Abstract
This study aims to examine the association between multimorbidity and depressive symptoms among rural middle-aged and older adults in Ningxia and the interaction effect of childhood socioeconomic status. On the basis of the 2022 Ningxia Rural Household Health Survey, we analysed a sample of 5567 individuals aged 45 years and older. Propensity score matching and ordered logit regression were conducted to estimate the relationship between multimorbidity and depression; multiple matching methods were used for robustness checks; and interaction effects were examined to assess childhood socioeconomic influences. After full adjustment, both ordered logit and propensity score matching models indicated that multimorbidity was significantly associated with more severe depressive symptoms (β = 0.295, P < 0.01; β = 0.288, P < 0.01). Heterogeneity analyses revealed variation in the association across subgroups, with the largest coefficient estimates observed for male (β = 0.491, P < 0.01), those aged 75 years and older (β = 0.708, P < 0.01), and individuals with no formal schooling (β = 0.418, P < 0.01). Childhood hunger experience (β = -0.627, P < 0.05) and father’s education (β = 0.692, P < 0.05) demonstrated a significant interaction effect in the relationship between multimorbidity and depressive symptoms. The negative interaction effect association was significant among individuals with childhood hunger, whereas a positive interaction effect association was observed in those with paternal education levels. Our findings emphazize a robust link between multimorbidity and depressive symptoms in later life, with childhood hunger and paternal education serving as significant moderators of this association.
Supplementary Information
The online version contains supplementary material available at 10.1038/s41598-026-37666-w.
Keywords: Multimorbidity, Depressive symptoms, Interaction effect, Middle-aged and older adults, Propensity score matching
Subject terms: Health care, Health policy, Depression
Introduction
The rapid aging of China’s population has led to a significant increase in the number of middle-aged and older adults. As of the end of 2019, China had 254 million individuals aged 60 years and above, accounting for 18.1% of the total population1. Projections suggest that by the end of 2020, the elderly population is anticipated to reach 264 million, accounting for 18.7% of the total population2. Moreover, the prevalence of depressive symptoms among middle-aged and older adults has been consistently increasing. According to the World Health Organization (WHO), depressive symptoms affect more than 300 million people worldwide, accounting for more than 4% of the global population3. In China, the prevalence of depressive symptoms among middle-aged and older adults was 42.92% in this population3,4. These statistics emphasize the growing prevalence of depressive symptoms among middle-aged and older adults, emphazizing their status as a significant public health concern affecting both their physical and mental well-being.
Multimorbidity, the simultaneous presence of two or more chronic diseases in an individual, is widespread among a significant portion of the elderly population5. Studies have shown that the prevalence of chronic diseases among older adults in China reaches 50.5%6. Studies have shown that multimorbidity significantly jeopardizes the physical and mental well-being of older individuals, potentially worsening depressive symptoms, specifically, individuals with chronic conditions such as cardiovascular disease, diabetes, and chronic respiratory issues present elevated rates of depressive symptoms7. Hence, exploring the association between multimorbidity and childhood socioeconomic status with depressive symptoms among middle-aged and older adults is of substantial theoretical and practical importance, especially the severe impact of late-life depression on quality of life, functional independence, healthcare utilization, and mortality, making it a critical public health concern worthy of investigation, even within the context of a complex, bidirectional relationship. While many studies have investigated the link between chronic diseases and depressive symptoms, most have concentrated on the direct impact of chronic diseases, overlooking other potential moderating factors.
Childhood is a critical stage in the physical and psychological maturation process, with events during this period significantly impacting an individual’s health later in life8,9. Early life circumstances, nutritional well-being, and experiences with illness can profoundly impact an individual’s physical and mental well-being10. Research has shown a link between health status in childhood and the later development of chronic diseases and mental health disorders in adulthood11,12. However, there is a lack of research examining interaction effect of childhood socioeconomic status within the relationship between multimorbidity and depressive symptoms.
Lower childhood socioeconomic status constitutes a important indicator of child poverty13. Although some scholars regard childhood socioeconomic status (SES) as a distal factor that decays over time14, in resource-scarce rural Ningxia—where intergenerational transmission is significant. Specifically, early-life psychosocial stress may trigger epigenetic modifications that exert distinct functional effects on brain plasticity and behavior15; and childhood health disadvantages deplete adult health human capital, intensifying postcomorbidity depression risk16. Proximate social determinants of health (SDOH), such as rural poverty and weak social ties are influential but are themselves partly shaped by childhood SES; the relationship is thus a chained progression rather than a simple substitution. Neglecting childhood SES risks overattributing outcomes to adult SDOH and thereby underestimating the persistent health effects of early inequality. Therefore, incorporating childhood SES into the moderation model serves a dual purpose: it tests whether “early disadvantage still leaves a significant health imprint on rural elders,” and it clarifies the boundary conditions of proximal SDOH, offering precise intervention targets for middle- and late-life depression grounded in life-course theory. By examining the interaction effect of childhood socioeconomic status through hunger experience, medical accessibility, and overall childhood health within the relationship between multimorbidity and depressive symptoms, this study seeks to uncover the structural mechanisms of health-inequality transmission across generations and to provide early-life theoretical foundations for health interventions amid population aging.
In recent years, numerous studies by both national and international scholars have investigated the link between comorbid chronic diseases and depressive symptoms. For example, JaeMoo et al.17 examined the associations between the number of disease multimorbidity, grip strength, various variables, and depressive symptoms in older Korean adults. The results revealed a significant increase in the prevalence of depressive symptoms with increasing numbers of disease multimorbidity. Additionally, Cheng et al reported that individuals with chronic diseases and multimorbidity were more prone to developing depressive symptoms in populations with a higher dietary inflammatory index (DII)18.
The literature review highlights limitations in current research on the link between multimorbidity and depressive symptoms. Numerous studies have been conducted in developed countries, and there is a lack of relevant research in developing countries such as China. Furthermore, research in China often concentrates on urban samples, neglecting rural areas. Moreover, current studies lack comprehensive analyses of mechanisms and do not address self-selection bias. Additionally, there is a dearth of research on whether childhood socioeconomic status serves as a moderator that indirectly influences the association between multimorbidity and childhood socioeconomic status. This study aims to conduct a propensity score matching approach to examine the association between multimorbidity and depressive symptoms, as well as the interaction effect of childhood socioeconomic status within this relationship. Robustness checks conducting various matching methods to analyze the heterogeneity in the relationship between multimorbidity and depressive symptoms across different genders, age groups, and educational levels. Finally, this study analyzed the interaction effect of childhood socioeconomic status within the relationship between multimorbidity and depressive symptoms in middle-aged and older adults. This study aims to examine the association between multimorbidity and depressive symptoms among rural middle-aged and older adults, thereby offering actionable insights for government agencies and relevant organizations to formulate targeted mental health intervention strategies.
