Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2022 Aug 1.
Published in final edited form as: Soc Sci Med. 2021 Jun 30;283:114188. doi: 10.1016/j.socscimed.2021.114188

Duration of subjective poverty in relation to subsequent cognitive performance and decline among adults aged ≥64 in China, 2005-2018

Xuexin Yu 1,2, Wei Zhang 1, Lindsay C Kobayashi 2
PMCID: PMC8325638  NIHMSID: NIHMS1721480  PMID: 34225036

Abstract

Background

The effects of late-life subjective poverty on brain health are understudied. We aimed to investigate the association between duration of subjective poverty after age 64 and subsequent cognitive function and decline in China.

Methods

Data were from 4,118 adults aged ≥64 at baseline in the population-based China Longitudinal Healthy Longevity Survey (CLHLS), 2005-2018. The duration of subjective poverty was measured from self-rated economic status relative to neighbors in 2005, 2008, and 2011 (never; one time point; two or three time points). Cognitive function was assessed by the Chinese Mini-Mental State Exam (CMMSE; range: 0-30) in 2011, 2014, and 2018. We fitted attrition-weighted, multivariable mixed-effect Tobit regression models to examine the relationship between duration of subjective poverty from 2005-2011 and subsequent cognitive function and decline from 2011-2018.

Results

A total of 2,675 (64.96%) participants never reported subjective poverty over the period 2005-2011, 930 (22.58%) participants reported subjective poverty at one time point, and 513 (12.46%) reported subjective poverty at two or three time points. Compared to those who never reported subjective poverty, participants experiencing subjective poverty at one time point (β=−0.95, 95% CI: −1.48 to −0.41) and two or three time points (β=−2.01; 95% CI: −2.73 to −1.29) had lower CMMSE scores in 2011, indicating a dose-response relationship. Individuals with a longer duration of subjective poverty had a slower rate of decline in CMMSE scores than those never in subjective poverty (β=1.44; 95% CI: 0.20 to 2.68 for 2018×Two or three time points).

Conclusion

Subjective poverty in late life may have unique and cumulative contributions to cognitive aging among older adults in China. The lower level of initial cognitive function but slower rate of cognitive decline observed for those with greater subjective poverty is consistent with theories of cognitive reserve and empirical evidence from Western settings on other socioeconomic markers.

Keywords: subjective poverty, longitudinal study, cognition, cognitive decline, older adults, China

1. Introduction

Measures of social and socioeconomic status are consistently associated with a wide range of health outcomes, with recent evidence indicating that the richest 1% of the United States population had an average life expectancy over 10 years longer than the poorest 1% of the population (Chetty et al., 2016; Kawachi et al., 1997; Pickett and Wilkinson, 2015). Consistent with cumulative disadvantage and fundamental cause theories, social inequality may accumulate over time to shape long-term health trajectories, resulting in heterogeneous and socially unequal health outcomes among older adults (Dannefer, 2003; Ferraro et al., 2009; Melo et al., 2019; Phelan et al., 2010). For example, individuals with high socioeconomic status (e.g., as often indicated by income, wealth, education, and occupation) may have access to flexible material resources (e.g., health care) and non-material resources (e.g., social cohesion) that help them accumulate health advantages over the life course (Adler and Ostrove, 1999; Kawachi et al., 1997; Melo et al., 2019; Phelan et al., 2010).

Independent of objectively measured traditional socioeconomic status markers, subjective social status (SSS) may also contribute to physical and mental health in late-life (Adler et al., 2000; Demakakos et al., 2008; Euteneuer, 2014; A. Singh-Manoux et al., 2003; Zahodne et al., 2018). SSS has been theorized as a comprehensive measure of social status that captures an individual's 'cognitive averaging' of multi-dimensional elements that represent past (e.g., education, childhood socioeconomic conditions), current (e.g., occupation, job control, income, and life satisfaction), and future social status (e.g., financial security, opportunity, and expectations) (Miyakawa et al., 2012; Archana Singh-Manoux et al., 2005). Despite substantial evidence on the role of objective socioeconomic status (e.g., education and income) in cognitive aging (Jefferson et al., 2011; Stern, 2009), only two studies have investigated the relationship between SSS and cognitive function and decline in later-life, with inconsistent results (Kim et al., 2021; Zahodne et al., 2018). Zahodne et al. (2018) and Kim et al. (2021) both found that lower SSS was associated with worse cognitive performance among adults aged 65 and over, indicating that SSS, like education and other objective socioeconomic markers, may be an indicator of cognitive reserve, which refers to the brain's ability to compensate for latent neuropathology (Stern et al., 2020). Zahodne et al. (2018) did not observe a longitudinal relationship between SSS and cognitive decline over time in the United States, while Kim et al. (2021) observed that SSS was associated with accelerated cognitive decline among older adults in Hong Kong.

Despite the sparse evidence in this area, biological plausibility for a role of SSS in cognitive aging outcomes is strong. The experience of occupying a low social position may lead to depressive symptoms and psychological stress, with suggestive evidence indicating a link between SSS and stress-related biomarkers such as resting heart rate, altered cortisol response, and immune defense (Adler et al., 2000; Diaz et al., 2014; Euteneuer, 2014; Santini et al., 2020). Many of these biomarkers have been recognized as important predictors of cognitive impairment and dementia among older adults (McNaull et al., 2010). Low SSS may also lead to loneliness and social isolation (Ayalon, 2019), potentially resulting in reduced cognitive stimulation, lower cognitive reserve, poor resilience of the brain to aging-related pathology, and decreased brain volumes, all of which are implicated in dementia etiology (Evans et al., 2019; Gianaros et al., 2007). As more than 70% of global dementia cases are projected to occur in low- and middle-income countries by 2050 (Prince et al., 2013), understanding the contributors to cognitive aging, especially in these countries, is essential to reducing the global dementia burden.

However, evidence on the potential role of SSS in cognitive aging is limited, as research to-date has measured SSS at only one point in time, resulting in a loss of exposure information (Kim et al., 2021; Zahodne et al., 2018). SSS is likely to be time-varying, and it is unknown whether it has a cumulative relationship with cognitive aging outcomes, as with objective measures of socioeconomic status (Haan et al., 2011; Kobayashi and Feldman, 2019; Zeki Al Hazzouri et al., 2017). In addition, prior research on both objective and subjective measures of social and socioeconomic status tend to under-represent middle-old (aged 75 to 84) and oldest-old adults (aged 85+), potentially losing information about key life course periods when socioeconomic fluctuations could impact cognitive aging trajectories (Garfein and Herzog, 1995; Livingston et al., 2020). Finally, most existing cognitive aging research has focused on older populations of high-income countries (Miyakawa et al., 2012; Archana Singh-Manoux et al., 2005; Zahodne et al., 2018). Little is known about the social epidemiology of cognitive aging in low- and middle-income countries, such as China, which has the largest population of older adults worldwide (Livingston et al., 2020; Zeng et al., 2017). Older adults in China are living in a context of rapid urbanization and socioeconomic development, which may exacerbate social stratification and income inequality (Jakovljevic, 2016; Jakovljevic et al., 2020; Tollman et al., 2016; Xie and Zhou, 2014).

