconomic conditions with healthy longevity at the oldest-old ages in China. Demography 44(3):497–518 Zhu H, Xie Y (2007) Socioeconomic differentials in mortality among the oldest old in China. Res Aging 29(2):125–143

Chapter 3

The Age, Gender, Urban-Rural and Regional Differences in Dynamic Changes of Activity of Daily Living Among the Chinese Oldest-Old Wenjuan Zhang and Meng Wei

3.1 Introduction According to the population projection by researchers in China and the United Nations, the population of oldest-old in China, those aged 80+, will increase from 11.5 million in 2000 to 27 million in 2020 and over 100 million in 2050. The average annual growth rate of this age group in China in the first half of the twenty-first century is 4.4%, which is twice the growth rate of the elderly aged 65 and over and 6.1 times the growth rate of the total population. The rapid growth of the oldest-old population has posed severe challenges to China’s health expenditure, old-age security and social services for the elderly. On the one hand, with improved healthcare and increasing investment in health resources in China, patients with chronic diseases have reduced morbidity risks but are now facing the severe challenges of disability. As a result, the declining mortality rate has increased, rather than lowered, the burden of disease, as the oldest-old adults have declining organ functions and often suffer from one or more diseases. On the other hand, the oldest-old individuals are financially vulnerable with a lower ability to provide for themselves. As their population grows, it requires more financial support to the oldest-old adults, which is undoubtedly a challenge for the current social security system. In addition, the elderly age group comprise distinct categories. One prominent difference in health is that most of the elderly under the age of 80 are able to take care of themselves, while the elderly aged 80 and over have the highest probability of living with chronic diseases or even being bedridden, who are in most needs for care, thus putting forward higher requirements for social services. Therefore, understanding the trends of the health status of the oldest-old population is of great practical significance for the rational W. Zhang (B) School of Sociology and Population Studies, Renmin University of China, Beijing, China e-mail: [email protected] M. Wei National Institute of Social Development, Chinese Academy of Social Sciences, Beijing, China © Science Press 2022 Y. Zeng et al. (eds.), Trends and Determinants of Healthy Aging in China, https://doi.org/10.1007/978-981-19-4154-2_3

37

38

W. Zhang and M. Wei

utilization and enhancement of elderly service resources, optimizing the allocation of healthcare resources, and improving old-age security. Independence in Activities of Daily Living (ADL) is a measure designed to objectively evaluate an individual’s capacity to engage in activities of daily living, which is commonly used to assess a person’s overall health and functional status. In this study, we used the Katz ADL Index (Katz et al. 1963), which includes six activities: feeding, continence, transferring, going to the toilet, dressing, and bathing. In this chapter, an elderly individual would be defined as disabled if he/she required assistance in one of the six activities. If a respondent is capable of performing all of the six activities, he/she would be defined as independent; otherwise, he/she would be defined as disabled. We used 2002 and 20111 data of Chinese Longitudinal Healthy Longevity Survey (CLHLS) collected by the Center for Healthy Aging and Development Studies (CHADS), Peking University, which covered 22 provinces and provincial-level municipalities. The CLHLS 2002 sample included 11,175 oldest-old individuals aged 80 years and over, and the CLHLS 2011–2012 follow-up survey tracked 6530 oldest-old adults. Data included family structure, marital status, independence in activities of daily living, living habits, social activities, diseases and health status, providing rich and valuable information for studying the health and trends among the oldest-old in China.

3.2 Basic Information of Respondents Tables 3.1 and 3.2 presented profiles of CLHLS 2002 and CLHLS 2011 respondents. Overall, the oldest-old individuals in both surveys were predominantly female and widowed. It was more common for respondents to live with their offspring, with younger family members such as children playing important roles as family caregivers. More oldest-old respondents lived in rural areas or eastern and central regions. Respondents were better off financially, despite lower education level.

3.3 Results of the Analyses This section presents a descriptive statistical analysis of disability in the activities of daily living (ADL) of the oldest-old respondents in CLHLS 2002 and 2011–2012 samples by ADL tasks, age, gender, urban/rural residence and by region, makes a comparison between 2002 and 2011–2012 waves.

1

CLHLS 2011–2012 Survey added Hainan Province (Chengmai County), covering in total 23 provinces.

17.46

46.83

East

Without spouse

82.54

Region

Marital status

With spouse

20.81

38.51

Central

12.72

14.66

West

28.54

80.35

Living with household members

46.86

13.99

53.14

Rural

67.2

5.66

Insufficient 19.79

80.17

8.38

7+

Sufficient

Financial status

24.42

1=6

Education level 0

Living in institutions

Residence Urban

Living alone

60.73

F

Living arrangements

39.27

19.04

18.89

100+

Sex 95–99

M

90–94

85–89

Age

80–84

Source Calculation based on Chinese Longitudinal Healthy Longevity Survey (CLHLS) 2002 data collected by Peking University

%

Profile

%

Profile

Table 3.1 Profile of CLHLS 2002 respondents (%)

3 The Age, Gender, Urban-Rural and Regional Differences … 39

22.68

47.08

East

Without spouse

77.32

Region

Marital status

With spouse

22.34

39.89

Central

14.92

13.03

West

22.31

79.52

Living with household members

47.09

17.72

52.91

Rural

68.33

2.77

Insufficient 20.1

79.9

6.98

7+

Sufficient

Financial status

24.69

1–6

Education level 0

Living in institutions

Residence Urban

Living alone

59.88

F

Living arrangements

40.12

20.14

20.29

100+

Sex 95–99

M

90–94

85–89

Age

80–84

Source Calculation based on Chinese Longitudinal Healthy Longevity Survey (CLHLS) 2011 data collected by Peking University

