Digital health literacy and self-efficacy among older adults: Mediating roles of social support and life satisfaction and the moderating role of health consciousness

DHL in later life

DHL extends beyond basic technical competence and encompasses functional, interactive, and critical skills for acquiring and applying health information from electronic sources.^5,6 This multidimensional construct is particularly salient for older adults who must manage multiple chronic conditions and complex treatment regimens. ¹⁹ Empirical studies document substantial variability in DHL within older populations, with differences in literacy associated with variation in health behaviors, patterns of healthcare use, and psychological outcomes. ²⁰

The developmental context of aging introduces specific considerations for DHL. Age-related changes in cognitive processing, accumulated technology experience, and evolving health information needs shape how older adults engage with digital health resources. ²¹ Some older adults face barriers related to limited prior exposure to digital technologies or age-related cognitive decline, ²² whereas others successfully use digital resources to support health management. ²³ This heterogeneity underscores the need to examine how DHL is linked to psychological resources in later life rather than treating older adults as a homogeneous group.

Recent theoretical developments in models of technology adoption among older adults offer useful frameworks for understanding DHL in later life. The Senior Technology Acceptance Model, for instance, identifies factors that influence technology use in older populations, including perceived usefulness, ease of use, and facilitating conditions, while incorporating age-specific factors. ²⁴ This model suggests that older adults’ adoption of digital health technologies is shaped not only by general perceptions of technology but also by gerontechnology self-efficacy, anxiety about technology use, and contextual supports. ²⁵ These insights highlight DHL as a situated capability that depends on both individual and environmental conditions.

Self-efficacy through social cognitive mechanisms

Self-efficacy is a core psychological resource across the life course. ²⁶ In later life, it plays a particularly important role in adaptation to age-related challenges, influencing health behaviors, emotional well-being, and functional independence. ²⁷ According to social cognitive theory, self-efficacy beliefs develop through four main sources: mastery experiences, vicarious experiences, social persuasion, and interpretations of physiological states. ²⁸

DHL may contribute to self-efficacy through each of these mechanisms. Successful navigation of digital health platforms provides mastery experiences that can strengthen efficacy beliefs. Exposure to health-related content featuring similar others managing chronic conditions offers vicarious experiences that inform perceived capability.^12,28 Digital communication with healthcare providers and peers supplies social persuasion that reinforces confidence in managing health. ²⁹ Finally, the use of health monitoring applications generates physiological feedback (e.g., blood pressure readings, step counts) that individuals interpret when forming judgments about their abilities.

These processes may operate in distinctive ways in later life. Older adults may benefit especially from mastery experiences that counter negative age stereotypes about technology use. ³⁰ Vicarious experiences may be particularly influential when they feature age-similar models successfully coping with health challenges. ³¹ In addition, social persuasion may carry greater weight because socioemotional selectivity processes increase the importance of close social relationships in later life. ³² Taken together, these considerations suggest that older adults with higher DHL are more likely to develop stronger self-efficacy in managing health and daily life. Therefore:

Hypothesis 1 (H1): DHL is positively associated with self-efficacy among older adults.

Social support as intermediary mechanism

Social support is a key resource for older adults as they navigate age-related challenges, providing informational, instrumental, and emotional assistance. ²⁹ DHL may be linked to social support through several mechanisms. Higher literacy can facilitate connection with geographically distant network members, enable participation in online health communities, and support maintenance of existing relationships through digital communication. ¹³

The association between social support and self-efficacy is well documented. Social networks provide informational support that enhances problem-solving, emotional support that buffers stress, and instrumental assistance that helps individuals achieve goals, all of which strengthen efficacy beliefs. ²⁹ Social support also fosters social integration and belonging, which contribute to positive self-perceptions and confidence in managing everyday challenges. ³³

For older adults, this relationship may be especially important in light of age-related changes in social networks and support needs. Although social networks tend to contract with age, older adults often maintain a smaller set of emotionally meaningful relationships that provide high-quality support. ³² Digital health engagement may complement this process by facilitating access to specialized support resources and by helping older adults maintain close relationships despite physical limitations or geographical distance. ¹³ These arguments suggest that social support mediates the relationship between older adults’ DHL and self-efficacy. Therefore:

Hypothesis 2 (H2): Social support mediates the relationship between DHL and self-efficacy.

