Data from the 4 waves of the CHARLS survey, which started in 2011 (wave 1) and was repeated in 2013, 2015, and 2018 (waves 2, 3, and 4), were used in this study. CHARLS, a longitudinal survey of Chinese residents aged ≥45 years, used multistage sampling to sample community inhabitants’ health, social, and economic conditions. All data were collected by the CHARLS interviewers through face-to-face computer-assisted personal interviews (CAPIs). The interviewers were trained at Peking University by CHARLS staff members, and the interviews took place in respondents’ homes using CAPI technology [47]. After follow-up, 25,589 adults constituted the longitudinal cohort. For each step of this investigation, participants were interviewed individually. According to the World Health Organization, older adults are defined as those over 60 years old, while those aged 45-59 years are middle-aged adults [48]. By 2021, the population aged 45-59 years accounted for 24.5% of the total population in China, and the population over 65 years old accounted for 18.9% [4]. Among these groups, the prevalence of chronic diseases was 62.3% in people over 65 years old and was also high in people aged 45-54 years and 55-64 years, at 31.3% and 48.4%, respectively [29]. Therefore, considering that the middle-aged population is in a transitional stage of aging and also has a high incidence of chronic diseases [49], we also included people aged 45-59 years, in addition to older adults aged ≥60 years, for a more comprehensive data analysis.

To address this first knowledge gap, we included wave 4 data of the CHARLS survey, which ended in 2018, for a cross-sectional study. After eliminating missing data (n=2540, 12.8%) and participants who were 44 years or younger (n=172, 0.9%), 17,117 individuals were included to determine the sociodemographic characteristics of middle-aged and older internet users. Longitudinal data, including 4 waves, were used to address the second and third knowledge gaps. First, after removing the data that did not meet the inclusion requirements (Figure 1), there were 20,113 participants remaining in the survey whose usage behavior was used to determine the trends in internet use over the past decade. For the third study point, we performed separate risk assessments for each chronic disease to determine the association between daily internet use and chronic diseases. Consequently, on the basis of these 20,113 individuals, we analyzed the longitudinal data of each disease and excluded baseline disease carriers. For example, in the case of hypertension, baseline hypertension (n=5632, 28%) and missing data (n=26, 0.1%) were removed before analysis, leaving 14,455 people for the next analysis. Figure 1 depicts the research process and chronic diseases covered.

R version 4.1.2 (R Foundation for Statistical Computing) was used for all statistical calculations, and 2-tailed P<.05 was considered statistically significant. We substituted surrounding follow-up data for each participant in place of the missing data to ensure that the fraction of missing variables was less than 5% [63]. Case-weighted percentages were generated to describe sociodemographic factors and level distributions among participants.

In the cross-sectional study, categorical variables were presented as numbers and percentages and compared using chi-square tests and the Fisher exact test. For ordinal categorical variables, the Mann–Whitney U test and Kruskal–Wallis tests were used to compare sociodemographic characteristics between groups. Binary logistic regression and ordinal logistic regression were used for multivariate analysis of internet use and internet use frequency, respectively. In logistic regression analysis, first, the variables having statistical significance in the univariate analysis were included, then stepwise regression was used to find and eliminate multicollinearity, and finally the ideal model was produced. Ordinal logistic regression passed the test of parallel lines. Odds ratios (ORs) with 95% CIs were used to analyze the relationship between internet use and individual characteristics.

In the trend analysis using longitudinal data, chi-square tests and the Fisher exact test were used to compare the prevalence and frequency of internet use by gender.

In the longitudinal population, daily internet use and chronic disease incidence were examined using a Cox proportional hazards model, with HRs and 95% CIs. Three models were created based on the degree of adjustment: model 1 was a univariate model of internet use; in model 2, sociodemographic factors, such as gender, education, nationality, work status, recruitment age, resident area, marital status, and health insurance, were considered; and model 3 was then further modified to account for health behaviors, including smoking status, drinking status, restless sleep times, and ADLs. Follow-up time (in years) from baseline was used as the timeline, and the endpoint for each participant was either the date of the last follow-up or the date of disease manifestation, whichever came first.

CHARLS was a survey approved by the Ethical Review Committee of Peking University (approval number IRB00001052–11015), and the study data were anonymous. Each participant provided signed informed consent at the time of participation. There was no requirement for additional ethics approval for approved data users.

