Original Research••

Triple digital divide and depressive symptoms among middle-aged and older Chinese adults: a disparity analysis

Abstract

Background The triple digital divide refers to the lack of internet access, use and knowledge among specific populations. In China, middle-aged and older adults and those living in rural areas or various regions of the country are more likely to have limited internet access and skills and, thus, have less accessibility to internet services. Few longitudinal studies have explored the association between the digital divide and the progression of depressive symptoms among middle-aged and older Chinese adults. Significantly, none of the existing studies have estimated this long-term relationship from a disparity perspective.

Aims This study investigates the association between the triple digital divide and depressive symptom trajectories among middle-aged and older adults in China during a 10-year follow-up period from 2011 to 2020.

Methods The sample for this secondary analysis comprises 3019 urban and 10 427 rural respondents selected from the China Health and Retirement Longitudinal Study baseline survey in 2011. Depressive symptoms were measured using the Center for Epidemiologic Studies Depression Scale. Employing longitudinal mixed-effects models, this study explored the association between the triple digital divide and depressive symptom trajectories among middle-aged and older Chinese adults by examining gender, rural–urban and regional disparities in this relationship.

Results Our findings revealed a significant association between the triple digital divide and increasing trajectories of depressive symptoms, showing significant disparities based on gender, rural–urban dwelling and regional location. Notably, for both male and female participants who resided in urban areas or the central region of the country, their ability to use the internet, coupled with enhanced internet skills and greater access to internet services, was found to have a mitigating effect on the increasing trajectories of depressive symptoms.

Conclusions To alleviate some of the confounding influences on the trajectory of depression in middle-aged and older adults, policymakers in China should continue to prioritise the development of internet technology, foster easy access to the internet to ensure it is ‘elder-friendly’, provide internet skill training platforms for this population and broaden access to various internet services appropriate for them. Additionally, the implementation of tailored interventions to address depression, especially targeting the more vulnerable cohorts, such as middle-aged and older women, those residing in rural areas and the western regions, is crucial. Such tailored approaches are essential for addressing the disparities and challenges associated with the triple digital divide.

What is already known on this topic

  • As the internet and information technology rapidly develops in China, the impact of the digital divide on depressive symptoms among middle-aged and older adults is becoming increasingly apparent. While certain studies have zeroed in on specific dimensions of this phenomenon, there remains a gap in research examining the associations between the triple digital divide and depressive symptoms trajectories among middle-aged and older Chinese adults, particularly from the perspective of health disparities.

What this study adds

  • This study revealed that middle-aged and older adults in China who used the internet, acquired enhanced internet skills and accessed more internet services over time exhibited a slower increase in depressive symptoms. Significant gender, rural–urban and regional disparities were observed across the recent decade.

How this study might affect research, practice or policy

  • This study builds upon and expands the existing research on the effect of internet use on depression, offering a deeper understanding of the subject. The findings of this study are valuable in guiding targeted interventions via the internet to promote mental health among middle-aged and older Chinese adults, especially for the more vulnerable groups.

Introduction

Depression accounts for the largest proportion of the disease burden throughout the world.1 As the Chinese population of middle-aged and older adults experiences an unprecedented surge, depression has emerged as a pressing mental health concern for the country. The disorder concurrently heightens vulnerability to various bodily pains, functional debilitation and psychiatric comorbidities.2–4 Given the rapid development of the internet and information technology in China, the profound impact of the digital divide on mental health issues, particularly depression, among middle-aged and older adults, has come to the fore. The triple digital divide theory comprises three main components: access gap, usage gap and knowledge gap. The access gap reflects disparities in internet accessibility, namely whether or not individuals use the internet. This dimension determines whether individuals are able to connect to the online world and highlights the physical and economic barriers that hinder specific individuals or groups from accessing the internet. The usage gap focuses on variations in internet skills among users, reflecting the varying degrees of digital literacy and proficiency in effectively navigating and using the online space. Even when the internet is accessible, individuals may lack the necessary skills to leverage its full potential. The third digital divide is known as the knowledge gap, indicative of differences in the quantity of services individuals acquire via the internet. This dimension underscores the disparities in accessing and comprehending the vast amount of information available in the digital realm, which can further exacerbate social and economic inequalities.5 6 By considering these three dimensions collectively, we can better comprehend the intricate nature of the digital divide and devise more focused and effective strategies to tackle it. Given the rapid advancements in information technology and the internet, the triple digital divide plays a pivotal role in mental health, particularly concerning depressive symptoms (see online supplemental figure 1: the theoretical framework).7