Life satisfaction in sequential mediation

Life satisfaction—a cognitive evaluation of one's life as a whole—is a central indicator of subjective well-being and has particular relevance in later life. ³⁴ Social support contributes to life satisfaction by providing emotional comfort, practical assistance, and a sense of belonging.^29,33 DHL may indirectly influence life satisfaction by enabling older adults to seek support and maintain relationships through digital tools. ¹⁰ In turn, higher life satisfaction may be associated with greater willingness to engage with health-promoting behaviors and digital resources as individuals navigate age-related challenges.

Positioning life satisfaction as a mediator rather than an outcome warrants theoretical justification. Although reciprocal relationships between life satisfaction and self-efficacy have been documented, ³⁵ the present study conceptualizes life satisfaction as preceding self-efficacy for three reasons. First, the broaden-and-build theory proposes that positive evaluative states create psychological conditions that facilitate the development of personal resources. ¹⁷ Second, empirical evidence indicates that positive affect and favorable life evaluations enhance cognitive flexibility and problem-solving confidence. ³⁶ Third, socioemotional selectivity theory emphasizes the primacy of emotional well-being in later-life goal hierarchies, ³² suggesting that satisfaction experiences may shape subsequent efficacy beliefs. This perspective differs from models that treat efficacy as an antecedent of well-being and instead emphasizes how positive life evaluations create a context in which confidence can grow.

The relationship between life satisfaction and self-efficacy is likely dynamic. Higher life satisfaction may increase confidence in managing future challenges, while greater efficacy may further enhance life evaluations over time.^35,36 In the present study, life satisfaction is modeled as a mediator to capture how affective evaluations channel the influence of DHL and social support into efficacy beliefs. Therefore:

Hypothesis 3 (H3): Life satisfaction mediates the relationship between DHL and self-efficacy.

The sequential pathway from DHL through social support and life satisfaction to self-efficacy is consistent with broaden-and-build theory. ¹⁷ Positive emotional experiences are assumed to broaden cognitive and behavioral repertoires while building enduring personal resources. Enhanced social support derived from digital engagement may improve life evaluations, which in turn foster broader perspectives and confidence in managing future challenges. ³⁷ Considering these relational and affective processes jointly, DHL is expected to be linked to self-efficacy through a sequential pathway from social support to life satisfaction. Therefore:

Hypothesis 4 (H4): DHL is sequentially associated with social support and life satisfaction in relation to self-efficacy.

Health consciousness as moderator

Health consciousness is an important individual-difference factor that may shape how older adults engage with digital health resources. ³⁸ Individuals with higher health consciousness tend to show greater motivation to acquire health information, more systematic processing of health-related content, and stronger intentions to implement health-promoting behaviors.^18,39

These characteristics are likely to influence how DHL relates to psychological resources. Older adults with higher health consciousness may derive greater benefit from digital health engagement because they search for information more strategically, process content more deeply, and are more motivated to apply what they learn. ³⁹ They may also be more likely to recognize the value of digital health tools for managing health, thereby strengthening the association between DHL and self-efficacy. ¹⁸ Overall, these considerations suggest that health consciousness shapes the extent to which older adults translate DHL into a stronger sense of capability.

Taken together, these points lead to the following hypothesis:

Hypothesis 5 (H5): Health consciousness moderates the direct relationship between DHL and self-efficacy, such that this relationship is stronger when health consciousness is higher.

The moderating role of health consciousness may also extend to indirect pathways involving social support and life satisfaction. Older adults with higher health consciousness may participate more actively in health-related social exchanges facilitated by digital platforms, thereby amplifying the social support benefits of DHL. ³⁹ They may also derive greater satisfaction from successful health management supported by digital resources, potentially strengthening the pathway from DHL through life satisfaction to self-efficacy. ¹⁸ If health consciousness shapes whether and how older adults use DHL to build social ties and evaluate their lives positively, it should also influence the strength of the indirect pathway from DHL to self-efficacy. Therefore:

Hypothesis 6 (H6): Health consciousness moderates the indirect sequential pathway from DHL to self-efficacy through social support and life satisfaction.