Methods
Data sources
The data for this study were collected from a pilot project on ‘Innovative Payment Systems to Enhance Healthcare Efficiency’ jointly conducted by Harvard University and Ningxia Medical University in 2009, 2011, and 2012 (https://www.hsph.harvard.edu/re-alignment-health-system-incentives/sample-page/policy-impact-and-media-coverage/)19. Subsequent follow-up surveys were funded by the National Natural Science Foundation of China in 2015, 2019, and 2022. This research utilized follow-up data from the ‘Rural Household Health Inquiry Survey’ conducted in 2022, which aimed to assess health status and healthcare service utilization comprehensively among residents in rural areas of Ningxia, Western China. The survey provides empirical data and scientific evidence to inform the development of precise and realistic healthcare policies.
A multistage sampling method was used to conduct surveys in four counties in Ningxia: Haiyuan, Yanchi, Pengyang, and Xiji. In the first stage, all administrative villages under each township in the four counties were classified into three levels—good, medium, and poor—on the basis of their economic development levels. In the second stage, a random number table method was used to randomly select 40% of the sample villages from each stratum. In the third stage, systematic sampling was conducted in each sample village, where one household was selected every five households, resulting in a total of 20–33 households being randomly selected.
In this survey, a total of 21,300 questionnaires were distributed, and all were retrieved, yielding an effective response rate of 97.75%, with 20,821 valid questionnaires. The study focused on middle-aged and older adults who had been living at home for more than six months and who were aged 45 years or older. The age restriction of 45 years and older was applied to align with the study’s focus on middle-aged and older adults. This criterion ensures that our sample captures the population most likely to be experiencing the onset or presence of multimorbidity and its potential consequences for mental health, which are more prevalent in this age group. After samples with missing or invalid values were excluded, 5,567 eligible rural middle-aged and older adults were included in the study.The specific inclusion and exclusion flowchart is shown in Supplementary Figure S1.
Model and variables
Explained variable (Y): depressive symptoms
Depressive symptom screening was performed via the Patient Health Questionnaire (PHQ-9)20. The PHQ-9 assesses symptoms experienced over the past 2 weeks, deriving its scoring system from the DSM-IV criteria for depressive disorders. Each of the nine items receives a score ranging from 0 (not at all) to 3 (nearly every day). It is utilized as a continuous score ranging from 0 (no depressive symptoms) to 27 (all symptoms occurring daily). Additionally, it is categorized into bands, with scores of 5–9, 10–14, 15–19, and 20 plus representing mild, moderate, moderately severe, and severe depressive symptoms, respectively21,22. Its internal consistency reliability was acceptable (Cronbach’s alpha coefficient = 0.882).
Moderating variable (M): childhood socioeconomic status
Drawing on the literature23 and capitalizing on variables already available in the database, this study conducts five indicators—“experience of hunger,” “prompt medical treatment for illness,” “childhood health status,” “father’s educational attainment,” and “mother’s educational attainment”—to construct a composite measure of childhood socioeconomic status. The variable definitions and coding schemes are detailed in Table S1 of the Supplementary Materials.
Explanatory variables (X): Multimorbidity
This study investigated eight chronic health issues, including high blood pressure, diabetes, spinal disc disease, stroke, chronic stomach inflammation, heart disease, rheumatoid arthritis, and chronic lung disease. The study documented the frequency of participants reporting these chronic conditions and identified individuals with two or more conditions as having multimorbidity.
Control variables
Our analysis adjusted for several covariates, including demographic and socioeconomic characteristics, lifestyles, and health status24,25. The demographic characteristics included sex, age, marital status, education, and occupation. Socioeconomic characteristics included total household income and household size. Health status and behaviours included smoking, drinking, exercise and self-rated health. The definitions and assignments of all the variables are shown in Table 1.
Table 1.
Variable definitions.
| Variables | Explanation | Code | Mean | SD |
|---|---|---|---|---|
| Explained variable (Y) | Depressive symptoms | No depressive symptoms=1,Mild depressive symptoms=2，Moderate depressive symptoms =3，Moderate-to-Severe depressive symptoms=4，Severe depressive symptoms=5 | 1.27 | 0.64 |
| Explanatory variable (X) | Multimorbidity | Yes = 1 And No = 0 | 0.17 | 0.38 |
| Control variable (C) | Sex | Male = 1 And Female = 0 | 0.54 | 0.50 |
| Age | 45–59 years old = 1,60–74 years old = 2, ≥ 75 years old = 3 | 1.58 | 0.66 | |
| Marital status | Unmarried = 1, Married = 2, Divorced/Widowed = 3 | 2.08 | 0.31 | |
| Education | No Formal Education = 1,Primary School = 2,Junior High School = 3,Senior High School And Above = 4 | 1.85 | 0.86 | |
| Occupation | Agricultural workers = 1; Other = 0 | 0.74 | 0.44 | |
| Self-rated health | Very Bad = 1,Bad = 2, Fair = 3, Good = 4, Very Good = 5 | 2.62 | 1.06 | |
| Smoking | Yes = 1 And No = 0 | 0.21 | 0.41 | |
| Drinking | Yes = 1 And No = 0 | 0.07 | 0.25 | |
| Exercise | Yes = 1 And No = 0 | 0.36 | 0.48 | |
| Total household income (log) | Continuous Variable(logarithm) | 10.14 | 0.78 | |
| Household size | 1–3 persons = 1, 4–5 persons = 2, ≥ 6 persons = 3 | 1.50 | 0.71 | |
| Moderator variable | Childhood hunger | Yes = 1 And No = 0 | 0.83 | 0.37 |
| Childhood illness treatment | Yes = 1 And No = 0 | 0.20 | 0.40 | |
| Physical health status in childhood | Bad = 1, Fair = 2, Good = 3 | 2.75 | 0.55 | |
| Father’s education | Yes = 1 And No = 0 | 0.12 | 0.32 | |
| Mother’s education | Yes = 1 And No = 0 | 0.05 | 0.23 |
Statistical analysis
In our research, we coded the data via EpiData. Before analysis, all survey data underwent thorough checks for missing values, outliers, and adherence to normality assumptions. Statistical analysis was performed with econometric software, including SPSS 26.0 and STATA version 17.0. Initially, the study addressed mixed factor issues via the propensity score matching (PSM) method. This included balance testing and evaluating the common support hypothesis. PSM regression and ordered logit regression analyses were subsequently conducted to assess the associations between multimorbidity and depressive symptoms. To ensure robust findings, we conduct various propensity score matching methods, including radius matching, kernel matching and mahalanobis distance matching26. Finally, a interaction effect analysis was conducted to explore the association between multimorbidity and depressive symptoms. All the statistical tests in this study adhered to a significance level of P < 0.05, utilizing a two-tailed approach.
