Synonyms

CHARLS

Definition

This entry provides an overview of the China health and retirement longitudinal study, focusing on its value in geropsychology research in China. The entry starts with an introduction on CHARLS Sampling and Implementation including the background, the sampling procedure and design, tracking protocol, data release, and demographics of the respondents. It then describes the contents of the questionnaire, followed by psychologic measurements. This entry is concluded with future plans.

Introduction

China has the largest aging population in the world and also one of the highest aging rates in the world today. It is projected that the proportion of those aged 60 or over will increase from 13% of the population in 2010 (National Bureau of statistics of China 2011) to 33% in 2050 (United Nations 2013), whereas the elderly support ratio (the number of prime-age adults aged 20–59 divided by the number of adults aged 60 or above) will drop from about 4.9:1 in 2010 to 1.4:1 in 2050 (United Nations 2013).

With the rapid aging of Chinese population, the problem of providing for the aged population is becoming increasingly important. One feature of rapid economic growth is that lifetime incomes for younger people tend to be considerably higher than they were for their elderly parents, making the elderly one of the largest disadvantaged groups in China. At the same time, China’s birth control policy means that China’s elderly today have fewer children to support them than in the past. How to deal with problems of support for the well-being of the elderly is one of the greatest challenges to the fast-booming Chinese society in the decades to come.

In face of challenges posed by population aging, the health status of the elderly population is of great importance. A healthy older population can not only reduce the financial and personal care needs but can also contribute to the family and society in the form of working or helping to take care of the young children.

Of all dimensions of health, psychological health is at least as important as physical health to the functionality of older persons. Depression is already listed as a major cause of death and disability in China (Yang et al. 2013; Phillips et al. 2002). In the United States, dementia or cognitive impairment has been shown to cause major caring burdens to the family (Hurd et al. 2013).

At present, scientific studies of China’s aging psychological health problems are still at an early stage, the greatest obstacle being a lack of sufficient micro-longitudinal data. The existing data tend to be small scale in parts of China, not collecting the breadth of data necessary for good social scientific analysis of psychological health of the older population. China Health and Retirement Longitudinal Study (CHARLS) is the first nationally representative survey of the older population that enables the study of psychological health of the older population in China patterned after the Health and Retirement Study (HRS) in the United States, English Longitudinal Study of Ageing (ELSA), and the Study of Health, Ageing, and Retirement in Europe (SHARE).

This entry will give a comprehensive introduction of the CHARLS data set, its sampling method, longitudinal tracking protocol, the content of the questionnaire especially existing psychological measures, and plans for future data collection.

CHARLS Sampling and Implementation

Baseline Sampling

CHARLS is a biennial survey that aims to be representative of the residents of China aged 45 and older, with no upper age limit. The CHARLS national baseline survey was conducted in 2011–2012 and wave 2 in 2013. CHARLS is a nationally representative survey that includes one person per household aged 45 years of age or older and their spouse, totaling 17,708 individuals in wave 1, living in 10,257 households in 450 villages/urban communities (Zhao et al. 2013, 2014). At the first stage, all county-level units were sorted (stratified) by region, within region by urban district or rural county, and by GDP per capita (Tibet was the only province not included). Region was a categorical variable based on the NBS division of province area. After this sorting (stratification), 150 counties or urban districts were chosen with probability proportional to population size (Zhao et al. 2013). For each county-level unit, three PSUs (villages and urban neighborhoods) are randomly chosen with probability proportional to population (Zhao et al. 2013). Hence, CHARLS is nationally represented for both rural and urban areas within China. Counties and districts in 28 provinces are included in the CHARLS sample (Zhao et al. 2013).

China Health and Retirement Longitudinal Study (CHARLS), Table 1 Response rates: 2011 Baseline, Wave 2, 2013
Full size table

In light of the outdated household listings at the village/community level due to population migration, CHARLS designed a mapping/listing software (Charls-GIS) that makes use of Google Earth map images to list all dwelling units in all residential buildings to create sampling frames.

The response rate for the baseline survey was 80.5%, 94% in rural areas and 69% in urban areas, lower in urban areas as is common in most surveys undertaken in developing countries (Table 1) (Zhao et al. 2013). A description of the sample for waves 1 and 2 is provided in Table 1. After applying sampling weights created using the sampling procedure, the CHARLS sample demographics mimics very closely that of population census in 2010 (Zhao et al. 2013).

China Health and Retirement Longitudinal Study (CHARLS), Table 2 Number and age/sex structure of individuals: 2011 Baseline and Wave 2, 2013
Full size table

In each sampled household, a short screening form was used to identify whether the household had a member meeting the age eligibility requirements. If a household had persons older than 39 and meeting the residence criterion, one of them will be randomly selected. If the chosen person is 45 or older, then he/she became a main respondent and his or her spouse was interviewed. If the chosen person was between ages 39 and 44, he/she was reserved for refresher samples for future waves. In wave 2, respondents who were aged 43–44 in wave 1 (plus their spouses) were added from the refresher sample. The same for wave 3 (4) will be done in 2015 (2017), out of those aged 41–42 (39–40) in wave 1. Starting in wave 5 (2019), a new mapping/sampling exercise will be conducted to replenish the sample with appropriate aged cohorts.