%

Profile

%

Profiles

Table 3.2 Profile of CLHLS 2011 respondents (%)

40 W. Zhang and M. Wei

3 The Age, Gender, Urban-Rural and Regional Differences …

41

3.3.1 Trends in the Oldest-Old by Each Basic Activity of Daily Living In CLHLS 2002 data, the proportions for ADL disability in the sample of the oldestold were 36.5% bathing, 17.7% dressing, 19.1% toileting, 15.4% transferring, 9.8% continence, and 11.2% feeding. 41.1% of respondents were disabled in at least one ADL task. In CLHLS 2011 data, the proportions for ADL disability were 32.8% bathing, 18.1% dressing, 18.7% toileting, 16.2% transferring, 9.1% continence, and 12.2% feeding. 36.3% of respondents were disabled in at least one ADL task (see Table 3.3 for details). Figure 3.1 shows the histogram of disability in basic activities of daily living among the CLHLS 2002 and 2011 oldest-old respondents. We can see that in both 2002 and 2011 data, the proportions of ADL disability increased as the ADL task gets more complex from continence, feeding, transferring, dressing, toileting to bathing. Compared with 2002, the proportion for disability in bathing decreased significantly in 2011; the proportions for disability in toileting and continence decreased slightly, and the proportions for disability in dressing, transferring, and feeding increased slightly. Among them, the proportion for disability in bathing was the highest, and its change was the largest. From the data, it can be inferred that improvement in living environment and convenience facilities has a significant contribution to the capacity of the oldest-old to perform bathing independently. Its effect in reducing disability in other activities of daily living that require higher physical fitness, such as dressing, transferring and feeding, is not obvious. Dependence level in basic activities of daily living reflects disability in performing one or more ADL tasks. The more ADL tasks which cannot be accomplished independently, the higher the dependence level. As shown in Fig. 3.2, in 2002, 16.81% of oldest-old respondents in the sample had difficulty in one ADL task, 6.19% in two ADL tasks, and 18.01% in three and more ADL tasks. Furthermore, 59% were able to accomplish independently all six ADL items. In 2011, 13.34% of oldest-old respondents in the sample had difficulty in one ADL task, 4.54% in two ADL tasks, and 17.92% in three and more ADL tasks. In the 2011 sample, the proportions of oldest-old respondents who had difficulty in one, two, three or more ADL tasks were all lower than those in 2002, and the proportion of oldest-old adults with no difficulty in all ADL tasks (64.2%) was 5.2 percentage points higher than that in 2002. Among oldest-old respondents who were disabled in one or more ADL tasks, in 2002 data, 41% were disabled in one ADL task, 15% in two, 11% in three, 11% in four, 12% in five, and 10% in all six ADL tasks. In the 2011 sample, 37% were disabled in one ADL task, 12% in two, 7% in three, 11% in four, 17% in five, and 13% in all six ADL tasks. Disability in one ADL task was most prevalent though finding respondents dependent on more than one ADL task was relatively common. Compared with 2002, the proportions of oldest-old respondents with disability in one, two or three ADL tasks decreased in 2011, and the proportions with disability in five or six ADL tasks increased (as shown in Fig. 3.3). It can be seen that in 2011, the proportions of respondents who had mild to moderate disability in ADL tasks

Toileting

Transferring

29.5

40.9

28.0

39.4

57.8

31.4

Aged 85–89

Aged 90–94

Aged 95+

Male

36.3

33.1

38.0

44.8

37.4

38.6

Rural

Eastern region

Central region

Western region

36.5

34.3

32.9

40.2

33.2

40.3

42.4

27.4

52.3

35.2

24.4

15.8

32.8

30.2

29.9

36.0

29.2

36.8

37.2

26.3

50.0

31.9

21.1

14.0

16.0

15.5

20.0

16.7

18.8

20.9

12.7

27.6

15.4

10.3

5.9

17.7

15.8

17.2

19.5

15.7

20.8

21.0

13.8

29.8

16.3

9.7

7.0

18.1

16.4

18.2

20.8

18.0

20.4

23.2

12.8

29.7

17.2

11.0

6.4

19.1

13.4

18.3

20.6

16.8

20.8

22.3

13.4

30.9

17.4

9.6

6.9

18.7

13.1

14.3

16.9

14.4

16.5

18.8

10.0

24.1

13.7

8.7

5.0

15.4

12.9

16.2

17.2

14.7

18.0

19.3

11.7

26.9

15.5

7.9

5.7

9.8

9.5

8.3

11.0

8.5

11.2

11.6

6.9

14.0

9.1

6.2

4.8

9.1

7.9

8.4

10.1

7.7

10.7

9.8

8.1

13.6

8.8

5.8

4.6

8.9

10.8

12.3

11.3

11.1

13.8

7.2

17.9

9.2

6.4

3.4

11.2

9.1

12.4

12.8

11.1

13.4

14.4

8.8

20.1

11.3

6.4

4.2

12.2

Source Calculation based on Chinese Longitudinal Healthy Longevity Survey (CLHLS) 2002 and 2011–2012 waves datasets. Same source for data used in other tables and graphs in this chapter

34.5

33.7

39.0

39.9

47.3

44.5

Female

Urban

54.37

35.1

24.4

16.1

41.1

19.4

Whole

Aged 80–84

Feeding

2011–2012 2002 2011–2012

Continence

16.2

Dressing 2011–2012 2002

Bathing

2002 2011–2012 2002 2011–2012 2002 2011–2012 2002 2011–2012 2002

Category Total

Table 3.3 Disability in activities of daily living (ADLs) among the oldest-old respondents in CLHLS 2002 and CLHLS 2011–2012 (%)

42 W. Zhang and M. Wei

3 The Age, Gender, Urban-Rural and Regional Differences …

43

Percentage 40

bathing dressing going to the toilet transferring continence feeding

35 30 25 20 15 10 5 0 2002

2011

Year

Fig. 3.1 Disability in basic activities of daily living among the oldest-old respondents Percentage 70 need no assistance

60

had difficulty in one task had difficulty in two tasks

50

had difficulty in three tasks 40

had difficulty in four tasks had difficulty in five tasks

30

had difficulty in six tasks

20 10 0 2002

2011

Year

Fig. 3.2 Dependence level in basic activities of daily living among the oldest-old respondents

decreased, but the proportion with severe ADL disability increased. The most crucial issue is to provide care for this oldest-old category who had severe disability.