Existing research on the association between the digital divide and depressive symptoms has yielded equivocal findings. Notably, empirical evidence suggests that using the internet can exert a mitigating effect on depression.7 According to the information processing model, individuals who use the internet can access valuable information and resources, enhance interpersonal interactions and improve their sense of control over life, thereby potentially reducing depressive symptoms.8 Among various internet skills, the use of mobile phones appears to play a particularly significant role in alleviating depression.7 Specific internet services, such as chatting, watching news and other entertaining activities, have the potential to trigger the release of dopamine in the striatum system, resulting in lower levels of depression.9 Nevertheless, a previous study asserted a lack of a direct linkage between depression and the use of the internet,10 while others have identified adverse impacts of internet use on depressive symptoms.11 12 Additionally, certain internet services, such as financial management and payments, may not necessarily correlate with diminished depressive symptoms.7 Furthermore, excessive and undesirable internet use, in line with technological stress theory and time displacement theory,11 can contribute to mental health issues that include depressive symptoms. Overall, while some studies have focused on the effects of specific dimensions of the digital divide on depression, few have examined the associations between this divide and the trajectories of depressive symptoms within its context.

The investigation into the intricate relationship between the digital divide and depressive symptoms among older Chinese adults over an extended duration deserves meticulous attention. China witnessed rapid internet development and other information technologies beginning in the 1990s. Over the following decades, numerous Chinese individuals entered middle or old age and were exposed to these newfangled technologies, but they did not know how to use them; hence, the term ‘digital refugees’ emerged to describe this cohort.11 Compared with their western counterparts, middle-aged and older adults in China may encounter distinct challenges and difficulties in using the internet because they had no exposure to such technology in their early education. Many of this group were only more recently introduced to the internet but had no formal means of learning how to navigate it, so they discovered its use gradually as taught by younger family members, neighbours or friends.

This phenomenon offers a unique context to comprehend the impact of the digital divide on depressive symptoms within this demographic. Because most of the relevant longitudinal research studies have been conducted predominantly in Europe and the USA, their findings cannot be generalised across different ethnicities and nationalities. Consequently, there is a notable gap in the existing literature, emphasising the need for such longitudinal studies within the Chinese context. Such research offers a more nuanced understanding of the association between the digital divide and the trajectories of depressive symptoms throughout the lifecycle in this specific demographic, filling a crucial void in our knowledge.

Health disparities are a significant social and economic issue that scholars in various disciplines have long studied and recognised. By understanding the underlying causes and mechanisms of these disparities, we can gain insights into the complex interactions between social, economic and environmental factors that shape individual and population health outcomes. Health disparities, particularly those rooted in gender, socioeconomic status and geographical location, represent a significant global challenge that cannot be ignored.13 Existing literature offers diverse insights into gender differences in the correlation between the digital divide and depression, highlighting the need for a nuanced understanding of these complexities.11 Furthermore, in China’s rapidly changing urban and rural structure, it is imperative to consider urban–rural disparities in the correlation between internet use and depression among middle-aged and older adults. These disparities can have significant implications for health outcomes and equitable healthcare services.14 Moreover, mainland China’s distinct economic regions—eastern, central and western—each possess unique socioeconomic landscapes and developmental trajectories.15 These regional disparities can influence the longitudinal impact of the digital divide on depressive symptoms, necessitating a regionalised approach to understanding and addressing these issues.16 Therefore, exploring gender, urban–rural and regional disparities in the longitudinal association between the digital divide and the trajectories of depressive symptoms assumes profound significance. By using theoretical frameworks such as the social determinants of health and digital divide theory, we hope to gain a deeper understanding of these disparities that could inform policies and interventions to reduce health inequities.

This study aimed to explore whether the digital divide correlates with depressive symptom trajectories among middle-aged and older Chinese adults over a 10-year follow-up period. We anticipated that the effect of using the internet, coupled with acquiring internet skills and internet services, would evolve and impact the trajectory of depressive symptoms. Additionally, we hypothesised that the association between the digital divide and the trajectory of depressive symptoms would vary across gender, rural–urban and regional groups. By conducting this investigation, we hoped to gain a deeper understanding of these complexities and their potential implications for mental health among this demographic.