Ethics approval and consent to participate
Ethical approval was granted by the Ethics Committee of Ningxia Medical University, approval number No. 2021-G152. All the participants provided signed informed consent at the time of participation. The study methodology was carried out in accordance with approved guidelines.
Results
Descriptive analyses of variables
The demographic distribution of depressive symptoms is presented in Table 2. The study included a total of 5567 individuals, showing variations in multimorbidity (χ2 = 148.244, P < 0.01), sex (χ2 = 58.321, P < 0.01), age (χ2 = 97.890, P < 0.01), marital status (χ2 = 69.025, P < 0.01), education (χ2 = 100.456, P < 0.01), occupation (χ2 = 35.866, P < 0.01), self-rated health (χ2 = 830.638, P < 0.01), smoking (χ2 = 51.288, P < 0.01), drinking (χ2 = 27.200, P < 0.01), exercise (χ2 = 9.726, P < 0.05), childhood hunger (χ2 = 45.481, P < 0.01), childhood illness treatment (χ2 = 47.136, P < 0.01), physical health status in childhood (χ2 = 29.145, P < 0.01), and mother’s education (χ2 = 13.627 , P<0.01) These factors demonstrated statistically significant differences in depressive symptoms among middle-aged and older adults.
Table 2.
Demographic characteristics and the distribution of depressive symptoms.
| Characteristics | No depressive symptoms | Mild depressive symptoms | Moderate depressive symptoms | Moderate-to-Severe depressive symptoms | Severe depressive symptoms | Chi-Square/F | P Value |
|---|---|---|---|---|---|---|---|
| Multimorbidity | 148.244 | < 0.01 | |||||
| Yes | 649(14.35) | 206(27.69) | 64(31.37) | 16(26.67) | 19(52.78) | ||
| No | 3874(85.65) | 538(72.31) | 140(68.63) | 44(73.33) | 17(47.22) | ||
| Sex | 58.321 | < 0.01 | |||||
| Male | 2538(56.11) | 326(43.82) | 81(39.71) | 28(46.67) | 16(44.44) | ||
| Female | 1985(43.89) | 418(56.18) | 123(60.29) | 32(53.33) | 20(55.56) | ||
| Age | 97.890 | < 0.01 | |||||
| 45–59 years old | 2444(54.03) | 324(43.55) | 65(31.86) | 24(40.00) | 13(36.11) | ||
| 60–74 years old | 1706(37.72) | 321(43.15) | 100(49.02) | 31(51.67) | 13(36.11) | ||
| ≥ 75 years old | 373(8.25) | 99(13.31) | 39(19.12) | 5(8.33) | 10(27.78) | ||
| Marital status | 69.025 | < 0.01 | |||||
| Unmarried | 54(1.19) | 4(0.54) | 1(0.49) | 1(1.67) | 0(0.00) | ||
| Married | 4120(91.09) | 649(87.23) | 162(79.41) | 46(76.67) | 29(80.56) | ||
| Divorced/widowed | 349(7.72) | 91(12.23) | 41(20.10) | 13(21.67) | 7(19.44) | ||
| Education | 100.456 | < 0.01 | |||||
| No formal education | 1731(38.27) | 387(52.02) | 124(60.78) | 29(48.33) | 21(58.33) | ||
| Primary school | 1739(38.45) | 251(33.74) | 52(25.49) | 23(38.33) | 10(27.78) | ||
| Junior high school | 814(18.00) | 86(11.56) | 24(11.76) | 7(11.67) | 4(11.11) | ||
| Senior high school and above | 239(5.28) | 20(2.69) | 4(1.96) | 1(1.67) | 1(2.78) | ||
| Occupation | 35.866 | < 0.01 | |||||
| Agricultural workers | 3429(75.81) | 507(68.15) | 128(62.75) | 41(68.33) | 25(69.44) | ||
| Other | 1094(24.19) | 237(31.85) | 76 (37.25) | 19(31.67) | 11(30.56) | ||
| Self-rated health | 830.638 | < 0.01 | |||||
| Very bad | 421(9.31) | 231(31.05) | 93(45.59) | 39(65.00) | 27(75.00) | ||
| Bad | 1445(31.95) | 324(43.55) | 84(41.18) | 15(25.00) | 8(22.22) | ||
| Fair | 1591(35.18) | 132(17.74) | 21(10.29) | 4(6.67) | 0(0.00) | ||
| Good | 819(18.11) | 49(6.59) | 6(2.94) | 2(3.33) | 1(2.78) | ||
| Very good | 247(5.46) | 8(1.08) | 0(0.00) | 0(0.00) | 0(0.00) | ||
| Smoking | 51.288 | < 0.01 | |||||
| Yes | 1033(22.84) | 104(13.98) | 23(11.27) | 7(11.67) | 2(5.56) | ||
| No | 3490(77.16) | 640(86.02) | 181(88.73) | 53(88.33) | 34(94.44) | ||
| Drinking | 27.200 | < 0.01 | |||||
| Yes | 343(7.58) | 29(3.90) | 4(1.96) | 0(0.00) | 1(2.78) | ||
| No | 4180(92.42) | 715(96.10) | 200(98.04) | 60(100.00) | 35(97.22) | ||
| Exercise | 9.726 | 0.045 | |||||
| Yes | 1683(37.21) | 251(33.74) | 69(33.82) | 14(23.33) | 10(27.78) | ||
| No | 2840(62.79) | 493(66.26) | 135(66.18) | 46(76.67) | 26(72.22) | ||
| Total household income (log) | 10.14 ± 0.78 | 1.550 | 0.1855 | ||||
| Household size | 10.365 | 0.240 | |||||
| 1–3 persons | 2826(62.48) | 489(65.73) | 138(67.65) | 46(76.67) | 22 (61.11) | ||
| 4–5 persons | 1103(24.39) | 162 (21.77) | 44(21.57) | 10(16.67) | 8(22.22) | ||
| ≥ 6 persons | 594(13.13) | 93(12.50) | 22(10.78) | 4 (6.67) | 6 (16.67) | ||
| Childhood hunger | 45.481 | < 0.01 | |||||
| Yes | 3696(81.72) | 663(89.11) | 192(94.12) | 54(90.00) | 31(86.11) | ||
| No | 827(18.28) | 81(10.89) | 12(5.88) | 6(10.00) | 5(13.89) | ||
| Childhood illness treatment | 47.136 | < 0.01 | |||||
| Yes | 979(21.64) | 108(14.52) | 21(10.29) | 2(3.33) | 3(8.33) | ||
| No | 3544(78.36) | 636(85.48) | 183(89.71) | 58(96.67) | 33(91.67) | ||
| Physical health status in childhood | 29.145 | < 0.01 | |||||
| Bad | 248(5.48) | 45(6.05) | 10(4.90) | 2(3.33) | 5(13.89) | ||
| Fair | 581(12.85) | 115(15.46) | 46(22.55) | 9(15.00) | 9(25.00) | ||
