Productive Activity Patterns and Functional Health: A Comparison of Urban and Rural Chinese Older Adults

Abstract

This study examines the association between productive activity patterns and functional health trajectories of Chinese older adults and whether this association varies by urban/rural residence. Using three waves of the China Health and Retirement Longitudinal Study from a sample of 7,503 older adults, we first performed latent class analysis (LCA) to identify productive activity patterns based on four activities (work, caregiving, informal help, and formal volunteering). Next, multilevel regression analyses were conducted to assess the association between the identified productive activity patterns and functional health trajectories among older adults. Four productive activity patterns are identified from LCA: nonengagers, working-caregivers, workers, and helpers. We find that participation in productive activities is associated with slower functional health decline. The moderation effects of urban/rural differences are prominent across identified groups. Our findings highlight the importance of the urban/rural context in understanding productive aging and its health consequences among Chinese older adults.

China has witnessed a rapid increase in the aging population in recent decades. People aged 60 and older are projected to account for 36.5% of the population by 2050 (United Nations, D. o. E. a. S. A & Population Division, 2015). The increasing prevalence of chronic conditions and physical disabilities among the expanding population of older adults poses tremendous challenges for the health care system and for society (Guo & Zheng, 2018; Wang & Chen, 2014; Wu & Dang, 2013). Against this backdrop, an increasing number of studies have begun to investigate social factors associated with functional disability to help Chinese older adults live a longer and healthier life.

Productive activities, which are activities that produce socially valued goods or services (Bass & Caro, 1993), have a beneficial effect on the health of the individual and broader society (Hinterlong et al., 2007). Existing studies from the United States and from European countries have found that older adults who frequently engage in productive activities report favorable physical health, have better cognitive function, and suffer less from functional impairment and depression (Choi et al., 2013; Hinterlong et al., 2007; Jung et al., 2009; Kim & Ferraro, 2014; Tang, 2009). A few studies have documented the protective effect of productive activities on the mental and cognitive health of Chinese older adults (Liu & Lou, 2017; Luo et al., 2019; Xu, 2019). Yet, little is known about the relationship between productive activities and functional health.

Our study aims to advance current understanding of productive activities and functional health research in three domains. First, previous studies conducted in the United States use discrete measurement or the total number of social roles a participant undertook to measure productive activities, ignoring the fact that older adults often engage in multiple productive activities in a patterned way (Baker et al., 2005; Hinterlong et al., 2007). We perform latent class analysis (LCA) to construct the profiles of productive activity patterns of Chinese older adults. Second, this study adopt a longitudinal design to estimate the association between productive activity patterns and functional health trajectories by using three waves of a nationally representative sample (the CHARLS dataset). Third, the rural-urban contextual differences in China are characterized by their distinct welfare systems, uneven distribution of health care resources, and differential opportunities for productive activities (Li et al., 2014; Liu & Lou, 2017). We examine whether urban-rural differences shape productive activity patterns and whether the link between productive activities and health varies by urban/rural residence.

Background

Productive Activities and Functional Health

There is no consensus on how to define productive activities. We follow the definition of Hinterlong et al. (2007, p. 349) to define productive activities as the activities that are “generating goods or services and for which the individual may or may not be paid.” In our study, we focus on four productive activities: paid/unpaid work, care-giving, informal help (providing help to people who live outside the household such as neighbors, friends, and relatives), and formal volunteering. Empirical studies across various countries have demonstrated health benefits of productive activities in later life (Morrow-Howell & Mui, 2014; Staudinger et al., 2016; Zhan et al., 2009). Studies of older U.S. adults have found that among people who are over 70 years old, engaging in moderate levels of volunteering, paid work, or informal help is associated with better self-rated health, lower risk of mortality, and slower increases in functional limitations (Hinterlong et al., 2007; Lum & Lightfoot, 2005; Luoh & Herzog, 2002). The beneficial effects of volunteering and working have also been found among older adults in Japan (Tomioka et al., 2017) and in urban China (Li et al., 2014). The literature on caregiving appears to be inconsistent. Some studies have shown a deleterious impact of caregiving, while other studies have demonstrated positive effects. Cross-sectional studies document a positive relationship between caregiving to grandchildren and functional limitations of American and Chinese grandparents (Li et al., 2011; Minkler & Fuller-Thomson, 2001). Studies examining the longitudinal patterns reveal that caregiving is beneficial to both urban- and rural- dwelling grandparents in China (Xu, 2019; Zhou et al., 2017).

Numerous empirical studies identify a general positive association between productive activities and functional health. Scholars have begun to investigate the social mechanisms underlying this association. Role theory provides us with a theoretical framework. Two theoretical perspectives—role strain (Goode, 1960) and role enhancement (Rozario et al., 2004)—have the potential to elucidate the interrelations between multiple productive roles and late-life functional health. The role strain perspective emphasizes excessive demands arising from role conflict or role overload that can be detrimental to older adults’ health (Gordon et al., 2012; Liu & Lou, 2017), while the role enhancement perspective posits that social roles could provide individuals with social connections and access to different resources. Occupying social roles promotes social integration and active lifestyles. Social integration is essential for older adults who are gradually losing their social roles (Rosow, 1974). Participation in productive activities preserves their social relationships, reduces stress-related responses and contributes to better health (Hinterlong et al., 2007; Luoh & Herzog, 2002). The role enhancement perspective emphasizes the underlying psychosocial pathways that relate productive activities to better health. More supportive evidence has been found for role enhancement than for role strain.

While previous studies have employed role theory to explain the protective effects of productive activities, the major limitation in existing research is the way of measuring productive activities. Researchers either focus on the influence of a single activity or use aggregate scores (e.g., the total number of productive roles) to measure overall engagement in productive activities. The aggregate indicator may not reflect the latent structure and patterns of activity profiles realistically (Burr et al., 2007). Recent studies using large-scale nationally representative samples reveal the existence of productive activity patterns (Burr et al., 2007; Liu & Lou, 2016). Morrow-Howell (2010) points out that a careful examination of productive activity patterns might predict health outcomes better than a single activity. Building upon previous literature, we constructed productive activity profiles to represent the patterns of late-life activity participation and to test their effects on older adults’ functional health trajectories.