Methods

Sample

The data in this secondary analysis study are from the China Health and Retirement Longitudinal Study (CHARLS), a national research initiative conducted in China. Commencing with the inaugural survey in 2011, subsequent surveys have been conducted biennially. The second cohort (wave 2) was interviewed in 2013, followed by wave 3 in 2015, wave 4 in 2018 and wave 5 in 2020.16 CHARLS aims to comprehensively gather participant information encompassing demographics, family structure, cognition, financial and housing wealth, income and consumption, employment and retirement, pension, health status and functioning, cognition and depression, and other pertinent domains. All participant information is meticulously gathered through face-to-face interviews.17

CHARLS employs a rigorous random sampling methodology. Initially, using the Probability Proportional to Size (PPS) technique, 150 districts and counties are collected randomly, considering regions, urban or rural designations, and per capita gross domestic product. Subsequently, three units are randomly chosen from each selected district or county using the PPS technique. Within each selected sampling unit, a comprehensive household list is generated by mapping all households, and 24 households are then randomly selected from the list as samples. Finally, one participant over 45 years and their spouse are chosen as respondents from each selected household. With a commendable response rate of 80.5%, the inaugural wave of the 2011 national survey successfully recruited 17 708 participants.18 This rigorous sampling approach ensures the representativeness and reliability of the data collected through CHARLS for examining the longitudinal relationship between the digital divide and depressive symptom trajectories among middle-aged and older adults in China.

The present study analysed data from 2011, 2013, 2015, 2018 and 2020, focusing on participants aged 45 years and over. The initial wave in 2011 included 17 708 individuals, with 13 446 retained in the final sample after applying a list-wise deletion method to address missing data. Similarly, the second wave in 2013 involved 18 612 participants, resulting in a final sample size of 15 097. The third wave in 2015 comprised 21 095 participants, narrowing down to 14 524 in the final sample. The fourth wave in 2018 included 19 816 participants, leading to a final sample size of 15 740 after list-wise deletion. Lastly, the fifth wave in 2020 involved 19 395 participants, and after list-wise deletion, the final sample consisted of 16 044 individuals. In aggregate analysis, the final sample for the longitudinal study encompassed 16 044 participants from wave 1 to wave 5. Of these, 13 446 individuals participated in all five waves, 1078 in four, 573 in three and 643 in two waves. The detailed processing of the sample is depicted in figure 1.

Figure 1
Request permissionrequest permission icon
Figure 1

Flowchart for sample processing in this study. CHARLS, China Health and Retirement Longitudinal Study.

Measurement

Self-reported depressive symptoms

Depressive symptoms were evaluated at each wave using the 10-item Center for Epidemiologic Studies Depression Scale (CES-D10). This condensed version, derived from the original 20-item version, exhibits comparable psychometric properties, as corroborated by studies.19 20 The CES-D10 Scale, incorporating eight negative and two positive questions, has been widely employed in the assessment of depressive symptoms among middle-aged and older adults.18 The scale gauges depressive symptoms by querying individuals about the frequency with which they have experienced specific symptoms associated with depression, such as feelings of depression, fear or loneliness, over the preceding 2 weeks. Participants were presented with four response options for each item: rarely or none of the time, some or a little of the time, occasionally or a moderate amount of time, and most or all of the time. Each response corresponds to a score ranging from 0 to 3, culminating in a total depressive symptoms score ranging from 0 to 30.16 A cut-off score of 10 is conventionally applied to this scale, with scores of 10 or higher indicative of a heightened likelihood of significant depressive symptoms, as supported by research findings.

Digital divide

Drawing from the digital divide framework encompassing the access, usage and knowledge gaps, as delineated in previous studies,6 this study operationalised the digital divide across three dimensions: whether or not individuals use the internet, their proficiency in internet skills and the extent of services accessed via the internet.