We aimed to investigate the association between duration of subjective poverty from 2005 to 2011 and subsequent cognitive performance and rate of decline from 2011 to 2018 among adults aged ≥64 years in China. We hypothesized that a longer duration of subjective poverty, independent of the duration of objectively measured poverty, would be associated with (a) a lower level of cognitive performance and (b) a faster rate of cognitive decline. This hypothesis is informed by cognitive reserve theory and empirical evidence from studies in mostly high-income, Western settings, indicating that socioeconomic markers of cognitive reserve are associated with higher cognitive performance and delayed onset of Alzheimer's disease in later life (Jefferson et al., 2011; Qiu et al., 2001; Stern, 2002, 2009; Stern et al., 1994).

2. Method

2. 1. Data source, Study design, and sample collection

Data were from the China Longitudinal Healthy Longevity Survey (CLHLS) 2005-2018 (Zeng, 2008). The CLHLS is a nationally representative population-based cohort study administered by Duke University and Peking University since 1998, with follow-up interviews conducted in 2000, 2004, 2006, 2008, 2011, 2014, and 2018. The first two CLHLS waves incorporated individuals aged >80, while those aged 64-79 and those aged 40-59 were enrolled beginning in 2002 and 2008, respectively. The CLHLS has conducted face-to-face interviews with over 20,000 individuals from 22 provinces in mainland China, with detailed methods available elsewhere (Zeng, 2008). All data from the CLHLS are publicly available at https://sites.duke.edu/centerforaging/. The CLHLS was approved by the ethics committee of the Peking University (IRB00001052–13074). Written informed consent was obtained from all participants.

The present study employed a longitudinal design with exposure to subjective poverty assessed from 2005-2011 and cognitive performance assessed from 2011-2018. After excluding individuals who died before 2011, were lost to follow-up before 2011, or, who had incomplete subjective poverty data from 2005-2011, we included 4,118 individuals who contributed 8,188 observations from 2011-2018 (Figure 1).

Figure 1. Study sample.

Figure 1

Open in a new tab

Note: All participants who survived and remained in the CLHLS study or their proxies were directly asked about subjective measured poverty. The CLHLS consisted of 15,638 participants in 2005, who were either newly enrolled in 2005 or who had been recruited in an earlier wave and retained in 2005. We excluded participants with missing subjective poverty data in 2005, 2008, or 2011 and those who died or who were lost to follow-up before 2011.

2. 2. Measures

2. 2. 1. Outcome: Cognitive function from 2011-2018

We measured cognitive function in 2011, 2014, and 2018 using the Chinese Mini-Mental State Exam (CMMSE), which has been validated in prior research (Folstein et al., 1975; ZM. Zhang, 2006; Z. Zhang et al., 2010). The CMMSE contains 24 items that measure seven cognitive domains: Orientation, Registration, Naming foods, Attention and Calculation, Copying a figure, Recall, and Language (ZM. Zhang, 2006). CMMSE scores ranged from 0 to 30, with higher scores indicating a higher level of cognitive performance.

2. 2. 2. Exposure: Duration of subjective poverty from 2005-2011

We measured the duration of subjective poverty according to the number of CLHLS time points from 2005, 2008, and 2011 in which individuals reported "poor" or "very poor" compared with their neighbors on the following study interview question: "How do you rate your economic status compared with others in your local area?" The available response options were: "very rich", "rich", "so-so", "poor", and "very poor". We categorized the subjective poverty duration variable as "never", "one time point", and "two or three time points" from 2005 to 2011.

2. 2. 3. Covariates

We included the duration of objective poverty from 2005-2011 to examine the effect of subjective poverty duration, independent of objective income status. We measured the duration of objective poverty according to the number of the CLHLS time points (2005, 2008, and 2011) that individuals' household income was below the mean value of the bottom income quintile within their region of residence by urban vs. rural regions (categories of never; one time point; two or three time points), consistent with prior research (Yu et al., 2021). In urban regions of China, the mean annual household income of the bottom quintile group was 4,290 CNY (~$613 USD) in 2005, 6,560 CNY (~$937 USD) in 2008, and 9,785 CNY (~$1,398 USD) in 2011. In rural regions of China, the mean annual household income of the bottom quintile group was 2,090 CNY (~$299 USD) in 2005, 3,072 CNY (~$439 USD) in 2008, and 4,421 CNY (~$632 USD) in 2011. Hence, we classified objective poverty separately for those living in urban and rural regions to account for income differences between urban and rural regions. These population income data are from the China Statistical Yearbooks, published by the National Bureau of Statistics of China and available at http://www.stats.gov.cn (Supplemental Table 1).

Further, we included as potential confounders sociodemographic characteristics, lifestyle behaviors, comorbid disease history, and depressive symptoms, all assessed at baseline in 2005 (Livingston et al., 2020; Zahodne et al., 2018; Zeki Al Hazzouri et al., 2017; Zeki Al Hazzouri et al., 2011). Sociodemographic characteristics included age (in years), sex (female; male), ethnicity (Han; others), marital status (married; widowed; others), residence (urban vs. rural regions), and education (years of schooling). Lifestyle behaviors included body mass index (BMI; categories of <18.5; 18.5-24.9; 24.9-29.9; >29.9), self-reported regular physical activity (yes; no), and smoking history (yes; no). Self-reported comorbid disease history included hypertension (yes; no), diabetes (yes; no), stroke (yes; no), and cardiovascular disease (yes; no). As the CLHLS did not have direct measures of depressive symptoms in 2005, consistent with the CLHLS data codebook and previous research (Yin et al., 2020), we used answers to the following questions to reflect potential depressive symptoms: a) Do you often feel fearful or anxious? (b) Do you often feel lonely and isolated? (c) Do you feel the older you get, the more useless? We summed scores (range 3-15) to these three answers, which are on a five-point Likert scale with "1" for "always" and "5" for "never", with higher scores indicating fewer depressive symptoms.