Productive Activities of Older Adults in China: Urban-Rural Differences

China’s urban-rural dual social system is relevant to understanding the significance of contexts in shaping productive activity patterns. Under this system, urban residents receive financial support from the social security and the pension system after retirement, while only 5% of rural seniors can rely on pensions or social security (National Bureau of Statistics of China (NBS), 2012). Most rural older residents perform agricultural work to sustain their livelihoods until they experience severe functional declines in old age (Cai et al., 2012). For urban older adults, late-life employment is an option rather than a necessity. For rural older adults, working in older age is crucial to sustain financial security and to maintain their quality of life. However, engaging in labor-intensive farming may harm their physical health. We know little about how continuous work into very old age might shape the health trajectories of rural elders who are more likely to stay longer in the labor market.

Expectations for family caregiving also differs between urban and rural older adults. Traditional family norms in Asian culture value mutual aid and collective family interests. Providing care to family members, especially grandparenting and helping with family chores, is common in China (Ku et al., 2013; Liu & Lou, 2017). Yet the effects of caregiving vary between urban and rural contexts. Xu (2019) found that urban grandparents benefit from caregiving whereas rural older caregivers garner health disadvantages in providing care. Rural Chinese older adults are more likely than their urban counterparts to become custodial caregivers of their grandchildren due to their adult children’s rural to urban migration (Cong & Silverstein, 2012). In contrast, urban older adults make limited grandparenting contributions. Taking full responsibility for grandparenting increases the health risks among those financially fragile rural older adults. Therefore, urban-rural differences are likely to moderate the effects of caregiving on the health of Chinese older adults.

In China, formal volunteering is more accessible in urban areas than in rural areas. Formal volunteering activities are mainly organized by the government (Luo et al., 2019). Government-led volunteering programs target urban “healthy young-old and retired older elites” (Chen & Adamek, 2017). Rural older adults have fewer opportunities to participate in these formal volunteer programs. Only a few studies have examined the health outcomes of participation in formal volunteering, and they have identified salutary effects of volunteering on the functional health of urban Chinese older adults (Li et al., 2014; Liu & Lou, 2017; Morrow-Howell & Mui, 2014). Additionally, informal helping such as providing help to neighbors and friends is also common in China but has been rarely studied.

Given the structural and contextual differences between urban and rural areas, it is intriguing to study how the relationships between productive activities and functional health vary between rural and urban China. Using three waves of the China Health and Retirement Longitudinal Survey, this study aims to construct a productive activity profile to assess: (a) the relationship between activity patterns and functional health trajectories of Chinese older adults; (b) whether urban and rural differences moderate the association between activity patterns and their functional health. Older adults who are healthier are more likely to participate in social activities and meanwhile, they may experience slower functional health decline. The relationship between social engagement and functional health is confounded by older adults’ health status. We also conducted a sensitivity analysis to test the robustness of our results to address this issue.

Method

Data

The data are derived from three waves of the China Health and Retirement Longitudinal Survey (CHARLS). CHARLS is a nationally representative longitudinal study that collects information from Chinese community-dwelling adults aged 45 years and older. The survey has interviewed participants every 2 years from 2011 to 2015. CHARLS employs a multistage cluster sampling design to obtain the study participants through a face-to-face computer-assisted personal interview system. The baseline data included 10,287 households and 17,705 individuals in 450 communities, among which 89.1% and 83.3% of the participants were followed up in 2013 and 2015, respectively. For this study, we only included participants aged 60 years and older, who accounted for 44% of the original sample. This yields 7,790 study participants. After excluding individuals who had missing data for functional health at baseline, our final analytic sample included 7,503 older adults aged 60 years and older who completed all baseline functional health measures. Among the study sample, 2,942 (39.2%) lived in urban areas and 4,561 (60.8%) lived in rural areas.

Measures

Functional Limitations

We assessed functional limitations with a scale of nine activities (Verbrugge & Jette, 1994). Participants were asked to indicate their difficulty in performing these tasks: running or jogging about 1 km; walking about 1 km; walking 100 m; getting up from a chair after sitting for a long period; climbing several flights of stairs without resting; stooping, kneeling, or crouching; extending arms above the shoulder; lifting weights over 10 jin (about 11 lbs.); and picking up a small coin from a table. Each task was measured from 0 (no difficulty) to 3 (cannot do). We summed all items to obtain a functional limitation score (Nagi, 1976). The average of the scores ranged from 5.1 to 6.4 (S D = 5.3–6.3) and Cronbach’s α for this scale ranged from 0.80 to 0.82 across three waves.

Productive Activities

We followed previous studies (Li et al., 2014; Liu & Lou, 2017; Luo et al., 2019) to include (a) paid/unpaid work, (b) family caregiving, (c) providing informal help, and (d) formal volunteering as the basic categories of productive activities in our study. Working status was assessed differently for urban and rural older adults. Rural older adults were asked if they were engaged in agricultural work for more than 10 days for a wage or for their own farm without a wage in the past year. Urban older adults were asked if they currently had non-agricultural jobs or if they were self-employed. We coded working status as a dichotomous variable (1 = yes, 0 = no) if we obtained affirmative responses for any of these questions. We measured caregiving through two questions asking participants whether they provided care to their parents and whether they provided care to grandchildren.¹ Participants who answered “yes” to either of these two questions were marked as providing care to families (1 = yes, 0 = no). Although caring for parents and caring for children are qualitatively different, there were only a small number of participants who were caregivers to their parents across all three waves. The numbers ranged from 380 at wave 1 to 0 at wave 3. We combined those two types of caregiving activities into a single indicator to represent caregiving status. Informal help reflected affirmative responses in providing help to either relatives, friends, neighbors, or disabled adults who did not live with the participants in the last month (1 = yes, 0 = no). As for volunteering, participants were asked whether they engaged in doing voluntary or charity work in the last month (1 = yes, 0 = no). Four dichotomous productive activities across three waves yielded 12 measures of productive activities.