Internet use

Data on whether or not participants used the internet were derived from the CHARLS question, ‘Have you used the internet in the last month?’ Selecting 1 indicated ‘yes’, while 0 indicated ‘no’.11

Internet skills

The level of internet skills was assessed with the CHARLS question, ‘Which types of devices do you use to access the internet?’ Possible answers included a desktop computer, laptop, tablet and cellphone. Each response was assigned a binary value (0 or 1), resulting in a total score ranging from 0 to 4.7

Internet services

Data on internet services were acquired through the CHARLS question, ‘What do you usually do on the internet?’ Responses included chatting, watching the news, watching videos, playing games and financial management. Like the previous dimension, each response was assigned a binary value (0 or 1), resulting in a total score ranging from 0 to 5.

Confounding variables

We considered a series of demographic, socioeconomic and health-related variables that could confound the relationship between the digital divide and depressive symptoms. Demographic confounding variables included gender (male, female), age (≥45 years), marital status (married, unmarried), residence status (urban, rural) and region (eastern, central, western). A prior study21 identified gender, age and marital status as potential confounders that may impact the correlation between the digital divide and depressive symptoms. Residential status is determined by the location of where one’s hukou or residence permit was issued. The hukou system classifies Chinese citizens into two main categories: agricultural (rural) and non-agricultural (urban). This classification determines access to various social and economic benefits, including education, healthcare, employment, social welfare and so on, leading to disparities in psychological health among urban and rural residents.2 Regions were divided based on geographical location and level of economic development.22 In CHARLS, metropolises such as Beijing, Tianjin and Shanghai and provinces such as Hebei, Jiangsu, Zhejiang, Fujian, Shandong, Guangdong and Liaoning are classified as the eastern region. Provinces such as Shanxi, Anhui, Jiangxi, Henan, Hubei, Hunan, Jilin and Heilongjiang are classified as the central region. The western region encompassed the metropolis of Chongqing and provinces such as Inner Mongolia, Guangxi, Sichuan, Guizhou, Yunnan, Shaanxi, Gansu, Qinghai and Xinjiang.

Socioeconomic confounding variables included household expenditure, education and employment status (employed, not working). Household expenditure, treated as a continuous variable, objectively reflected the financial resources among middle-aged and older adults, encompassing expenses for food, communication, medical care, fuel, clothing, daily necessities, entertainment and more.18 Education was categorised into four levels: illiterate, elementary school, high school and above high school. Employment status was measured as a binary variable by asking participants if they were engaged in agricultural work, self-employment, unpaid family assistance or paid work. Those who participated in any of these types of work were classified as employed.

The health-related confounding variable was self-reported health status. Participants were asked to assess their health status and were subsequently grouped into five categories: very good, good, fair, poor and very poor. Self-reported health status provides valuable insights into individuals’ physical and mental well-being, including depressed mood.4

Statistical analyses

The study employed a linear mixed-effects model to analyse the intricate relationship between the digital divide and the trajectory of depressive symptoms. This statistical method, renowned for its adaptability in handling complex data structures, has found widespread application in mental health studies.23 24 The linear mixed-effects model offers a versatile framework that accommodates multiple outcome variables for each participant. It allows diverse outcomes to share parameters under the same set of covariates.23 Notably, this model is instrumental in capturing changes at both the individual and population levels over time, modelling participant-specific trends by leveraging information from individuals with similar characteristics.24 The linear mixed-effects model has several distinctive features. First, it accommodates unbalanced data, where participants may be measured at different times throughout the study.25 Second, the model allows for the inclusion of participants who may be missing at a given wave, without imposing any restrictions on the number of observations for each individual. A third crucial feature is the model’s capability to incorporate both time-varying and time-invariant covariates. This enables the estimation of the evolving relationship between covariates and outcomes across the study period.24 The linear mixed-effects model’s flexibility and robustness make it well suited for capturing the nuanced dynamics of the relationship between dependent variables and independent variables over a long period of time.