Current study

Guided primarily by social cognitive theory, the present study uses broaden-and-build theory and socioemotional selectivity theory in a targeted way to clarify how DHL may shape self-efficacy through social support, life satisfaction, and health-related motivation in later life. It responds to gaps in prior work in which DHL has mainly been linked to health behaviors or quality of life, often without treating self-efficacy as the central psychological outcome for older adults or simultaneously modeling socio-relational, affective, and motivational processes. Within this focused framework (Figure 1), the study examined whether DHL is positively associated with self-efficacy among older adults (H1), whether perceived social support mediates this association (H2), whether life satisfaction also mediates the association (H3), and whether DHL is related to self-efficacy through a sequential pathway from social support to life satisfaction (H4). In addition, health consciousness was conceptualized as a key motivational boundary condition and was expected to strengthen the direct association between DHL and self-efficacy (H5) as well as the sequential indirect pathway from DHL to self-efficacy via social support and life satisfaction (H6).

OPEN IN VIEWER

Figure 1.

Research hypothesis model.

Research design and sampling procedure

This investigation used a sequential explanatory mixed-methods design conducted in community settings across six provinces in China between March 2024 and November 2024. The quantitative phase (March–August 2024) consisted of a cross-sectional survey of 1016 community-dwelling older adults, followed by a qualitative phase (September–November 2024) in which 30 participants completed in-depth interviews. All data were collected in community meeting rooms within residential neighborhoods to provide familiar and accessible environments for older adults. Study sites were located in eastern coastal provinces (Guangdong, Zhejiang), central provinces (Hubei, Anhui), and western provinces (Sichuan, Guangxi), which were purposively selected to capture variation in economic development, healthcare infrastructure, and digital technology penetration. This design allowed robust hypothesis testing while also generating deeper insight into the processes linking DHL with psychological outcomes in later life. ⁴⁰

The sampling procedure employed multi-stage probability techniques to enhance the representativeness of the target population. The first stage involved purposive selection of six provinces representing eastern (Guangdong, Zhejiang), central (Hubei, Anhui), and western (Sichuan, Guangxi) regions of China to capture geographic and socioeconomic diversity. Within each province, three communities were randomly selected from stratified lists organized by socioeconomic indicators (high, medium, low) based on official census data. From community registries, households with older adults were selected using systematic random sampling, and one eligible participant per household was randomly selected using the Kish grid method.

Inclusion criteria were (1) age ≥60 years; (2) no severe cognitive impairment (Mini-Mental State Examination score ≥24); (3) no terminal illness; and (4) voluntary informed consent. Exclusion criteria included (1) communication barriers or inability to complete questionnaires independently; and (2) psychiatric treatment within the past six months. Community committees assisted with participant recruitment, and data collection occurred in community meeting rooms.

The final quantitative sample (N = 1016, 92.36% response rate) demographically paralleled the older adult population in the selected regions, with no significant differences in age distribution (M = 67.11, SD = 5.92), gender ratio (50.39% male), or educational attainment when compared to census data (χ²(3) = 4.28, p = .23). Table 1 presents the geographic distribution of the final sample across provinces, demonstrating proportional representation with no significant regional differences in demographic characteristics (χ²(10) = 14.23, p = .16).

OPEN IN VIEWER

Table 1.

Geographic distribution of final quantitative sample after data screening.

Region	Province	Communities sampled	Initial contacted	Final sample	Response rate (%)
Eastern	Guangdong	3	192	178	92.7
	Zhejiang	3	188	171	91.0
Central	Hubei	3	184	169	91.8
	Anhui	3	181	167	92.3
Western	Sichuan	3	186	172	92.5
	Guangxi	3	169	159	94.1
Total		18	1100	1016	92.4

Note. Response rates calculated as (final sample/initial contacted) × 100. Non-response primarily due to incomplete questionnaires (n = 58) or withdrawal during data collection (n = 26). Regional distribution maintained proportional representation.

A priori power analysis using G*Power 3.1 indicated that a minimum sample of 395 participants would provide 95% power to detect small-to-medium effects (f² = 0.08) in the complex structural equation model with 22 observed variables at α = 0.05; thus, our sample substantially exceeded requirements, ensuring adequate statistical power for subgroup analyses.

For the qualitative phase, a theoretically driven purposive sampling strategy was used to select 30 interview participants stratified by health consciousness (15 high, 15 moderate, based on quantitative scores). ⁴¹ Sampling was further balanced by gender (male = 14, female = 16), age (60–70 years = 18, >70 years = 12), and educational background (lower = 17, higher = 13) to maximize variation in perspectives while maintaining the theoretical focus. The target sample size was informed by the concept of information power, ⁴² which proposes that smaller samples are adequate when study aims are specific, the theoretical framework is strong, interview dialogue is rich, and cross-case analysis is employed—all of which applied to the present study.

