The Mediation Effect of Socioeconomic Status in Childhood Parental Literacy and Cognitive Function in Middle-aged and Older People Based on "Guangzhou Biobank Cohort Study"

Chai Zhihao¹, Zhang Weisen²*, Luo Jiaoling¹, Zhu Feng², Zhu Tong², Jin Yali², Pan Jing², Lu Yingjun², Jiang Chaoqiang²

¹Molecular Epidemiology Research Center, the Affiliated Guangzhou Twelfth People's Hospital of Guangzhou Medical University, Guangzhou 510620, Guangdong, China
²Team of Guangzhou Biobank Cohort Study, Guangzhou Twelfth People's Hospital, Guangzhou 510620, Guangdong, China

Corresponding author: ZHANG Weisen, Chief physician; E-mail: zwsgzcn@163.com

[Abstract]
Background: Cognitive impairment is one of the serious risk factors leading to disability and death in the elderly. Some studies have found an association between parental literacy and cognitive function, but research on the mediating role of socioeconomic status (SES) between the two is relatively scarce.
Objective: To examine the mediating role of SES between childhood parental literacy (CPL) and late-life cognitive function.
Methods: Based on the baseline data in phase 3 of Guangzhou Biobank Cohort Study (GBCS), 8,891 Guangzhou residents aged 50 and above were included. Data on general demographic characteristics, SES, lifestyle, CPL, health conditions, and cognitive function were collected. R 4.1.1 software was used to analyze the mediating effects of SES and its representative indicators (education, family income, and occupation) between CPL and cognitive function. Pearson correlation analysis was employed to examine the associations among CPL, SES, and cognitive function. Multivariable linear regression was used to examine the association between CPL and cognitive function, and logistic regression model was used to examine the associations of CPL with MCI and MMI.
Results: The prevalence of mild cognitive impairment among the study subjects was 12.60% (1,120/8,891). Correlation analysis showed that childhood parental literacy was positively correlated with SES (rs=0.312, P<0.001), cognitive function (rs=0.306, P<0.001), and short-term memory function (rs=0.218, P<0.001); and SES was positively correlated with cognitive function (rs=0.337, P<0.001) and short-term memory function (rs=0.240, P<0.001). After adjustment for covariates, multivariable linear and logistic regression analysis showed that, compared with both parents couldn't read/write, only father/mother could read/write and both parents could read/write were associated with higher cognitive function [β(95%CI)=1.00(0.84~1.14), 1.46(1.30~1.62)] and short-term memory function [β(95%CI)=0.44(0.33~0.56), 0.64(0.52~0.77)] (all P<0.001), and the risks of MCI [OR(95%CI)=0.48(0.41~0.57), 0.27(0.22~0.34)] and MMI [OR(95%CI)=0.65(0.56~0.74), 0.50(0.43~0.59)] (P<0.001) were lower, the increases and decreases exhibit a certain trend (Ptrend<0.001). The proportions of the mediating effects of SES and its representative indicators (education, family income, and occupation) between CPL and cognitive function were 21.77% (18.79%~24.77%), 18.44% (15.49%~21.50%), 6.56% (5.08%~8.24%) and 6.67% (5.07%~8.34%), respectively, and were 25.51% (21.10%~30.61%), 23.65% (19.25%~29.03%), 7.03% (5.14%~9.85%) and 7.02% (4.74%~9.68%), respectively, between CPL and short-term memory function.
Conclusion: SES, mainly education, may partially mediate the association between childhood parental literacy and cognitive function. Strengthening early education and socioeconomic status (especially education) could help delay and prevent the occurrence of cognitive impairment in middle and old age.