Urban Versus Rural Residence

Urban versus rural residence was the moderator for this study. In CHARLS, the measure of urban-rural residence was incorporated in the sampling design. The administrative villages (Cun) in rural areas and neighborhoods (Shequ or Juweihui) were the primary sampling units (see more detailed information in Zhao et al., 2014). Thus, the place of residence (urban versus rural) was assessed with the question answered by interviewers on “types of the neighborhood” based on the address of the participants. Because the geographic information of participants is confidential, we relied on this question to identify the place of residence. We coded 1 (rural village) as residing in a rural area and 0 (urban community) as residing in an urban area.

Health

Health conditions were measured based on self-rated health and chronic conditions. Self-rated health was assessed by asking participants to evaluate their health status on a five-point scale, ranging from very poor to very good. We recoded responses as a three-category variable (1 = poor/very poor, 2 = fair, and 3 = good/very good). Chronic conditions were assessed as a count of 13 diagnosed conditions: hypertension, dyslipidemia, diabetes, cancer, lung diseases, liver disease, heart problems, stroke, kidney disease, stomach disease, Parkinson’s disease, asthma, and arthritis.

Socioeconomic Status

Socioeconomic status was measured based on participants’ years of education and household income. The majority of older adults did not receive formal schooling. We recoded education into four categories, 1 = no formal education; 2 = primary school; 3 = middle school; 4 = high school and above. Household income was measured in Chinese yuan (¥); 1 yuan is equivalent to about $0.15. We extracted the household income variable from the harmonized CHARLS dataset developed by the Gateway to Global Aging Data.² Because the household income variable contained 14% missing values, we used multiple imputation to impute missing values based on household assets, education, working status, and household registration (hukou) status. We used logged household income in the final analysis.

Covariates

We included covariates that were correlated with late-life functional status: gender (0 = male, 1 = female), marital status (0 = not married, 1 = married), age, squared age, and living arrangement (1 = multiple-generation dwelling, 2 = living with spouse; 3 = living with children).³

Analytic Strategies

Latent Class Aanlysis

We performed LCA to identify patterns of productive activities based on working, caregiving, volunteering, and providing informal help across three waves. LCA is a statistical method to identify latent class membership among subjects based on categorical variables (Vermunt & Magidson, 2004). LCA uses the maximum likelihood estimation to assign each participant to distinct classes depending on the estimated posterior membership probabilities. We estimated two- to five-class LCA models, controlling for age, urban/rural residence, and self-rated health. We presented model fit indices of the χ² test, the Akaike information criterion (AIC), and the Bayesian information criterion (BIC) for each model. Smaller AIC and BIC indicate a better model fit.

Multilevel Regression

We estimated multilevel regression models with predicted productive activity patterns from LCA and covariates to predict the functional limitations of participants. To facilitate the efficiency of estimation, we centered age and numbers of chronic conditions at their grand means. We specified a two-level model that estimated within-person trajectories as a function of age at level 1 and estimated the level 2 model with productive activity patterns and covariates. To test the moderation effects of urban and rural differences on participants’ functional health, we included productive activity patterns by urban/rural interaction term, allowing the effects of productive activity patterns to vary by urban/rural areas.

Missing Data and Selection Bias

In longitudinal studies, attrition can potentially bias estimations, which may lead to an overestimation of the effects of productive activity patterns on functional health decline. In our analysis, we included a variable of “lost to follow-up” to control nonrandom attrition in this sample. Another issue to consider is the sample selection bias. Adults who were older and sicker were more likely to be selected into a nonengager group. To adjust for potential selection bias, we performed a Heckman two-stage modeling adjustment (Heckman, 1979). After we obtained the predicted activity classes, we dichotomized our sample into an engagement group versus a non-engagement group based on their predicted activity patterns. Then we performed the two-stage bias correction. In the first stage, we estimated a probit model assessing factors leading to selection into non-engagement. The dependent variable was whether the participant was a nonengager. The predictors were age, health status, family structure, and socio-demographic characteristics. We generated the inverse Mills ratio (IMR) after performing the probit model. In the second stage, we estimated the multilevel regression models controlling for the IMR from the first stage to account for possible selection bias. Lastly, we conducted a series of sensitivity analyses to ensure the robustness of the analysis and to control the potential reversed causality and selection bias. The results for sensitivity analyses are presented in the appendices.

Results

Descriptive Statistics

Table 1 presents descriptive statistics for functional limitations, productive activities, and all covariates at wave 1 among total, urban, and rural samples. We performed χ² test for categorical variables and t-test for continuous variables to compare urban-rural differences. We found that rural older adults experienced more functional disabilities, and were less likely to report good or very good self-rated health, while they reported fewer chronic conditions than urban older adults. Urban older adults also had better socioeconomic status, such as higher household income and more formal education. For the productive activities, about 64% of the rural older adults engaged in paid/unpaid work in the past year, whereas the percentage was 30% among urban older adults. There were 33.1% urban older adults who were caregivers compared with 29.5% of rural participants. No significant differences were found for informal help between urban and rural samples. Although, we noticed an extremely low prevalence of volunteering in both the urban and the rural areas at the first wave, with 29 participants engaged in volunteering and charity work. The total number of older adults who participated in volunteer/charity work at least once during the study period is 470. Indeed, volunteering is more prevalent in urban areas. Since 40% participants in this data represented an urban sample, we expected limited numbers of older adults to be engaged in volunteering and charity work. The information on volunteering/charity activities still contributed to the estimation in LCA. Thus, we retained this group in our analysis to gain a better understanding of productive activity patterns among Chinese older adults.