The linear mixed-effects model was used to explore the dynamic interplay between the triple digital divide and depressive symptom trajectories among middle-aged and older adults over a specified 10-year span in China. The analyses were stratified according to gender, rural–urban residence and regional subgroups. First, we explored the relationship between internet use and depressive symptom trajectories over 10 years. This involved analysing subgroups of males and females, urban and rural residents, and those residing in the eastern, central and western regions. The interaction terms of internet usage and the specific wave periods were incorporated to assess whether the association strengthened over time. Second, we explored the relationship between internet skills and depressive symptom trajectories, considering gender, rural–urban and regional subgroups. Notably, data regarding internet skills were available only for waves 3 (2015), 4 (2018) and 5 (2020). Interaction terms of internet skills and wave were included to explore trajectories over 6 years. Lastly, we examined the correlation between internet services and depressive symptom trajectories within gender, rural–urban and regional subgroups. Data about internet services were collected during wave 4 (2018) and wave 5 (2020). Interaction terms of internet services and each wave were incorporated to investigate trajectories over 3 years. All analyses were conducted using Stata V.17, providing a comprehensive understanding of the intricate relationship between the three dimensions of the digital divide and depressive symptom trajectories across diverse demographic contexts and time frames. We used the Strengthening the Reporting of Observational Studies in Epidemiology cohort checklist when writing our report.26

Results

Sample characteristics by rural–urban group

Online supplemental table 1 presents descriptive statistics for the sample. A total of 13 446, 15 097, 14 524, 15 740 and 16 044 participants provided data on the digital divide, depressive symptoms and several confounding variables. The average scores of depressive symptoms were 8.43, 7.83, 7.92, 8.45 and 8.62 from wave 1 to wave 5, respectively. Regarding the digital divide, the proportion of participants using the internet was 2.73%, 5.02%, 6.67%, 15.03% and 44.27% from wave 1 to wave 5, respectively. Furthermore, the average level of internet skills increased notably, registering 0.10, 0.19 and 0.49 at wave 3, wave 4 and wave 5, respectively. Additionally, the mean quantity of internet services at wave 4 and wave 5 was 0.36 and 0.97, respectively.

The trajectories of depressive symptoms by internet use in gender, rural–urban and regional subgroups

Online supplemental table 2 displays the interaction effects of using the internet and each wave on depressive symptom trajectories across gender, rural–urban and regional subgroups. After adjustments for potential confounders, significant effects were observed for the interaction terms of using the internet and each wave within all seven subgroups. Notably, individuals who did not use the internet exhibited a faster increase in depressive symptom trajectories compared with those who used the internet, as illustrated in figure 2.

Figure 2
Request permissionrequest permission icon
Figure 2

The trajectories of depressive symptoms and internet use in gender, rural–urban and regional subgroups over a 10-year follow-up. (A) The trajectories of depressive symptoms and internet use among middle-aged and older men. (B) The trajectories of depressive symptoms and internet use among middle-aged and older women. (C) The trajectories of depressive symptoms and internet use among middle-aged and older adults in urban areas. (D) The trajectories of depressive symptoms and internet use among middle-aged and older adults in the eastern region. (E) The trajectories of depressive symptoms and internet use among middle-aged and older adults in the central region.

Regarding gender differences, using the internet had a short-term effect on the trajectory of depressive symptoms among females, whereas it exhibited a long-term effect among males (online supplemental table 2). Specifically, during the transition from wave 1 to wave 2, middle-aged and older women who used the internet (β=1.396, 95% CI 0.45 to 2.34) experienced a steeper increase in depressive symptoms compared with their male counterparts (β=0.971, 95% CI 0.21 to 1.73). Subsequently, female internet users reported higher levels of depressive symptoms than non-users at wave 2, whereas among males, the reverse was true (figure 2A,B). When considering residential differences, using the internet had a long-term effect (waves 1–4) on the trajectory of depressive symptoms among the urban group, whereas no significant long-term effect was observed among the rural group (online supplemental table 2). Notably, among urban groups, internet users reported lower levels of depressive symptoms than non-users over a prolonged period (figure 2C). Regarding regional differences, a long-term effect of internet use on the trajectory of depressive symptoms was found among the eastern and central groups, but this long-term effect was absent in the western group (online supplemental table 2). Specifically, using the internet was associated with lower levels of depressive symptoms among the eastern and central groups (figure 2D,E).