[Key words] Cognitive function; Childhood parental literacy; Socioeconomic status; Middle-aged and older people; Guangzhou Biobank Cohort Study; Mediation effects

As global population aging intensifies, health issues among older adults have become increasingly prominent. Cognitive impairment and dementia are closely associated with disability and mortality in older adults. Cognitive impairment refers to varying degrees of functional damage in one or more cognitive domains caused by various factors. The global prevalence of mild cognitive impairment (MCI) exceeds 15%, and the number of dementia patients has surpassed 55 million, with projections indicating a doubling by 2050. In China, the prevalence of MCI among adults is approximately 15.5%, with about 20% of MCI patients progressing to dementia over time. Cognitive impairment and dementia not only affect the quality of life of middle-aged and older patients but also impose heavy economic burdens on society and families, representing an important public health issue that urgently needs to be addressed. Therefore, early identification and intervention of potential risk factors for cognitive decline are crucial for delaying and preventing the occurrence of cognitive impairment and dementia in middle-aged and older adults.

Research has shown that both childhood parental literacy (CPL) and socioeconomic status (SES) are closely associated with cognitive function. However, few studies have examined whether the association between CPL and cognitive function is mediated by SES and what the magnitude of this mediation effect might be. Therefore, this study utilized data from the Guangzhou Biobank Cohort Study (GBCS) to explore the associations among CPL, SES, and cognitive function in middle-aged and older adults in Guangzhou from a life-course perspective. We further analyzed the mediating effects of SES and its representative indicators—education level, annual family income, and occupation—on the relationship between CPL and cognitive function in middle-aged and older adults, aiming to provide scientific evidence for preventing and delaying cognitive impairment and dementia through targeted interventions on modifiable factors.

1.1 Study Participants

GBCS is an ongoing population-based prospective cohort study conducted through collaboration among Guangzhou Twelfth People's Hospital, the University of Hong Kong, and the University of Birmingham. All participants were recruited from the Guangzhou Elderly Health and Joy Association, the largest elderly civic organization in Guangzhou with branches in all districts. The baseline survey was conducted from September 2003 to January 2008 in three phases. Phase 3 baseline survey recruited 10,088 participants and added the Mini-Mental State Examination (MMSE) to assess multidimensional cognitive function based on the Delayed 10-Word Recall Test (DWRT). This study selected participants recruited in Phase 3 baseline survey. Inclusion criteria were: Guangzhou residents aged 50 years or older who voluntarily participated and could travel to Guangzhou Twelfth People's Hospital for questionnaire survey and free physical examination. Exclusion criteria were: individuals with mental or psychological disorders who could not communicate normally, those who could not cooperate to complete the questionnaire, those lacking data on CPL, SES, MMSE, or DWRT, and those with missing values in relevant covariates. Finally, 8,891 middle-aged and older adults were included. This study was approved by the Guangzhou Medical Ethics Committee (Approval No. 2017002). All participants understood the study content, voluntarily enrolled, and signed informed consent forms.

1.2 Methods

Baseline survey used standardized electronic questionnaires administered through face-to-face interviews by 6-8 research assistants who had undergone rigorous training and passed qualification assessments. Survey content included general demographic characteristics (gender, age, marital status, etc.), socioeconomic status (education level, annual family income, occupation), lifestyle (smoking, alcohol consumption, physical activity, etc.), health status, and cognitive function. Physical examinations were conducted using standard methods to measure height, weight, and other indicators.

1.3 Variables

1.3.1 Outcome Variables (Cognitive Function)

The MMSE and DWRT are widely used for cognitive function assessment due to their reliability and feasibility. The MMSE includes four dimensions: orientation, memory, attention and calculation, and language ability, comprising 11 items with scores ranging from 0 to 30. Higher scores indicate better cognitive function, with scores below 25 defined as MCI. The DWRT assesses short-term memory function through a 10-word delayed recall method, with scores ranging from 0 to 10. Higher scores indicate better short-term memory function, with scores below 5 defined as mild short-memory impairment (MMI).

1.3.2 Independent Variable (Childhood Parental Literacy)

In the historical context when most parents lacked access to formal education, the ability to read/write was a primary indicator of literacy. This study used childhood parental literacy (CPL) as the independent variable, categorized into three groups: both parents illiterate (0 points), one parent literate (1 point), and both parents literate (2 points). The score range was 0-2, with higher scores representing higher CPL.