2. 3. Statistical analyses

A total of 4,335/8,188 (52.94%) observations in our final analytic sample had missing CMMSE items. We performed multiple imputation by chained equations (MICE) on the full CLHLS 2005-2018 dataset to fill missing CMMSE values, consistent with previous studies (Gao et al., 2017; Godin et al., 2017). MICE is a sequential multivariable regression imputation approach that fills the missing values conditional upon all observed variables (Azur et al., 2011). We produced five imputed datasets and conducted MICE at the item-level (Burns et al., 2011; Godin et al., 2017; Zeki Al Hazzouri et al., 2011). All imputation regression models were conditional upon all variables in the final analytic models (Burns et al., 2011; Godin et al., 2017; Haan et al., 2011; Zeki Al Hazzouri et al., 2011). We calculated the final CMMSE scores after performing the item-level imputation (Godin et al., 2017).

To compare baseline characteristics of the study sample by the duration of subjective poverty, we performed Pearson chi-square tests, analysis of variance (ANOVA), and Kruskal-Wallis rank-sum tests in the imputed dataset. We calculated Spearman correlation coefficients to examine the collinearity between duration of subjective poverty and duration of objective poverty. To account for ceiling effects in the CMMSE, whereby levels of cognitive function beyond the measurement range of the CMMSE cannot be observed, we fitted mixed-effects Tobit models to examine the association between duration of subjective poverty and subsequent cognitive function and decline. Tobit regression, also called censored regression, assumes that the observed range of the dependent variable y is either left- or right-censored (as with a ceiling effect) and represents an unmeasured latent variable, y*, such that we only observe y = max(0, y*) (Amemiya, 1984; McDonald and Moffitt, 1980). Tobit regression coefficient allows estimation and inference of the effect of x on the uncensored latent y* via maximum likelihood, assuming that y* = xβ + ε, ε∣x ~ Normal(0, σ2). We included statistical interaction terms between year and duration of subjective poverty to examine the association between duration of subjective poverty and the rate of changes in CMMSE scores over time. Three models were sequentially adjusted for (1) duration of objective poverty from 2005-2011, baseline age, sex, ethnicity, marital status, and urban vs. rural residence (objective poverty and demographic adjustment); (2) years of schooling (objective poverty, demographic, and education adjustment); and (3) lifestyle behaviors and comorbid disease history (fully adjusted).

To minimize any potential bias from selective study attrition and mortality over the follow-up period, we created wave-specific inverse probability weights (IPWs) that jointly incorporated the probabilities of survival and retaining in the study (Glymour et al., 2012; Weuve et al., 2012). Full methodological details for the creation of the IPWs are provided in Supplemental Table 2 in the Supplemental Material.

We performed several sensitivity analyses to examine the robustness of our findings. First, as the duration of objective poverty in our analyses may not fully reflect income status over the entire bottom quintile of the population income distribution, we repeated the analyses conducting models adjusting for cumulative income levels from 2005-2011, rather than duration of objective poverty, to reflect absolute income status during the exposure period. Second, we repeated modeling analyses with the IPWs trimmed at the 99th percentiles (4.63) to exclude potential outsized effects of IPW outliers. Third, we repeated our analyses restricting the study sample to individuals with baseline CMMSE scores >28 in 2005 to rule out reverse causality, whereby individuals with pre-existing poor cognitive function may be more likely to experience and perceive sustained poverty. Finally, we repeated our analyses using the dataset without imputed CMMSE values to compare against the results with imputation. All analyses were performed with Stata/SE 15.0 (StataCorp, TX, USA).

3. Results

Overall, this longitudinal study included 4,118 individuals (mean [SD] age: 78.02 [9.38] in 2005). Table 1 provides baseline characteristics of the study sample, overall and according to the duration of subjective poverty from 2005-2011. A total of 2,675 (64.96%) individuals never reported subjective poverty during the period 2005-2011, 930 (22.58%) individuals reported subjective poverty at one time point, and 513 (12.46%) reported subjective poverty at two or three time points from 2005, 2008, and 2011. Supplemental Table 3 describes the proportion of individuals retained in the study from 2011-2018 according to duration of subjective poverty. Supplemental Table 4 compares the baseline characteristics of excluded individuals and the included study sample.

Table 1.

Baseline characteristics by duration of subjective poverty, CLHLS, China, 2005-2018 (N=4,118)

Baseline characteristics Total
(N=4,118)		 Duration of subjective poverty P value
Never
(n=2,675)		 One time point
(n=930)		 Two or three time
points (n=513)
CMMSE scores, mean (SE) 27.17 (0.06) 27.51 (0.08) 26.66 (0.15) 26.33 (0.20) <.001*
Duration of objective poverty
  Never 1,394 (33.85) 1,151 (43.03) 204 (21.94) 39 (7.60) <.001‡
  One time point 1,513 (36.74) 989 (36.97) 389 (41.83) 135 (26.32)
  Two or three time points 1,211 (29.41) 535 (20.00) 337 (36.24) 339 (66.08)
Age, mean (SD) 78.02 (9.38) 77.87 (9.32) 78.22 (9.55) 78.44 (9.41) .33*
Age, median 76 76 76 77 -
Age, range 64-108 64 -108 65-105 65-108 -
Men (vs. women), n (%) 1,918 (46.58) 1,293 (48.34) 402 (43.23) 223 (43.47) <.001†
Ethnicity (Han vs. others), n (%) 3,831 (93.03) 2,520 (94.21) 847 (91.08) 464 (90.45) <.001†
Marital status, n (%)
  Married 2,052 (49.83) 1,384 (51.74) 430 (46.24) 238 (46.39) <.001†
  Widowed 1,920 (46.62) 1,218 (45.53) 465 (50.00) 237 (46.20)
  Other 146 (3.55) 73 (2.73) 35 (3.76) 38 (7.41)
Urban (vs. rural) residence, n (%) 1,595 (38.73) 1,155 (43.18) 290 (31.18) 150 (29.24) <.001†
Years of schooling, mean (SE) 2.55 (0.06) 3.03 (0.08) 1.76 (0.10) 1.52 (0.11) <.001*
Regular physical activity (yes vs. no), n (%) 1,520 (36.91) 1,097 (41.01) 294 (31.61) 129 (25.15) <.001†
Smoking history (yes vs. no), n (%) 1,472 (35.75) 984 (36.79) 314 (33.76) 174 (33.92) .17†
BMI, n (%)
  <18.5 1,155 (28.05) 667 (24.93) 298 (32.04) 190 (37.04) <.001‡
  18.5 - 24.9 2,400 (58.28) 1,607 (60.07) 522 (56.13) 271 (52.83)
  24.9 - 29.9 482 (11.70) 348 (13.01) 90 (9.68) 44 (8.58)
  >29.9 81 (1.97) 53 (1.98) 20 (2.15) 8 (1.56)
Hypertension (yes vs. no), n (%) 855 (20.76) 571 (21.35) 168 (18.06) 116 (22.61) .06†
Diabetes (yes vs. no), n (%) 106 (2.57) 80 (2.99) 19 (2.04) 7 (1.36) .05†
Stroke (yes vs. no), n (%) 207 (5.03) 127 (4.75) 57 (6.13) 23 (4.48) .08†
Cardiovascular disease (yes vs. no), n (%) 394 (9.57) 276 (10.32) 77 (8.28) 41 (7.99) .33†
Depressive symptoms, mean (SE) 10.30 (0.03) 10.39 (0.03) 10.26 (0.07) 9.93 (0.10) <.001†
Open in a new tab

Note:

*

ANOVA, analysis of variance.

