This cross-sectional study was conducted in Shanghai, a typical Chinese mega city, in 2024, aiming to assess social media use and its association with mental well-being among retired older adults. This study was conducted in 2 phases using convenience sampling. In the first phase, the questionnaire was distributed through universities for older adults in Shanghai. In the second phase, after conducting a preliminary analysis of the questionnaires collected in the first stage, the areas with a smaller number of collected questionnaires were identified. Questionnaires were distributed through primary health care institutions, namely, community health service centers (CHSCs), to areas that either were not initially covered or had a smaller number of responses. All the questionnaires were distributed online. The inclusion criteria were as follows: (1) being retired, (2) no terminal illnesses, (3) ability to speak and read Chinese, and (4) no severe concurrent psychological or psychiatric diseases or other physical diseases that impact physical activities or mental health (self-report diagnosis). The exclusion criteria were as follows: (1) severe cognitive impairment, (2) active psychiatric disorders, and (3) inability to use a smartphone independently. A total of 15,986 valid online questionnaires were collected, and the questionnaire response rate was 99.95% (15,986/15,994).

Short versions of anxiety and depression scales were used to minimize the time required to complete the questionnaires for older adults. The 2-item Generalized Anxiety Disorder scale (GAD-2) and the 2-item Patient Health Questionnaire (PHQ-2) were administered. The GAD-2 and PHQ-2 each consist of 2 questions, with each question scored from 0 to 3. The total possible score for each scale is 6 points, with a score of 2 or higher indicating clinically significant symptoms warranting further diagnostic evaluation [21,22]. While we acknowledge that they have a lower sensitivity than that of the full scales (the 9-item Patient Health Questionnaire and 7-item Generalized Anxiety Disorder scale), these instruments demonstrate adequate validity for population-level screening in older adults (area under the curve=0.78-0.84), particularly given the optimized acceptability in older cohorts with cognitive fatigue concerns [23-25].

Social media use was represented using 2 variables: time spent on social media and social media addiction. Time spent on social media was measured using the self-reported amount of time spent on social media each day, categorized into 4 groups: ≤1 hour, 2 to 3 hours, 4 to 5 hours, and ≥6 hours. Social media addiction was assessed using the Bergen Social Media Addiction Scale, with a cutoff score of 19. A score above 19 indicates addiction to social media [26].

The covariates in this study included sex (male=1; female=2), age, educational level (senior high school and lower=1; junior college=2; university and higher=3), marital status (married and living with spouse=1; married but not living with spouse=2; divorced or widowed=3; unmarried=4), urban or rural residence, living status (living alone=1; not living alone=2), and monthly income (≤¥6000 [US $842.05]=1; ¥6001‐¥9000 [US $842.19-$1263.07]=2; ≥¥9001 [US $1263.21]=3) of the active older individuals. Health behavior–related variables encompassed physical activity (regular exercise=1; no regular exercise=2), sedentary hours (<0.5=1; 0.5‐4=2; >4=3), sleep duration (≤6 hours=1; 7‐8 hours=2; ≥9 hours=3), and social engagement levels (hardly ever socialize=1; sometimes socialize=2; often socialize=3). All the covariates were based on the self-reports of the research subjects.

For the descriptive statistics, we used the mean (with SE) for continuous variables and the number (with weighted percentage) for categorical variables. To identify association factors for anxiety and depression in older adults, we conducted logistic regression to calculate odds ratios (ORs) with 95% CIs. Adjusted ORs were obtained through multivariate regression analyses, using selection methods to account for potential confounding variables. Subgroup analyses were conducted to explore the associations between social media use and mental health among individuals of different sexes, ages, educational levels, marital statuses, monthly incomes, urban and rural residences, and living conditions. Database management and all statistical analyses were conducted using the Survey package in R (version 4.3.1; R Foundation for Statistical Computing) and ArcMap (version 10.8; Esri), which are software for statistical computing and plotting, respectively. A 2-sided P value of <.05 was considered statistically significant.

This study was approved by the Ethics Committee of Public Health and Nursing Research, Shanghai Jiao Tong University School of Medicine (approval: SJUPN-2024-041-KS1). Written informed consent was obtained from all participants, with procedures for illiterate subjects involving witnessed verbal consent documented using institutional review board–approved forms. For secondary data analyses, original consent permitted anonymized reuse. All data were deidentified through replacement of personal identifiers with unique codes and aggregated for analysis to ensure confidentiality. There was no charge for participating in the study, and no compensation was provided to the participants. No identifiable images or personal information appear in this manuscript or the supplementary materials. This study was conducted in accordance with the Declaration of Helsinki.