3.3.2 Differences and Trends in ADL Performance by Age Groups Among Oldest-Old Respondents Among the oldest-old respondents, the ratio of disability went up with advancing age across all six ADL tasks. In 2002, the proportions for ADL disability were

44

W. Zhang and M. Wei

2002

2011-2012

6 items 5 items 10% 12%

1 item 41%

4 items 11%

3 items 11%

6 items 13% 5 items 17%

4 items 11%

2 items 15%

1 item 37%

2 items 12%

3 items 7%

Fig. 3.3 Proportions of oldest-old respondents who cannot accomplish one to six activities of daily living (ADL) tasks

19.4% in the 80–84 age group, 28% in 85–89, 39.4% in 90–94, and 57.8% in 95 and over. In 2011 data, the proportions for ADL disability were 16.1% in the 80–84 age group, 24.4% in the 85–89 age group, 35.1% in 90–94 and 54.37% in 95 and over. As shown in Fig. 3.4, the ratio of ADL disability in all age groups among oldest-old respondents decreased in 2011. For one thing, this trend may be partly due to the increased awareness of health and improved health status of the oldest-old in recent years, consistent with the pattern of disability compression for the elderly. And for another, it may also be due to the improvement of the living environment and the widespread use of convenience facilities, which allowed older adults to accomplish ADL tasks independently with the same physical conditions. It remains to be investigated which factor plays a leading role. Percentage 70 60 50 40 in 2002

30 20

in 2011

10 0 80-84

85-89

90-94

95+

Age group

Fig. 3.4 Activities of daily living (ADL) disability by age groups among the oldest-old respondents

3 The Age, Gender, Urban-Rural and Regional Differences …

45

Percentage 70 60 50 40 male in 2002 30

female in 2002

20

male in 2011

10

female in 2011

0 80-84

85-89

90-94

95+

Age group

Fig. 3.5 Trends in ADL disability among oldest-old respondents by age and gender

3.3.3 Differences in ADL Performance by Gender Among Oldest-Old Respondents Figure 3.5 shows the age-specific trends in ADL disability among oldest-old respondents by gender in 2002 and 2011. In the 2002 oldest-old sample, the rate of basic ADL disabilities among male adults was 31.4%, and the rate among female adults was 47.3%, while in 2011, the rate was 29.5% among men and 40.9% among women. The disability rate of women was higher than that of men in all age groups among the oldest-old respondents in 2002 and 2011, indicating that women, though with higher life expectancy, have worse overall health than men. In 2011, there was a slight decrease in the ADL disability rate of men and a larger decrease in the disability rate of women. Thus, women’s overall health improved compared to the 2002 sample. In general, the ADL disability rates of both men and women conformed to the pattern of disability compression for the elderly, consistent with the overall ADL disability trend in the whole older population. The decrease in ADL disability rate among the overall older population was mainly due to the decline in disability rate among women.

3.3.4 Differences and Trends in ADL Performance by Urban/Rural Residence Among Oldest-Old Respondents In 2002, the rates of basic ADL disabilities were 44.5% among oldest-old urban respondents and 38.0% among respondents living in rural areas. In 2011, the rates decreased to 39.9% in the urban area and 33.1% in the rural area. Figure 3.6 shows the

46

W. Zhang and M. Wei

Percentage 70 60 50 40 30

urban in 2002 rural in 2002 urban in 2011 rural in 2011

20 10 0 80-84

85-89

90-94

95+ Age group

Fig. 3.6 Trends in activities of daily living (ADL) disability among oldest-old respondents by age and urban/rural residence

age-specific trends in ADL disability in 2002 and 2011 by urban/rural residence. The rates of disabilities were higher in all age groups for urban respondents than their rural counterparts. The reasons may be that, on the one hand, older people in rural areas have better self-care capacity because they are more likely to engage in agricultural activities. On the other hand, oldest-old adults with urban residence usually are able to live with disabilities due to better healthcare facilities. Rural oldest-old individuals and those in the central and western regions have a lower probability of living with disabilities due to untimely detection of diseases and inadequate treatment. In other words, due to the selective role of death, oldest-old adults in the rural area who live longer have better overall health. Compared with 2002, the rates of basic ADL disabilities in 2011 were lower among both urban and rural respondents, in line with the pattern of disability compression. On the one hand, this may attribute to expansion of healthcare coverage, increasing pension income, and improved health awareness, which has led to a reduced probability of disability in ADL and better overall health than before among the oldest-old respondents. On the other hand, it may be due to the improvement of the living environment and convenient facilities. Compared with the respondents in urban areas, the rate of basic ADL disability among rural ones declined more slowly, indicating that there is still a gap in healthcare conditions and convenient facilities between rural and urban areas.