†

Pearson chi-square test.

‡

Kruskal-Wallis rank-sum tests. Significance tests were performed by using dataset after imputation. Missing rate before imputation: duration of objective poverty (22.07%), years of schooling (0.19%), BMI (0.50%), smoking history (0.06%), hypertension (5.09%), diabetes (6.47%), stroke (6.41%), cardiovascular disease (5.75%), and depressive symptoms (3.89%). Spearman correlation coefficient between duration of subjective poverty and duration of objective poverty: 0.34, P<.001.

Table 2 provides results from attrition- and mortality-weighted, multivariable mixed-effects Tobit models. Compared to those never in subjective poverty from 2005-2011, individuals in subjective poverty at one time point (β=−1.17, 95% CI=−1.70 to −0.64) and two or three time points (β=−2.21, 95% CI=−2.95 to −1.46) had lower CMMSE scores in 2011 (Model 1), indicating a dose-response relationship (Table 2). These associations were slightly attenuated to - 0.95 (95% CI=−1.48 to −0.41, for subjective poverty at one time point vs. never in subjective poverty) and −2.01 (95% CI=−2.73 to −1.29, for subjective poverty at two or three time points vs. never in subjective poverty) but remained statistically significant in the fully adjusted Model 3 (Table 2).

Table 2.

Attrition-weighted multivariable mixed-effects Tobit analyses of CMMSE scores, CLHLS, China, 2005-2018 (N=4,118)

Characteristics Model 1
Coefficient (95% CI)		 Model 2
Coefficient (95% CI)		 Model 3
Coefficient (95% CI)
Year
  2011 ref. ref. ref.
  2014 −1.00 (−1.46 to −0.54) −1.00 (−1.46 to −0.54) −1.01 (−1.47 to −0.55)
  2018 −2.32 (−2.83 to −1.80) −2.33 (−2.84 to −1.81) −2.33 (−2.85 to −1.82)
Duration of subjective poverty
  Never ref. ref. ref.
  One time point −1.17 (−1.70 to −0.64) −0.98 (−1.51 to −0.45) −0.95 (−1.48 to −0.41)
  Two or three time points −2.21 (−2.95 to −1.46) −2.01 (−2.75 to −1.28) −2.01 (−2.73 to −1.29)
Interaction terms
  2014×One time point −0.23 (−1.11 to 0.66) −0.21 (−1.10 to 0.67) −0.20 (−1.09 to 0.69)
  2014×Two or three time points 0.94 (−0.11 to 1.99) 0.94 (−0.11 to 1.99) 0.94 (−0.10 to 1.99)
  2018×One time point 0.25 (−0.69 to 1.19) 0.28 (−0.66 to 1.22) 0.29 (−0.66 to 1.24)
  2018×Two or three time points 1.45 (0.21 to 2.69) 1.45 (0.22 to 2.69) 1.44 (0.20 to 2.68)
Open in a new tab

Note: IPWs were included.

Model 1 adjusted for duration of objective poverty, baseline age, sex, ethnicity, marital status, and urban vs. rural residence.

Model 2 adjusted for duration of objective poverty, baseline age, sex, ethnicity, marital status, urban vs. rural residence, and years of schooling.

Model 3 adjusted for duration of objective poverty, baseline age, sex, ethnicity, marital status, urban vs. rural residence, years of schooling, physical activity, smoking history, BMI, hypertension, diabetes, stroke, cardiovascular disease, and depressive symptoms.

As suggested by the statistical interaction terms between year and duration of subjective poverty, individuals with a longer duration of subjective poverty from 2005-2011 had, on average, a slower rate of decline in CMMSE scores from 2011-2018 (Table 2 and Figure 2). Specifically, individuals in subjective poverty at two or three time points from 2005-2011 experienced a slower average rate of decline in CMMSE scores compared to those never in subjective poverty, although the statistical interaction term between year 2014 and subjective poverty at two or three time points was not statistically significant (Model 3: β=0.94, 95% CI=−0.10 to 1.99 for 2014×Two or three time points; β=1.44, 95% CI=0.20 to 2.68 for 2018×Two or three time points).

Figure 2. Predicted trajectories of CMMSE scores from 2011 to 2018 by duration of subjective poverty.

Figure 2.

Open in a new tab

Note: CMMSE scores were predicted by using estimates from Model 3 in Table 2. Covariates in Model 3 were set to the following levels: never in objective poverty, 82 years old, male, non-Han, married, rural region of residence, zero years of schooling, no smoking history, physical inactivity, BMI <18.5, no hypertension, no diabetes, no stroke, no cardiovascular disease, depressive symptoms = 3.

Results from the sensitivity analyses were consistent with our main findings (Supplemental Tables 5-9). First, results from models adjusting for cumulative income levels from 2005 to 2011 did not differ from those in our main models, which adjusted for duration of objective poverty (Supplemental Tables 5-6). Second, results from models trimmed at 99% percentiles of IPWs (4.63) were negligibly different from our main results (Supplemental Table 7). In addition, results from models restricted to cognitively healthy individuals at baseline (CMMSE scores >28) were consistent with our main findings and helped rule out potential reverse causality, whereby individuals with pre-existing poor cognitive function may have higher possibilities of experiencing and perceiving poverty (Supplemental Table 8). Finally, results from models using the non-imputed dataset were similar to those in our main analyses, but the estimates were less precise due to the restricted sample size (Supplemental Table 9).

4. Discussion

In this longitudinal study of adults aged ≥64 in China from 2005-2018, we observed that a longer duration of subjective poverty in late life, independent of objectively measured poverty duration, was associated with lower subsequent cognitive function in a dose-response fashion. We also observed that a longer duration of subjective poverty might lead to a slower rate of aging-related decline in CMMSE scores among middle-old and oldest-old adults in China. These results were independent of a range of sociodemographic, lifestyle, and health-related factors, most of which did not appear to be strong confounders of the observed relationships. Further research on the role of subjective poverty duration on subsequent cognitive performance and decline from diverse global settings is warranted.