Table 1

Baseline Descriptive Statistics of Chinese Older Adults for the Total Sample, Urban Sample, and Rural Sample.

	Total sample	Urban sample	Rural sample
	Mean (SD)/Freq. (%)	Mean (SD)/Freq. (%)	Mean (SD)/Freq. (%)
Functional limitations	5.11 (5.29)	4.59 (5.23)	5.45 (5.30)	***
Female	3,741 (49.90)	1,519 (51.60)	2,222 (48.70)	*
Currently married	5,561 (74.10)	2,219 (75.40)	3,342 (73.30)	*
Age	68.50 (7.07)	68.70 (7.14)	68.37 (7.03)	n/s
Household income (logged)	9.12 (1.56)	9.77 (1.38)	8.70 (1.53)	***
Education
No formal education	2,798 (37.30)	767 (26.10)	2,031 (44.50)
Primary school	3,027 (40.30)	1,149 (39.10)	1,878 (41.20)	***
Middle school	1,073 (14.30)	553 (18.8)	520 (11.40)
High school and above	605 (8.10)	473 (16.10)	132 (2.90)
Living arrangement
Multiple- generation household	2,646 (35.30)	1,026 (34.90)	1,620 (35.50)
Living with spouse	3,546 (47.30)	1,431 (48.60)	2,115 (46.40)	n/s
Living with children	1,311 (17.50)	485 (16.50)	826 (18.10)
Self-rated health
Poor	2,518 (33.60)	789 (26.80)	1,729 (37.90)
Fair	3,556 (47.40)	1,519 (51.60)	2,037 (44.70)	***
Good	1,429 (19.00)	634 (21.50)	795 (17.40)
Chronic conditions	1.62 (1.48)	1.70 (1.53)	1.58 (1.44)	***
Rural	4,561 (60.80)	--	--
Productive activities^a
Paid/unpaid work	3,778 (50.40)	878 (29.80)	2,900 (63.60)	***
Caregiving	2,317 (30.90)	973 (33.10)	1,344 (29.50)	**
Informal help	330 (4.40)	136 (4.60)	194 (4.30)	n/s
Volunteering	29 (0.40)	18 (0.60)	11 (0.20)	**
N	7,503	2,942	4,561

Note. Frequencies with percentages for categorical variables and means with standard deviations for continuous variables are reported. χ² test (for categorical variables) and t test (for continuous variables) are performed to compare urban/rural differences.

^aWe report the frequencies and percentages of the original measures of productive activities. For example, 50.4% of participants engaged in paid/unpaid work in the total sample. Correspondingly, 49.6% of the participants didn’t participate in paid/unpaid work in the total sample.

^* p < .05. ^** p < .01. ^*** p < .001. n/s = not significant.

Productive Activity Patterns

We performed LCA to identify productive activity patterns. We tested four solutions from a two-class to five-class model to identify the best fit. Model-fit indices indicated that a four-class solution (AIC = 52629.96, BIC = 53066.11) had the best fit of all models (see Appendix Table A-1). We labeled each class based on the item-response probabilities. It can be understood as the relative scores of each observed activity for a certain class (see Appendix Table A-2). We labeled class one as “nonengagers” who scored higher in “no” for all activities across three waves (19%). Class two was labeled as “workers.” Participants in this class had high scores of participating in paid/unpaid work but low scores for the remaining activities (29%). Class three was labeled as “working caregivers.” This group of older people had high probabilities in working and caregiving for all waves but scored low in informal help and volunteering (34%). The last class was “helpers,” including older adults who were characterized as having relatively low probabilities of employment but moderate probabilities in caregiving, informal help, and volunteering (18%).

Table 2 presents functional status, demographics, and socioeconomic characteristics of the four identified productive activity patterns. Bivariate tests indicated that the differences in all characteristics among these patterns were statistically significant. The nonengagers were more likely to be female, older, not in a marital relationship, poorly educated, and to report more chronic diseases and poor health compared with respondents in the other groups. Workers and working caregivers were more likely to be males and married. They had lower household income but better health status. Helpers were a group of older people with the highest socioeconomic status. About 17% of them had completed high school or higher education, and they had the highest household income. Their health status was better than the nonengagers and slightly worse than the two working groups.

Table 2

Functional Limitations and Demographic Characteristics by Productive Activity Patterns.

	Nonengager	Worker	Working caregiver	Helper
	Mean (SD)/ Freq. (%)	Mean (SD)/ Freq. (%)	Mean (SD)/ Freq. (%)	Mean (SD)/ Freq. (%)
Functional limitations W1	7.74 (6.3)	3.77 (3.8)	3.03 (3.4)	4.50 (4.9)
Functional limitations W2	8.88 (7.0)	4.24 (4.5)	3.34 (3.7)	5.40 (5.7)
Functional limitations W3	9.61 (7.1)	5.05 (5.3)	4.13 (4.6)	6.22 (6.1)
Female	1,451 (57.1)	996 (45.5)	570 (40.2)	724 (53.4)
Currently married	1,569 (61.7)	1,718 (78.6)	1,203 (84.9)	1,071 (79.0)
Age	74.44 (7.0)	66.89(5.2)	63.45 (3.3)	65.21 (4.6)
Household income (logged)	9.08 (1.67)	8.77 (1.42)	9.13 (1.44)	9.75 (1.48)
Education
No formal education	1,206 (47.4)	823 (37.6)	436 (30.8)	333 (24.6)
Primary school	823 (32.4)	1,010 (46.2)	706 (49.8)	488 (36.0)
Middle school	277 (10.9)	278 (12.7)	211 (14.9)	307 (22.6)
High school and above	237 (9.3)	76 (3.5)	64 (4.5)	228 (16.8)
Living arrangement
Multiple-generation household	739 (29.1)	605 (27.7)	699 (49.3)	603 (44.5)
Living with spouse	1,470 (57.8)	1,213 (55.5)	366 (25.8)	497 (36.7)
Living with children	334 (13.1)	369 (16.9)	352 (24.8)	256 (18.9)
Self-rated health
Poor	1,074 (42.2)	620 (28.3)	366 (25.8)	458 (33.8)
Fair	1,066 (41.9)	1,091 (49.9)	723 (51.0)	676 (49.9)
Good	403 (15.8)	476 (21.8)	328 (23.1)	222 (16.4)
Chronic conditions	1.83 (1.6)	1.45 (1.4)	1.36 (1.3)	1.81 (1.5)
Rural	1,278 (50.3)	1,760 (80.5)	1,077 (76.0)	446 (32.9)
N	2,543	2,187	1,417	1,356

Note. N = 7,503. Frequencies with percentages for categorical variables and means with standard deviations for continuous variables are reported. Bivariate tests indicate that all group differences are statistically significant at the p < .001 level.

