This cross-sectional study used secondary data from 2014 to 2020 from an ongoing community-based study of aging and health in Zhejiang Province, eastern China. The program was initiated in 2014 and included 11 randomly selected counties/districts out of a total of 90: Tongxiang, Haishu, Changshan, Jingning, Haiyan, Yuhuan, Yuecheng, Yiwu, Binjiang, Nanxun, and Putuo. For each county/district, between one and three towns/sub-districts were randomly selected according to their population density. Several villages/neighborhoods were randomly selected for each town/sub-district. All residents of the selected villages/neighborhoods who met the following inclusion criteria were invited to participate in the study: (1) permanent residents aged 60 years and over, and (2) lived in the area for at least 6 months of the previous year. Multiple individuals from the same household could participate if they each met the inclusion criteria. At least 1,500 participants were randomly recruited in 10 counties/districts; 600 were enrolled in the remaining district because of the sparse population. Seven counties/districts completed the baseline survey in 2014–2015 and four in 2018. New older adults who met the inclusion criteria were invited each year and completed the baseline survey.

The baseline survey was conducted by trained public health practitioners and nurses and involved face-to-face interviews with each participant. Written informed consent was obtained from all participants. Individuals who were unable to read had the informed consent information read to them by their interviewers and they signed with their fingerprints.

By 2020, 21,367 participants had completed the baseline survey. Among the participants, one had missing sex data, eight had missing age data, 23 had missing arthritis data, and 32 had missing depressive symptom data. In total, 40 participants had missing data; 21,327 participants had complete data. Participants ranged in age from 60 to 100 years, with a mean age of 68.21 years; 47.12% were men and 52.88% were women. Compared to participants with incomplete data, those with complete information tended to be younger, married, and working (p < 0.05). No significant differences were observed in sex, ethnicity, or perception of economic status (p > 0.05). A total of 1,081 participants with self-reported doctor-diagnosed arthritis were included in this study.

Sociodemographic characteristics, health conditions, health behaviors, and social support factors were examined in terms of their univariate associations with depressive symptoms. The chi-squared test, t-test, and chi-square test for trend were used for univariate analysis, as appropriate. Exact p-values were calculated when necessary. Sex, age, education, and variables showing a statistical association with depressive symptoms in the univariate analyses were further included in the multivariate analysis. Multilevel logistic regression with random intercepts was performed to examine factors associated with depressive symptoms in patients with arthritis. The county/district was treated as a group variable in the regression model, considering the area differences in depressive symptoms.

All analyses were performed using STATA software version 16 (StataCorp LLC, College Station, TX, United States). All statistical tests were two-sided. Statistical significance was set at p < 0.05.

The mean age of older patients with arthritis was 69.16 ± 7.13 years, with a gender ratio of 0.75:1 (men:women). Depressive symptoms were experienced by 14.99% (95% confidence interval: 12.91–17.26%) of these individuals, a significantly greater proportion than that of older individuals without arthritis (8.49%, p < 0.001) in the 11 selected counties/districts of Zhejiang Province.

The sociodemographic characteristics and their univariate associations with depressive symptoms are shown in Table 1. A greater proportion of participants with depressive symptoms than those without depressive symptoms were women (68.52% vs. 55.06%, p = 0.001). Individuals with depressive symptoms were more likely to be illiterate (58.75% vs. 46.93%, p = 0.006), single/widowed/divorced (27.78% vs. 18.72%, p = 0.008), and self-report as having a poor economic status (19.75% vs. 5.66%, p < 0.001) than those without depressive symptoms. No statistical difference was found in age, ethnicity, or employment status between patients with arthritis with and without depressive symptoms (p > 0.05).