4.1. Comparison with existing studies

Our findings are consistent with previous studies on the association between SSS and health outcomes such as self-reported health, depressive symptoms, diabetes, and psychological distress (Adler et al., 2000; Demakakos et al., 2008; Euteneuer, 2014; Miyakawa et al., 2012; Archana Singh-Manoux et al., 2005; Zahodne et al., 2018). Our findings indicate that the health effects of subjective and objective socioeconomic measures are not interchangeable with respect to cognitive function and rate of decline, although there are few studies investigating the role of SSS in cognitive aging (Kim et al., 2021; Zahodne et al., 2018). While the associations between objectively measured socioeconomic trajectories from childhood to mid-life and later-life brain health have been established (Grasset et al., 2019; Haan et al., 2011; Kobayashi and Feldman, 2019; Zeki Al Hazzouri et al., 2017; Zeki Al Hazzouri et al., 2011), our findings expand on existing knowledge by suggesting that improvement in subjective economic status relative to neighbors, even in late life, could help to maintain cognitive health and well-being among older adults.

Our results should be cautiously compared with the two existing studies of SSS and cognitive aging because of differences between our subjective poverty exposure measure and SSS as measured in these prior studies (Kim et al., 2021; Zahodne et al., 2018). We measured subjective poverty relative to neighbors. In contrast, these two studies employed the 10-rung MacArthur Scale, which asks participants to imagine that a 10-rung ladder represents "society" (or another reference frame) with the top ladder representing the people on top of society and the bottom ladder representing the people on the bottom of society, and to rank which ladder rung they feel they sit on (Adler et al., 2000). Zahodne et al. measured SSS in relation to United States society, finding that lower SSS was associated with lower baseline memory function but not rate of memory decline over time, partially consistent with our findings (Zahodne et al., 2018). In the context of Hong Kong, a highly urbanized city in China, Kim et al. measured SSS in relation to one's community and all of Hong Kong, finding that lower SSS-community, but not lower SSS-Hong Kong, was associated with worse cognitive function among older adults, consistent with our findings, although, in contrast to ours, they also found that lower SSS-community predicted more cognitive decline at a four-year follow-up (Kim et al., 2021). Taken together, these findings indicate that experiencing a higher subjective social position in late life may contribute to brain health, independently of objectively measured socioeconomic status.

To date, evidence on the relationships of objective and subjective measures of social or socioeconomic status and subsequent rate of cognitive decline is conflicting. Although higher socioeconomic status (e.g., education) has been associated with reduced risk and delayed onset of Alzheimer's disease (Qiu et al., 2001; Stern et al., 1994), our study identified that a longer duration of subjective poverty might lead to a slower rate of cognitive decline, in contrast to our second hypothesis and existing evidence on SSS in relation to cognitive decline (Kim et al., 2021; Zahodne et al., 2018). However, our findings are in line with evidence indicating that higher education is not protective of cognitive decline (Wilson et al., 2009; Zahodne et al., 2011), but rather associated with accelerated cognitive decline once brain degeneration is evident (Alley et al., 2007; Andel et al., 2006; Cadar et al., 2017; Clouston et al., 2020; Mungas et al., 2018; Yu et al., 2021). These empirical studies support the cognitive reserve hypothesis, whereby individuals with high cognitive reserve, which is often proxied as high education, are thought to experience delayed but accelerated cognitive deterioration relative to individuals with lower levels of cognitive reserve (Jefferson et al., 2011; Stern, 2002, 2009). Our findings support the compensation hypothesis of cognitive reserve theory, which posits that aging brains with accumulating neuropathology may work harder to compensate (Reuter-Lorenz and Cappell, 2008; Stern et al., 2020). An interpretation of our findings in line with this hypothesis suggests that the participants in our study who perceived longer durations of subjective poverty might have recruited brain structures and neural networks earlier in the cognitive aging process than those experiencing less subjective poverty, thus slowing down their rate of cognitive decline (Reuter-Lorenz and Cappell, 2008; Stern, 2002; Stern et al., 2020). Subjective measures of socioeconomic conditions may be a useful marker of cognitive reserve in middle-income settings, deserving further investigation.

4.2. Limitations and Strengths

This study has limitations. First, changes in CMMSE scores may not fully capture subtle declines in cognitive performance, especially at the higher end of the range of function. The MMSE is designed to detect cognitive impairment and demonstrates ceiling effects (Anstey and Christensen, 2000). We used Tobit regression models, which treat the outcome distribution as right censored, to help account for these ceiling effects. Second, we were unable to analyze multiple cognitive domains, potentially missing any domain-specific association between SSS and cognitive function (Glymour et al., 2012). Third, the association between subjective poverty duration and cognitive performance could be confounded by negative affectivity, which refers to the stable tendency to experience negative emotions regardless of the situation (Watson and Clark, 1984). However, previous studies have indicated that negative affectivity could be a mediator of the relationship between SSS and health, rather than a confounder (Adler et al., 2000; Miyakawa et al., 2012). Further, the CLHLS did not collect comprehensive data on personality, so we were unable to adjust for personality traits, such as agreeableness, extroversion, neuroticism, and conscientiousness, which have been associated with dementia risk and could be potential confounders of the relationship between subjective poverty and cognitive performance (Aschwanden et al., 2021; Terracciano et al., 2017; Wang et al., 2009). However, evidence on these associations is limited to Western populations, and the role of personality in cognitive performance among Chinese older adults is unknown. Our findings may be subject to selective survival bias, as we required participants to maintain survival in order to have complete SSS data from 2005-2011. Meanwhile, attrition bias may exist due to selective loss to follow-up or mortality as the study sample consisted of older adults. Our observed association between a longer duration of subjective poverty and a slower rate of cognitive decline could be biased, as individuals in sustained subjective poverty who survived over the follow-up could be more resilient than those who died earlier. However, we minimized this bias at best possible by applying inverse probability of attrition weights.

Our study has several strengths. To the best of our knowledge, this is one of the first longitudinal studies on the duration of subjective poverty in relation to subsequent brain health among older adults. Our sensitivity analysis restricting the study sample to cognitive healthy participants at baseline helped rule out reverse causality. Our findings expand on current knowledge about the role of subjective economic status after age 64, an understudied life course exposure period, and may provide valuable insight into promoting health equity in the context of global demographic transition to an aging population. To-date, the majority of the global evidence base on cognitive aging represents populations from high-income, Western countries. Our data improve representation in the cognitive aging evidence base by including middle-old and oldest-old adults in China, a middle-income country, providing new evidence on the social epidemiologic pattern of cognitive aging in this region (Livingston et al., 2020).