In regard to functional limitations, comparing functional health across different groups, nonengagers demonstrated the highest levels of functional disabilities. Working caregivers had the lowest levels of functional problems. This pattern was consistent across three waves. As for the changes of functional health within each group, nonengagers experienced rapid health decline. Working caregivers showed the lowest levels of health decline compared with other groups. These findings may indicate some beneficial effect of productive activity participation that helps to sustain a slower health decline at old age. Next, using the four patterns as predictors, we conducted multilevel analyses to examine the effects of productive activity patterns on older people’s functional health, adjusting for confounding factors and covariates.

Functional Health and Productive Activity Patterns

In Table 3, we present the results of the multilevel analyses, adjusting for demographic characteristics, socioeconomic status, health status, attrition, and selection. Model 1 shows the impact of urban-rural differences and productive activity patterns on the health trajectories of older adults. Rural respondents reported higher levels of functional limitations (β = .98, p < .000) compared with their urban counterparts. Three groups that were characterized by various productive activity engagement on average had significantly lower levels of functional limitations compared with nonengagers. The interaction term between age and productive activity patterns indicated that health declines were significantly slower among workers (β = −0.09, p < .000) in comparison with nonengagers.

Table 3

Multilevel Regression Analysis of Productive Activity Patterns on Functional Limitations.

	Model 1		Model 2
	b	SE	b	SE
Fixed effects
Region (ref. urban)
Rural	0.98***	(0.10)	1.85***	(0.21)
Productive activity group(ref. Nonengager)
Worker	−1.96***	(0.16)	−1.11***	(0.24)
Working caregiver	−1.92***	(0.19)	−1.03**	(0.31)
Helper	−.71***	(0.15)	−.67***	(0.19)
Growth rates of activity groups (ref. Nonengager*age)
Worker*age	−.09***	(0.02)	−.15***	(0.04)
Working caregiver*age	−.05	(0.03)	−.11*	(0.05)
Helper*age	−.02	(0.03)	−.09**	(0.03)
Interaction effects
Average effect (ref. Urban Nonengager)
Rural*Worker	--	--	−1.40***	(0.28)
Rural*Working caregiver	--	--	−1.46***	(0.35)
Rural*Helper	--	--	0.04	(0.29)
Growth rate (ref. NonengagerUrbanage)
Rural* Worker*age	--	--	0.12**	(0.04)
Rural* Working caregiver*age	--	--	0.12*	(0.06)
Rural* Helper*age	--	--	0.15**	(0.05)
Intercept	10.86***	(0.37)	10.14***	(0.38)
Random effect
Level 1 Residual	3.36***	(0.02)	3.36***	(0.02)
Level 2 Intercept	2.56***	(0.05)	2.54***	(0.05)
Level 2 Slope	0.22***	(0.01)	0.22***	(0.01)

Note. N = 7,503. Coefficients with standard errors (SE) in parentheses are reported, adjusting for gender, marital status, education, income, age, age(squared), living arrangement, chronic conditions, self-rated health, attrition, and selection (IMR).

*p < .05. ^** p < .01. ^*** p < .001.

In model 2, we further included the residence-by-productive activity pattern interaction terms to formally test the moderation effect of urban-rural differences on functional health trajectories. The estimated effects are illustrated in Figure 1. On average, rural older adults demonstrated higher levels of functional declines for all four groups. The predicted functional limitations for urban and rural nonengagers converged, whereas we observe diverged functional limitations for workers, the working caregivers, and for the helpers. Specifically, the initial health gap was large for the nonengagers. This gap narrowed with the aging of this group and crossed at the very old age (over 93 years). The predicted functional health at the upper end (right side of the x-axis) should be interpreted with caution. Since only very few participants (less than 1%) survived over age 95, the estimates of functional limitations of the very old people might be highly variable. This also applies to the other three groups. There was a wide health gap between urban and rural helpers when these older adults were relatively young. The health gap became even wider with increases of age. The patterns of health declines were similar for workers and for working caregivers. The functional capacities of this group overlapped at the onset of the study and they experienced slight increases of health decline as they aged.

Figure 1

Urban-rural differences in the trajectories of functional limitations by productive activity patterns.

Discussion

This study investigated the association between productive activity patterns and functional health trajectories among Chinese older adults. Distinct from previous research, our study examined the latent structure of productive activities through LCA. We identified four productive activity patterns: nonengagers, working-caregivers, workers, and helpers. These patterns capture the variation of role changes of older people and better reflect the profiles of productive activities compared to single or aggregate measures (Liu & Lou, 2017).

We found significant differences in functional health trajectories between the nonengaged group and the engaged groups (workers, working-caregivers, and helpers). Older adults who participated in productive activities reported fewer functional limitations over a 5-year period. This finding provides evidence to support the role enhancement perspective. Social roles augment individuals’ cognitive capacity, social resources, and emotional gratification through psychosocial pathways, which proffers individuals avenues for sustaining functional health. First, participation in productive activities benefits the cognitive health of older adults. Performing work-related tasks, caregiving, or volunteering requires cognitive skills such as memorizing, reasoning, and decision making. An engaged lifestyle preserves cognitive abilities and slows down cognitive deterioration later in life (Luo et al., 2019). Second, active engagement in social activities, such as volunteering, expands older people’s social network and social capital (McNamara & Gonzales, 2011). The social resources embedded in social ties buffer against stress responses and reduce mental distress (Norstrand & Xu, 2012). Third, Chinese culture also plays a role in the association between productive activities and health. Chinese society values collective well-being in which mutual help is the cultural norm. Engaging in productive activities and producing socially valued goods label people as valuable social members. The gratification received via productive activity participation enhances the overall health of older adults (Xu, 2019).