Health conditions and health behaviors and their univariate associations with depressive symptoms are shown in Table 2. Participants with depressive symptoms tended to be diagnosed with arthritis earlier (p = 0.022) and had a longer duration of arthritis (p = 0.041) than those without depressive symptoms. Participants with depressive symptoms were more likely to have multimorbidity (43.21% vs. 29.27%, p < 0.001), cognitive impairment (19.14% vs. 12.40%, p = 0.020), and limitations in activities of daily living (16.05% vs. 6.31%, p < 0.001) than those without depressive symptoms. Significant differences were observed in smoking (p = 0.037) and alcohol consumption (p = 0.009) between participants with and without depressive symptoms. Participants with depressive symptoms were less likely to be tea drinkers (18.52% vs. 30.25%, p = 0.002) and were more sedentary (3.09 vs. 2.64 h/day, p = 0.007) than those without. No significant differences in the history of stroke, BMI, blood pressure control, physical activity, vegetable consumption, or fruit consumption were observed between patients with arthritis with and without depressive symptoms (p > 0.05).

Social support factors and their univariate associations with depressive symptoms are presented in Table 3. There was a significantly greater proportion of eating alone (21.60% vs. 11.21%, p < 0.001) and living alone (20.37% vs. 9.58%, p < 0.001) among participants with depressive symptoms than among those without. Having a child did not significantly affect the prevalence of depressive symptoms (p < 0.05).

Sex, age, education level, marital status, perception of economic status, age at diagnosis of arthritis, arthritis duration, multimorbidity, cognitive impairment, limitations in activities of daily living, smoking, alcohol consumption, tea consumption, sedentary hours, living alone, and eating alone were further included in a multilevel logistic regression analysis. Associations between the selected factors and depressive symptoms are shown in Table 4. The perception of poor economic status showed a strong statistical association with depressive symptoms compared to the perception of a high economic status (odds ratio [OR] = 5.52, p < 0.001). Age at arthritis diagnosis was negatively associated with depressive symptoms, with a later diagnosis corresponding to lower odds of depressive symptoms (OR = 0.67, p = 0.046). Multimorbidity (OR = 1.96, p = 0.001), limitations in activities of daily living (OR = 2.36, p = 0.004), and living alone (OR = 3.13, p = 0.026) were all significantly associated with depressive symptoms. All remaining variables failed to reach statistical significance (p > 0.05).

The prevalence of depressive symptoms, although high in patients with arthritis, varies according to the type of arthritis, method of assessing depressive symptoms, and population studied (25). A meta-analysis published in 2020 revealed that patients with RA had a 47% higher risk of depression than older adults without RA (8). A separate meta-analysis in the same year showed that 17% of individuals with psoriatic arthritis experienced depression, 1.68 times higher than the proportion of those without psoriatic arthritis (26). A meta-analysis in 2016 estimated that the pooled prevalence of depressive symptoms was 19.9% among patients with OA, with a relative risk of 1.17 compared to individuals without OA (27). When including all forms of arthritis, the national prevalence of depressive symptoms classified based on a frequency question and an intensity question of depressed feeling in the United States between 2015 and 2017 was 12.1% among adults with arthritis and 4.7% among those without arthritis (28). The present study on community-dwelling older adults revealed a prevalence of depressive symptoms of nearly 15.0% in patients with arthritis, approximately 1.8 times higher than in those without arthritis. A similar gap was observed in a longitudinal study on older Chinese patients with arthritis (29). Evidence has therefore consistently demonstrated that patients with arthritis experience a more significant burden of depressive symptoms than their arthritis-free counterparts.

The current study identified perception of economic status, multimorbidity, limitations in activities of daily living, living alone, and age at arthritis diagnosis as factors significantly associated with depressive symptoms in older patients with arthritis.

Previous studies have also associated poor economic status with depression, as evaluated by the Hamilton Depression Scale-17 and PHQ-9 in patients with arthritis (17, 30). The relationship between economic status and depressive symptoms can be partly explained by arthritis-related costs, which are more difficult to bear for those with lower incomes, as indicated by a qualitative study performed in Australia (31). In addition, depressive symptoms as assessed using the CESD-10 have been associated with higher healthcare costs and productivity loss among Chinese individuals with 12 chronic diseases (32), which may impact economic status. Furthermore, lower socioeconomic status has been related to multimorbidity patterns of cardiovascular and metabolic disorders (33), indicating that economic conditions may affect depressive symptoms partly through multimorbidity.