1.3.3 Mediator Variable (SES)

SES is a comprehensive concept that cannot be fully captured by a single indicator. This study used three indicators—education level, annual family income, and occupation—to measure SES. These indicators are interrelated yet distinct, reflecting different facets of socioeconomic status in middle-aged and older adults. Following previous studies, education was dichotomized into junior high school and below (0 points) versus high school and above (1 point); annual family income was dichotomized into <30,000 yuan (0 points) versus ≥30,000 yuan (1 point); and occupation was dichotomized into manual labor or other (0 points) versus non-manual labor (1 point). The sum of these three indicators formed the SES score (0-3 points), with higher scores representing higher SES. Education, annual family income, occupation, and overall SES were analyzed as mediator variables in the relationship between CPL and cognitive function/short-term memory function.

1.3.4 Covariates

This study also included general demographic characteristics (gender, age, marital status), lifestyle (smoking status, alcohol consumption, physical activity), and health status (BMI, cardiovascular disease, diabetes, self-rated health, depressive symptoms) as covariates. Physical activity was classified into three groups—insufficient, moderate, and active—based on activity days and metabolic energy values using the International Physical Activity Questionnaire short form. BMI was calculated as weight (kg) divided by height (m) squared. Self-rated health was assessed through the question: "Overall, how would you rate your health status?" Responses of "very good" or "good" were classified as good self-rated health, while "poor" or "very poor" were classified as poor self-rated health. Depressive symptoms were evaluated using the validated Chinese version of the 15-item Geriatric Depression Scale (GDS-15), with scores ranging from 0 to 15. Higher scores indicate more severe depressive symptoms, with total scores <5 considered normal and ≥5 indicating depressive symptoms.

1.4 Statistical Methods

SPSS 25.0 software was used for statistical analysis. Categorical data were described as frequencies (percentages) and compared between groups using chi-square tests. Continuous data were normally distributed and described as means (standard deviations), with comparisons between two groups using independent samples t-tests. Pearson correlation analysis was used to explore associations among CPL, SES, and cognitive function. Multivariable linear regression models examined the association between CPL and cognitive function, while logistic regression models examined associations with MCI and MMI. Mediation analysis was conducted using the "mediation" and "bootstrap" packages in R 4.1.1 software. Sensitivity analysis was further performed by excluding participants with poor cognitive and short-term memory function (below the 5th percentile). A 95% confidence interval (CI) not containing 0 indicated statistically significant mediation effects. P<0.05 was considered statistically significant.

2.1 Basic Characteristics of Study Participants

This study included 8,891 middle-aged and older adults, comprising 2,226 males (25.0%) and 6,665 females (75.0%), with a mean age of 59.8±7.7 years. Among them, 7,771 had normal cognitive function and 1,120 had MCI. The two groups showed statistically significant differences in age, gender, marital status, childhood parental literacy, education level, annual family income, occupation, alcohol consumption, physical activity, prevalence of poor self-rated health, and depressive symptoms (P<0.05), but no significant differences in BMI, smoking status, or diabetes prevalence (P>0.05). For short-term memory function, 7,085 participants had normal function and 1,806 had MMI. The two groups showed statistically significant differences in age, gender, marital status, childhood parental literacy, education level, annual family income, occupation, smoking status, alcohol consumption, physical activity composition, and prevalence of poor self-rated health and depressive symptoms (P<0.05), but no significant differences in BMI or prevalence of cardiovascular disease and diabetes (P>0.05). See [TABLE:1].

2.2 Correlation Analysis

Correlation analysis showed that childhood parental literacy was positively correlated with SES (rs=0.312, P<0.001), cognitive function (rs=0.306, P<0.001), and short-term memory function (rs=0.218, P<0.001). SES was also positively correlated with cognitive function (rs=0.337, P<0.001) and short-term memory function (rs=0.240, P<0.001). See [TABLE:2].