Among wave 4 participants in our study, 14,805/17,117 (86.5%) did not use the internet, while 2312/17,117 (13.5%) did. Details of the characteristics of the study population are presented in Table 1. The binary logistic regression model showed that women, those with higher education, those not working, the relatively young, urban residents, nonsmokers, alcohol drinkers, those rarely experiencing restless sleep, and those with unimpaired ADLs were more likely to use the internet in daily life (Table 2). On this basis, we further conducted univariate analysis and ordinal logistic regression analysis according to the frequency of internet use (Table S2 in Multimedia Appendix 1). Ordinal logistic regression analysis showed that middle-aged and older adults were more likely to use the internet more regularly if they were women, well-educated, relatively young, living in an urban area, not smoking, drinking, and rarely experiencing restless sleep (Table 2).

Overall, internet use has grown rapidly in recent years, from 2% in 2011 to 12.3% in 2018. In this study, the increase was larger for men (12.3%) than for women (8.4%). Across all age categories, there was significant growth in internet use. When compared to other age groups, individuals aged 45-54 years had the biggest rise in internet use, up by 21.4%, followed by those aged 55-64 years, while those aged over 75 years had the smallest growth (Figure 2). Figure 3 is a density plot showing the distribution of internet use in each age group for men and women over the past decade. For both men and women, the 45-54-year age group was the largest group of middle-aged and older adults using the internet, and with ascending age, the internet usage rate steadily fell. Figure 4 is a histogram showing the distribution of the internet use frequency in the past decade for each age group of men and women. As can be seen from the change in the frequency ratio in the past 10 years, the frequency of internet use among middle-aged and older adults steadily rose and was sustained at a high frequency. This rising tendency was more pronounced in men. Even though there were relatively fewer older adults using the internet—especially those aged over 75 years—this group had the largest proportion of internet users who used the internet almost daily, which means they had a higher frequency of internet use than the relatively younger middle-aged adults in recent years. Figure S1 in Multimedia Appendix 1 shows that over time, an increasing number of age groups exhibited gender differences in internet use, with men having slightly higher rates of internet use than women. In 2011, there was a gender difference only among participants aged 55-64 years, but by 2015 and 2018, there were gender differences in all 4 age groups. Similarly, the change in use frequency also shows that the gender difference in internet use has become increasingly obvious over time, with men using the internet slightly more frequently than women (Figure S2 in Multimedia Appendix 1). Especially in the age group of 45-54 years, there were gender differences in the 3 frequency groups in 2015 and 2018.

Interestingly, Cox proportional hazards modeling revealed that using the internet was associated with lower rates of most chronic diseases. Figure 5 shows that in the adjusted model 3, daily internet use was associated with a reduced risk of hypertension (HR 0.78, 95% CI 0.65-0.95), chronic lung disease (HR 0.74, 95% CI 0.57-0.97), stroke (HR 0.69, 95% CI 0.50-0.94), digestive disease (HR 0.73, 95% CI 0.58-0.91), memory-related disorders (HR 0.58, 95% CI 0.37-0.91), arthritis or rheumatism (HR 0.60, 95% CI 0.48-0.76), asthma (HR 0.52, 95% CI 0.33-0.84), depression (HR 0.80, 95% CI 0.71-0.89), and VI (HR 0.83, 95% CI 0.74-0.93) compared to internet nonuse. Model 2 showed that in addition to these chronic diseases, internet use also was associated with decreased incidences of heart disease as well as emotional and psychiatric disease (Table S3 in Multimedia Appendix 1).

The risk of chronic diseases, such as hypertension (HR 0.91, 95% CI 0.85-0.98), digestive disorders (HR 0.87, 95% CI 0.80-0.95), memory-related disorders (HR 0.83, 95% CI 0.71-0.99), arthritis or rheumatism (HR 0.82, 95% CI 0.75-0.89), asthma (HR 0.79, 95% CI 0.66-0.95), depression (HR 0.92, 95% CI 0.88-0.96), and VI (HR 0.94, 95% CI 0.89-0.98), also decreased as internet use frequency increased in the adjusted model 3 (Figure 6). Model 2 added stroke, chronic lung disease, and emotional and psychiatric diseases (Table S4 in Multimedia Appendix 1).

