Prevalence and associated factors of mental health among older adults in Guangxi, China: insights from depression, anxiety and cognitive function assessments

To the editor:

China is undergoing a profound demographic transition, with the proportion of adults aged 65 and older reaching 14.9% in 2022 and projected to continue rising due to declining birth rates and increased life expectancy.¹ This ageing trend has brought mental health in late life into sharp focus as a growing public health issue. Depression, anxiety and cognitive decline are especially common among older adults, often underdiagnosed, and substantially contribute to the global disease burden.² Studies have shown that advanced age, chronic diseases, social isolation and living arrangements are key risk factors for poor mental health in old age.³ For example, individuals living alone or in nursing homes are more likely to experience mood disorders.³ Cognitive decline—particularly early-stage dementia—adds further challenges to independent living and quality of life.

In response, the Chinese government launched the ‘14th Five-Year Plan for Healthy Aging’, a policy blueprint aiming to improve services for older adults between 2021 and 2025.⁴ A core component of the plan involves establishing dedicated mental health centres for older adults in both urban and rural communities, expanding access to early detection and intervention services. Each county or district is expected to host at least one such centre by 2025.⁴ While national surveys such as the China Health and Retirement Longitudinal Study⁴ have provided useful insights into depressive symptoms and cognition among middle-aged and older adults, they have not focused specifically on the 65+ population or on underserved western provinces such as Guangxi.

To address this gap, the current study aimed to assess the prevalence and associated factors of depression, anxiety and cognitive decline among older adults in Guangxi. We focused on 35 communities (15 urban and 20 rural) in Guangxi Zhuang Autonomous Region that were selected as part of the national ‘Psychological Care for the Elderly’ initiative under the 14th Five-Year Plan for Healthy Aging.⁴ As part of this programme, provinces were instructed to select subdistricts or townships based on stratified criteria, including the percentage of older adults, number of administrative units and urbanisation level derived from the China Statistical Yearbook (2021).⁵ Final community sites were chosen by provincial health authorities with consideration of local demographic stability and infrastructure capacity. Within each selected site, we used cluster sampling to enrol all residents aged 65 and older. A combination of centralised, household and telephone surveys was used to collect responses between 2022 and 2023. All community surveyors completed standardised training over 2–3 days and passed written or online assessments prior to data collection. Surveys were conducted using a digital questionnaire system. Random audits were performed to ensure data integrity, and any issues were addressed through retraining and community liaison meetings.

Participants reported their age, gender, residence (urban/rural), years of education, marital status, living arrangements and chronic disease history. We also identified vulnerability indicators, including economic hardship (eligibility for social assistance), empty-nest status (living alone or with a spouse but without children), disability (impairment in one or more activities of daily living), dementia diagnosis, living alone at age ≥80 without support from children and one-child policy family (parents affected by the one-child policy). Each respondent was classified as having 0, 1 or ≥2 vulnerability factors.

Mental health outcomes were assessed using three widely validated tools: the Patient Health Questionnaire-9 (PHQ-9) for depressive symptoms, the Generalized Anxiety Disorder-7 (GAD-7) for anxiety symptoms, and the Ascertain Dementia-8 (AD-8) for cognitive screening. The PHQ-9 and GAD-7 are self-administered, while the AD-8 is completed by a caregiver. Participants were categorised into general, borderline or high-risk groups according to standard cut-off scores for each tool.

Descriptive statistics were used to summarise sample characteristics. Given the non-normal distribution of most outcome and predictor variables, non-parametric tests were employed to assess group differences in PHQ-9, GAD-7 and AD-8 scores across categorical variables. Specifically, Mann-Whitney U tests and Kruskal-Wallis tests were conducted, followed by Dunn’s post hoc tests with Bonferroni correction where appropriate. Spearman’s rank correlation was used to examine associations between education and mental health scores.

Variables that demonstrated statistically significant associations in the univariate analyses were subsequently included in the multinomial logistic regression models to estimate adjusted associations between demographic and health-related factors and mental health outcomes. Bonferroni correction was applied across the three regression models, resulting in an adjusted significance threshold of p<0.017 (ie, 0.05/3). Odds ratios (ORs) with corresponding 95% confidence intervals (CIs) were reported. Multinomial logistic regression analyses were performed using the nnet package in R software (V.4.3.0).