3 The Age, Gender, Urban-Rural and Regional Differences …

47

3.3.5 Differences and Trends in ADL Performance by Region (East/Central/West) Among Oldest-Old Respondents In terms of regional differences, respondents in the eastern region have the highest proportion of ADL disability, followed by the central region and then the western region, across all ADL tasks. In 2002, the proportion of ADL disability was 44.8% among the oldest-old respondents living in the eastern region, 37.4% in the central region, and 38.6% in the western region which only includes Chongqing, Sichuan, and Shaanxi. In 2011, the proportions of ADL disability were 39.0% in the eastern region, 33.7% in the central region, and 34.5% in the western region, which were all lower than those in 2002. Figure 3.7 reflects the age group-specific trends of ADL disability by region of residence in 2002 and 2011. Overall, the rates of basic ADL disabilities among all age groups, except for 80–84 years old, were higher in the eastern region than in the central and western regions. In 2011, the rates of ADL disability decreased in all regions compared to 2002, in line with the pattern of disability compression, except for an increase in the disability rate of the 80–84 age group in the western region. This result is attributed to the increasing health awareness among the oldest-old respondents as well as the improvement of the living environment and widespread use of convenient facilities. The reasons for the higher ADL disability rate in the eastern region may include the following. First, the “rich men’s diseases” such as diabetes and hyperlipidemia are spreading fast due to poor living habits. Pollution in the urban environment and excessive work pressure in early life also contributed to the result. Second, with an Percentage 70 eastern region in 2002 60

central region in 2002 western region in 2002

50

eastern region in 2011 central region in 2011

40

western region in 2011

30 20 10 0 80-84

85-89

90-94

95+ Age group

Fig. 3.7 Trends in activities of daily living (ADL) disability among oldest-old respondents by age and region of residence

48

W. Zhang and M. Wei

improved allocation of healthcare resources and healthcare security in the eastern region, the oldest-old respondents are more likely to survive severe illness and live with disabilities. The reasons for the lower ADL disability rate in the central and western regions include a good natural environment and healthy living habits. Due to the selective role of death, the oldest-old respondents in the central and western regions who manage to live longer are themselves in better overall health than peers.

3.4 Conclusion Our study shows that between 2002 and 2011, due to socioeconomic development and advances in healthcare technology, the rates of basic ADL disabilities among the oldest-old respondents have significantly decreased as they live longer. Our data analysis is summarized as follows. 1. The disability rate among urban oldest-old people measured by the presence of at least one ADL disability is higher than that in rural areas, and the rate of ADL disability in eastern regions is higher than that in central and western regions. To a large extent, this is due to better allocation of health resources and healthcare security in the urban area and eastern regions, which helps the oldest-old individuals to live with disabilities. At present, there are considerable gaps in economic and social development as well as the usage and investment of healthcare resources between urban and rural areas and between the eastern, central and western regions in China. Priorities are given to urban areas over rural areas and to the east over the central and western regions, leading to a wide gap in health resources between them. Although the situation has improved in recent years, it is still difficult to meet the basic needs for healthcare services in rural areas and central and western regions. The health status of people there is not optimistic. Therefore, we need to adjust health policies, change the unbalanced and unreasonable health resource distribution, increase research on rational allocation of health resources, strengthen investment in health resources in rural area and in central and western regions, improve the efficiency of health resources investment in rural and in central and western regions, and close the gap in health human resources. At the same time, it is necessary to vigorously strengthen the construction of grassroots health institutions, give full play to the role of community service platforms, establish a strict and effective healthcare regulatory system, and build a better healthcare system instead of a purely curative model. 2. The rate of basic ADL disabilities among female oldest-old respondents were significantly higher than that of male ones. Due to their higher rate of widowhood and poorer overall health, women have a greater need for care by family or social services. But at the same time, women’s ability to pay for social care is more limited due to poorer financial situations. Therefore, it is necessary for the

3 The Age, Gender, Urban-Rural and Regional Differences …

49

government to assist this vulnerable group by offering subsidies for the oldestold women to stay in institutions or purchasing community services to meet their care needs. 3. The decrease in the rates of basic ADL disabilities among the oldest-old respondents is largely due to the improvement of the living environment, as more and better convenience facilities for the elderly can relatively improve their capacity for performing ADL tasks. Among them, the improvement of the living environment has the most significant effect on reducing the disability rate in bathing and toileting. Within the oldest-old respondents with disabilities, the disability rate in bathing is the highest, and its decline is the fastest in the past 10 years. It is recommended to strengthen investment in research and development to develop facilities that assist the elderly to take care of daily routines independently, especially facilities to address bathing and toileting difficulties, for example, making bathroom facilities simpler, installing height-adjustable water sinks and toilets, or developing tools to help the elderly take a bath. After solving bathing and toileting difficulties, we believe that the rates of basic ADL disabilities will continue to decrease, helping to save on care costs. 4. Compared with 2002, the proportions of oldest-old respondents with mild and moderate ADL disabilities decreased in 2011. Still, the proportion with severe disability increased, which means that more oldest-old adults need high intensity and high level of long-term care (such as multiple cares, longer average daily care time, and high-quality caregivers). The care burden on the family is bound to increase. Therefore, it is necessary to improve the community-based home care and institutional care, to realize the collaboration and coordination among home, community and institutions in caring for the elderly, to improve the quality of caregivers, and to reduce the care burden of family members while providing adequate and quality care for the oldest-old adults. In addition, long-term care insurance should be developed at the right time to address the burden of long-term care costs for the elderly themselves, their families and society. In conclusion, data analysis in this chapter shows that in the past 10 years, the rates of basic ADL disabilities among all age groups in the oldest-old respondents have decreased for both men and women, in urban and rural areas, and in the eastern, central, and western regions in China. In some way, it proves the theory of disability compression. This outcome is due to the improvement in some aspects of the overall health of the elderly, driven by a better living environment and convenience facilities as a result of social and economic development and rising standards of living. However, more oldest-old people live with illness thanks to the improvement of medical technology, which may have a negative effect on the average level of other health status indicators (such as physical and cognitive functions). This issue needs to be given great attention by society and the government. Therefore, it is not advisable to be overly optimistic about the overall decline in the rates of basic ADL disabilities among the oldest-old respondents in China in the last decade. This topic will be further analyzed and discussed in this chapter.