| Good | 3694(81.67) | 584(78.49) | 148 (72.55) | 49(81.67) | 22(61.11) | ||
| Father’s education | 9.121 | 0.058 | |||||
| Yes | 567(12.54) | 76(10.22) | 17(8.33) | 4(6.67) | 2(5.56) | ||
| No | 3956 (87.46) | 668(89.78) | 187(91.67) | 56 (93.33) | 34 (94.44) | ||
| Mother’s education | 13.627 | <0.01 | |||||
| Yes | 266(5.88) | 23(3.09) | 10(4.90) | 1(1.67) | 0(0.00) | ||
| No | 4257(94.12) | 721 (96.91) | 194 (95.10) | 59(98.33) | 36(100.00) | ||
The propensity score matches the hypothesis of common support
To effectively implement the propensity score matching method, two critical conditions must be met: the common support hypothesis and the parallel hypothesis. To assess the common support assumption in this study, we utilized kernel density function plots27,28. Figure 1 clearly shows a disparity in the propensity score kernel density distributions between the two groups (the experimental group with multimorbidity and the control group without multimorbidity) before the matching process. However, after matching, there is a significant alignment in the kernel density distribution curves of both groups. This alignment signifies a successful matching effect, confirming the common support hypothesis. The results of the propensity score matching common-support test and standardized deviation plots for the covariates are presented in the supplementary material Figure S2.
Fig. 1.
Balance test results before and after matching between the treatment group and the control group.
Processing the self-selection problem
We conducted propensity score matching to assess the association between multimorbidity and depressive symptoms in rural middle-aged and older adults, aiming to minimize potential bias resulting from self-selection. In Table 3, implementing nearest-neighbor matching significantly reduced the standard deviations in the samples by less than 10%29,30. This reduction indicates a more consistent variation in observed characteristics within the sample, demonstrating effective control over confounding variables. The results obtained through propensity score matching demonstrate a noteworthy level of explanatory power in understanding the association between multimorbidity and depressive symptoms in rural middle-aged and older adults.
Table 3.
Sample balance test of propensity score matching.
| Variable | Unmatched | Mean | Standard deviation % | Deviation reduction % | T test | ||
|---|---|---|---|---|---|---|---|
| Matched | Treated | Control | T value | P value | |||
| Gender | Before | 0.445 | 0.556 | −22.2 | −6.24 | < 0.01 | |
| After | 0.445 | 0.437 | 1.7 | 92.4 | 0.37 | 0.712 | |
| Age | Before | 65.130 | 59.499 | 60.5 | 16.76 | < 0.01 | |
| After | 65.130 | 65.271 | −1.5 | 97.5 | −0.32 | 0.747 | |
| Marital status | Before | 2.124 | 2.070 | 16.6 | 4.92 | < 0.01 | |
| After | 2.124 | 2.123 | 0.2 | 99.0 | 0.03 | 0.974 | |
| Education | Before | 1.688 | 1.885 | −23.7 | −6.45 | < 0.01 | |
| After | 1.688 | 1.652 | 4.3 | 81.7 | 0.99 | 0.322 | |
| Occupation | Before | 0.672 | 0.756 | −18.8 | −5.44 | < 0.01 | |
| After | 0.672 | 0.676 | −0.8 | 95.7 | −0.17 | 0.864 | |
| Self-rated health | Before | 2.011 | 2.747 | −75.3 | −20.31 | < 0.01 | |
| After | 2.011 | 2.018 | −0.7 | 99.0 | −0.18 | 0.858 | |
| Smoking | Before | 0.134 | 0.226 | −24.0 | −6.34 | < 0.01 | |
| After | 0.134 | 0.116 | 4.8 | 80.0 | 1.21 | 0.226 | |
| Drinking | Before | 0.029 | 0.076 | −20.9 | −5.19 | < 0.01 | |
| After | 0.029 | 0.026 | 1.4 | 93.2 | 0.42 | 0.676 | |
| Exercise | Before | 0.408 | 0.355 | 10.9 | 3.08 | <0.01 | |
| After | 0.408 | 0.394 | 2.9 | 73.1 | 0.63 | 0.528 | |
| Total household income (log) | Before | 10.012 | 10.161 | −19.3 | −5.38 | < 0.01 | |
| After | 10.012 | 10.034 | −2.8 | 85.5 | −0.60 | 0.551 | |
| Household size | Before | 1.355 | 1.526 | −24.7 | −6.75 | < 0.01 | |
| After | 1.355 | 1.375 | −2.8 | 88.6 | −0.64 | 0.523 | |
Comprehensive evaluation of the associations between Multimorbidity and depressive symptoms among rural middle-aged and older adults
The analysis in Table 4, specifically in columns (1)-(4), explores the association between multimorbidity and depressive symptoms among rural middle-aged and older adults. Columns (1) and (2) conduct ordered logit regression, whereas columns (3) and (4) use propensity score matching (PSM). In Column (1), multimorbidity was significantly associated with increased depressive symptoms before adjustment for control variables. In columns (2) and (4), even after other factors are considered, the findings persist, highlighting that multimorbidity was significantly associated with increased depressive symptoms. The control variables in Column (4) show statistically significant coefficients for education, occupation, self-rated health, and exercise at the 1% and 5% levels, respectively. The linear regression analysis of the PHQ-9 score is provided in Supplementary Table S2.
Table 4.