5. Conclusion

In this longitudinal study of adults aged 64 and over in China, we found that a longer duration of subjective poverty relative to neighbors in late life was associated with a lower subsequent level of cognitive function and a slower rate of cognitive decline over time. Our findings suggest that subjective economic status in late life may have unique and cumulative contributions to cognitive aging, independent of objectively measured economic status, in this middle-income setting. Future research from diverse global settings and to identify potential mechanisms is warranted.

Supplementary Material

1

Highlights.

  • One of the first longitudinal studies of subjective poverty and cognitive aging

  • This study focuses on the middle-old and oldest-old adults in China

  • A longer subjective poverty duration was associated with worse cognitive function

  • However, a longer subjective poverty duration may lead to slower cognitive decline

Acknowledgments

We thank Peking University and Duke University for providing CLHLS data. We would like to thank Dr. Fan Xia at the Department of Neurosurgery, West China Hospital, Sichuan University for modifying Figure 2.

Financial support

WZ was supported by 1·3·5 project for disciplines of excellence, West China Hospital, Sichuan University [grant number ZYJC18010]. LCK was supported by the National Institute on Aging at the National Institutes of Health [grant numbers R01AG069128, R01AG070953, and P30AG012846].

Footnotes

Disclaimer

The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Declaration of competing interest

None.

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Data availability

All data from the CLHLS are publicly available at https://sites.duke.edu/centerforaging/.