We found significant urban-rural differences in the association between functional health and productive activities. First, urban-rural residence shapes the productive activity patterns of older adults (Table 2). Our findings revealed that rural older adults frequently participate in economic activities. Long-term underinvestment in social infrastructure and lack of sufficient pension support in rural areas place rural older adults in a vulnerable position of facing higher financial risks especially in later life (Norstrand & Xu, 2012). Rural older adults stay in economic activities longer to sustain financial security even after they reach the age of 60 years (Pang et al., 2004). In contrast, urban older adults are more likely to be involved in non-economic activities such as volunteering and helping. This finding is consistent with previous studies, which indicate the distinct opportunity structures of access to volunteering activities between urban and rural areas (Li et al., 2014; Luo et al., 2019).

Second, functional health trajectories presented distinct patterns by activity groups. Overall, urban older adults with different productive activity patterns experienced less severe functional strain and slower declines in their health. Considering group differences, rural nonengagers demonstrated the highest levels of functional limitations and their health deteriorated rapidly. Although urban nonengagers were comparatively healthier, disengagement from social activities may be related to their later rapid functional health decline. The health gap between urban and rural nonengagers narrowed as they aged. This implies a universal negative effect of disengagement from meaningful social roles on functional health. The disengaged people are most likely to be a group of older adults who lack health care and financial resources (Morrow-Howell & Mui, 2014), suggesting a greater challenge for them to benefit from participation in productive activities (Liu & Lou, 2017). For the caregivers, workers, and helpers, our findings indicated a widening urban and rural health gap. The dual social system creates barriers for rural older adults to gain access to health care and social services but benefits urban older adults (Cai et al., 2012; Zhang et al., 2017). For urban residents, health care facilities are concentrated in metropolitan areas and big cities. The pension and health care system serve as safety nets that help them to access health-promoting resources and participate in activities such as volunteering, helping, or charities. Engagement in those activities can slow down their health decline and, reciprocally, promotes their functional health. The effect of productive activities is subject to the macro-social environment and the urban-rural division in China. Unequal access to social resources and health care services between urban and rural contexts contributes to urban-rural health disparities.

We also conducted sensitivity analyses to address the potential selection bias in our study. Functional limitations might also determine whether older adults are capable of engaging in productive activities. To rule out this possibility, we selected “healthy” participants at baseline and repeated previous analyses (see Appendix Table A-3). We examined the association between productive activity patterns and functional health among three subset samples: participants who report “good” self-rated health; participants without any chronic conditions; and participants who have no limitations on the Activity of Daily Living (ADL) scale. The results from these subset samples confirmed our main findings. All coefficients retain the same direction with slightly reduced effect sizes and significance levels.

Several limitations of our study need to be noted. One limitation of this study is that we did not examine the frequency and intensity of each social activity. How frequently older adults engage in productive activities have an impact on their functional health. Future research could collect information about the frequency, duration, and roles in productive activities to evaluate the association between these factors and individuals’ functional health trajectories. Another limitation is the measurement of volunteering or charity engagement. CHARLS only asked if participants engaged in volunteering or charity work in the last month. Compared to working or caregiving, which usually lasts for a significantly longer period of time, participation in volunteering or charity work may be underreported by older people due to a shorter observational time window. In our sample, only a small number of participants reported engagement in volunteering or charity work. The lack of variation on this type of activity may potentially weaken the power to make definitive conclusions about the association between volunteering engagement and later life functional health among Chinese older adults. Lastly, we were not able to identify potential pathways that mediate the association between productive activity patterns and functional health. Further studies are needed to test how social capital, cognitive capacity, and cultural factors influence the association between productive activities and functional health in the Chinese context.

Conclusion

Consistent with studies from the Western countries, our findings illustrate a negative association between productive activity participation and functional health decline for Chinese older adults. We find significant urban-rural differences in productive activity patterns, which are associated with distinct functional health trajectories of rural and urban older adults. Rural nonengagers are the most vulnerable group and their health deteriorates rapidly as they age. Urban older adults have advantageous resources and they derive health benefits from the macro-level environment.

Our findings have implications for researchers and policymakers who aim to promote productive activities in China. Engaging in productive activities can be a viable way to reduce rapid health decline and to promote active lifestyles for older adults. Unequal access to productive activities and the resulting differential health outcomes in urban and rural areas call for a policy design tailored to heterogeneous subgroups of Chinese older people. Rural older adults should benefit from an expansion of the current pension system. Creating programs to encourage volunteering and other forms of productive activities among rural older adults may reduce urban-rural health disparities. Continued efforts to investigate various productive activities and their impacts on late-life health are essential in rapidly aging societies.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Jiao Yu

Notes

Author Biographies

Jiao Yu is a PhD candidate in sociology department at Case Western Reserve University. Her research interests include aging and life course, health disparities and medical sociology.

Yulin Yang is a doctoral candidate in sociology at the University at Buffalo. Her research interests are aging and life course, health inequality, and medical sociology.

Eva Kahana, PhD, is Distinguished University professor and Robson professor of Sociology, Nursing Medicine and Applied Social Sciences at Case Western Reserve University. Her research interests are in Successful Aging, Coping with Late Life Disability and Health Communication.

Appendices

Table A-3

Latent Class Probabilities of Four-Class Model of Productive Activities.