Results of ordered logit and PSM.
| Variable | (1) | (2) | (3) | (4) |
|---|---|---|---|---|
| Multimorbidity(reference: No) | ||||
| Yes | 0.907*** | 0.295*** | 0.349*** | 0.288*** |
| (0.0793) | (0.0880) | (0.0949) | (0.101) | |
| Gender(reference: Female) | ||||
| Male | −0.146* | −1.61e-05 | ||
| (0.0852) | (0.115) | |||
| Age (reference:45–59 years old) | ||||
| 60–74 years old | 0.0293 | −0.0228 | ||
| (0.0865) | (0.121) | |||
| ≥ 75 years old | −0.0803 | −0.0579 | ||
| (0.135) | (0.169) | |||
| Marital status(reference: Unmarried) | ||||
| Married | 0.490 | −0.0389 | ||
| (0.456) | (0.616) | |||
| Divorced/widowed | 0.980** | 0.368 | ||
| (0.467) | (0.628) | |||
| Education(reference: No formal education) | ||||
| Primary school | −0.273*** | −0.379*** | ||
| (0.0846) | (0.116) | |||
| Junior high school | −0.284** | −0.190 | ||
| (0.123) | (0.173) | |||
| Senior high school and above | −0.511** | −0.666* | ||
| (0.230) | (0.361) | |||
| Occupation(reference: Other) | ||||
| Agricultural workers | −0.210** | −0.278** | ||
| (0.0855) | (0.113) | |||
| Self-rated health(reference: Very bad) | ||||
| Bad | −1.158*** | −1.131*** | ||
| (0.0896) | (0.111) | |||
| Fair | −2.142*** | −1.923*** | ||
| (0.113) | (0.164) | |||
| Good | −2.426*** | −1.791*** | ||
| (0.158) | (0.282) | |||
| Very good | −3.171*** | −3.286*** | ||
| (0.369) | (1.022) | |||
| Smoking(reference: No) | ||||
| Yes | −0.312*** | −0.259 | ||
| (0.119) | (0.175) | |||
| Drinking(reference: No) | ||||
| Yes | −0.213 | −0.201 | ||
| (0.203) | (0.339) | |||
| Exercise(reference: No) | ||||
| Yes | −0.152* | −0.234** | ||
| (0.0780) | (0.104) | |||
| Total household income (log) | 0.0501 | 0.00282 | ||
| (0.0514) | (0.0700) | |||
| Household size(reference:1–3 persons) | ||||
| 4–5 persons | −0.0188 | −0.160 | ||
| (0.0966) | (0.143) | |||
| ≥ 6 persons | −0.139 | −0.265 | ||
| (0.121) | (0.174) | |||
| N | 5567 | 5567 | 2192 | 2192 |
| R2 | 0.0171 | 0.1158 | 0.0037 | 0.0824 |
| Prob > F | < 0.01 | < 0.01 | < 0.01 | < 0.01 |
Notes: Column (1) is an ordered logit regression that includes multimorbidity, and Column (2) is an ordered logit regression that includes multimorbidity, demographic characteristics, socioeconomic characteristics and health status and behaviours characteristics. Column (3) shows the PSM+ordered logit regression including multimorbidity. Column (4) shows the PSM+ordered logit regression including multimorbidity, demographic characteristics, socioeconomic characteristics and health status and behaviours characteristics. ***P < 0.01, **P < 0.05, *P < 0.1. The numbers in the brackets are the standard errors of coefficient robustness.
Robustness test based on the matching approach
To ensure the accuracy of the findings, this study conducts various propensity score matching techniques, including radius matching, kernel matching and mahalanobis distance matching. As shown in Table 5, multimorbidity was significantly associated with increased depressive symptoms, irrespective of the specific propensity score matching method used. The average treatment effects for different propensity score-matched samples are reported in the supplementary material Table S3.
Table 5.
Robustness test based on the matching approach.
| Variable | Radius matching | Kernel matching | Mahalanobis distance matching |
|---|---|---|---|
| Multimorbidity | 0.297*** | 0.297*** | 0.372*** |
| (0.088) | (0.088) | (0.118) | |
| Control variable | Yes | Yes | Yes |
| N | 5545 | 5545 | 1596 |
| R2 | 0.1148 | 0.1148 | 0.0795 |
| Prob > F | < 0.01 | < 0.01 | < 0.01 |
Notes: ***P < 0.01, **P < 0.05, *P < 0.1. The numbers in the brackets are the standard errors of coefficient robustness.
Heterogeneity
This study analysed the relationships between multimorbidity and depressive symptoms among middle-aged and older adults in different populations to provide more precise policy interventions. The sample was stratified by sex, age, and education. The results are presented in Table 6. In terms of sex, multimorbidity was significantly associated with increased depressive symptoms in males (β = 0.491, P < 0.01). With respect to age, the estimated results revealed that multimorbidity was significantly associated with increased depressive symptoms in individuals aged 75 years and above (β = 0.708, P < 0.01). In terms of education, the results indicate that multimorbidity was significantly associated with increased depressive symptoms is mainly observed in the group with no formal education (β = 0.418, P < 0.01).
Table 6.
Heterogeneity analysis of gender, age, and education.
| Variable | Male | Female | 45–59 | 60–74 | ≥ 75 | No formal education | Primary school | Junior high school | Senior high school and above |
|---|---|---|---|---|---|---|---|---|---|
| Multimorbidity | 0.491*** | 0.139 | 0.379** | 0.127 | 0.708*** | 0.418*** | 0.129 | 0.382 | −0.629 |
| (0.134) | (0.117) | (0.156) | (0.125) | (0.218) | (0.121) | (0.154) | (0.275) | (0.680) | |
| Control variable | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes |
| N | 2989 | 2578 | 2870 | 2171 | 526 | 2292 | 2075 | 935 | 265 |
| R2 | 0.1206 | 0.1035 | 0.1145 | 0.1104 | 0.1183 | 0.1126 | 0.0964 | 0.1297 | 0.2360 |
| Prob > F | < 0.01 | < 0.01 | < 0.01 | < 0.01 | < 0.01 | < 0.01 | < 0.01 | < 0.01 | < 0.01 |
Notes: ***P < 0.01, **P < 0.05, *P < 0.1. The numbers in the brackets are the standard errors of coefficient robustness.
Analysis of the interaction effect analysis
To explore the interactive effect of childhood socioeconomic status within the association between multimorbidity and depressive symptoms among middle-aged and elderly people in rural Ningxia, this study carried out a interaction effect analysis, and the results are presented in Table 7. The introduction of the interaction term between multimorbidity and childhood hunger experience yielded a statistically significant negative coefficient (β = −0.627, P < 0.05), indicating a significant interaction effect. The interaction terms between multimorbidity and the level of childhood physical health status (fair or good, with ‘bad’ as the reference) were not statistically significant (P > 0.05)and multimorbidity was nota statistically significant predictor of depressive symptoms (β= 0.345, P > 0.05). When the interaction term between multimorbidity and timely treatment of childhood illness was added, it was not statistically significant (β = 0.190, P > 0.05). Multimorbidity remained significantly associated with depressive symptoms (β = 0.268, P < 0.01), whereas timely treatment itself was a significant predictor. The interaction term between multimorbidity and father’s education level was statistically significant and positive (β = 0.692, P < 0.05). The positive association between multimorbidity and depressive symptoms was stronger for individuals with a higher level of paternal education. The analysis of mothers’ education level as a moderator revealed that the interaction term between multimorbidity and mothers’ education was not statistically significant (β = 0.328, P > 0.05). The results of the interactions are shown in Figure S3 of the Supplementary Materials.