References

  1. Adler NE, Epel ES, Castellazzo G, et al. (2000). Relationship of subjective and objective social status with psychological and physiological functioning: preliminary data in healthy white women. Health Psychol, 19(6), 586–592. doi: 10.1037//0278-6133.19.6.586 [DOI] [PubMed] [Google Scholar]
  2. Adler NE, & Ostrove JM (1999). Socioeconomic status and health: what we know and what we don't. Ann N Y Acad Sci, 896, 3–15. doi: 10.1111/j.1749-6632.1999.tb08101.x [DOI] [PubMed] [Google Scholar]
  3. Alley D, Suthers K, & Crimmins E (2007). Education and cognitive decline in older Americans: results from the AHEAD sample. Res Aging, 29(1), 73–94. doi: 10.1177/0164027506294245 [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Amemiya T (1984). Tobit models: a survey. J Econom, 24(1), 3–61. doi: 10.1016/0304-4076(84)90074-5 [DOI] [Google Scholar]
  5. Andel R, Vigen C, Mack WJ, et al. (2006). The effect of education and occupational complexity on rate of cognitive decline in Alzheimer's patients. J Int Neuropsychol Soc, 12(1), 147–152. doi: 10.1017/s1355617706060206 [DOI] [PubMed] [Google Scholar]
  6. Anstey K, & Christensen H (2000). Education, activity, health, blood pressure and apolipoprotein E as predictors of cognitive change in old age: a review. Gerontology, 46(3), 163–177. doi: 10.1159/000022153 [DOI] [PubMed] [Google Scholar]
  7. Aschwanden D, Strickhouser JE, Luchetti M, et al. (2021). Is personality associated with dementia risk? A meta-analytic investigation. Ageing Res Rev, 67, 101269. doi: 10.1016/j.arr.2021.101269 [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Ayalon L (2019). Subjective social status as a predictor of loneliness: the moderating effect of the type of long-term care setting. Res Aging, 41(10), 915–935. doi: 10.1177/0164027519871674 [DOI] [PubMed] [Google Scholar]
  9. Azur MJ, Stuart EA, Frangakis C, et al. (2011). Multiple imputation by chained equations: what is it and how does it work? Int J Methods Psychiatr Res, 20(1), 40–49. doi: 10.1002/mpr.329 [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Burns RA, Butterworth P, Kiely KM, et al. (2011). Multiple imputation was an efficient method for harmonizing the Mini-Mental State Examination with missing item-level data. J Clin Epidemiol, 64(7), 787–793. doi: 10.1016/j.jclinepi.2010.10.011 [DOI] [PubMed] [Google Scholar]
  11. Cadar D, Robitaille A, Clouston S, et al. (2017). An international evaluation of cognitive reserve and memory changes in early old age in 10 European countries. Neuroepidemiology, 48(1-2), 9–20. doi: 10.1159/000452276 [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Chetty R, Stepner M, Abraham S, et al. (2016). The association between income and life expectancy in the United States, 2001-2014. JAMA, 315(16), 1750–1766. doi: 10.1001/jama.2016.422 [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Clouston SAP, Smith DM, Mukherjee S, et al. (2020). Education and cognitive decline: an integrative analysis of global longitudinal studies of cognitive aging. J Gerontol B Psychol Sci Soc Sci, 75(7), e151–e160. doi: 10.1093/geronb/gbz053 [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Dannefer D (2003). Cumulative advantage/disadvantage and the life course: cross-fertilizing age and social science theory. J Gerontol B Psychol Sci Soc Sci, 58(6), S327–S337. doi: 10.1093/geronb/58.6.S327 [DOI] [PubMed] [Google Scholar]
  15. Demakakos P, Nazroo J, Breeze E, et al. (2008). Socioeconomic status and health: the role of subjective social status. Soc Sci Med, 67(2), 330–340. doi: 10.1016/j.socscimed.2008.03.038 [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Diaz OV, Guendelman S, & Kuppermann M (2014). Subjective social status and depression symptoms: a prospective study of women with noncancerous pelvic problems. Womens Health Issues, 24(6), 649–655. doi: 10.1016/j.whi.2014.07.003 [DOI] [PubMed] [Google Scholar]
  17. Euteneuer F (2014). Subjective social status and health. Curr Opin Psychiatry, 27(5), 337–343. doi: 10.1097/yco.0000000000000083 [DOI] [PubMed] [Google Scholar]
  18. Evans IEM, Llewellyn DJ, Matthews FE, et al. (2019). Living alone and cognitive function in later life. Arch Gerontol Geriatr, 81, 222–233. doi: 10.1016/j.archger.2018.12.014 [DOI] [PubMed] [Google Scholar]
  19. Ferraro KF, Shippee TP, & Schafer MH (2009). Cumulative inequality theory for research on aging and the life course. In Handbook of Theories of Aging, 2nd ed. New York: Springer Publishing Company. [Google Scholar]
  20. Folstein MF, Folstein SE, & McHugh PR (1975). "Mini-mental state": a practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res, 12(3), 189–198. doi: 10.1016/0022-3956(75)90026-6 [DOI] [PubMed] [Google Scholar]
  21. Gao M, Kuang W, Qiu P, et al. (2017). The time trends of cognitive impairment incidence among older Chinese people in the community: based on the CLHLS cohorts from 1998 to 2014. Age Ageing, 46(5), 787–793. doi: 10.1093/ageing/afx038 [DOI] [PubMed] [Google Scholar]
  22. Garfein AJ, & Herzog AR (1995). Robust aging among the young-old, old-old, and oldest-old. J Gerontol B Psychol Sci Soc Sci, 50(2), S77–87. doi: 10.1093/geronb/50b.2.s77 [DOI] [PubMed] [Google Scholar]
  23. Gianaros PJ, Horenstein JA, Cohen S, et al. (2007). Perigenual anterior cingulate morphology covaries with perceived social standing. Soc Cogn Affect Neurosci, 2(3), 161–173. doi: 10.1093/scan/nsm013 [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Glymour MM, Tzourio C, & Dufouil C (2012). Is cognitive aging predicted by one's own or one's parents' educational level? results from the three-city study. Am J Epidemiol, 175(8), 750–759. doi: 10.1093/aje/kwr509 [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Godin J, Keefe J, & Andrew MK (2017). Handling missing Mini-Mental State Examination (MMSE) values: results from a cross-sectional long-term-care study. J Epidemiol, 27(4), 163–171. doi: 10.1016/j.je.2016.05.001 [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Grasset L, Glymour MM, Elfassy T, et al. (2019). Relation between 20-year income volatility and brain health in midlife: the CARDIA study. Neurology, 93(20), e1890–e1899. doi: 10.1212/WNL.0000000000008463 [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Haan MN, Zeki Al-Hazzouri A, & Aiello AE (2011). Life-span socioeconomic trajectory, nativity, and cognitive aging in Mexican Americans: the Sacramento Area Latino Study on Aging. J Gerontol B Psychol Sci Soc Sci, 66 Suppl 1, i102–110. doi: 10.1093/geronb/gbq071 [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Jakovljevic M (2016). Comparison of historical medical spending patterns among the BRICS and G7. J Med Econ, 19(1), 70–76. doi: 10.3111/13696998.2015.1093493 [DOI] [PubMed] [Google Scholar]
  29. Jakovljevic M, Sugahara T, Timofeyev Y, et al. (2020). Predictors of (in)efficiencies of healthcare expenditure among the leading Asian economies - comparison of OECD and non-OECD nations. Risk Manag Healthc Policy, 13, 2261–2280. doi: 10.2147/RMHP.S266386 [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Jefferson AL, Gibbons LE, Rentz DM, et al. (2011). A life course model of cognitive activities, socioeconomic status, education, reading ability, and cognition. J Am Geriatr Soc, 59(8), 1403–1411. doi: 10.1111/j.1532-5415.2011.03499.x [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Kawachi I, Kennedy BP, Lochner K, et al. (1997). Social capital, income inequality, and mortality. Am J Public Health, 87(9), 1491–1498. doi: 10.2105/ajph.87.9.1491 [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Kim JH, Sumerlin TS, Goggins WB, et al. (2021). Does low subjective social status predict cognitive decline in Chinese older adults? a 4-year longitudinal study from Hong Kong. Am J Geriatr Psychiatry. doi: 10.1016/j.jagp.2021.01.014 [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Kobayashi LC, & Feldman JM (2019). Employment trajectories in midlife and cognitive performance in later life: longitudinal study of older American men and women. J Epidemiol Community Health, 73(3), 232–238. doi: 10.1136/jech-2018-211153 [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Livingston G, Huntley J, Sommerlad A, et al. (2020). Dementia prevention, intervention, and care: 2020 report of the Lancet Commission. Lancet, 396(10248), 413–446. doi: 10.1016/S0140-6736(20)30367-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. McDonald JF, & Moffitt RA (1980). The uses of Tobit analysis. Rev Econ Stat, 62(2), 318–321. doi: 10.2307/1924766 [DOI] [Google Scholar]
