This study was based on a community-based cross-sectional survey conducted among older adults aged 60 years and older in the Lanzhou New Area of Gansu Province, China. Lanzhou New Area is a national-level administrative district that oversees 7 community health service centers and 80 family physician teams. All 7 centers participated in the survey. Each family physician team was required to complete at least 20 household-based interviews. Among more than 30,000 older adults registered in the catchment areas of these centers, approximately 10% were randomly selected, yielding 2486 recruited participants in March 2025.

A standardized data collection procedure was implemented to maintain consistency across centers. Family physicians received centralized training before fieldwork. Questionnaires completed in less than 600 seconds were excluded, removing 290 responses. Another 86 individuals submitted duplicate entries and were therefore excluded. The final analytic sample consisted of 2110 participants..

Demographic and socioeconomic variables controlled for in the regression models included sex (male or female), age in years (60‐69 years, 70‐79 years, or ≥80 years), marital status (single, married, or partnered), education (less than high school or high school or above), employment status (not working or working), self-reported health status (bad, general, or good), chronic diseases (no or yes), smoking (no or yes), and alcohol consumption (no or yes).

Descriptive analyses were conducted to characterize the study sample. Continuous variables were presented as medians together with IQRs. Categorical variables were summarized using counts (n) and proportions (%).

Multiple linear regression models were fitted to examine the association between loneliness and quality of life. To evaluate the moderating role of eHealth literacy, interaction terms between loneliness and eHealth literacy were introduced. The primary analyses were conducted using the total eHealth literacy score as a continuous variable. Analyses based on item groupings of eHealth literacy were exploratory in nature and are presented in Multimedia Appendix 1. Prior to constructing the interaction terms, continuous variables involved in the interaction analyses were mean-centered to improve interpretability and reduce potential multicollinearity. To facilitate interpretation of the moderation effect, predicted values of quality of life were estimated at +1 or −1 SD of eHealth literacy, and the results were visualized. In addition, simple slope analyses were conducted to estimate the association between loneliness and quality of life at different levels of the continuous moderator. Model linearity assumptions were assessed by visually inspecting residual plots of continuous variables against regression residuals. The plots did not indicate substantial deviations from linearity. All analyses were performed using Stata (version 17.0; StataCorp). A 2-sided P value of <.05 was regarded as statistically significant.

To assess the robustness of the primary interaction findings based on the total eHealth literacy score, 2 sets of sensitivity analyses were performed. First, the outcome variable was replaced with depressive symptoms. Interaction models were reestimated with depressive symptoms as the outcome to examine the consistency of the moderating relationship. Second, stratified analyses were carried out with sex as the grouping variable. Robustness was evaluated based on whether the direction or statistical significance of the associations remained consistent after conducting stratified analyses [27].

Formal ethics approval was granted by the institutional review board of Peking University Third Hospital (20250224-01025-0129). Electronic informed consent was obtained from all study participants and assisting family physicians before participation. All data were collected and stored in a deidentified and anonymized form to protect participants’ privacy and confidentiality. Access to the data was restricted to authorized members of the research team only. Participation in this study was voluntary, and participants were informed of their right to withdraw at any time without consequences. No financial compensation was provided to participants.

Figure 1 presents a flowchart of participant inclusion. Table 1 presents the baseline characteristics of the 2110 participants. Overall, loneliness, quality of life, eHealth literacy, and depressive symptoms were distributed at low to moderate levels. Of the 2110 participants, 1033 (48.9%) were female and 1077 (51.1%) were male. Most participants were aged 60 to 69 years (n=1132, 53.7%), followed by those aged 70 to 79 years (n=798, 37.8%). A large proportion were married or partnered (n=1778, 84.3%) and had education lower than high school (n=1802, 85.4%). The majority were not working (n=1919, 90.9%). Self-reported health was most commonly rated as general (n=1247, 59.1%). Chronic diseases were reported by 1232 (58.4%) participants. Most participants did not smoke (n=1536, 72.8%) and did not consume alcohol (n=1677, 79.5%).