Data saturation was systematically monitored throughout the qualitative data collection process using established criteria. ⁴³ Saturation assessment involved (1) ongoing analysis of transcripts by two independent coders after every five interviews; (2) documentation of new themes emerging from each subsequent interview; (3) determination that theoretical saturation was achieved when three consecutive interviews (interviews 28–30) yielded no new substantive themes or conceptual insights beyond those identified in earlier interviews. The saturation determination was confirmed through team consensus meetings where all qualitative researchers reviewed coded transcripts and thematic maps. This systematic approach to saturation assessment enhanced confidence that the sample size was sufficient to capture the range of experiences and perspectives relevant to the research questions.

Ethical considerations and informed consent procedures

This study received ethical approval from the Institutional Review Board (Approval No. 20231126; date of approval: February 1, 2024). All procedures involving human participants were conducted in accordance with the Declaration of Helsinki and relevant institutional guidelines and regulations. Participation in both the survey and the qualitative interviews was entirely voluntary, and written informed consent was obtained from all participants before data collection.

The consent process involved (1) providing written information sheets describing study objectives, procedures, potential risks and benefits, data confidentiality, and withdrawal rights; (2) offering opportunities for questions and clarification; (3) obtaining signatures from participants or their legally authorized representatives; and (4) giving participants copies of the signed consent form for their records. For participants with visual impairments, consent forms were read aloud by trained research assistants, and verbal consent was audio-recorded and subsequently transcribed. All consent materials and procedures were reviewed and approved by the ethics committee prior to data collection.

Measures and cultural adaptation

All instruments underwent rigorous cross-cultural adaptation following Beaton's six-stage process ⁴⁴ : (1) forward translation by two independent bilingual experts; (2) synthesis of translations; (3) back-translation by two different translators; (4) expert committee review for semantic, idiomatic, experiential, and conceptual equivalence; (5) cognitive interviewing with 12 older adults; and (6) psychometric evaluation in a pilot sample (n = 50). This systematic process supported the cultural appropriateness and measurement equivalence of all instruments.

Permissions for questionnaires

All constructs were assessed with established questionnaires whose full item content and response formats have been reported in previous publications. According to information provided in the original articles and on the publishers’ or developers’ websites, these instruments may be used for non-commercial academic research without additional written permission. In the present study, all questionnaires were reproduced and administered solely for non-commercial academic purposes in line with these conditions.

Qualitative data collection procedures

Semi-structured interviews averaging 52 min (range = 41–68) were conducted by trained interviewers using a theoretically grounded interview guide with open-ended questions exploring each pathway in the hypothesized model, followed by systematic probes for clarification and elaboration. The interview protocol (available upon request) focused on three primary domains: (1) digital health information seeking and application strategies, (2) social interactions facilitated by digital health engagement, and (3) perceived impacts on life satisfaction and self-confidence.

Sample questions included: “How do you typically search for and evaluate health information using digital devices?”, “In what ways has your use of digital health resources affected your social connections?”, and “How has your ability to use digital health resources influenced your confidence in managing health challenges?” Interviews were conducted in private rooms at community centers, audio-recorded with permission, and transcribed verbatim. Transcripts were verified for accuracy by comparing audio recordings with transcribed text and reviewed by participants to ensure interpretive validity.

Analytical strategy

Our analytical strategy combined SEM for hypothesis testing with thematic analysis of interview data, using methodological triangulation to enhance the validity of findings. ⁴⁰ This approach enabled examination of the hypothesized relationships and, simultaneously, exploration of the mechanisms through which DHL relates to psychological outcomes in later life. Given the cross-sectional design, all reported associations are interpreted as non-causal. Although the theoretical model specifies directional relationships based on social cognitive theory, socioemotional selectivity theory, and broaden-and-build theory, the present data cannot establish temporal precedence or rule out alternative causal orderings.

Composite score construction and handling

All multi-item scales were modeled as latent variables in the SEM analyses rather than as summed or averaged composite scores. This approach offers several advantages: (1) it accounts for measurement error in observed indicators; (2) it permits explicit assessment of construct validity via confirmatory factor analysis; (3) it enables tests of measurement invariance across groups; and (4) it yields more accurate estimates of structural relationships. ⁴⁹ Each latent construct (DHL, self-efficacy, social support, life satisfaction, health consciousness) was specified with its respective items as indicators, with factor loadings estimated from the data. For descriptive statistics and correlations in Table 2, scale scores were computed as arithmetic means of item responses to aid interpretability. Missing item responses (<0.8% across all items) were handled using an expectation-maximization algorithm before averaging. All hypothesis tests, however, were conducted using latent variable models rather than composite scores, thereby maximizing psychometric rigor and statistical power.