Our work focused on assessing whether there is a correlation between daily internet use and the incidence of chronic diseases, in contrast to past studies that have focused on eHealth and mHealth. This longitudinal study reveals a relationship between daily internet use and a gradual decline in the incidence of numerous chronic diseases among middle-aged and older Chinese people and shows that the risk for some chronic diseases decreases as internet use frequency increases.

Daily internet use is not as purposeful as eHealth and mHealth in terms of obtaining health knowledge and health services, but it is, nevertheless, how most older adults access the internet, using it a social and entertainment medium [38]. A few factors support our hypothesis that daily internet use may lower the chances of developing certain chronic diseases.

First, middle-aged and older adults’ physical and mental health may be directly affected by the internet because through its effect on the volume of the globus pallidus, internet use is associated with positive effects on cognitive function [64]. Previous studies have reported that digital gameplay for older adults may slow cognitive decline [65,66]. In addition, there is ample proof that regular internet use, such as social interaction and communication, may satisfy people’s needs for information and social interaction, significantly lessen loneliness for individuals, improve psychological well-being, and benefit the mental health of middle-aged and older adults [42,67,68]. Moreover, online entertainment, such as watching videos and playing mobile games using the internet, may bring users more happiness and further improve life satisfaction [69,70].

Second, middle-aged and older adults are more concerned with their health and may be more inclined to use the internet regularly to obtain the health information they need to positively modify their behavior [5,71]. Well-established evidence shows that an unhealthy lifestyle accounts for more than two-thirds of chronic diseases [72]; therefore, changing one’s unhealthy lifestyle using the internet may prevent these diseases. For example, in groups at higher risk for cardiovascular disease, internet use may reduce the risk by encouraging them to adopt a balanced diet and physical exercise [73]. New prospects for the dissemination of online health information are presented through short videos and papers on official accounts and search engines with low-threshold access and high interaction. For instance, short health videos distributed through social media may influence middle-aged and older adults’ everyday activities and social interactions, providing them with health and medical information more efficiently and easily [74]. Notably, to prevent the spread of false health information, excessive entertainment, and other issues, the popularization of health knowledge in this way is still in progress and must be improved from the perspectives of content, form, and oversight mechanisms [75].

Finally, sharing through forwarding or recounting is a hallmark of daily internet use at the interpersonal level [74]. Chatting over the internet may boost the interactions and communication between middle-aged and older adults, enhancing their interpersonal relationships, which may subsequently have a direct or indirect impact on their health-related behaviors and status [76]. By exposing middle-aged and older adults to environments with web-based health knowledge dissemination, sharing health information may expand the reach of influence and make them more likely to participate in these initiatives and subsequently receive preventive intervention, which is challenging to do in other ways.

Nevertheless, in addition to the aforementioned reasons, there are several factors that may influence the occurrence of chronic diseases. For example, the dopamine transporter (DAT1) gene has been implicated in addiction, including excessive internet use [77]. This may indicate that internet use may affect the physical condition of the elderly through dopamine levels. Internet use may also be a marker of higher income levels for individuals and groups, which is related to better health care access and healthy lifestyles [78]. Moreover, internet use also indicates a level of improvement of people’s living environments. This may affect their health through direct passive impact, such as effects of indoor environmental quality or by influencing behaviors that can affect health [79]. However, the specific pathways of these influencing factors remain to be further studied.

Previous research has revealed that frequency, rather than the time spent, may be a better measure of the motive for using the internet [80]. Our study showed that with the increasing frequency of internet use, the risk of some chronic diseases decreases. This may be because more frequent internet use can increase life satisfaction and enhance the quality of life [43,81]. Previous studies have also found that more frequent internet use is related to easier adoption of preventive behaviors [82]. In addition, studies have shown that more frequent internet use is positively related to the overall preferences for obtaining health information and decision-making autonomy associated with higher health literacy levels [21,83,84]. However, excessive internet use could also have a negative impact on health [85,86]. For instance, Liao et al [58] found that overuse of the internet may be associated with more severe depression. Therefore, although we believe that frequent internet use may further lower the risk for certain chronic diseases, it is also important to maintain good online behaviors, notably avoiding using the internet for extended periods of time.