Table 2 summarizes the results of the multiple linear regression models. In model 1, loneliness was significantly associated with lower quality of life (β=−0.66, 95% CI −1.03 to −0.29; P<.001). After adding eHealth literacy in model 2, this association remained significant and became slightly stronger (β=−0.83, 95% CI −1.18 to −0.49; P<.001). eHealth literacy was positively associated with quality of life (β=0.41, 95% CI 0.35-0.47; P<.001).

Exploratory analyses based on author-defined item groupings of eHealth literacy were conducted to examine their associations with quality of life. The results showed broadly consistent patterns with the primary analyses based on the total eHealth literacy score. Detailed results are presented in Table S5 in Multimedia Appendix 1.

Table 3 shows the results of the interaction models. In model 6, the interaction between loneliness and eHealth literacy was significant (β=−0.04, 95% CI −0.08 to −0.002; P=.04), indicating that higher eHealth literacy was associated with a steeper negative association between loneliness and quality of life. The inclusion of eHealth literacy and its interaction with loneliness increased the explained variance in quality of life from 6.58% (model 1, R²=0.0658) to 15.37% (model 6, R²=0.1537), indicating improved model explanatory power (ΔR²=0.0879). When comparing the model with main effects to the interaction model, the incremental contribution of the interaction term was 0.0016 (model 6, R²=0.1537; model 2, R²=0.1521; ΔR²=0.0016), and the interaction effect was statistically significant. Additional exploratory analyses examined interaction effects between loneliness and the author-defined item groupings of eHealth literacy. The patterns were consistent with the primary interaction findings based on the total score. Detailed results are presented in Table S6 in Multimedia Appendix 1.

To further interpret the interaction between loneliness and eHealth literacy, predicted values of quality of life were estimated at different levels of the continuous moderator. Figure 2 presents the interaction plot based on mean-centered eHealth literacy, with predicted values shown at low (mean−1 SD), middle, and high (mean+1 SD) levels of eHealth literacy. The results indicate that quality of life declined as loneliness increased across all levels of eHealth literacy. Simple slope analyses showed that the association between loneliness and quality of life was not statistically significant at low levels of eHealth literacy (simple slope=−0.46; P=.13), but became significantly stronger at the middle levels (simple slope=−0.83; P<.001) and high levels of eHealth literacy (simple slope=−1.21; P<.001). These findings suggest that the negative association between loneliness and quality of life becomes more pronounced as eHealth literacy increases.

Table S7 in Multimedia Appendix 1 shows the results of the interaction models when depressive symptoms were used as the outcome. Loneliness remained strongly associated with higher depressive symptoms across all models. The interaction between loneliness and eHealth literacy was significant and positive, suggesting that higher eHealth literacy was associated with a steeper positive association between loneliness and depressive symptoms (model 10). Figure S1 in Multimedia Appendix 1 illustrates the simple slopes of loneliness on depressive symptoms at different levels of the continuous eHealth literacy moderator. Depressive symptoms increased with higher levels of loneliness across all levels of eHealth literacy (low eHealth literacy, simple slope=1.62; P<.001 vs middle eHealth literacy, simple slope=1.84; P<.001 vs high eHealth literacy, simple slope=2.06; P<.001). The slope was steeper at higher levels of eHealth literacy, indicating that the positive association between loneliness and depressive symptoms became stronger as eHealth literacy increased. These results indicate that the moderating role of eHealth literacy remained stable when the outcome variable was replaced, supporting the robustness of the study findings.

Table S8 in Multimedia Appendix 1 presents the stratified results. In both male and female participants, eHealth literacy continued to show a positive association with quality of life. The interaction between loneliness and eHealth literacy was significant among female participants (β=−0.07, 95% CI −0.13 to −0.01; P=.02). Overall, the stratified models showed that the interaction results were consistent with the primary interaction findings based on the total eHealth literacy score, indicating the robustness of the findings.