OPEN IN VIEWER

Table 2.

Descriptive statistics and correlations among study variables.

Variable	M	SD	1	2	3	4	5
1. Digital health literacy	2.83	0.44	—
2. Social support	3.26	0.82	0.26**	—
3. Life satisfaction	2.56	0.44	0.43**	0.42**	—
4. Health consciousness	3.08	0.96	0.49**	0.51**	0.53**	—
5. Self-efficacy	3.35	0.67	0.19*	0.41**	0.38**	0.56**	—

Note: N = 1016.

*p < .05. **p < .01.

Measurement model evaluation

Confirmatory factor analysis supported a five-factor measurement model with excellent fit (χ²(220) = 497.06, p < .001, χ²/df = 2.26, CFI = 0.94, TLI = 0.93, RMSEA = 0.035, 90% CI [0.031, 0.039]). All standardized factor loadings were statistically significant (range = 0.62–0.89, p < .001), and inter-factor correlations were below 0.60, indicating adequate discriminant validity. Alternative models, including a four-factor model combining social support and life satisfaction (χ²(224) = 789.34, p < .001, χ²/df = 3.52, CFI = 0.83, TLI = 0.81, RMSEA = 0.063) and a single-factor model (χ²(230) = 1457.21, p < .001, χ²/df = 6.33, CFI = 0.57, TLI = 0.53, RMSEA = 0.112), showed substantially poorer fit, further supporting the discriminant validity of the constructs.

Measurement invariance tests across health-consciousness groups (high vs. moderate, based on a median split) supported configural invariance (χ²(440) = 753.89, p < .001, CFI = 0.93, RMSEA = 0.038), metric invariance (Δχ²(18) = 24.31, p = .15, ΔCFI = 0.004), and scalar invariance (Δχ²(18) = 29.27, p = .09, ΔCFI = 0.007, ΔRMSEA = 0.005). These findings indicate that the measurement structure was equivalent across groups, allowing meaningful comparison of structural parameters.

Descriptive statistics and correlations

Descriptive statistics indicated moderate levels of DHL (M = 2.83, SD = 0.44) and generally favorable psychological profiles (Table 2). DHL was positively correlated with social support (r = .26, p < .01), life satisfaction (r = .43, p < .01), health consciousness (r = .49, p < .01), and self-efficacy (r = .19, p < .05). Social support correlated positively with life satisfaction (r = .42, p < .01), health consciousness (r = .51, p < .01), and self-efficacy (r = .41, p < .01). Life satisfaction was positively associated with health consciousness (r = .53, p < .01) and self-efficacy (r = .38, p < .01). Health consciousness also correlated positively with self-efficacy (r = .56, p < .01).

To examine potential demographic correlates, we conducted additional analyses of differences in key variables by education level, gender, and urban versus rural residence. Education was significantly associated with DHL (F(3, 1012) = 18.42, p < .001), with higher education corresponding to higher DHL scores. Urban–rural differences were observed for DHL (t(1014) = 4.26, p < .001) and social support (t(1014) = 2.85, p < .01), with urban residents reporting higher levels of both. These patterns underscore the importance of considering demographic factors when interpreting the associations between DHL and psychological outcomes.

Structural model and mediation analysis

The hypothesized structural model demonstrated excellent fit (χ²/df = 2.27, CFI = 0.93, TLI = 0.92, RMSEA = 0.03). To evaluate explanatory power, we examined the squared multiple correlations (R²) for the endogenous constructs. The model accounted for 16% of the variance in social support (R² = .16), 31% in life satisfaction (R² = .31), and 48% in self-efficacy (R² = .48). These effect sizes indicate that the integrated digital health model had substantial predictive utility for psychological outcomes in older adults.