The trajectories of depressive symptoms by internet skills in gender, rural–urban and regional subgroups

Online supplemental table 2 presents the interaction effects of internet skills and the respective waves on trajectories of depressive symptoms within gender, rural-urban and regional subgroups. In terms of gender differences, middle-aged and older women (β=−0.459, 95% CI −0.82 to –0.10) exhibited a faster increase in the trajectory of depressive symptoms with higher levels of internet skills compared with middle-aged and older men (β=−0.397, 95% CI −0.67 to –0.12). Among the male subgroup, the level of depressive symptoms among people with more internet skills was lower than among people with less internet skills (figure 3A). Conversely, among the female subgroup, while the initial level of depressive symptoms was comparable between people with more and less internet skills at wave 3, the former subsequently exhibited a slower increase, resulting in lower levels than the latter (figure 3B). Regarding rural–urban differences, internet skills had a significant effect on the depressive symptom trajectory among the urban group (β=−0.527, 95% CI −0.78 to –0.27), whereas no such effect was observed among the rural group (online supplemental table 2). In terms of regional distinctions, the effect of internet skills on the depressive symptom trajectory was evident among the eastern group (β=−0.599, 95% CI −0.93 to –0.27) and the central group (β=−0.511, 95% CI −0.88 to –0.14), but it did not exist among the western group (online supplemental table 2).

Figure 3
Request permissionrequest permission icon
Figure 3

The trajectories of depressive symptoms and internet skills in gender, rural–urban and regional subgroups over a 6-year follow-up. (A) The trajectories of depressive symptoms and internet skills among middle-aged and older men. (B) The trajectories of depressive symptoms and internet skills among middle-aged and older women. (C) The trajectories of depressive symptoms and internet skills among middle-aged and older adults in urban areas. (D) The trajectories of depressive symptoms and internet skills among middle-aged and older adults in the eastern region. (E) The trajectories of depressive symptoms and internet skills among middle-aged and older adults in the central region.

The trajectory of depressive symptoms by internet services in gender, rural–urban and regional subgroups

As shown in online supplemental table 2, regarding gender subgroups, significant effects of the interaction terms between internet services and each wave were observed for both male group (β=−0.216, 95% CI −0.35 to –0.08) and female group (β=−0.179, 95% CI −0.33 to –0.03). The depressive symptom trajectory of those who had acquired more internet services increased faster among females than males (figure 4A,B). In terms of the rural–urban subgroups, the interaction terms of internet services and each wave had a significant impact on the trajectory of depressive symptoms among the urban group (β=−0.267, 95% CI −0.40 to –0.13), while no such effect was evident among the rural group. Regarding regional subgroups, the interaction terms of internet services and each wave had a significant impact on the trajectory of depressive symptoms in the central region group (β=−0.314, 95% CI −0.48 to –0.15), with no remarkable effect observed among the eastern and western region groups. The trajectories of depressive symptoms as related to internet services were consistent over time for these four subgroups (male and female groups, urban group, central region group) (see figure 4A–D).

Figure 4
Request permissionrequest permission icon
Figure 4

The trajectories of depressive symptoms and utilisation of internet services in gender, rural–urban and regional subgroups over a 3-year follow-up. (A) The trajectories of depressive symptoms and utilisation of internet services among middle-aged and older men. (B) The trajectories of depressive symptoms and utilisation of internet services among middle-aged and older women. (C) The trajectories of depressive symptoms and utilisation of internet services among middle-aged and older adults in urban areas. (D) The trajectories of depressive symptoms and utilisation of internet services among middle-aged and older adults in central region.

Discussion

Main findings

In this longitudinal study, we found increasing trajectories of depressive symptoms concerning the digital divide among middle-aged and older adults in China, and there were gender, rural–urban and regional disparities over the recent decade. First, our findings unveiled notable gender disparities, with the beneficial effects of internet usage and improved internet skills on depressive symptoms being more evident among men compared with women over time. Among women, those using the internet exhibited higher levels of depressive symptoms than internet non-users at wave 2, a phenomenon absent among men. Second, in terms of rural–urban disparities, the depressive symptom trajectories of those who used the internet, acquired more internet skills and accessed more internet services changed over time among the urban group. Third, regarding regional disparities, a slower increase in depressive symptom trajectories linked to internet skills was evident only in the eastern and central region groups. Additionally, among the central region group, an increase in usage of internet services was associated with a slower rise in depressive symptoms.