50

W. Zhang and M. Wei

Acknowledgements The study was sponsored by the National Natural Science Foundation of China (71110107025, 71233001, 71490732).

Reference Katz S, Ford AB, Moskowitz RW, Jackson BA, Jaffe MW (1963) Studies of illness in the aged: the index of ADL: a standardized measure of biological and psychosocial function. J Am Med Assoc 185(12):914–919

Chapter 4

Trends of Dynamic Changes in Activities of Daily Living, Physical Performance, Cognitive Function and Mortality Rates Among the Oldest-Old in China Yi Zeng and Qiushi Feng

4.1 Introduction Population aging is one of the major challenges facing most countries in the world, including China. The accompanied dramatic increase in numbers of oldest-old (individuals older than 80 years) is of particular concern, presenting a major challenge for health and social care systems, because the oldest-old often need daily assistance and medical care (Zeng and George 2010). Two contrasting scenarios of health trends in aging populations have been proposed. One view states that advances in medical technology, improvements in lifestyle and socioeconomic development will postpone the onset of disability and chronic diseases among the elderly, so that morbidity will be compressed in old age (Fries 1980; Christensen et al. 2009; Vaupel 2010). This concept is linked to the benefits of success—i.e., that people are living longer (success) and in better health at older ages than they were previously (benefits). By contrast, in the alternative scenario, reduced mortality is hypothesised to result in an increased number of frail elderly surviving with health problems, thus worsening the overall health of the elderly population. This concept is often referred to as expansion of morbidity (Gruenberg 1977; Waidmann et al. 1995), closely linked to that of costs of success, which specifically means that people’s lifespans are lengthening (success) but with worse health at older ages than previously (costs). In reality, these two trends

This chapter is done by Yi Zeng and Qiushi Feng, based on the following paper: Zeng et al. (2017). Y. Zeng (B) National School of Development, Peking University, Beijing, China e-mail: [email protected] Center for Studies of Aging and Human Development, Duke University, Durham, NC, USA Q. Feng Department of Sociology and Anthropology, National University of Singapore, AS1 04-30, 11 Arts Link, 117570 Singapore, Singapore © Science Press 2022 Y. Zeng et al. (eds.), Trends and Determinants of Healthy Aging in China, https://doi.org/10.1007/978-981-19-4154-2_4

51

52

Y. Zeng and Q. Feng

might coexist and interplay (Manton 1982), and the concept of dynamic equilibrium has been introduced to help understand the association between morbidity and increasing life expectancy (Robine and Michel 2004). Trends in the overall health status of the elderly are generally positive in highincome societies (Freedman et al. 2002). However, several reports support the opposite trends for some major health indicators. For example, findings from a Swedish study showed that the objective function tests of physical capacity, lung function, and cognition were significantly worse in 2002 compared with 1992 in individuals older than 77 years (Parker et al. 2005). Although dementia incidence has fallen in some European countries (Matthews et al. 2016) and the USA (Satizabal et al. 2016), findings from nine large Japanese studies have suggested that prevalence of all-cause dementia and Alzheimer’s disease are increasing in Japan (Dodge et al. 2012). Investigators building on the work of several studies (including two nationally representative surveys) reported opposing trends of improvement in disability measures, alongside an expansion of morbidity in chronic diseases and functional impairments, among Swedish oldest-old (Parker et al. 2005; Parker and Thorslund 2007). Several studies have reported that the prevalence of disability according to activities of daily living among Chinese elderly has decreased in the past two decades (Liang et al. 2015; Martin et al. 2014). However, Wu and colleagues (2014) concluded that dementia prevalence among elderly individuals aged 70 years or older was generally increasing, on the basis of an evaluation of 70 prevalence studies of dementia in Mainland China, Hong Kong and Taiwan from 1980 to 2012. Similarly, Chan and colleagues (2013) reported that the prevalence of all forms of dementia at ages 65– 69 and 95–99 years in China in 2010 had increased 44.4% and 43.7%, respectively, compared to 1990. The existing scientific literature provides empirical support for both compression and expansion of morbidity, but little research so far has investigated the mixed effects of these two opposing trends in a single study with a sufficient sample size of the oldest-old. The exception is a Danish study of a cohort born in 1905 and assessed at age 93 in 1998, compared with a later cohort born in 1915 and assessed at age 95 in 2010 (Christensen et al. 2013). This study provided some support for the mixed effects of both compression of morbidity and expansion of morbidity. However, whether these mixed effects also exist among the oldest-old in low-income or middle-income countries such as China is unclear. We aimed to address this research question by comparatively analysing cohorts of the oldest-old born in 1909–18 versus 1919–28 (aged 80–89 years in 1998 vs. 2008), born in 1899–1908 versus 1909–18 (aged 90– 99 years in 1998 vs. 2008), and born in 1893–98 versus 1903–08 (aged 100–105 years in 1998 vs. 2008). To our knowledge, this study is the first to assess this important issue in a low-income or middle-income country, and uses the largest dataset of oldest-old cohorts in the world.