Tracking Protocol

Respondents and spouses will be tracked if they exit the original household. While the original CHARLS sample is of the noninstitutionalized elderly population, if a respondent becomes institutionalized, such as entering a nursing home or hospital for a long stay, CHARLS follows them. This potentially matters for obtaining prevalence rates for dementia since it might be that some of the population with dementia is institutionalized. However, in China, the institutionalized population is very small, so in practice for CHARLS, this is unlikely to be an important issue.

Main respondents and spouses in the baseline survey are followed throughout the life of CHARLS or until they die. If a main respondent or spouse remarries, the new spouse is interviewed so long as they are still married to the baseline respondent at the time of the specific wave. In wave 2, only 25 couples split up because of divorce.

For respondents in the baseline, after deaths, 91% of them were recontacted (Table 1). Four hundred twenty-seven exit interviews were conducted on respondents who died between the baseline and wave 2 (464 deaths), including verbal autopsies using the 2012 version from the World Health Organization. In addition, the households which were not found in the baseline were revisited. One thousand one hundred twenty-nine of these (51.6% of those households who had age-eligible members living in nonempty dwellings) were contacted. The households that split because of divorce or moving were also followed. The total household size in wave 2 is 10,832 households with a total of 18,648 individuals (main respondents plus spouses). The age distribution of respondents in baseline and wave 2 is shown in Table 2.

Data Release

The national baseline data and documentation were released publicly, on the CHARLS website (www.charls.ccer.edu.cn/en), in early February 2013, less than 1 year after the fieldwork was completed. The second wave of the national CHARLS sample was fielded in the summer and through the fall of 2013. It was released publicly at the end of this January.

Demographics of the CHARLS Sample

Table 2 describes the age/sex composition of the CHARLS sample. There are 17,708 individuals in the national baseline sample, of which 52.1% are female. While most of the samples are the younger old, 40% are aged 60 years and older. Of the sample, 91.3% were directly interviewed and 8.7% were interviewed by proxy respondent (Table 2).

Content of the Household Survey

Household Survey Instruments

The core survey consists of the following sections: (1) demographics; (2) family structure/transfer; (3) health including biomarkers; (4) health insurance and healthcare utilization; (5) work, retirement, and pension; (6) relative income; (7) family income, wealth, and expenditures; (8) personal income, assets; and (9) housing characteristics. All interviews are conducted using the computer-assisted personal interview (CAPI) technology. The health modules will be described in detail.

Health Status: Self-Reports and Assessments

The self-reports start with the respondent rating health on a scale of excellent, very good, good, fair, and poor or instead very good, good, fair, poor, and very poor. As in HRS, respondent’s self-assessment is asked twice, using each scale, once at the start of the module and once at the end of the sub-module asked randomly determined within CAPI. This is followed by questions asking about diagnoses by doctors of a set of chronic diseases, including stroke and separately psychology diseases, and the timing of diagnoses of specific conditions. Where relevant, current medications and treatments for each specific condition are also collected. Questions about eyesight, hearing, and dental health are asked next and then questions on hedonic well-being. The CHARLS team follows this subsection with a section to obtain information on activities of daily living (ADLs), instrumental activities of daily living (IADLs), and physical functioning. For those who have been identified as having difficulties in ADL or IADL, the care givers are collected. Up to three names are chosen from all of list of family members. Time of care and financial arrangement are asked. Sections on depressive symptoms and cognition follow.

In addition to self-reported health outcome variables, information is collected on several health behaviors. This includes detailed information on smoking, drinking, and physical activities.

Health Status: Biomarkers

Following ELSA and HRS, detailed biomarkers, blood and non-blood, were collected. Non-blood biomarkers such as anthropometrics and blood pressure were collected in waves 1 and 2 and will again be in wave 3. Then the blood biomarkers was collected in wave 1 and will be collected in every other wave, to harmonize with HRS and other aging surveys. In CHARLS the data are collected on height, lower leg and upper arm lengths (useful to get measures related to height not contaminated by shrinkage), waist circumference, blood pressure (measured 3 times), grip strength (measured by a dynamometer two times for each hand), lung capacity measured by a peak flow meter, and doing a timed sit to stand (5 times starting from a full sit position on a common, plastic stool). The balance tests are also conducted, just the same as those used in HRS, and a timed walk at normal speed for 2.5 m again follows HRS.

Healthcare Utilization and Insurance

Indicators of curative and preventive healthcare utilization and health insurance coverage are collected in this module. A separate section on health insurance is asked to collect details of current and past coverage and whether coverage was lost. Healthcare utilization of outpatient care for the last 1 month is asked, with details about last visit. Inpatient utilization over the past 1 year is asked, with details about last visit. The questions include from whom and at what location medical care was received, how much was total cost, what was out of pocket cost, whether insurance was used, if others help pay for the care, whom, and how far respondents traveled.