Our analysis revealed that participants who used the internet, who acquired advanced internet skills and who accessed more internet services exhibited a slower increase in depressive symptoms over the follow-up period, highlighting the impact of the digital divide. This phenomenon can be attributed, in part, to the happiness incentive fostered by internet use.27 On one hand, the internet transcends social and spatial barriers, enabling individuals to maintain connections with family, friends and the wider world.9 By facilitating increased external interactions, the internet strengthens social support, alleviates social isolation and neglect, and boosts individuals’ self-esteem, sense of belonging and control. This, in turn, has a mitigating effect on depressive symptomatology.14 On the other hand, various internet services provide individuals with enhanced information and knowledge resources, enriching life experiences and reinforcing self-efficacy. This mechanism contributes to the alleviation and reduction of depressive symptoms and stress.28 For instance, entertainment services via the internet could stimulate the release of dopamine, thereby mitigating depression.7

In terms of gender disparity, the depressive symptoms among internet users were observed to exceed those of non-users at wave 2, with this pattern notably absent among men. This gender difference can be partly attributed to the existing digital divide between men and women. As the internet emerged, men tended to adopt technology earlier and display greater enthusiasm for novel technological advancements, such as the internet, while women may have been more likely to exhibit technophobia, delaying their integration and proficiency in navigating the internet.29 This gender-based digital divide appeared to have an alleviating effect on depressive symptoms among male participants, who experienced its benefits at an earlier stage compared with their female counterparts. Moreover, male participants benefited more from using the internet and possessing advanced internet skills compared with their female counterparts. Previous studies have reported gender differences in the relationship of the digital divide with depressive symptoms, with varying results.11 30 This study supports the idea of a larger discrepancy in depressive symptoms between internet users and non-users among middle-aged and older men compared with their female counterparts. Since middle-aged and older women are less educated and skillful than middle-aged and older men in new technologies, they tend to report more internet anxiety and less self-efficacy than men, leading to the lessened positive effect of internet use on depressive symptom trajectory among middle-aged and older women.11

Furthermore, the study uncovered a significant divergence in depressive symptom scores among older individuals residing in rural and urban settings, particularly regarding internet usage, advanced internet skills and access to internet services. In the rural–urban disparity, using the internet, enhanced internet skills and accessing more internet services appeared to slow the upward trend of depressive symptoms among the urban group. However, this mitigating effect was not observed among the rural group. While some prior studies have suggested that internet use may have a more pronounced impact on depression among older rural residents compared with their urban counterparts,31 our research underscores a heightened positive effect on depression to internet use and enhanced internet skills and services specifically among the urban group. This trend aligns with previous findings and can be explained by the rural–urban development divide.11 This phenomenon may be ascribed to the delayed development of the internet in rural areas and the additional challenges faced by middle-aged and older adults in the countryside due to lower education and skill levels, resulting in a non-significant effect of internet use on the trajectory of depressive symptoms among rural residents. Additionally, rural citizens often encounter greater difficulties in adapting to new technologies, and the availability and quality of internet services in rural regions lag significantly behind those in urban areas, further limiting the potential benefits of internet use in alleviating depression among middle-aged and older adults.11

With respect to regional disparity, the study found that a slower increase in depressive symptom trajectories with respect to more internet skills or internet services was observed over time solely among participants in the eastern and central regions. This observation highlights the historical and developmental context of China, particularly the impact of the Seventh Five-Year Plan (1986–1990), which initially classified the nation into three development tiers: the eastern, the central and the western regions. Subsequent sequential developmental trajectories, beginning with the eastern region followed by the central and western regions, have led to an imbalanced and centralised pattern of regional advancement.32 Significant disparities exist in the development of information technology and internet usage levels across these three regions in China, exhibiting a gradual increase in the capacity of internet usage from the western to the central and then the eastern regions. Compared with the western region, the eastern and central regions boast superior high-speed network infrastructure, internet penetration and broadband service quality. Furthermore, residents from the eastern and central regions typically possess higher educational levels, resulting in higher internet utilisation efficiency, better literacy and proficiency in internet usage, and enhanced transformation efficiency of internet usage into social participation which could significantly relieve depressed mood.7 Consequently, internet use in these more developed regions has a more pronounced impact on the trajectory of depressive symptoms. In contrast, less extensive internet usage, limited internet skills and fewer internet services within the western region exert a lessened effect of the internet on depressive symptom trajectories compared with the other regions.

