Samples
The data for this study were obtained from the Chinese Longitudinal Healthy Longevity Survey (CLHLS)[18]. CLHLS collected longitudinal data coordinated by the Center for Healthy Aging and Development Studies of National School of Development at Peking University. The baseline survey was conducted in 1998 and the follow-up surveys were conducted in 2000, 2002, 2005, 2008-2009, 2011-2012, 2014 and 2017-2018 in randomly selected about half of the counties and city districts in 23 Chinese provinces. The questionnaire data collected provides information on family structure, living arrangements and proximity to children, activities of daily living (ADL), the capacity of physical performance, self-rated health, self-evaluation of life satisfaction, cognitive functioning, chronic disease prevalence, care needs and costs, social activities, diet, smoking and drinking behaviors, psychological characteristics, economic resources, and caregiving and family support among elderly respondents and their relatives[19]. All the survey data are publicly available. This study used the cross-sectional data from the latest follow-up survey (2017-2018). Initially, 15,874 samples were included. After excluding duplicate values, missing values, and outliers with logical problems, 7595 participants at age 65 or older were obtained. The samples were well representative.
Variables and instruments
Depressive symptoms
The 10-item of the Center for Epidemiologic Studies Depression Scale (CES-D-10) was used to measure depressive symptoms in this study. The CES-D-10 consisted of 10 items using a 4-point Likert scale. For the two positive questions, “I was happy” and “I felt hopeful about the future,” answers were reversely coded before summation. We then coded all answers from 0 to 3 as “rarely” to “most of the time,” respectively. The total range of CES-D-10 scores in this study was 0–30, with higher scores indicating greater severity of depressive symptoms. A person is considered to have depressive symptoms if he/she scored no <10 on the CESD-10[20]. Previous studies have confirmed the reliability and effectiveness of CESD-10 in measuring depressive symptoms in older adults[21].
Socioeconomic status
We measured SES by the participant's education level, occupation, and economic status. Education level was classified as illiterate (no schooling), primary and junior high school (schooling for 1-9 years) and high school or above (schooling for 10 years or above). The occupation was classified as senior practitioners (professional and technical personnel/doctors/teachers, administration or military), intermediate practitioners (general staff/service staff/workers), and general practitioners (freelancer, farmer, domestic worker or unemployed person) according to the participant's occupation before the age of 60 years. We used self-rated economic status to represent the economic status of the participants, which was classified as poor, fair, and good. We used factor analysis to synthesize the three indicators to provide a composite score to represent the socioeconomic status of the participants. SES composite scores in this study ranged from -1.46 to 2.59, with higher scores indicating a higher socioeconomic status of the participants.
Cognitive function
We used the Mini Mental State Examination (MMSE) from the CLHLS questionnaire to assess the cognitive function of the participants. The scale included 24 items in five aspects: general ability, reaction ability, attention and calculation ability, recall ability and language ability, understanding and self-coordination ability. The total score of the MMSE scale in this study ranged from 0 to 30, with higher scores indicating a better cognitive function of the participants. The validity and reliability of the MMSE scale have been verified in several previous studies[22, 23].
Lifestyle
The lifestyle score of the participants was constructed by collecting information on diet, smoking, alcohol consumption, exercise, and sleep in the CLHLS questionnaire. Diet was defined by 2 questions— “Do you often eat fresh vegetables?” and “Do you often eat fresh fruit?”. A score of 0 was assigned if the participant answered “rarely or never” or “sometimes”; A score of 1 was assigned if the participant answered “eat every day/almost every day” or “eat often”. Smoking status of the participants was categorized as smoking (assigned a score of 0) or not smoking (assigned a score of 1). Drinking status was categorized as drinking (assigned a score of 0) or not drinking (assigned a score of 1). Exercise status was divided into regular exercise (assigned a score of 1) and infrequent exercise (assigned a score of 0). Participants were classified as either less than 7 hours or more than 9 hours (assigned a score of 0) or between 7 and 9 hours (assigned a score of 1). The total score of six items was used to reflect the lifestyle of the participants, with higher scores indicating a healthier lifestyle for older adults.
Social participation
According to the communicative or interactive dimensions of the social participation of older adults and whether it needs to generate social bonds, we divided the social participation of older adults into two main forms of "self-recreation" and "group interaction". We selected the “self-recreation” type of social participation item (doing housework, planting flowers and keeping pets, reading books and newspapers, raising poultry and livestock, watching TV and listening to the radio) and the "group interaction" type of social participation item (outdoor activities, playing cards or mahjong, participating in organized social activities) from the CLHLS questionnaire to evaluate the level of social participation of older adults. Responses to these items included "1= almost every day; 2= not every day, but at least once a week; 3= not weekly, but at least once a month; 4= Not every month, but sometimes; 5= not participating ". We calculated the total score of all items after reverse-scoring the responses to each item, with higher scores indicating higher social participation of older adults.
Covariates
Information about other sociodemographic statuses and health conditions of the participants were controlled in our study, including age (65~, 75~, 85~ or 95~), gender (male or female), residence (town or countryside), marital status (without spouse, with spouse), activities of daily living (impaired or normal), self-rated health status (good, fair or poor), hearing status (difficult or normal), serious illness in the past two years (yes, no).
Statistical analysis
Stata 16.0 software was used to clean the 2017-2018 CLHLS data, and SPSS 21.0 was used for statistical analysis. Pearson’s 2 test was used to analyze the differences between groups in categorical variables such as sociodemographic characteristics, health status, and socioeconomic status-related indicators, and P < 0.05 was considered statistically significant. The SPSS macro program (PROCESS) developed by Hayes[24] was used for mediating effect analysis to explore the mediating effect of cognitive function, lifestyle and social participation on SES and depression in older adults. The biasing corrected percentile Bootstrap method (repeated sampling 5000 times) and 95% Confidence Interval (95%CI) were used to infer the significance of the mediating effect. If the 95%CI did not include 0, the mediating effect was significant.