The present study revealed that multimorbidity increased the odds of depressive symptoms among older patients with arthritis nearly two-fold. Our results are consistent with those of two previous studies. The United States National Health Interview Survey (2015–2017) found that the prevalence of depressive symptoms classified based on a frequency question and an intensity question of depressed feeling in patients with arthritis and more than two chronic conditions was double that of patients with one or two conditions (28). The Korea National Health and Nutrition Examination Survey revealed that patients with RA with three or more comorbidities had five-fold higher odds of depression as measured by the PHQ-9 than those without (34). In addition to its relationship with depressive symptoms, multimorbidity has been found to be related to limitations in activities of daily living in China, India, and the United States (35–37). This suggests that multimorbidity may be indirectly linked to depressive symptoms through functional limitations.

Our results associating limitations in activities of daily living with depressive symptoms in older patients with arthritis is consistent with the findings of many previous studies. A study in Korea found that patients with RA were three times more likely to experience depression, as assessed by the PHQ-9, if they also had limitations in activities of daily living (34). A study in Taiwan Province of China also associated higher functional ability with lower levels of depression measured by the Geriatric Depression Scale (38). A longitudinal study conducted in China revealed a reciprocal relationship between depressive symptoms, defined by the CESD-10, and mobility disability in patients with arthritis (5), reflecting their complex interaction. Depressive symptoms, assessed using the Geriatric Depression Scale, can lead to functional limitations by impairing mobility capacities and exacerbating feelings of pain (39). In addition, limitations in activities of daily living can reduce independence and participation in social activities, thus increasing the risk of depressive symptoms (40).

In the current study, living alone was associated with depressive symptoms, as expected. Living alone can result in limited social and emotional support, and the latter was seen among patients with arthritis in relation to depression assessed by a modified version of the CESD in the United States (41) and the Beck Depression Inventory II in Japan (42). It has also been reported that limitations in activities of daily living were significantly more prevalent in older Korean adults living alone than in those living with relatives (43). People living alone might be more affected by limitations in activities of daily living because they receive less help from others. Therefore, we speculate that living alone has a greater influence on depressive symptoms in patients with arthritis and limitations in activities of daily living than in those without.

We found that arthritis diagnosed at an earlier age was significantly associated with depressive symptoms, even after controlling for disease duration. Age at diagnosis provides information on disease duration and reflects the impact of the disease at different life stages. It may be an important screening factor for depressive symptoms in patients with arthritis; however, supporting evidence is limited. There are several possible explanations for this finding. First, arthritis is debilitating and influences the ability of patients to work, thus affecting their economic status. Individuals diagnosed at an earlier age would be more affected by this. Second, patients with early-onset arthritis have a longer interval between disease onset and diagnosis (44), leading to negative psychological effects. Third, patients with late-onset arthritis may be better emotionally and mentally equipped to cope with the disease because they may already have comorbidities (44) and therefore have experience coping with illness. Fourth, patients of different ages at diagnosis are expected to have different clinical characteristics and thus depressive symptoms. Previous studies have reported conflicting results based on different indicators. Radiographic damage progression and time to remission did not differ between patients with early- and late-onset RA (45, 46). However, patients with late-onset RA had greater inflammation and bone erosion (47–49). Further large-scale research on socioeconomic and clinical factors is needed to examine these associations.

Smoking and alcohol consumption were not associated with depressive symptoms in our study, consistent with a recent meta-analysis of 11 studies involving older Chinese adults (40). Physical activity has been shown to be inversely associated with depressive symptoms as determined by the CESD among Chinese individuals, such as older veterans (50), middle-aged and older city dwellers (51), and older women (52). However, the current study did not observe such an association among older patients with arthritis. This may be due to different target populations and physical activity evaluations. According to the assessment of CESD-10, the association between physical activity and depressive symptoms depends on the intensity of the physical activity (53, 54). The pain and functional limitations experienced by patients with arthritis may not allow for activity of sufficient intensity to combat depressive symptoms.