	Model 1		Model 2		Model 3
	Self-rate health= “good”		Chronic conditions = 0		ADL = 0
	b (SE)		b (SE)		b (SE)
Rural	1.81	***	2.02	***	1.19	***
	(0.46)		(0.41)		(0.21)
Productive activity group (ref. Nonengager)
Worker	−.40		−.82	*	−.82	***
	(0.36)		(0.37)		(0.21)
Working caregiver	−.84		−1.27	**	−.49
	(0.45)		(0.48)		(0.27)
Helper	0.04		−.40		−.05
	(0.32)		(0.34)		(0.17)
Growth rates of activity groups (ref. Nonengager*age)
Worker*age	−.17	**	−.13		−.12	***
	(0.07)		(0.07)		(0.04)
Working caregiver*age	−.28	***	−.17	*	−.10	*
	(0.08)		(0.09)		(0.05)
Helper*age	−.13	*	−.09		−.08	**
	(0.06)		(0.06)		(0.03)
Urban-rural differences
Average effect (ref. Urban*Nonengager)
Rural*Worker	−1.19	*	−1.74	***	−.60	*
	(0.54)		(0.50)		(0.27)
Rural*Working caregiver	−1.00		−1.53	*	−.78	*
	(0.63)		(0.60)		(0.33)
Rural*Helper	−.86		−1.49	**	−.19
	(0.66)		(0.56)		(0.29)
Growth rate (ref. Nonengagerurbanage)
RuralWorkerage	0.08		0.13		0.07
	(0.08)		(0.07)		(0.04)
RuralWorking caregiverage	0.22	*	0.17		0.10
	(0.10)		(0.10)		(0.06)
RuralHelperage	0.19		0.18	*	0.12	*
	(0.11)		(0.09)		(0.05)
Intercept	8.60	***	9.53	***	7.32	***
	(0.69)		(0.64)		(0.35)
Level 1 Residual	2.66	***	3.05	***	3.11	***
	(0.04)		(0.04)		(0.02)
Level 2 Intercept	1.52	***	1.91	***	1.91	***
	(0.09)		(0.09)		(0.05)
Level 2 Slope	0.23	***	0.20	***	0.16	***
	(0.02)		(0.02)		(0.01)
N	1,429		1,908		5,641

Note: Coefficients with standard errors (SE) in parentheses are reported, adjusting for gender, marital status, education, income, age, age(squared), living arrangement, chronic conditions, self-rated health, attrition, and selection (IMR). ^* p < .05. ^** p < .01. ^*** p < .001.

References

Baker

L. A.

Cahalin

L. P.

Gerst

Burr

J. A

. (2005). Productive activities and subjective well-being among older adults: The influence of number of activities and time commitment. Social Indicators Research, 73(3), 431–458.doi:10.1007/s11205-005-0805-6

Bass

S. A.

Caro

F. G

. (1993). Achieving a productive aging society. Auburn House Pub. Co.

Burr

J. A.

Mutchler

J. E.

Caro

F. G

. (2007). Productive activity clusters among middle-aged and older adults: Intersecting forms and time commitments. The Journals of Gerontology Series B: Psychological Sciences and Social Sciences, 62(4), S267–S275.doi:10.1093/geronb/62.4.S267

Cai

Giles

O’Keefe

Wang

. (2012). Old-Age support in rural China: Challenges and prospects. The World Bank.

Chen

Adamek

. (2017). Civic engagement of older adults in mainland China: Past, present, and future. The International Journal of Aging and Human Development, 85(2), 204–226.

Choi

K.-S.

Stewart

Dewey

. (2013). Participation in productive activities and depression among older Europeans: Survey of health, ageing and retirement in Europe (share). International Journal of Geriatric Psychiatry, 38(11), 1157–1165.doi:10.1002/gps.3936

Cong

Silverstein

. (2012). Caring for grandchildren and intergenerational support in rural China: A gendered extended family perspective. Ageing and Society, 32(3), 425–450.doi:10.1017/S0144686X11000420

Goode

W. J

. (1960). A theory of role strain. American Sociological Review, 25(4), 483–496.doi:10.2307/2092933

Gordon

J. R.

Pruchno

R. A.

Wilson-Genderson

Murphy

W. M.

Rose

. (2012). Balancing caregiving and work: Role conflict and role strain dynamics. Journal of Family Issues, 33(5), 662–689.

10.

Guo

Zheng

. (2018). Health challenges and opportunities for an aging China. American Journal of Public Health, 108(7), 890–892.doi:10.2105/AJPH.2018.304444

11.

Heckman

J. J

. (1979). Sample selection bias as a specification error. Econometrica, 47(1), 153–161.doi:10.2307/1912352

12.

Hinterlong

J. E.

Morrow-Howell

Rozario

P. A

. (2007). Productive engagement and late life physical and mental health: Findings from a nationally representative panel study. Research on Aging, 29(4), 348–370.

13.

Jung

Gruenewald

T. L.

Seeman

T. E.

Sarkisian

C. A

. (2009). Productive activities and development of frailty in older adults. Journals of Gerontology Series B: Psychological Sciences Social Sciences, 65(2), 256–261.

14.

Kim

Ferraro

K. F

. (2014). Do productive activities reduce inflammation in later life? Multiple roles, frequency of activities, and C-reactive protein. The Gerontologist, 54(5), 830–839.doi:10.1093/geront/gnt090

15.

L.-J. E.

Stearns

S. C.

Van Houtven

C. H.

Lee

S.-Y. D.

Dilworth-Anderson

Konrad

T. R

. (2013). Impact of caring for Grandchildren on the health of grandparents in Taiwan. The Journals of Gerontology Series B: Psychological Sciences and Social Sciences, 68(6), 1009–1021.doi:10.1093/geronb/gbt090

16.

Chi

Krochalk

P. C.

. (2011). Widowhood, family support, and self-rated health among older adults in China. International Journal of Social Welfare, 20(s1), S72–S85.doi:10.1111/j.1468-2397.2011.00818.x

17.