Table 7.
Interaction effects of childhood socioeconomic status on multimorbidity-depressive symptoms.
| Variables | Coefficient | SE | Z | 95% CI | ||
|---|---|---|---|---|---|---|
| Lower | Upper | |||||
| depressive symptoms symptoms(Y) | ||||||
| Independent variable(X) | Multimorbidity(reference: No) | 0.858*** | 0.272 | 3.16 | 0.326 | 1.391 |
| Moderator variable(W) | Childhood hunger(reference: No) | 0.516*** | 0.136 | 3.79 | 0.249 | 0.782 |
| Interaction(X × W) | Multimorbidity×Childhood hunger(reference: No) | -0.627** | 0.283 | -2.21 | -1.182 | -0.072 |
| Control variable | Yes | |||||
| Prob > F | < 0.01 | |||||
| Independent variable(X) | Multimorbidity(reference: No) | 0.268*** | 0.098 | 2.73 | 0.075 | 0.461 |
| Moderator variable(W) | Childhood illness treatment(reference: No) | -0.531*** | 0.119 | -4.46 | -0.765 | -0.298 |
| Interaction(X × W) | Multimorbidity×Childhood illness treatment(reference: No) | 0.190 | 0.247 | 0.77 | -0.294 | 0.673 |
| Control variable | Yes | |||||
| Prob > F | < 0.01 | |||||
| Independent variable(X) | Multimorbidity(reference: No) | 0.345 | 0.347 | 1.00 | -0.335 | 1.025 |
| Moderator variable(W) | Physical health status in childhood(reference: Bad) | |||||
| Fair | 0.294 | 0.204 | 1.44 | -0.106 | 0.695 | |
| Good | 0.222 | 0.182 | 1.22 | -0.134 | 0.577 | |
| Interaction(X × W) | Multimorbidity×Physical health status in childhood(reference: Bad) | |||||
| Multimorbidity×Fair | 0.095 | 0.409 | 0.23 | -0.706 | 0.896 | |
| Multimorbidity×Good | -0.090 | 0.360 | -0.25 | -0.795 | 0.615 | |
| Control variable | Yes | |||||
| Prob > F | < 0.01 | |||||
| Independent variable(X) | Multimorbidity | 0.238** | 0.095 | 2.50 | 0.051 | 0.424 |
| Moderator variable(W) | Father’s education | -0.057 | 0.143 | -0.40 | -0.337 | 0.223 |
| Interaction(X × W) | Multimorbidity×Father’s education | 0.692** | 0.309 | 2.24 | 0.086 | 1.297 |
| Control variable | Yes | |||||
| Prob > F | < 0.01 | |||||
| Independent variable(X) | Multimorbidity | 0.283*** | 0.093 | 3.05 | 0.101 | 0.465 |
| Moderator variable(W) | Mother’s education | -0.330 | 0.220 | -1.50 | -0.761 | 0.100 |
| Interaction(X × W) | Multimorbidity×Mother’s educationr | 0.328 | 0.534 | 0.62 | -0.718 | 1.375 |
| Control variable | Yes | |||||
| Prob > F | < 0.01 | |||||
Notes: ***P < 0.01, **P < 0.05, *P < 0.1.
Discussion
Depressive symptoms are common in individuals with chronic illnesses, and multimorbidity is significantly associated with increased depressive symptoms31–33. Recently, there has been a growing emphasis on investigating the associations between multimorbidity and depressive symptoms. With the aging of the global population and the increasing prevalence of chronic diseases, investigations of multimorbidity and depressive symptoms are crucial. This research is crucial for preventing and treating depressive symptoms and improving the quality of life of individuals with chronic diseases. Earlier studies revealed a positive association between multimorbidity and depressive symptoms. The present study revealed that multimorbidity was significantly associated with increased depressive symptoms among rural middle-aged and older adults. This may result from the limited financial resources of rural middle-aged and older adults34, coupled with the substantial costs of treating and rehabilitating multimorbidity, leading to a significant financial burden35. Long-term financial strain may contribute to the onset of depressive symptoms. Furthermore, improving mental health education for individuals with chronic diseases is recommended to enhance their understanding of their condition and their capacity for independent management. Emphasis should be placed on establishing social support networks for patients.
Further analysis of heterogeneity revealed that multimorbidity was significantly associated with increased depressive symptoms among males, those aged over 75 years, and individuals with no formal education. The significant association between multimorbidity and depressive symptoms among men can be attributed to societal expectations in specific rural areas, where men are expected to shoulder greater familial and social responsibilities36. Poor health can hinder their ability to work and earn income37, leading to feelings of loss and depressive symptoms. Moreover, living in rural areas may encounter barriers in accessing mental health services because of their reluctance to seek help or lack awareness of accessing such services38. The risk of depressive symptoms linked to multimorbidity increases among individuals aged over 75 years, possibly because of the loss of spouses, friends, or loved ones as they age, resulting in feelings of isolation, helplessness, and sorrow39. Moreover, elderly individuals may face challenges such as financial difficulties and substandard living conditions, contributing to increased vulnerability to depressive symptoms. The risk of depressive symptoms increases among individuals with no formal education, especially those with multimorbidity. This risk is apparent across different educational levels. One potential explanation is that individuals with lower education levels may lack health education and disease management skills40, causing challenges in effectively handling the symptoms and progression of chronic diseases and ultimately increasing the risk of depressive symptoms.