Initially, a total of 12 190 participants were surveyed, of whom 608 were excluded due to incomplete data, leaving 11 582 for analysis. Among the 10 370 participants assessed using the PHQ-9, 9629 (92.9%) were categorised as the general population, 580 (5.6%) as the borderline group and 161 (1.6%) as the high-risk group. Of the 10 369 participants assessed with the GAD-7, 9901 (95.5%) were in the general population, 368 (3.5%) in the borderline group and 100 (1.0%) in the high-risk group. For the 11 582 participants assessed using the AD-8, 9392 (81.1%) were classified as the general population, 1539 (13.3%) as the borderline group and 651 (5.6%) as the high-risk group. Detailed demographic characteristics and associated factors for participants assessed with PHQ-9, GAD-7 and AD-8 are provided in online supplemental tables S1-S3.

Most participants were aged 65–74 years (61.8%, mean=69.3), while 29.1% were 75–84 (mean=78.7) and 9.1% were ≥85 (mean=88.4). Females accounted for 55.8%, and 53.7% lived in urban areas. The majority were married (68.9%) and had low education levels (mean=5.2 years). Regarding living arrangements, 34.8% lived with children, 31.7% with a spouse, while 6.7% lived alone and 0.2% resided in nursing homes. Vulnerability factors were reported by 31.7%, including economic hardship (16.0%), empty-nest status (9.7%) and living alone at an advanced age (2.6%). Chronic illness was prevalent (69.9%). Mean scores on PHQ-9, GAD-7 and AD-8 were low, suggesting mild symptoms overall. Detailed distributions appear in supplemental table S4.

Significant between-group differences were observed across multiple variables (figure 1, supplemental figure S1). Females had consistently higher depression, anxiety and cognitive decline scores. Rural residents scored higher on GAD-7 and AD-8. Chronic illness was associated with elevated scores across all domains. Age differences were notable: individuals aged ≥85 years had the highest AD-8 scores, while those aged 65–74 had the lowest PHQ-9 and GAD-7 scores. Widowed individuals exhibited higher scores than those who were married. Living alone and residing in nursing homes were linked to greater psychological and cognitive distress, whereas cohabiting with children or spouses was relatively protective. Vulnerability factors like economic hardship, disability and dementia showed strong associations with worse outcomes; in particular, dementia was linked to the highest AD-8 scores.

Figure 1. Comparison of cognitive function, anxiety and depression scores by sex, chronic disease and age group. A, B and C: PHQ-9, GAD-7 and AD-8 scores by sex; D, E and F: PHQ-9, GAD-7 and AD-8 scores by residence; G, H and I: PHQ-9, GAD-7 and AD-8 scores by chronic disease; J, K and L: PHQ-9, GAD-7 and AD-8 scores by age group. AD-8, Ascertain Dementia-8; GAD-7, Generalized Anxiety Disorder-7; PHQ-9, Patient Health Questionnaire-9.

A significant negative correlation was found between years of education and all mental health outcomes. Higher educational attainment was associated with lower PHQ-9 (p=−0.112), GAD-7 (p=−0.068) and AD-8 (p=−0.211) scores (all p<0.001), underscoring education’s protective role across depression, anxiety and cognitive decline.

Multinomial logistic regression analyses (table S5) identified several significant factors associated with depression, anxiety and cognitive decline among older adults. Higher educational attainment was consistently associated with lower odds across all outcomes, including the borderline depression group (OR=0.95; 95% CI 0.92 to 0.97), the borderline anxiety group (OR=0.95; 95% CI 0.92 to 0.98), and the cognitive decline risk (OR range 0.94 to 0.91 across different groups).

Female participants had lower odds of being in the borderline or high-risk depression (OR range 0.66 to 0.76) and anxiety groups (OR=0.63) and were less likely to have borderline cognitive impairment (OR=0.72). Age ≥85 was associated with increased risk across all domains, particularly cognitive decline (high-risk group OR=4.84; 95% CI 3.77 to 6.20). Living in rural areas was associated with elevated depression risk (OR=1.22), while chronic disease was a robust risk factor for all three outcomes (eg, high-risk depression OR=3.89; cognitive decline OR=3.41).