Chi

Guo

. (2014). Participation in productive activities and health outcomes among older adults in urban China. The Gerontologist, 54(5), 784–796.doi:10.1093/geront/gnt106

18.

Liu

Lou

W. Q

. (2016). Patterns of productive activity engagement among older adults in urban China. European Journal of Ageing, 13(4), 361–372.doi:10.1007/s10433-016-0387-y

19.

Liu

Lou

V. W. Q

. (2017). Patterns of productive activity engagement as a longitudinal predictor of depressive symptoms among older adults in urban China. Aging & Mental Health, 21(11), 1147–1154.doi:10.1080/13607863.2016.1204983

20.

Lum

T. Y.

Lightfoot

. (2005). The effects of volunteering on the physical and mental health of older people. Research on Aging, 27(1), 31–55.doi:10.1177/0164027504271349

21.

Luo

Pan

Zhang

. (2019). Productive activities and cognitive decline among older adults in China: Evidence from the China health and retirement longitudinal study. Social Science & Medicine, 229, 96–105.doi:10.1016/j.socscimed.2018.09.052

22.

Luoh

M.-C.

Herzog

A. R

. (2002). Individual consequences of volunteer and paid work in old age: Health and mortality. Journal of Health and Social Behavior, 43(4), 490–509.doi:10.2307/3090239

23.

McNamara

T. K.

Gonzales

. (2011). Volunteer transitions among older adults: The role of human, social, and cultural capital in later life. The Journals of Gerontology Series B: Psychological Sciences and Social Sciences, 66B(4), 490–501.doi:10.1093/geronb/gbr055

24.

Minkler

Fuller-Thomson

D. E

. (2001). Physical and mental health status of American grandparents providing extensive child care to their grandchildren. Journal of the American Medical Women’s Association, 56(4), 199–205.

25.

Morrow-Howell

. (2010). Volunteering in later life: Research frontiers. The Journals of Gerontology Series B: Psychological Sciences and Social Sciences, 65B(4), 461–469.doi:10.1093/geronb/gbq024

26.

Morrow-Howell

Mui

A. C

. (2014). Productive engagement in later life: A global perspective. Routledge.

27.

Nagi

S. Z

. (1976). An epidemiology of disability among adults in the United States. The Milbank Memorial Fund Quarterly. Health and Society, 54(4), 439–467.doi:10.2307/3349677

28.

National Bureau of Statistics of China (NBS) . (2012). Tabulation on the 2010 Population Census of the People’s Republic of China. China Statistical Press.

29.

Norstrand

J. A.

. (2012). Social capital and health outcomes among older adults in China: The urban-rural dimension. The Gerontologist, 52(3), 325–334.doi:10.1093/geront/gnr072

30.

Pang

De Brauw

Rozelle

. (2004). Working until you drop: The elderly of rural China. The China Journal, 52, 73–94.

31.

Rosow

. (1974). Socialization to old age. University of California Press.

32.

Rozario

P. A.

Morrow-Howell

Hinterlong

J. E

. (2004). Role enhancement or role strain: Assessing the impact of multiple productive roles on older caregiver well-being. Research on Aging, 26(4), 413–428.

33.

Staudinger

U. M.

Finkelstein

Calvo

Sivaramakrishnan

. (2016). A global view on the effects of work on health in later life. The Gerontologist, 56(Suppl 2), S281–S292.doi:10.1093/geront/gnw032

34.

Tang

. (2009). Late-Life volunteering and trajectories of physical health. Journal of Applied Gerontology, 28(4), 524–533.doi:10.1177/0733464808327454

35.

Tomioka

Kurumatani

Hosoi

. (2017). Association between social participation and 3-year change in instrumental activities of daily living in community-dwelling elderly adults. Journal of the American Geriatrics Society, 65(1), 107–113.doi:10.1111/jgs.14447

36.

United Nations, Department of Economic and Social Affairs, Population Division . (2015). World Population Prospects: The 2015 revision, key findings and advance tables. Working Report No. ESA/P/WP.241. https://population.un.org/wpp/Publications/Files/Key_Findings_WPP_2015.pdf

37.

Verbrugge

L. M.

Jette

A. M

. (1994). The disablement process. Social Science & Medicine, 38(1), 1–14.doi:10.1016/0277-9536(94)90294-1

38.

Vermunt

J. K.

Magidson

. (2004). Latent class analysis. The Sage Encyclopedia of Social Sciences Research Methods, 2, 549–553.

39.

Wang

X.-Q.

Chen

P.-J

. (2014). Population ageing challenges health care in China. The Lancet, 383(9920), 870.doi:10.1016/S0140-6736(14)60443-8

40.

Dang

. (2013). Blue book of aging: China report of the development on aging cause. Social Sciences Academic Press (Chinese).

41.

. (2019). Physical and mental health of Chinese grandparents caring for grandchildren and great-grandparents. Social Science & Medicine, 229, 106–116.doi:10.1016/j.socscimed.2018.05.047

42.

Zhan

Wang

Liu

Shultz

K. S

. (2009). Bridge employment and retirees’ health: A longitudinal investigation. Journal of Occupational Health Psychology, 14(4), 374–389.doi:10.1037/a0015285

43.

Zhang

Dupre

M. E.

Qiu

Zhou

Zhao

. (2017). Urban-Rural differences in the association between access to healthcare and health outcomes among older adults in China. BMC Geriatrics, 17(1), 151.doi:10.1186/s12877-017-0538-9

44.

Zhao

Smith

J. P.

Strauss

Yang

. (2014). Cohort profile: The China health and retirement longitudinal study (CHARLS). International Journal of Epidemiology, 43(1), 61–68.doi:10.1093/ije/dys203

45.

Zhou

Mao

Lee

Chi

. (2017). The impact of caring for Grandchildren on grandparents’ physical health outcomes: The role of intergenerational support. Research on Aging, 39(5), 612–634.doi:10.1177/0164027515623332