Life Histories

A special wave to collect life histories was fielded in 2014. Life histories can greatly add to aging surveys because they help to fill in very important details regarding earlier periods in the respondent’s life, which are germane to understanding outcomes when older. Ways to minimize recall error have been greatly improved primarily through the use of calendars that are anchored to key lifetime or calendar events (both national events, like the Cultural Revolution and local, like a major flood) that are salient to respondents’ memory. Such calendars have been developed.

The CHARLS life histories are developed using as a base the ELSA and SHARE life histories, the most complete life histories of the HRS-type aging surveys. The CHARLS life history includes retrospectives on domains that cover family background when the respondent was a child, child health and health care, work and retirement, marriage, childbirths, migration, some retrospective information on income, wealth and poverty status when young, and schooling is collected. Some special history issues germane to China are also included, such as experiences during the Cultural Revolution and the Great Famine and during local events such as a major local flood. These life histories will be especially useful for linkage with the CHARLS ADAMS 2 data.

Community Survey Instrument

One special feature of CHARLS that is new to the HRS-type surveys is to collect detailed panel data from community-level informants (e.g., formal and informal community leaders). Basic community information is collected on, for example, land and its allocation, population, and the most populous surnames and their numbers. More standard information is also collected, such as details about local infrastructure and public facilities such as roads, electrification, water and sanitation infrastructure, and the availability of schools; health insurance and health facilities; and pensions and prices. In addition, the Policy Questionnaire collects details of social welfare programs such as pensions and health insurance, In addition, at the county level.

Psychological Health Measures

Depression

CHARLS uses the ten-question version of the Center for Epidemiologic Study depression (CES-D) battery (The CES-D ten questions are reported in Appendix Table 3, and CHARLS uses the Chinese translation provided at the Center for Epidemiologic Studies website). The answers for CES-D are on an f-scale metric, from rarely, to some days (1–2 days), to occasionally (3–4 days) to most of the time (5–7 days).

Lei et al. (2014a) provides a descriptive analysis of the depressive symptoms as revealed in CHARLS. They scored these answers using the metric suggested by Radloff (1977). Numbers from 0 for rarely to 3 for most of the time are used for negative questions such as “do you feel sad.” For positive questions such as “do you feel happy,” the scoring is reversed from 0 for most of the time to 3 for rarely. A validation exercise of answers to these questions indicates a reasonable level of internal consistency. Lei et al. (2014b) report that in 2011/12 a high fraction of Chinese people 45 and older, both men and women, are suffering from high levels of depressive symptoms, with some 30% of men and 43% of women having CES-D scores 10 and over (out of 30 as a maximum). Rural residents have substantially higher levels of depressive symptoms than urban residents.

Cognition

In the first two waves, CHARLS used a reduced form of the Telephone Interview for Cognitive Status, TICS (Brandt et al. 1988). This includes recognition of date: month, day, year, season (lunar calendar is allowed in addition to Gregorian calendar), day of the week, how the respondent rates their own memory on an excellent, very good, good, fair, poor scale, and serial subtraction of 7s from 100 (up to five times). The respondent is asked to redraw a picture of overlapping pentagons. In addition, immediate and delayed word recall is used, using ten nouns randomly chosen from a list of four groups of words, with approximately 5 min between the immediate and delayed answers. The words will not be read out a second time before the delayed recall (the word lists are reported in Appendix Table 4).

China Health and Retirement Longitudinal Study (CHARLS), Table 3 CES-D questions
Full size table
China Health and Retirement Longitudinal Study (CHARLS), Table 4 Word recall list, English and Mandarin
Full size table

CHARLS shows a steep decline of cognitive functions with age (Lei et al. 2014b). There exist large sex-related differences in cognition to the disadvantage of women, with the large sex-related gap in education being the primary reason for this. These sex-related disparities are eliminated in younger cohorts.

Future Plans

Starting in wave 3 (2015), CHARLS will be introducing a number series test of fluid intelligence, patterned on the HRS number series test (Fisher et al. 2013; Prindle and McArdle 2013).

In CHARLS wave 4, it is scheduled to diagnose dementia and impaired of cognition among the CHARLS respondents aged 65 and older. This will be done in two steps. First, a formal validation sample will be collected from which both interviewer assessment and doctor diagnosis will be conducted. From these data, a statistical model will be built to use interview tests to predict dementia and CIND. This information will be used to inform the final choice of tests and the estimation of weights and cutoff points specific to China with which to classify CHARLS respondents as having dementia and CIND. Among the tests currently planned are the mini-mental state exam (MMSE); immediate and delayed word recall; a measure of verbal fluency, animal naming; the symbol digit modalities test; and backwards digit span.

Cross-References