Living alone markedly increased risk for depression (OR=3.79), anxiety (OR=3.29) and cognitive impairment (OR=1.44). Nursing home residence conferred especially high odds for cognitive decline (OR=10.07). Widowed or separated marital status was associated with increased mental health risks; in particular, separated marital status was associated with higher odds of borderline depression (OR=4.31) and cognitive decline (OR=3.71). Vulnerability factors such as disability and dementia greatly elevated odds across all domains. Economic hardship and parents affected by the one-child policy were also significantly associated with increased high-risk classification, particularly for cognitive decline.

These findings collectively reveal that mental health risks among older adults are shaped by a complex web of demographic, social and health-related factors. Notably, higher education emerged as a consistent protective factor across all domains. This supports the cognitive reserve theory, suggesting that education enhances resilience to neurodegeneration and enables better coping with life stressors. Access to information, stronger health literacy and greater engagement in mentally stimulating activities may also mediate these effects. Moreover, those with more education may benefit from better social support, financial stability and healthcare utilisation.⁶

Contrary to global trends reporting higher depression and anxiety rates among women, this study found that female gender was associated with lower odds of depressive, anxious and cognitive symptoms. This may reflect gendered coping styles and social networks, with older women in China potentially more engaged in reciprocal caregiving or community roles.² Alternatively, it may indicate underreporting among men due to stigma around mental illness. Public health strategies should thus include gender-sensitive mental health screening and interventions tailored to older men, who may be less likely to seek help voluntarily.

As expected, advanced age was a strong risk factor for cognitive decline and, to a lesser extent, for depression and anxiety. Both biological ageing processes (eg, neurodegeneration, frailty) and psychosocial factors (eg, bereavement, role loss) may underlie these associations.⁷ Targeted outreach to individuals in advanced age—especially those living alone or in institutional care—should prioritise social connectedness, multimorbidity management and cognitive training.

Chronic disease emerged as one of the strongest predictors of all outcomes. These findings reflect the bidirectional relationship between physical and mental health.⁸ Chronic illnesses, particularly cardiovascular and metabolic disorders, contribute to systemic inflammation and cerebral vascular damage that can exacerbate cognitive and emotional symptoms. Primary care integration of physical and mental health screening could greatly improve detection and early intervention among high-risk groups.

Social isolation and precarious living arrangements such as living alone, with other relatives (ie, relatives other than children or spouses), or in nursing homes were associated with worse mental and cognitive outcomes. While institutional settings may offer basic care, they often lack stimulation and autonomy.⁹ Conversely, multigenerational or spousal living arrangements were associated with better mental health. Programmes fostering intergenerational bonding, community participation and improved nursing home environments may help buffer these risks.

Economic hardship and disability, though less prevalent in the sample, were strong predictors of adverse outcomes.⁹ Individuals facing these challenges often lack access to care, face physical limitations and experience greater psychological stress. Strengthening the safety net through subsidies, mental health services and home-based support for disabled older adults is essential.

Although this study offers a comprehensive and regionally representative view of mental health among older adults in Guangxi, it has limitations. Its cross-sectional design precludes causal inference. Additionally, unmeasured lifestyle variables (eg, diet, physical activity), social engagement and medication use may confound the associations observed.¹⁰ Around 5% of participants were excluded due to missing data and limited access to item-level responses precluded formal imputation or missingness analysis, raising the possibility of selection bias. Finally, as this was a cross-sectional study, all associations are correlational in nature and cannot establish causal relationships.

In conclusion, this study provides evidence of significant disparities in mental and cognitive health among elderly adults in Guangxi. Education, gender, age, health status and social context emerged as key determinants. These findings suggest that future ageing policies should prioritise vulnerable subgroups through enhanced outreach, integrated care models and socially embedded interventions. By addressing the multifactorial nature of ageing and mental health, China can better support its rapidly growing ageing population.

Biography

Zuxing Wang graduated with a Master of Medicine from Sichuan University, China in 2019. Since 2019, he has been affiliated with the Sichuan Provincial Center for Mental Health at Sichuan Provincial People's Hospital (a teaching hospital of the University of Electronic Science and Technology of China), where he currently holds the position of consultant psychiatrist. His primary research focuses on the neuropathogenesis, epidemiology, and therapeutic interventions for psychiatric conditions, particularly major depressive disorder and anxiety disorders.