4 Trends of Dynamic Changes in Activities of Daily …

53

4.2 Methods 4.2.1 Study Design and Participants This study draws on data from the oldest-old participants (i.e. those aged 80– 105 years) from the 1998 and 2008 waves of the Chinese Longitudinal Healthy Longevity surveys (CLHLS). The CLHLS is a nationwide survey done in a randomly selected half of the counties and cities in 22 of the 31 provinces, covering about 85% of the total population of China. The CLHLS attempted to interview all centenarians who voluntarily agreed to participate in the study in the sampled counties and cities. The CLHLS also adopted a targeted random-sample design to ensure representativeness, through interviews with approximately equal numbers of male and female nonagenarians, octogenarians and young-old (aged 65–79 years) living near the centenarians (i.e., in the same village or street, if available, or in the same sampled county or city). This design serves well our aim of investigating determinants of healthy longevity of different age and sex groups who live in the same social and natural environment (Zeng 2012). The Research Ethics Committees of Peking University and Duke University granted approval for the Protection of Human Subjects for the Chinese Longitudinal Healthy Longevity Survey, including collection of the data used for present study. The survey respondents gave informed consent before participation.

4.2.2 Procedures and Variables The CLHLS was initially designed to facilitate international comparative analyses, and its questionnaire was translated from the instruments of the Danish longevity survey analysed by Christensen and colleagues (2013). The instruments were adapted to the Chinese culture and socioeconomic context. A wide variety of international and domestic studies have confirmed that age reporting of the Han Chinese oldest-old is in general reasonably accurate, due to the cultural tradition of memorising one’s date of birth to determine dates of important life events such as engagement and marriage (Coale and Li 1991; Wang et al. 1998). The CLHLS 1998 and 2008 surveys used almost exactly the same ascertainment and assessment protocols. No proxy was used for objective questions such as assessment of cognitive function and physical performance. The survey was administered in the participants’ homes by trained interviewers from the local centres for disease prevention and control for university students. 1. Mean annual death rates: As described above, information on date of death were collected for the interviewees who were interviewed in 1998 or 2008, but died in the inter-wave period 1998–2000 or 2008–2011. We estimated the age-sexspecific weighted mean annual death rate for each of the cohorts by dividing the

54

Y. Zeng and Q. Feng

weighted total number of deaths among the cohort members in the follow-up period by the weighted total number of person-years lived by all of the cohort members (including those who survived and died). 2. Activities of daily living (ADL) disability: The ADL functional status of six daily activities of eating, dressing, transferring, using the toilet, bathing, and continence were used to measure the elders’ status of independence in daily living. A score of 0 was given to participants needing assistance with the activity, a score of 1 if no help was needed, resulting in a range of 0–6 for the ADL scores. ADL is a good measurement of functional capacity and a proxy of health status widely used in healthy aging studies (Katz et al. 1983; Spitzer 1987; Wiener et al. 1990). In this study, we follow the ADL capacity group classification widely adopted in the other studies on oldest-old (Christensen et al. 2013): if none or one of the six ADL activities is impaired, the oldest-old is classified as “normal”; if two activities are impaired, the oldest-old is classified as “moderately disabled”; “severely disabled” refers to those elders who have three or more activities impaired. 3. Physical performance in three tests: Self-reported subjective measures of disability in activities of daily living have been criticised for their potential to be affected by both differences in availability of associated facilities and perceptions of the participants. The objective performance-based tests are highly recommended as complementary measures in examining physical functions (Daltroy et al. 1995; Melzer et al. 2004; Reuben et al. 2004). In the Chinese elderly population, the objective performance-based tests have been recently valued as important complementary measures for routinely-used ADL, which help clarify the intrinsic physiological impairment of the elderly and environmental barriers of their daily activities (Feng et al. 2010; Purser et al. 2012). Three objective physical performance tests were administrated in the CLHLS surveys. The first task asked the respondent to stand from a chair. This test has three levels of outcomes, i.e. “can without using arms” (coded as 1), “can using arms” (coded as 0.5), and “cannot” (coded as 0). The second task is to pick up a book from the floor, and respondents are “can while standing” (coded as 1), “can while sitting” (coded as 0.5), and “cannot” (coded as 0). The last task is to test whether the respondent is able to turn around 360° without help (yes vs. no, coded as 1 or 0). 4. Cognitive function measured by Mini-Mental State Examination (MMSE): The MMSE, a global assessment test of cognitive function (Folstein et al. 1975; Christensen et al. 2013), was adapted to the Chinese cultural context and was carefully tested in the pilot survey (Zeng and Vaupel 2002). The testing protocol includes 24 items regarding orientation, registration, attention, calculation, recall and language, with a total score ranging from 0 to 30. Following the practice widely adopted in the other studies (Shen 2010), we use the MMSE cutoffs to define cognitive function as: severe impairment (0–17), mild impairment (18– 22), normal (23–27) and maximum (28–30). Note that a zero score was given to those items to which the interviewee was not able to answer or perform the test, purely due to his or her mental or physical impairment (rather than not willing to answer or perform the test), and no proxy was allowed in performing the MMSE tests.

4 Trends of Dynamic Changes in Activities of Daily …

55

The missing values for all the variables analysed in this article were mostly less than 1%. Only among the centenarians, a few variables such as MMSE score and physical performances had relative higher missing rates of 2–3%. Due to such low missing rates, we did not impute the missing values. In the statistical analyses, we deleted the cases with missing values, and the results have no significant difference compared to those with imputation.