  36. McNaull BB, Todd S, McGuinness B, et al. (2010). Inflammation and anti-inflammatory strategies for Alzheimer's disease-a mini-review. Gerontology, 56(1), 3–14. doi: 10.1159/000237873 [DOI] [PubMed] [Google Scholar]
  37. Melo S, Guedes J, & Mendes S (2019). Theory of cumulative disadvantage/advantage. In Encyclopedia of Gerontology and Population Aging (pp. 1–8). Cham: Springer International Publishing. [Google Scholar]
  38. Miyakawa M, Magnusson Hanson LL, Theorell T, et al. (2012). Subjective social status: its determinants and association with health in the Swedish working population (the SLOSH study). Eur J Public Health, 22(4), 593–597. doi: 10.1093/eurpub/ckr064 [DOI] [PubMed] [Google Scholar]
  39. Mungas D, Gavett B, Fletcher E, et al. (2018). Education amplifies brain atrophy effect on cognitive decline: implications for cognitive reserve. Neurobiol Aging, 68, 142–150. doi: 10.1016/j.neurobiolaging.2018.04.002 [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Phelan JC, Link BG, & Tehranifar P (2010). Social conditions as fundamental causes of health inequalities: theory, evidence, and policy implications. J Health Soc Behav, 51 Suppl, S28–40. doi: 10.1177/0022146510383498 [DOI] [PubMed] [Google Scholar]
  41. Pickett KE, & Wilkinson RG (2015). Income inequality and health: a causal review. Soc Sci Med, 128, 316–326. doi: 10.1016/j.socscimed.2014.12.031 [DOI] [PubMed] [Google Scholar]
  42. Prince M, Bryce R, Albanese E, et al. (2013). The global prevalence of dementia: a systematic review and metaanalysis. Alzheimers Dement, 9(1), 63–75.e62. doi: 10.1016/j.jalz.2012.11.007 [DOI] [PubMed] [Google Scholar]
  43. Qiu C, Bäckman L, Winblad B, et al. (2001). The influence of education on clinically diagnosed dementia incidence and mortality data from the Kungsholmen Project. Arch Neurol, 58(12), 2034–2039. doi: 10.1001/archneur.58.12.2034 [DOI] [PubMed] [Google Scholar]
  44. Reuter-Lorenz PA, & Cappell KA (2008). Neurocognitive aging and the compensation hypothesis. Curr Dir Psychol Sci, 17(3), 177–182. doi: 10.1111/j.1467-8721.2008.00570.x [DOI] [Google Scholar]
  45. Santini ZI, Jose PE, York Cornwell E, et al. (2020). Social disconnectedness, perceived isolation, and symptoms of depression and anxiety among older Americans (NSHAP): a longitudinal mediation analysis. Lancet Public Health, 5(1), e62–e70. doi: 10.1016/S2468-2667(19)30230-0 [DOI] [PubMed] [Google Scholar]
  46. Singh-Manoux A, Adler NE, & Marmot MG (2003). Subjective social status: its determinants and its association with measures of ill-health in the Whitehall II study. Soc Sci Med, 56(6), 1321–1333. doi: 10.1016/s0277-9536(02)00131-4 [DOI] [PubMed] [Google Scholar]
  47. Singh-Manoux A, Marmot MG, & Adler NE (2005). Does subjective social status predict health and change in health status better than objective status? Psychosom Med, 67(6), 855–861. doi: 10.1097/01.psy.0000188434.52941.a0 [DOI] [PubMed] [Google Scholar]
  48. Stern Y (2002). What is cognitive reserve? Theory and research application of the reserve concept. J Int Neuropsychol Soc, 8(3), 448–460. [PubMed] [Google Scholar]
  49. Stern Y (2009). Cognitive reserve. Neuropsychologia, 47(10), 2015–2028. doi: 10.1016/j.neuropsychologia.2009.03.004 [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Stern Y, Arenaza-Urquijo EM, Bartrés-Faz D, et al. (2020). Whitepaper: defining and investigating cognitive reserve, brain reserve, and brain maintenance. Alzheimers Dement, 16(9), 1305–1311. doi: 10.1016/j.jalz.2018.07.219 [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Stern Y, Gurland B, Tatemichi TK, et al. (1994). Influence of education and occupation on the incidence of Alzheimer's disease. JAMA, 271(13), 1004–1010. [PubMed] [Google Scholar]
  52. Terracciano A, Stephan Y, Luchetti M, et al. (2017). Personality traits and risk of cognitive impairment and dementia. J Psychiatr Res, 89, 22–27. doi: 10.1016/j.jpsychires.2017.01.011 [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. Tollman SM, Norris SA, & Berkman LF (2016). Commentary: the value of life course epidemiology in low- and middle-income countries: an ageing perspective. Int J Epidemiol, 45(4), 997–999. doi: 10.1093/ije/dyw109 [DOI] [PMC free article] [PubMed] [Google Scholar]
  54. Wang HX, Karp A, Herlitz A, et al. (2009). Personality and lifestyle in relation to dementia incidence. Neurology, 72(3), 253. doi: 10.1212/01.wnl.0000339485.39246.87 [DOI] [PubMed] [Google Scholar]
  55. Watson D, & Clark LA (1984). Negative affectivity: the disposition to experience aversive emotional states. Psychol Bull, 96(3), 465–490. [PubMed] [Google Scholar]
  56. Weuve J, Tchetgen Tchetgen EJ, Glymour MM, et al. (2012). Accounting for bias due to selective attrition: the example of smoking and cognitive decline. Epidemiology, 25(1), 119–128. doi: 10.1097/EDE.0b013e318230e861 [DOI] [PMC free article] [PubMed] [Google Scholar]
  57. Wilson RS, Hebert LE, Scherr PA, et al. (2009). Educational attainment and cognitive decline in old age. Neurology, 72(5), 460–465. doi: 10.1212/01.wnl.0000341782.71418.6c [DOI] [PMC free article] [PubMed] [Google Scholar]
  58. Xie Y, & Zhou X (2014). Income inequality in today's China. Proc Natl Acad Sci U S A, 111(19), 6928–6933. doi: 10.1073/pnas.1403158111 [DOI] [PMC free article] [PubMed] [Google Scholar]
  59. Yin S, Yang Q, Xiong J, et al. (2020). Social support and the incidence of cognitive impairment among older adults in China: findings from the Chinese Longitudinal Healthy Longevity Survey Study. Front Psychiatry, 11, 254–254. doi: 10.3389/fpsyt.2020.00254 [DOI] [PMC free article] [PubMed] [Google Scholar]
  60. Yu X, Zhang W, & Kobayashi LC (2021). Duration of poverty and subsequent cognitive function and decline among older adults in China, 2005–2018. Neurology, 97(7), 1–8. doi: 10.1212/wnl.0000000000012343 [DOI] [PMC free article] [PubMed] [Google Scholar]
  61. Zahodne LB, Glymour MM, Sparks C, et al. (2011). Education does not slow cognitive decline with aging: 12-year evidence from the victoria longitudinal study. J Int Neuropsychol Soc, 17(6), 1039–1046. doi: 10.1017/S1355617711001044 [DOI] [PMC free article] [PubMed] [Google Scholar]
  62. Zahodne LB, Kraal AZ, Zaheed A, et al. (2018). Subjective social status predicts late-life memory trajectories through both mental and physical health pathways. Gerontology, 466–474. doi: 10.1159/000487304 [DOI] [PubMed] [Google Scholar]
  63. Zeki Al Hazzouri A, Elfassy T, Sidney S, et al. (2017). Sustained economic hardship and cognitive function: the Coronary Artery Risk Development in Young Adults Study. Am J Prev Med, 52(1), 1–9. doi: 10.1016/j.amepre.2016.08.009 [DOI] [PMC free article] [PubMed] [Google Scholar]
  64. Zeki Al Hazzouri A, Haan MN, Kalbfleisch JD, et al. (2011). Life-course socioeconomic position and incidence of dementia and cognitive impairment without dementia in older Mexican Americans: results from the Sacramento area Latino study on aging. Am J Epidemiol, 173(10), 1148–1158. doi: 10.1093/aje/kwq483 [DOI] [PMC free article] [PubMed] [Google Scholar]
  65. Zeng Y (2008). Introduction to the Chinese Longitudinal Healthy Longevity Survey (CLHLS). In Healthy Longevity in China. Demographic Methods and Population Analysis. Dordrecht: Springer. [Google Scholar]
  66. Zeng Y, Feng Q, Hesketh T, et al. (2017). Survival, disabilities in activities of daily living, and physical and cognitive functioning among the oldest-old in China: a cohort study. Lancet, 389(10079), 1619–1629. doi: 10.1016/S0140-6736(17)30548-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
  67. Zhang Z (2006). Gender differentials in cognitive impairment and decline of the oldest old in China. J Gerontol B Psychol Sci Soc Sci, 61(2), S107–S115. doi: 10.1093/geronb/61.2.S107 [DOI] [PubMed] [Google Scholar]
  68. Zhang Z, Gu D, & Hayward MD (2010). Childhood nutritional deprivation and cognitive impairment among older Chinese people. Soc Sci Med, 71(5), 941–949. doi: 10.1016/j.socscimed.2010.05.013 [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

1
NIHMS1721480-supplement-1.docx (62.3KB, docx)

Data Availability Statement

All data from the CLHLS are publicly available at https://sites.duke.edu/centerforaging/.

RESOURCES