Internet use among middle-aged and older adults is linked to socioeconomic and demographic characteristics. Our study supports previous studies showing that middle-aged and older adults who are women, young, educated, unemployed, and residing in urban areas are more likely to use the internet because they accept it more [5,45,87]. It should be noted that in this study, the results of univariate analyses of gender, work, and area were not consistent with the results of multivariate analysis, which may be due to the effects of other confounding factors. Physical issues that make learning digital skills difficult may also prevent older adults from using the internet [88]. Similarly, restless sleep times and impaired ADLs have been attributed to internet use. Consistent with Duplaga’s and Peng’s results [89,90], internet use is associated with more frequent drinking (implying more social opportunities) and the healthier habit of less frequent smoking. Studying the factors that influence middle-aged and older adults using the internet may help us better understand the challenges they face and improve the internet’s capacity to slow the progression of chronic diseases.

In this study, 12.3% middle-aged and older adults in China used the internet, comparable to that reported by Lyu and Sun [91] in China but lower than that reported in high-income countries [46], which may result from varied levels of economic progress. Middle-aged and older people have steadily increased and maintained a high level of internet use in recent years, and over half of the respondents use the internet every day [5]. It is interesting to note that older adults used the internet more regularly. This may due to the rapid popularization of the internet and the fact that retired people have more free time than middle-aged people. This pattern suggests that access to reliable and scientific health information over the internet may significantly improve the health of middle-aged and older adults. Given the differences in daily internet use between urban and rural areas, governments should strengthen the network infrastructure and increase the equality in internet access among older people. Studies have found that middle-aged and older adults living in rural areas have fewer medical and health resources than those living in metropolitan areas, possibly exacerbating the health gap acquired through the internet. Elderly individuals in rural areas may thus benefit more from increased internet aid [51].

If future research further validates the causal relationship between daily internet use and chronic diseases, alternative strategies may be considered to promote daily internet use among middle-aged and older adults, considering the distinctive qualities of each group. Family digital feedback may be implemented to encourage the younger members of the family to teach older people about digital technology and stimulate their interest in the internet. Considering older people have high dependence on and trust in traditional media, such as television and newspapers, these channels can also be used for publicity [92,93]. Additionally, internet service providers should make efforts to improve themselves to fulfill the requirements of middle-aged and older adults who use the internet on a daily basis. For instance, they should use charts rather than large fonts and include operation prompts, voice aid, and other features to make the presented material as simple and user friendly as possible to improve the use of internet devices among older people [93]. Meanwhile, to address the challenge of combating false information during daily internet use, a range of approaches may be used, including technical measures, such as artificial intelligence algorithms for authenticity verification, timely official government surveillance, and public education on how to critically evaluate the credibility of internet-based information [75,93-95]. However, privacy concerns must also be considered when implementing such measures [93].

The study offers several advantages. To thoroughly explore the association between daily internet use and the incidence of chronic diseases and provide empirical evidence in favor of preventing rather than treating chronic diseases, middle-aged and older adults without chronic diseases were our target group. Second, in this paper, we chose to focus on daily internet use among older populations, which has more practical implications for the prevention of chronic diseases. Four waves of CHARLS cohort data were included in this study, which represents the relationship between daily internet use and the risk for chronic diseases in middle-aged and older adults more accurately, thoroughly, and precisely over time.

There are several limitations to be aware of in this study. First, the sample was restricted to Chinese participants due to data availability, which may limit the generalizability of the findings to other countries or populations. Additionally, the use of self-reported data for quantification, such as chronic disease status, may be subject to reporting bias, which could affect the accuracy of the results. Furthermore, the study did not control for sociodemographic factors, such as income and living environment, or clinical metrics, such as genetics, which may be related to the occurrence of chronic diseases. Lastly, detailed information about the participants’ daily internet use, such as internet access and internet characteristics, was not collected. It is imperative to conduct future prospective research, such as randomized controlled trials and large cohort studies, to authenticate the findings. These studies should consider controlling for these additional factors when exploring the effects of different patterns of daily internet use on middle-aged and older adults.

USING the CHARLS survey, this study found that middle-aged and elderly Chinese people who use the internet daily have a reduced risk of developing chronic diseases and provided specific recommendations on how to improve daily internet use. With middle-aged and older adults using the internet more than ever before in China, it has gradually become a near necessity of their daily life. The findings illustrate that promoting safe and appropriate daily internet use may be a potential strategy for improving health outcomes for adults, but further robust and prospective studies are reqiured to validate the findings prior to implementing corresponding measures.