This study examined loneliness, eHealth literacy, and quality of life among older adults in China and assessed the moderating role of eHealth literacy. Loneliness showed a clear negative association with quality of life. Higher eHealth literacy was positively associated with quality of life. Interaction analyses indicated that higher eHealth literacy was associated with a steeper negative association between loneliness and quality of life among older adults.

This study found a consistent negative association between loneliness and quality of life among older adults. This result aligns with evidence from cohort studies and systematic reviews showing that loneliness is linked to depression, anxiety, physical decline, cardiovascular events, and higher all-cause mortality, as well as sustained reductions in subjective well-being and health-related quality of life [10,15,17,28]. Persistent loneliness may impair psychological and social functioning through pathways involving functional deterioration [29], negative emotional accumulation, and reduced social participation, thereby undermining quality of life. These findings highlight the importance of prioritizing loneliness reduction in healthy aging policies and community interventions aimed at improving quality of life in later life.

We observed a robust positive association between eHealth literacy and quality of life in older adults. Existing evidence indicates that higher eHealth literacy enhances the ability of older adults to obtain and understand online health information [30], supports the use of digital health services, and promotes self-management [31], which can improve both psychological and physical health and ultimately enhance quality of life [32]. In China, where health care resources are unevenly distributed and accessing offline services can be burdensome, older adults with higher eHealth literacy are more likely to secure timely information and support through digital channels, translating digital access advantages into better quality of life [30,33].

Our study also showed that higher eHealth literacy was associated with a steeper negative association between loneliness and quality of life. This pattern differs from prior evidence, particularly from studies conducted during the COVID-19 pandemic among younger or student populations, where higher eHealth literacy was more commonly found to attenuate the negative associations between fear, loneliness, depressive symptoms, and health-related quality of life [34-37]. These findings suggest that eHealth literacy may have dual effects in digital health environments. Higher eHealth literacy may be accompanied by more intensive health information seeking and greater reliance on self-directed health management, which can increase cognitive load and decision burden [18,31]. Under conditions of loneliness and limited emotional support, greater engagement with online health information may translate into increased tendencies toward cyberchondria and psychological strain, thereby strengthening the association between loneliness and reduced quality of life [38,39]. In addition, in the absence of adequate family support or professional guidance, older adults with high eHealth literacy may also need to interpret complex health information on their own, which can intensify confusion, stress, and uncertainty and create a cumulative negative effect between loneliness and digital information exposure [40]. Furthermore, the cross-sectional design precludes conclusions regarding temporal ordering. It is possible that older adults with poorer well-being or heightened health concerns are more motivated to engage with online health information and to develop or report higher eHealth literacy, while loneliness simultaneously contributes to lower quality of life. This possibility highlights the need for longitudinal and experimental research to clarify directionality and underlying mechanisms. These findings highlight the need to further investigate the context-dependent functions of eHealth literacy and to examine potential mediating mechanisms, including information overload, health anxiety, cyberchondria, and decision burden, through longitudinal and experimental designs that can clarify temporal ordering and causal pathways. Additionally, strategies aimed at enhancing digital competence among older adults may require complementary forms of structured support, such as assistance in interpreting complex online health information and facilitated linkage to trusted community or primary care resources, particularly for individuals experiencing social vulnerability.

This study found that loneliness was consistently associated with lower quality of life among older adults and that higher eHealth literacy was associated with a steeper negative association between loneliness and quality of life. These results suggest that eHealth literacy functions as a dual-edged resource in later life. Although higher literacy is linked to better quality of life overall, it may intensify the challenges faced by lonely older adults when navigating complex or risk-oriented online health information. Given the cross-sectional and observational nature of the study, these interpretations should be considered exploratory. Future research is needed to examine the underlying psychological mechanisms, including potential roles of health anxiety, information overload, or emotional regulation processes, and to determine whether complementary support strategies could mitigate unintended negative consequences in digitally engaged but socially vulnerable populations.