Table 1 shows the descriptive statistics of the sample. The average age of the respondents was 68.49 (SD 7.6) years. Men constituted 51.0% (8155/15,986) of the sample. A significant portion (5633/15,986, 35.2%) had received education at the university level or higher. Most respondents (13,854/15,986, 86.7%) were married and living with their spouse. A substantial number (9669/15,986, 60.5%) had a monthly income exceeding ¥9000 (US $1263.07), and 67.8% (10,833/15,986) resided in urban areas. The prevalence rates of anxiety and depression were 20.1% (3218/15,986) and 21.7% (3467/15,986), respectively. Regarding daily social media use time, 15.4% (2463/15,986) spent no more than 1 hour, 52.3% (8361/15,986) spent between 2 and 3 hours, 25.9% (4134/15,986) spent between 4 and 5 hours, and 6.4% (1028/15,986) spent more than 6 hours. The prevalence of social media addiction was 20.3% (3253/15,986). The social media platforms most used by older adults were WeChat, Douyin, and Kuaishou (Multimedia Appendix 1), and the most viewed type of content was current news and health-related information (Multimedia Appendix 1).

Among the older adults with longer social media use times, a higher proportion exhibited higher scores for anxiety and depression (284/1028, 27.6% and 304/1028, 29.6%, respectively). Compared to those who used social media for 1 hour or less per day, those who used social media for 6 hours or more per day had 1.44 times the odds of anxiety (P<.001) and 1.50 times the odds of depression (P<.001). Among the older adults with social media addiction, a higher proportion had higher scores for anxiety and depression (1140/3,253, 35.0% and 1146/3253, 35.2%, respectively). The odds of anxiety were 2.81 times higher in those with addiction versus those without (P<.001), whereas their odds of depression were 2.51 times higher (P<.001; Table 2)

We observed that, in Shanghai, a typical Chinese megacity, older adults seem to have entered a new digital era. Our research found that over 98% of retired older adults in Shanghai (15,807/15,986, 98.8%) had ever used social media, particularly platforms such as WeChat, Douyin, and Kuaishou, with over 30% of them (5162/15,986, 32.3%) using it for more than 4 hours a day and over 20% (3253/15,986, 20.4%) being addicted to it. The frequency of social media use and the addiction rate among older adults in Shanghai have reached or even surpassed those of young people. Riehm et al [27] reported that approximately 80% of adolescents use social media for no more than 3 hours. In a study conducted by Eichenberg et al [28], approximately 20% of students were identified as being addicted to social media. On buses or subways in China, an increasing number of older adults being “phubbers” (people who are constantly looking at their phones) can be observed, mainly browsing short videos and social media feeds such as Moments on WeChat. Current social media platforms offer a variety of social features for older adults, including adding new friends and creating social circles, having voice and video conversations, sharing and commenting on daily life, and browsing and interacting with short videos. However, Jin et al [29] pointed out that more older adults are using social media such as WeChat and short videos, which can affect their daily life and mental health, possibly leading to addiction.

In our study, a positive correlation was found indicating that longer durations of social media use were associated with increased anxiety. Specifically, we found that social media use duration became highly associated with anxiety once it exceeded 6 hours, and such difference was not observed among those who used social media for less than 1 hour, 2 to 3 hours, or 4 to 5 hours. Similarly, the relationship between time spent on social media and depression was also positive and more prominent than that for anxiety. Existing research, which is limited, has provided contradictory evidence. Yan et al [30] investigated the association between internet exclusion and depressive symptoms among older adults in 32 countries and found that internet use was negatively associated with depressive symptoms among older adults. Due to limitations in public databases, the aforementioned study only examined the relationship between internet use (yes or no) and depression among older adults, not the duration of social media use or social media addiction. In contrast, our analytic framework aligns with a growing body of evidence that documents a positive association between social media use frequency and intensity and depressive symptoms in later life [31,32]. Two factors appear to underlie this positive association among “young-old” retirees. First, active, healthy older adults increasingly replace offline social, physical, and community engagements with prolonged online activity, eroding face-to-face support networks and diminishing the mood-regulating benefits of in-person interaction. Second, older users predominantly engage in passive content consumption rather than active creation, resulting in frequent upward social comparisons and algorithm-amplified exposure to negative information—such as health-related rumors and sensationalized news [33]. These passive, comparison-driven experiences exacerbate feelings of inadequacy, anxiety, and hopelessness, thereby translating extended daily social media use into heightened depressive and anxiety symptomatology.

We attempted to explain the negative impact of social media on the mental health of older adults, and we infer that it may be related to their use habits. When older adults use social media, they often scroll and browse, such as watching short videos on platforms such as Douyin or scrolling through their Moments feed on the WeChat app, which is classified as passive social media use. According to Verduyn et al [34], passive social media use is defined as monitoring or observing content and behavior on social media platforms without direct exchanges with others (eg, scrolling and browsing), whereas active use involves more direct interactions such as posting, commenting, and sending private messages [35]. A substantial body of research has revealed a positive relationship between passive social media use and anxiety or depression among younger individuals [36]. Passive social media use is considered a cause of stress and anxiety, largely due to upward social comparison triggered by prevailing positive self-presentations [37,34]. In addition, the older adult group involved in this study was in the process of entering old age, with many having just started retirement or experiencing the empty nest stage. This cohort of older adults is unique in that they are part of the internet era, being familiar with smartphones and internet browsing. Spending a long time on social media tends to decrease their opportunities for face-to-face communication with family and friends, and passive use of social media further limits their chances to express themselves and engage in interactive communication. Further research should be conducted to delve into the pathways and mechanisms through which prolonged social media use among these younger members of the older population leads to anxiety and depression.