2.3 Effects of Childhood Parental Literacy on Cognitive Function

Using childhood parental literacy as the independent variable (scored as: both parents illiterate=0, one parent literate=1, both parents literate=2) and cognitive function (MMSE score, actual value), short-term memory function (DWRT score, actual value), MCI (no=0, yes=1), and MMI (no=0, yes=1) as dependent variables, multivariable linear regression and logistic regression analyses were performed. After adjusting for general demographic characteristics (gender, age, marital status), socioeconomic status (education level, annual family income, occupation), lifestyle (smoking, alcohol consumption, physical activity), and health status (BMI, cardiovascular disease, diabetes, self-rated health, depressive symptoms), results showed that compared with the group where both parents were illiterate, groups where one parent was literate and where both parents were literate had higher cognitive function (MMSE scores) and short-term memory function (DWRT scores) (P<0.001), and lower risks of MCI and MMI (P<0.001). These increases and decreases exhibited significant trends (Ptrend<0.001). See [TABLE:3].

2.4 Mediation Effect Analysis

After adjusting for the aforementioned covariates, the mediation effects of SES and its representative indicators—education level, annual family income, and occupation—on the relationship between childhood parental literacy and cognitive function (MMSE score) were all statistically significant (P<0.001), accounting for 21.77% (18.79%~24.77%), 18.44% (15.49%~21.50%), 6.56% (5.08%~8.24%), and 6.67% (5.07%~8.34%) of the total effect, respectively. Consistent results were observed for short-term memory function (DWRT score), with mediation effects accounting for 25.51% (21.10%~30.61%), 23.65% (19.25%~29.03%), 7.03% (5.14%~9.85%), and 7.02% (4.74%~9.68%), respectively. The mediation pathways from childhood parental literacy to SES to middle-aged and older adult cognitive function and short-term memory function are shown in [FIGURE:1].

2.5 Sensitivity Analysis

After excluding participants with poor cognitive and short-term memory function (scores below the 5th percentile), sensitivity analysis results remained consistent. Compared with the group where both parents were illiterate, groups where one parent was literate and where both parents were literate had higher cognitive function levels (MMSE scores), with β(95%CI) of 0.56(0.44~0.67) and 0.92(0.79~1.04), respectively, and lower risks of MCI (MMSE<25), with OR(95%CI) of 0.51(0.42~0.61) and 0.33(0.26~0.42) (P<0.001), Ptrend<0.001. Consistent results were also observed for short-term memory function (DWRT score) and MMI, with β(95%CI) of 0.23(0.13~0.33) and 0.37(0.27~0.48), and OR(95%CI) of 0.71(0.59~0.85) and 0.59(0.48~0.73) (P<0.001), Ptrend<0.001.

This study, based on GBCS Phase 3 baseline data, analyzed the association between childhood parental literacy and cognitive function/short-term memory function in middle-aged and older adults, adjusted for relevant covariates, and used two cognitive scales (MMSE, DWRT) to corroborate findings, yielding objective and reasonable results. We further examined the mediating effects of education level, annual family income, occupation, and overall SES. Results showed that childhood parental literacy was positively associated with cognitive function and short-term memory function in middle-aged and older adults, consistent with previous studies and emphasizing the importance of early education for long-term cognitive development. Additionally, SES and its representative indicators partially mediated these relationships, with SES accounting for 21.77% and 25.51% of the mediation effects, respectively. Education level showed the strongest mediating effect (18.44% and 23.65%), while annual family income (6.56%) and occupation (6.67%) showed relatively smaller mediation effects. After including mediator variables and adjusting for covariates, childhood parental literacy still had direct effects on cognitive function and short-term memory function in middle-aged and older adults, with approximately one-fifth and one-quarter of the effects mediated through SES, suggesting that SES may play a key regulatory role in long-term cognitive development by integrating various resources and environmental factors.