In contrast to previous studies on the general population (55, 56), vegetable and fruit consumption were not associated with depressive symptoms in the present study. This may be due to the different types of vegetable and fruit consumed and the frequency of consumption. A study in Korea found an association between physician diagnosed depression and a high consumption of green vegetables and fruits, whereas no association was observed when the consumption of vegetables or fruit was moderate (57). A study in Japan found an association between tomato consumption and depressive symptoms assessed by the Zung Self-Rating Depression Scale, but not the consumption of any other vegetables (58). Further studies are required to determine the types and frequencies of vegetable and fruit consumption associated with depressive symptoms.

Tea consumption has previously been associated with lower odds of depressive symptoms, as assessed by the CESD-10 and Geriatric Depression Scale, among older Chinese adults (59, 60). The antidepressant effects of tea can be partially attributed to its antioxidant and anti-inflammatory constituents (59). The social aspects of tea drinking may also contribute to this effect (60, 61). However, tea consumption was not related to depressive symptoms in the present study. Regular tea consumption may be insufficient for its biomedical effect on depressive symptoms in patients with arthritis who are in a state of inflammation. Patients with arthritis may also have fewer opportunities for social tea consumption, as they are more likely to have functional disabilities.

The present study revealed a more significant burden of depressive symptoms in older adults with arthritis than their arthritis-free counterparts. It highlights the importance of recognizing and addressing the mental health needs of older patients with arthritis, which could further improve their health outcomes.

Our findings identified multidimensional associated factors of depressive symptoms among older patients with arthritis, including poor economic status, age at arthritis diagnosis, multimorbidity, limitations in activities of daily living, and living alone. These factors can be developed into a screening protocol for depressive symptoms and applied in both routine clinical arthritis care and community-based programs, making it possible for healthcare providers to prioritize resources and support for those who might be most likely to experience depressive symptoms. Furthermore, the inverse relationship between age at arthritis diagnosis and depressive symptoms observed in the study suggests that early intervention may play a crucial role in the management of mental health in older patients with arthritis. This could involve developing educational materials and support groups specifically for newly diagnosed patients aimed at helping them deal with depressive symptoms that might occur.

In conclusion, this study provides valuable insights into public health practice and clinical care for older patients with arthritis. By recognizing the heavy burden of depressive symptoms among older patients with arthritis and the associated multidimensional factors, healthcare systems could develop more comprehensive strategies to better support this vulnerable population, improving not only their physical health but also their mental well-being.

Our data were collected from a community-dwelling population and should therefore have better generalizability than previous studies performed in clinical settings. Many factors were included in the analysis, offering a more comprehensive picture of depressive symptoms in patients with arthritis. Nevertheless, our study has certain limitations. First, we did not examine the temporal association between arthritis and depressive symptoms. However, emerging evidence indicates that their correlation is probably bidirectional, as they impact each other and share mechanisms such as inflammation (62). The factors identified in the present study could serve as screening indicators for depressive symptoms among patients with arthritis in the community and play an important role in identifying those who require additional mental health care. Second, the presence of arthritis was determined by asking patients whether they had been previously diagnosed; verification of the diagnosis of arthritis was not performed, and misclassification of arthritis status may have occurred. A nationally representative study in older Chinese adults found that the prevalence of self-reported doctor-diagnosed arthritis was close to that of symptom-based diagnosed arthritis, with the prevalence of undiagnosed arthritis approximately 5% (63). Therefore, the potential number of misclassified cases, if any, is believed to be low. Third, disease intensity, such as disease activity and pain (15, 64, 65), have previously been shown to be associated with depressive symptoms in patients with arthritis; however, this information was not available in our study. Therefore, their association with depressive symptoms could not be analyzed. Fourth, treatment has also been suggested to be related to depressive symptoms among patients with arthritis (66), but was not included in the current analysis. The prevalence of untreated arthritis was estimated by a large-scale study of older Chinese individuals in 2007–2010 to be over 30% (63). Missing therapy data should be considered when interpreting these results. However, the prevalence of treated but uncontrolled arthritis has been reported to be up to 77% in older patients in six middle-income countries, including China (67), indicating unsatisfactory treatment outcomes of patients with arthritis. Therefore, the supposed benefits of arthritis therapy on depressive symptoms may be affected by its low efficacy.