Implication

The findings of this study carry significant implications for public health policy. First, it is crucial to establish and enhance high-speed network infrastructure, broadband service quality and a diverse range of internet services throughout the country. This will not only elevate the level of internet usage among middle-aged and older adults but also enhance the protective effects of the internet on mental health. Second, implementing strategies aimed at boosting internet knowledge and skills among this vulnerable demographic while simultaneously guiding them to use the internet in a responsible manner is paramount.8 Such approaches could mitigate the potentially harmful effects of internet use on mental health. Lastly, it is urgent that policy efforts prioritise providing greater access to the internet and its various services to middle-aged and older adults in rural areas and the western region. Furthermore, supporting women in enhancing their internet skills is vital, as this would address their disadvantages in using the internet.

Limitations

Several limitations characterise this study. A primary limitation arises from the relatively modest sample size of middle-aged and older rural adults who engaged with the internet. A closer analysis of the CHARLS dataset reveals that, out of a total cohort of 10 427 participants from rural areas, only 0.58% reported using the internet during the baseline survey. This figure gradually increased to 9.09% by wave 4 and 35.2% by wave 5. This under-representation of middle-aged and older adults in rural areas who use the internet necessitates a cautious interpretation of our findings and underscores the need for a more exhaustive investigation into the influence of internet use on the trajectory of depressive symptoms among this subgroup. Second, disentangling the causal relationship between the digital divide and depressive symptom trajectories presents a formidable challenge. In drawing definitive conclusions, it is crucial to carefully consider whether the observed associations are due to the direct effects of the digital divide or confounded by reverse causality. Despite this limitation, this study provides valuable evidence that using the internet and having enhanced internet skills and more access to internet services could potentially slow the increasing trajectories of depressive symptoms among middle-aged and older Chinese adults. This is because individuals with stronger internet skills and access to more services are able to effectively transform their social participation and more easily acquire useful information and other resources, which, in turn, can boost their self-efficacy and alleviate depressive symptoms. Third, there is a relatively high proportion of missing values in the data at wave 1 and wave 3. Moreover, the majority of the missing data from the five waves in this study were due to non-response in the depressive symptom questions, which was our main outcome of interest. Despite this limitation, the CHARLS data used in this study are large-scale, and our research incorporates data from the five waves. Given the large sample size, the missing values are insufficient to impact the research findings. Future studies could use higher-quality data to validate our results.

Conclusions

This study addresses a significant gap in research by exploring the relationship between the triple digital divide and the depressive symptom trajectories of middle-aged and older adults in China. The key findings suggest a retarding effect of using the internet on increasing trajectories of depressive symptoms over a 10-year period. The trajectory of depressive symptoms increased more slowly over time among individuals who engaged with the internet, possessed higher internet skills and accessed more internet services. Furthermore, the study highlights specific vulnerable groups, including middle-aged and older women, those residing in rural areas and individuals living in the western region. The implications of the study suggest that policymakers should prioritise the development of internet infrastructures vigorously in underserved areas and enhance access to a diverse range of internet services. Additionally, it is imperative to popularise internet education among middle-aged and older adults, enabling them to acquire more internet skills and harness the power of the internet as a tool for alleviating depressive symptoms. Meanwhile, it is also important to offer guidance to middle-aged and older adults on how to use the internet in a reasonable manner and prevent them from becoming addicted to the internet excessively. Given the identified vulnerabilities among specific groups, targeted strategies should be implemented to reduce regional disparities in information technology development and bridge the gender gap in internet usage. By addressing these issues, we can work towards a more inclusive and digitally literate society that supports the mental well-being of its older population.

Danxia Liu obtained an undergraduate degree from Anhui University School of Management, China in 2015. From 2015 to 2022, she attended Huazhong University of Science and Technology School of Sociology in China for her master's and doctoral degrees. Currently, she works at Tsinghua University School of Public Policy & Management as a postdoctoral fellow. During her master's and doctoral studies, she focused on social welfare for people with disabilities and older adults and the mental health of older adults. Now, her main research interests include geriatric health and the mental health of older adults. She is also interested in affective disorders and psychiatry.


Biography image

Overview

Article metrics
Altmetricopen-url
Dimensionsopen-url