4.2.3 Statistical Analyses We divided individuals into oldest-old born in 1909–18 versus 1919–28 (aged 80– 89 years in 1998 vs. 2008), born in 1899–1908 versus 1909–18 (aged 90–99 years in 1998–2008), and born in 1893–98 versus 1903–08 (aged 100–105 years in 1998 vs. 2008). We compare the following three pairs of cohorts of octogenarians, nonagenarians and centenarians, and the two cohorts in each pair of the comparison born ten years apart, with the same age at the time of the assessment in the CLHLS 1998 and 2008 surveys: • Octogenarians: comparison between the cohort born in 1909–1918 (assessed at ages 80–89 with mean age 83.1 in 1998 survey, n = 3235) and the cohort born in 1919–1928 (assessed at ages 80–89 with mean age 83.0 in 2008 survey, n = 4053). • Nonagenarians: comparison between the cohort born in 1899–1908 (assessed at ages 90–99 with mean age 92.1 in the 1998 survey, n = 2896) and the cohort born in 1909–1918 (assessed at ages 90–99 with mean age 92.2 in the 2008 survey, n = 4338). • Centenarians: comparison between the cohort born in 1893–1898 (assessed at ages 100–105 with mean age 101.1 in 1998 survey, n = 2197) and the cohort born in 1903–1908 (assessed at ages 100–105 with mean age 101.7 in 2008 survey, n = 2809). We compared annual mortality, self-reported disability according to the activities of daily living scale, physical performance in three tests, and cognitive function measured by Mini Mental State Examination scores for men and women separately and for both sexes combined. We did standard statistical χ2 tests (one-sided) or Z tests (two-sided) for categorical data, and t tests (two-sided) for continuous data. We also did multivariate regression analyses to explore the changes in mortality, physical function, and cognitive function between the oldest-old cohorts born 10 years apart, adjusted for the covariates of age, rural or urban residence, marital status, and education, which are the major demographic and socioeconomic factors affecting the mortality and health of elderly people in China. We based the mortality analysis on parametric survival models with Weibull distribution, while the Weibull assumption was satisfied. All other analyses were based on logistic regression models or linear regression models. We used STATA version 13.1 for the statistical analyses.

56

Y. Zeng and Q. Feng

4.2.4 Role of Funding Source The study funders provided financial support for data collection and analysis, but had no role in the writing of the report, interpretation of the results, or submission for consideration of publication. The corresponding author had full access to all the data in the study and had final responsibility for the decision to submit for publication.

4.3 Results We included 19,528 individuals in our study, comprising 7288 octogenarians, 7234 nonagenarians, and 5006 centenarians, interviewed in 1998 and 2008 (in view of the very high mortality at advanced ages, only 2.8% of the oldest-old participants were interviewed in both 1998 and 2008 surveys). Table 4.1 presents the basic demographic characteristics of the cohorts. Tables 4.2, 4.3 and 4.4 present the detailed results of cross-cohort changes in physical and cognitive function and death during follow-up for men, women, and both sexes combined. Figures 4.1, 4.2 and Table 4.5 present the summary results. Age-specific and sex-specific mortality among Chinese oldest-old aged 80–89, 90–99, and 100–105 years were all reduced in the later cohorts compared with the cohorts born 10 years earlier (Fig. 4.1; Tables 4.2, 4.3 and 4.4). All of the nine sets of comparisons of age-specific mortality between the different cohorts of the oldest-old showed reductions of −0.2 to −1.3% in annual mortality during followup (Table 4.5). Adjusted for covariates of age, sex, education, and rural or urban residence, the cross-cohort reduction in age-specific and sex-specific mortality was statistically significant in sex-combined centenarians (p = 0.0032) and female centenarians (p = 0.0163), and not significant in sex-combined octogenarians and nonagenarians, male or female octogenarians, male or female nonagenarians, or male centenarians (Tables 4.2, 4.3 and 4.4). Disability as measured through activities of daily living of the Chinese oldestold was significantly reduced in the later cohorts compared with the earlier cohorts (Fig. 4.1; Tables 4.2, 4.3 and 4.4). All of the nine sets of comparisons between different cohorts of the oldest-old showed substantial reductions in annual rates of disability, ranging from −0.8 to −2.8% (Table 4.5). Adjusted for the covariates, the cross-cohort reductions in the mean score of activities of daily living disability were statistically significant (p < 0.0001) for nonagenarians and centenarians (both sexes combined) and female nonagenarians and centenarians, significant (ranging from p = 0.0023 to p = 0.0082) in sex-combined octogenarians and male nonagenarians, significant (ranging from p = 0.0257 to p = 0.0290) in female and male octogenarians, and not significant in male centenarians (p = 0.0604; Tables 4.2, 4.3 and 4.4). The scores in objective physical performance tests (standing up from a chair, picking up a book from the floor, and turning-around 360° among the Chinese

4 Trends of Dynamic Changes in Activities of Daily …

57

Table 4.1 Comparisons of the basic demographic characteristics of cohorts of the oldest-old born ten years apart with the same ages when they were assessed in 1998 or 2008 Years of birth

Years of birth

Years of birth

1909–1918 (n = 3235)

1919–1928 (n = 4053)

1899–1908 (n = 2896)

1909–1918 (n = 4338)

1893–1898 (n = 2197)

1903–1908 (n = 2809)

Ages and year of interview

80–89 in 1998

80–89 in 2008

90–99 in 1998

90–99 in 2008

100–105 in 1998

100–105 in 2008

Mean age (standard division)

83.07 (2.59) 82.98 (2.57) 92.11 (2.13) 92.24 (2.19) 101.15 (1.34)

101.72 (1.55)

Women, n (%)

1995 (61.7%)

2362 (58.3%)

2102 (72.6%)

3144 (72.5%)

1652 (75.2%)

2254 (80.2%)

Rural residence (urban = 0) Both sex, n (%)

2135 (66.0%)

2186 (53.9%)

1770 (61.1%)

2314 (53.3%)

1342 (61.1%)

1466 (52.2%)

Women, n (%)

1058 (66.4%)

1108 (54.8%)

1007 (60.9%)

1377 (54.5%)

1019 (57.9%)

1237 (56.0%)

Men, n (%)

1072 (65.3%)

1072 (52.8%)

767 (61.7%)

911 (50.3%)

310 (70.6%)

220 (36.7%)

Both sex, n (%)

902 (27.9%)

1424 (35.1%)

280 (9.7%)