Research on the cumulative impact mechanism of SES on cognitive function in middle-aged and older adults from a life-course perspective is receiving increasing attention, but studies on the mediating role of SES between childhood parental literacy and cognitive function are scarce and have mostly focused on single SES indicators (e.g., education, income, or occupation), lacking in-depth exploration of comprehensive SES indicators. Studies in different populations show significant regional variations. A cohort study of British civil servants found no direct effect of childhood SES (including mother's education, father's education, father's social class, and childhood family economic status) on cognitive function, but significant indirect effects through education level and adult SES (including income and occupation). The mediating effects of education level were significantly higher than those of occupation. A cross-sectional study using the Latino Aging Study and the Mexican Health and Aging Study found that education level mediated 71.4% and 81.0% of the relationship between maternal and paternal education and short-term memory function, respectively. A cross-sectional study using data from the China Health and Retirement Longitudinal Study found that when the number of childhood adverse experiences was 1, cognitive reserve (including education, occupation, and leisure activities) mediated 31.19% of the relationship between childhood adverse experiences and cognitive function; when the number of adverse experiences was ≥2, the mediation effect was 40.27%. Another cohort study using Chinese survey data found that education level mediated 77.94% of the relationship between childhood SES (including childhood family economic status, parental education, household registration type at age 12, and occupational status) and cognitive function. These differences may be related to study population characteristics (e.g., age, gender, cultural background), research methods (cross-sectional vs. cohort studies), and selection of mediator variables (e.g., considering only education, income, or occupation vs. comprehensive indicators).

The "cognitive reserve hypothesis" suggests that early education can enhance the brain's adaptive capacity to degenerative changes. Parental educational investment in early life may positively influence early cognitive stimulation, promoting changes in brain structure and function, such as increased cortical thickness and dendritic branching. These changes occur during critical periods of brain development and neural maturation (from infancy to late adolescence), and the resulting advantages may help delay or mitigate cognitive decline in later life. Therefore, the early educational environment created by parents may have important implications for long-term cognitive health. Meanwhile, the "cumulative disadvantage theory" emphasizes that childhood is a critical life stage for the formation of later social inequalities. From a risk chain perspective, childhood adversity (e.g., poor early education, economic disadvantage, and traumatic stress) may affect individual development, resource acquisition, and social status, leading to disadvantages in socioeconomic status such as education, income, and occupation, ultimately exerting long-term negative effects on cognitive function in later life. Based on this, from a public health perspective, greater emphasis should be placed on early education, promotion of equitable educational resource allocation, and special attention to middle-aged and older adults with lower socioeconomic status through measures such as promoting re-employment of older adults, improving the social security system, and increasing pension levels to reduce or prevent the occurrence of cognitive impairment in middle-aged and older adults.

This study has several limitations. First, the independent variable (childhood parental literacy) was obtained through self-report, which may be subject to recall bias. However, this study excluded individuals with mental or psychological disorders who could not communicate normally, which may reduce errors caused by recall bias to some extent. The consistent results from sensitivity analysis after excluding participants with poor cognitive and short-term memory function further support this. Second, genetic factors that may affect cognitive function were not considered and should be examined in future research.

In conclusion, childhood parental literacy is positively associated with cognitive function and short-term memory function in middle-aged and older adults, and SES and its representative indicators (education level, annual family income, occupation) partially mediate these associations. From a life-course perspective and considering individual, family, and societal levels, strengthening early education and improving socioeconomic status in middle-aged and older adults (especially education level) will help delay and prevent the occurrence of cognitive dysfunction.

Funding: National Key R&D Program of China (2017YFC0907105); Guangdong Medical Science and Technology Research Fund (B2024194)

Citation: CHAI Z H, ZHANG W S, LUO J L, et al. The mediation effect of socioeconomic status in childhood parental literacy and cognitive function in middle-aged and older people based on "Guangzhou Biobank Cohort Study" [J]. Chinese General Practice, 2025. DOI: 10.12114/j.issn.1007-9572.2025.0177. [Epub ahead of print].

Author Contributions: CHAI Zhihao was responsible for data cleaning, analysis, and manuscript writing and revision; ZHANG Weisen was responsible for data collection, research conceptualization, guidance on manuscript writing and revision, and final version editing; LUO Jiaoling was responsible for data analysis and manuscript writing; ZHU Feng, ZHU Tong, JIN Yali, PAN Jing, LU Yingjun, and JIANG Chaoqiang were responsible for data collection and manuscript review and revision.

Conflict of Interest: The authors declare no conflict of interest.

CHAI Zhihao https://orcid.org/0009-0009-4598-3577
ZHANG Weisen https://orcid.org/0000-0002-6691-4409