This study further examined the interaction effect of childhood socioeconomic status within the relationship between multimorbidity and depressive symptoms. The results revealed that childhood hunger increase the risk of depression, yet the interaction between multimorbidity and hunger attenuates this risk. Hunger in childhood, as an early adverse life event, engenders enduring psychological pressure and fosters negative cognitive patterns and coping styles, rendering individuals more susceptible to depressive affect when confronted with subsequent stressors—a finding that is consistent with prior research41–44. Conversely, the interaction effect for childhood hunger presented a counterintuitive pattern. We did not find a straightforward explanation for this result in the literature. This unexpected finding emphasizes the complex and potentially nonlinear nature of how early-life stressors may influence health outcomes later in life. Further investigations are needed to replicate these findings and explore the underlying mechanisms, which may involve complex psychosocial or biological factors not measured in the present study. Our study also revealed that receiving timely medical treatment for childhood illness alleviates the risk of later depression. Prompt treatment during childhood illness implies early access to superior healthcare resources and familial support, exerting a positive influence on both physical recovery and psychological well-being. Research has demonstrated that early life factors is associated with the experiences of pain in later life45. Moreover, the interaction between multimorbidity and paternal education unexpectedly increased the risk of depression. This may be because, in rural Ningxia, educated fathers often hold elevated expectations for their children’s family roles. When offspring develop multimorbidity, the consequent shift in family roles can precipitate role conflict and expectation violations, intensifying internal contradictions and thereby increasing depressive affect.
Research strengths and limitations
This study has three distinct strengths. First, it is based on a population-based dataset that provides ample sample size and statistical power to examine the association between multimorbidity and depressive symptoms. Second, a propensity score matching (PSM) model is used to alleviate endogeneity arising from selection bias in the multimorbidity context, thereby enhancing the robustness of the empirical findings. Third, the investigation probes the interaction effect of childhood socioeconomic status in the relationship between multimorbidity and depressive symptoms, furnishing novel empirical evidence that further clarifies their interplay.
This study has several limitations. First, its cross-sectional design permits only the examination of associations between multimorbidity and depressive symptoms, precluding any causal inferences. Second, the assessment of multimorbidity was based on self-reports, which introduces potential recall bias. Third, because the sample was drawn from a single province, the findings may not be generalizable to other regions, as economic status and medical resources vary across provinces.
Conclusion
The main goal of this study was to examine the interaction effect of childhood socioeconomic status within the relationship between multimorbidity and depressive symptoms among rural middle-aged and older individuals. Using data from the 2022 family health survey of rural residents in Ningxia, this study used a propensity score matching technique. The results revealed that multimorbidity was significantly associated with increased depressive symptoms among rural middle-aged and elderly populations. To ensure the robustness of the findings, this study was rigorously tested by utilizing various matching methods. Furthermore, this study explored heterogeneity in factors such as sex, age, and education. The analyses revealed that multimorbidity was significantly associated with increased depressive symptoms among rural middle-aged and older adults, especially males, individuals aged over 75 years, and those with no formal education. Finally, this study explored the interaction effects, revealing how childhood socioeconomic status influences the relationship between multimorbidity and depressive symptoms among rural middle-aged and older adults. The findings revealed significant moderators, such as childhood hunger and fathers’ education, which play crucial roles in moderating the associations between multimorbidity and depressive symptoms in rural middle-aged and older adults.
Supplementary Information
Below is the link to the electronic supplementary material.
Acknowledgements
This study is a population-based survey, and we thank all the respondents who volunteered to participate in the study.
Author contributions
Hui Qiao conceptualized the research idea and design. Ning Xu participated in the research design, drafted the manuscript, and analysed and interpreted the data. Ximin Ma helped revise the manuscript and interpreted the data. Qi Hu, Jiancai Du, Jiahui He, and Wenlong Wang revised the manuscript and helped clean the data. All authors contributed to the revision and edits of subsequent drafts. All the authors contributed to the article and approved the submitted version.
Funding
This work was supported by the National Natural Science Foundation of China (grant numbers 72164033 and 72364031), the Natural Science Foundation of Ningxia (grant numbers 2023AAC03241, 2023AAC03224, 2024AAC03259 and 2022AAC02036), the Scientific Research Projects in Ningxia Colleges and University (grant number NYG2022049, XZ2024018) and the Major Scientific and Technological Projects of Ningxia Medical University, China (grant number XJKF240314).
Data availability
The data for this study are part of the overall project and are not publicly available. Access to the datasets of this study can be directed to the corresponding authors.
Declarations
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
These authors contributed equally to this work: Ning Xu and Ximin Ma.
References
- 1.National Bureau of Statistics of China. China Statistical Yearbook. (2020). Available at: https://www.stats.gov.cn/sj/ndsj/2020/indexch.htm
- 2.Main data of the seventh national census of the people’s Republic of China. Available at:https://www.stats.gov.cn/sj/pcsj/rkpc/d7c/
- 3.Depression-Fact Sheets. Available online at: https://www.who.int/news-room/fact-sheets/detail/depression [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Wang, J. et al. Effects of social participation and its Diversity, Frequency, and type on depression in Middle-Aged and older persons: evidence from China. Front. Psychiatry. 13, 825460 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Akker MVD, Buntinx F, Knottnerus JA. Comorbidity or multimorbidity: What’s in a name ? A review of literature. Eur J Gen Pract.2, 65–70 (1996).
- 6.Hu, X. et al. Prevalence and pattern of Multimorbidity in china: a cross-sectional study of 224,142 adults over 60 years old. Front. Public. Health. 12, 1349418 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Read, J. R., Sharpe, L., Modini, M. & Dear, B. F. Multimorbidity and depression: A systematic review and meta-analysis. J. Affect. Disorders. 221, 36 (2017). [DOI] [PubMed] [Google Scholar]
- 8.Chang, Y. H., Yang, M. H., Yao, Z. F., Tsai, M. C. & Hsieh, S. The Mediating Role of Brain Structural Imaging Markers in Connecting Adverse Childhood Experiences and Psychological Resilience. Children-Basel.10 (2023). [DOI] [PMC free article] [PubMed]
- 9.McGuinn, L. A. et al. The influence of maternal anxiety and cortisol during pregnancy on childhood anxiety symptoms. Psychoneuroendocrino139, 105704 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Kamiya, Y., Doyle, M., Henretta, J. C. & Timonen, V. Early-life circumstances and later-life loneliness in Ireland. Gerontologist54, 773 (2014). [DOI] [PubMed] [Google Scholar]
- 11.Kumari, V. et al. Reduced thalamic volume in men with antisocial personality disorder or schizophrenia and a history of serious violence and childhood abuse. EUR. PSYCHIAT. 28, 225 (2013). [DOI] [PubMed] [Google Scholar]
- 12.Pandey, R. et al. Childhood maltreatment and its mental health consequences among Indian adolescents with a history of child work. Aust Nz J. Psychiat. 54, 496 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Wang, L., Li, F., Meng, K. & Dunning, K. H. Childhood socioeconomic status and adult subjective wellbeing: the role of hope and sense of control. Front. Psychol.13, 879132 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Chai, Y. et al. The relationship between childhood socioeconomic status and depression level in older adults: the mediating role of adult socioeconomic status and subjective well-being. BMC Geriatr.24, 138 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Vaiserman, A. M. & Koliada, A. K. Early-life adversity and long-term neurobehavioral outcomes: epigenome as a bridge? Hum. Genomics. 11, 34 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Tian, F., Meng, S. S. & Qiu, P. Childhood adversities and mid-late depressive symptoms over the life course: evidence from the China health and retirement longitudinal study. J. Affect. Disorders. 245, 668 (2019). [DOI] [PubMed] [Google Scholar]
- 17.JaeMoo, L. & Edward, J. R. The relationship between muscular strength and depression in older adults with chronic disease comorbidity. Int. J. Env Res. Pub He. 17, 6830 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Cheng, J. et al. Dietary inflammatory index and depression risk in patients with chronic diseases and comorbidity. J. Affect. Disorders. 301, 307 (2022). [DOI] [PubMed] [Google Scholar]
- 19.Ma, X. et al. A study on the factors influencing the vulnerability of women of childbearing age to health poverty in rural Western China. Sci. Rep-UK. 14, 13219 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.K, K., B, W. J. & L, S. R. & The PHQ-9: validity of a brief depression severity measure. J. GEN. INTERN. MED.16, 606 (2001). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Hua, H. et al. Depression and anxiety symptoms and associated factors among community residents in Shenzhen, china: a cross-sectional study. BMC Public. Health. 25, 2148 (2025). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Lowe, B., Unutzer, J., Callahan, C. M., Perkins, A. J. & Kroenke, K. Monitoring depression treatment outcomes with the patient health questionnaire-9. Med. Care. 42, 1194 (2004). [DOI] [PubMed] [Google Scholar]
- 23.Xinyuan Du, Ya Fang, & Yanbing Zeng. A life-course perspective on the impact of socioeconomic status on the health of older adults in China. Medicine and Society. 36,19 (2023).