497 (11.5%)

74 (3.4%)

90 (3.2%)

Women, n (%)

225 (14.1%)

451 (22.3%)

47 (2.8%)

139 (5.5%)

5 (0.3%)

25 (1.2%)

Men, n (%)

822 (50.3%)

1077 (53.1%)

345 (27.7%)

490 (27.1%)

55 (12.6%)

70 (11.6%)

Married

Education for both sexes Not educated, n (%)

2006 (62.2%)

2525 (62.3%)

2129 (73.8%)

3233 (74.8%)

1763 (81.1%)

2389 (85.4%)

Primary school, n (%)

900 (27.9%)

1160 (28.6%)

593 (20.6%)

856 (19.8%)

335 (15.4%)

319 (11.4%)

Above primary school, n (%)

320 (9.9%)

362 (8.9%)

164 (5.7%)

231 (5.4%)

76 (3.5%)

90 (3.2%)

1269 (80.7%)

1439 (87.4%)

2180 (86.6%)

1594 (91.8%)

2045 (92.8%)

Education of women Not educated, n (%)

1286 (81.0%)

(continued)

58

Y. Zeng and Q. Feng

Table 4.1 (continued) Years of birth

Years of birth

Years of birth

1909–1918 (n = 3235)

1919–1928 (n = 4053)

1899–1908 (n = 2896)

1909–1918 (n = 4338)

1893–1898 (n = 2197)

1903–1908 (n = 2809)

Primary school, n (%)

227 (14.3%)

315 (15.6%)

174 (10.6%)

283 (11.2%)

118 (6.8%)

127 (5.8%)

Above primary school, n (%)

75 (4.8%)

76 (3.8%)

34 (2.0%)

55 (2.2%)

25 (1.4%)

32 (1.4%)

Education of men Not educated, n (%)

525 (32.0%)

748 (36.9%)

469 (37.8%)

792 (43.9%)

214 (49.1%)

328 (55.2%)

Primary school, n (%)

815 (49.8%)

951 (47.0%)

582 (46.9%)

764 (42.4%)

180 (41.2%)

204 (34.4%)

Above primary school, n (%)

198 (18.2%)

327 (16.1%)

189 (15.3%)

246 (13.7%)

43 (9.7%)

62 (10.5%)

Note The results are weighted averages

oldest-old were all significantly worsened in the later cohorts compared with the earlier cohorts (Fig. 4.2; Tables 4.2, 4.3 and 4.4). All of the 27 sets of comparisons of physical performance tests between different cohorts of the oldest-old showed substantial reductions in annual rates, from −0.4 to −3.8% (Table 4.4). Adjusted for the covariates, the cross-cohort differences in objective physical performance were highly significant in octogenarians, nonagenarians, and centenarians for men, women, and both sexes combined (p < 0.0001) in 22 comparisons, ranging from p = 0.0004 to p = 0.0064 in four comparisons, and p = 0.0184 in one comparison; Tables 4.2, 4.3 and 4.4). The cognitive function measured by the Mini-Mental State Examination test scores of the Chinese oldest-old was significantly worse in the later cohorts compared with the earlier cohorts (Fig. 4.2; Tables 4.2, 4.3 and 4.4). All of the nine sets of comparisons of cognitive function between different cohorts of the oldest-old showed significant reductions in annual rates, ranging from −0.7 to −2.2% (Table 4.5). Adjusted for the covariates, the cross-cohort differences in cognitive functional scores were statistically significant (p < 0.0001) in all of the nine comparisons for octogenarians, nonagenarians, and centenarians, for both sexes and for men and women separately (Tables 4.2, 4.3 and 4.4). Tables 4.2, 4.3 and 4.4 show male–female comparisons of the six pairs of oldestold cohorts aged 80–89, 90–99, and 100–105 years in 1998 and 2008; men had substantially and consistently higher age-specific mortality than did women, but

10.3%

Annual death rate

18 (0.6%)

Missing, n (%)

2974 (92.4%)

80 (2.5%)

164 (5.1%)

0–1

2

≥3

Grouped results, n (%)

0.36 (1.06)

Mean (range = 0–6)

ADL disability score

83.07 (2.59)

Mean age (standard division)

178 (4.4%)

51 (1.3%)

3823 (94.4%)

1 (0.0%)

0.28 (1.01)

9.6%

82.98 (2.57)

0.0015**

0.0006*

0.3204**

0.1516*

0.0054

0.0023

0.0599

0.2493

73 (4.5%)

33 (2.0%)

1525 (93.5%)

10 (0.6%)

0.32 (1.01)

12.5%

82.87 (2.51)

1909–1918 (n = 1641)

Unadjusted

1909–1918 (n = 3235)

Adjusted***

Years of birth

Years of birth

1919–1928 (n = 4053)

Men

p-value of changes

Two sexes combined

73 (3.6%)

28 (1.4%)

1927 (95.0%)

1 (0.0%)

0.24 (0.93)

10.9%

82.81 (2.50)

1919–1928 (n = 2030)

0.2077**

0.0219*

0.1325**

0.4955*

Unadjusted

0.0708

0.0290

0.0671

0.3279

Adjusted***

p-value of changes

86 (5.5%)

44 (2.8%)

1456 (91.8%)

8 (0.5%)

0.39 (1.08)

9.0%

83.20 (2.64)

1909–1918 (n = 1594)

Years of birth

Women

100 (5.0%)

23 (1.2%)

1899 (93.9%)

0 (0.0%)

0.30 (1.06)

8.7%

83.11 (2.61)

1919–1928 (n = 2023)

0.0055**

0.0163*

0.7523**

0.3055*

Unadjusted

0.0293

0.0260

0.264

0.5796

Adjusted***

p-value of changes

0.0425

0.0120