Additionally, our study provides clear evidence of a strong association between social media addiction and mental health issues among older adults. The odds of anxiety and depression for older individuals with social media addiction were 2.80 and 2.51 times higher, respectively, than for those without addiction (P<.001). This positive correlation has been consistently found in existing studies targeting adolescent populations [38,39]. Relatively few studies have focused on social media addiction among older adults; however, the results consistently show that social media addiction is highly correlated with mental health problems [40]. Some studies also suggest that mental health problems may lead to social media addiction among older adults [41]. Further exploration is needed to understand the causal relationship between the two.

Our study further identified that female, younger, married, and non–solo-dwelling older individuals are more likely to experience anxiety and depression when using social media. Sex and age factors have been partially explained in existing literature involving adolescents, such as female and younger individuals being more sensitive and active [40]. Özbek and Karaş [41] pointed out that sex roles may restrict women socially, so women may prefer to use social media to meet their social needs at home. Older individuals who are married and living with family or partners have also been identified (this was derived from 2 questions in the questionnaire. One was about marital status, and the other was about who one currently lives with, suggesting that, even when surrounded by others, they may still be at risk because social media can supplant direct human interaction, which in turn can be detrimental to their mental well-being [41]. In addition, we found that the association between social media use and mental well-being varied significantly by SES, with anxiety and depression being more prevalent among individuals with lower SES (lower income and educational level). Wang et al [42] identified area-level SES as a moderator between internet use and depressive symptoms, with an explanation that individuals residing in lower-SES regions were constrained by limited resources and social support, which could exacerbate the association between internet use and depressive symptoms. However, we also noticed that highly educated older adults who were addicted to social media were more likely to experience depression. We believe that the causal relationship requires further investigation. In addition, the mental health of older people in cities is more susceptible to the influence of social media, which is consistent with previous studies. Xu and Zhang [43] believe that although urban older people have easier access to the internet, information overload, online social comparison, and substitution of offline social interaction might be the underlying causes of higher levels of depression.

Drawing on our findings, we propose the following evidence-based policy recommendations. First, regulatory authorities should prioritize the development of age-friendly digital ecosystems by mandating algorithmic reforms that promote health-related content and enforce periodic use break reminders. This addresses our critical finding that older adults’ social media engagement has reached levels comparable to those of younger users. Second, community health services should implement offline alternative programs such as subsidized “screen-free hour” activities and intergenerational interaction hubs to mitigate the mental health risks associated with prolonged screen time. Finally, resource allocation should be precisely targeted toward the high-risk subgroups identified in our stratified analyses—specifically, urban women aged 60 to 74 years, individuals with low SES, and non–solo dwellers—through neighborhood-based digital literacy workshops, mental health screenings at CHSCs, and family-mediated device management initiatives. These measures aim to foster a healthier integration of the digital and physical aspects of life for China’s aging population.

Some limitations remain. First, although we targeted the older population, we mainly focused on the relatively younger and more vibrant segment of this demographic as they are the primary group of older social media users. However, we are not clear on the situation of more advanced-age older adults. Critically, we excluded working older people—omitting 8.9% of China’s population aged >65 years, who may experience distinct digital stressors. Meanwhile, despite efforts to balance district-level representation through supplementary CHSC sampling, participants from these 2 sources differed in age and education profiles. Caution is warranted when generalizing the findings to isolated older adult populations beyond educational institutions. Second, self-reported data may incur recall bias, and the GAD-2 and PHQ-2 are screening tools, not clinical diagnoses. Our cutoff scores (≥3) may not generalize across subpopulations. Third, we merely identified the relationship between social media use and mental health, but the causality between the two remains unclear. A cohort study is recommended in the future, which could further clarify this causal relationship. Fourth, the links between social media use and mental health issues are not direct, and various factors such as prolonged sitting, sleep, and social support might play a mediating role, which needs further exploration. Fifth, despite our screening procedures (self-report diagnosis), residual confounding from undiagnosed clinical disorders may persist. Individuals with subclinical anxiety or depression might overuse social media and report poorer mental health, inflating the observed associations. Future studies should incorporate structured clinical interviews to better control for this confounder. Finally, exploratory subgroup analyses did not include formal interaction testing due to sample constraints.

We focused on retired older adults instead of on adolescents and explored the impact of social media use on their mental well-being, yielding some interesting findings. We found that older adults in Shanghai have a high frequency of social media use, with over 30% of them (15,807/15,986, 32.3%) spending more than 4 hours per day on social media and over 20% (3253/15,986, 20.4%) exhibiting signs of social media addiction. Significant associations between social media use, including time spent on and addiction to social media, and mental health, including anxiety and depression, were identified. Furthermore, subgroup analysis revealed that older individuals with lower SES who were female, younger, married, and not living alone were more prone to experiencing anxiety and depression through their engagement with social media.