- 24.Wang, J. et al. The moderating role of psychological resilience in the relationship between falls, anxiety and depressive symptoms. J. Affect. Disorders. 341, 211 (2023). [DOI] [PubMed] [Google Scholar]
- 25.Ying, M., Qin, X., Chaoyang, Y., Hui, L. & Jing, W. Relationship between chronic diseases and depression: the mediating effect of pain. BMC Psychiatry. 21, 436 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Rosenbaum, P. R. & Rubin, D. B. The central role of the propensity score in observational studies for causal effects. Biometrika70, 41 (1983). [Google Scholar]
- 27.Zengxin, X. & Bowei, L., The Impact of the Integration of Urban and Rural Medical Insurance on Migrant Workers’ Overwork: Evidence From China. Front Public Health 10 934524 (2022). [DOI] [PMC free article] [PubMed]
- 28.Zhao, Z., Yihua, M., Yueyao, S., Ruyu, D. & Yuchen, H. Do community Home-Based elderly care services improve life satisfaction of Chinese older adults? An empirical analysis based on the 2018 CLHLS dataset. Int. J. Env Res. Pub He. 19, 15462 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Pasupula, D. K. et al. Combined atrial fibrillation ablation and left atrial appendage occlusion procedure in the united states: a propensity score matched analysis from 2016–2019 National readmission database. Europace25, 390 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Saito, Y. et al. High dose of dexamethasone attenuates docetaxel-induced fluid retention in breast cancer treatment. Sci. Rep-UK. 13, 9247 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Peng, S., Wang, S. & Feng, X. L. Multimorbidity, depressive symptoms and disability in activities of daily living amongst middle-aged and older chinese: evidence from the China health and retirement longitudinal study. J. Affect. Disorders. 295, 703 (2021). [DOI] [PubMed] [Google Scholar]
- 32.Dong, P. et al. Impact of Multimorbidity and cognitive impairment on depressive symptoms in community-dwelling older adults and their interaction effects. Sci. Rep-UK. 15, 15033 (2025). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Sinaga, I. O. Y. et al. Depression is associated with the increase risk of Multimorbidity among the general population in Indonesia. J. Multidiscip Health. 15, 1863 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Lai, Y. et al. Health poverty reduction effect of medical insurance in china’s middle-aged and elderly populations: a cross-sectional study. BMJ Open.15, e85226 (2025). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Miao, W. et al. Multi-dimensional vulnerability analysis on catastrophic health expenditure among middle-aged and older adults with chronic diseases in China. BMC Med. Res. Methodol.22, 151 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Deng, J., Kong, L. & Li, W. Gender disparities in healthy ageing in china: current status and future prospects. Front. Public. Health. 13, 1587922 (2025). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Pellicano, E., Hall, G. & Ying Cai, R. Autistic adults’ experiences of financial wellbeing: Part II. Autism.28, 1090 (2024). [DOI] [PMC free article] [PubMed]
- 38.Emma, K. et al. Engaging the agricultural community in the development of mental health interventions: a qualitative research study. BMC Psychiatry. 23, 399 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Middleton, G. et al. The health and Well-being impacts of community shared meal programs for older populations: A scoping review. Innov. Aging. 6, c68 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Riemann, L., Lubasch, J. S., Heep, A. & Ansmann, L. The role of health literacy in health Behavior, health service Use, health Outcomes, and empowerment in pediatric patients with chronic disease: A systematic review. Int J. Env Res. Pub He18 ,12464 (2021). [DOI] [PMC free article] [PubMed]
- 41.Perna, L. et al. Childhood exposure to hunger: associations with health outcomes in later life and epigenetic markers. Epigenomics-UK12, 1861 (2020). [DOI] [PubMed] [Google Scholar]
- 42.Halmdienst, N. & Winter-Ebmer, R. Long-run relations between childhood shocks and health in late Adulthood-Evidence from the survey of health, Ageing, and retirement in Europe. CESIFO Econ. Stud.60, 402 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43.Roseboom, T. J., Painter, R. C., van Abeelen, A. F., Veenendaal, M. V. & de Rooij, S. R. Hungry in the womb: what are the consequences? Lessons from the Dutch famine. Maturitas70, 141 (2011). [DOI] [PubMed] [Google Scholar]
- 44.Brown, A. S., Susser, E. S., Lin, S. P., Neugebauer, R. & Gorman, J. M. Increased risk of affective disorders in males after second trimester prenatal exposure to the Dutch hunger winter of 1944-45. Brit J. Psychiat. 166, 601 (1995). [DOI] [PubMed] [Google Scholar]
- 45.Ahmed, W. et al. Early life factors associated with the experiences of pain in later life: evidence from a population based study in India. BMC Public. Health. 23, 968 (2023). [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
Data Availability Statement
The data for this study are part of the overall project and are not publicly available. Access to the datasets of this study can be directed to the corresponding authors.