With respect to specific pathways, DHL was directly associated with self-efficacy (β = 0.21, p < .05), supporting H1. To evaluate alternative theoretical specifications, we estimated competing models, including a model with only direct effects (χ²/df = 3.45, CFI = 0.85, TLI = 0.83, RMSEA = 0.059) and a model with reversed directionality in which self-efficacy predicted DHL (χ²/df = 3.62, CFI = 0.84, TLI = 0.82, RMSEA = 0.061). Both alternatives showed substantially poorer fit, lending support to the hypothesized model.

Bootstrap analyses revealed significant indirect effects of DHL on self-efficacy through social support (β = 0.10, 95% CI [0.06, 0.14]) and life satisfaction (β = 0.17, 95% CI [0.12, 0.22]), supporting H2 and H3. Decomposition of the total effect (β = 0.53) indicated that the direct effect accounted for 39.62% of the total effect, while the indirect effect through social support accounted for 18.87% and the indirect effect through life satisfaction accounted for 32.08%.

Sequential mediation analysis

Bootstrap analysis further supported the sequential mediation pathway from DHL to self-efficacy through social support and life satisfaction (β = 0.05, 95% CI [0.01, 0.09]), which accounted for 9.43% of the total effect and supported H4. The full model, including all direct and indirect paths, showed a better fit than simpler mediation models, with specific path coefficients displayed in Figure 2.

OPEN IN VIEWER

Figure 2.

Structural equation model showing direct and indirect pathways from digital health literacy to self-efficacy. Note: Standardized coefficients are presented.

The sequential pathway operated as follows. DHL was positively associated with social support (β = 0.25, p < .05). Social support was positively associated with life satisfaction (β = 0.38, p < .01) while controlling for the direct association between DHL and life satisfaction (β = 0.32, p < .01). Life satisfaction was positively associated with self-efficacy (β = 0.55, p < .01) while controlling for the direct paths from DHL (β = 0.21, p < .05) and social support (β = 0.41, p < .01) to self-efficacy. Together, these results support the hypothesized sequential mediation process whereby DHL relates to social support, which in turn relates to life satisfaction and ultimately to self-efficacy.

Moderation analysis

Multi-group SEM, conducted after confirming scalar measurement invariance (ΔCFI = 0.007, ΔRMSEA = 0.005), showed that health consciousness moderated the direct association between DHL and self-efficacy (Δχ²(1) = 7.42, p < .01). The association was stronger in the high health consciousness group (β = 0.32, p < .01) than in the moderate group (β = 0.09, p < .05), supporting H5. Examination of specific paths in the moderation model (Table 3) also revealed significant group differences in the paths from DHL to self-efficacy, DHL to social support, social support to life satisfaction, life satisfaction to self-efficacy, and the sequential indirect pathway from DHL to self-efficacy through social support and life satisfaction.

OPEN IN VIEWER

Table 3.

Path coefficients by health consciousness level.

Path type	High health consciousness	Moderate health consciousness
DHL → self-efficacy	0.32**	0.09*
DHL → social support	0.26**	0.12*
Social support → life satisfaction	0.29**	0.05**
Life satisfaction → self-efficacy	0.30***	0.22**
DHL → social support → life satisfaction → self-efficacy	0.06**	0.01*

Note: DHL = digital health literacy.

*p < .05. **p < .01. ***p < .001.

Moderated mediation analysis

The sequential indirect effect of DHL on self-efficacy through social support and life satisfaction was significantly stronger at high levels of health consciousness (β = 0.06, 95% CI [0.02, 0.09]) than at moderate levels (β = 0.01, 95% CI [0.001, 0.03]). The index of moderated mediation was significant (Index = 0.05, 95% CI [0.01, 0.08]), supporting H6. These results suggest that health consciousness not only moderates the direct association between DHL and self-efficacy but also conditions the strength of the sequential mediation pathway through social support and life satisfaction.

Qualitative findings

Thematic analysis yielded three overarching themes that elaborated on the quantitative findings: (1) behavioral transformation pathways linking digital literacy to efficacy beliefs, (2) socio-emotional support mechanisms facilitating resource accumulation, and (3) cognitive-motivational processes underlying the moderating role of health consciousness.

Integrated mixed-methods findings

Joint display analysis revealed strong convergence between quantitative and qualitative findings, with participant narratives providing experiential context for statistical relationships (Table 4). The integrated analysis revealed that while quantitative findings established the presence and magnitude of relationships between DHL, mediating factors, and self-efficacy, qualitative findings illuminated the experiential mechanisms through which these relationships operate in daily life.

OPEN IN VIEWER

Table 4.

Integration of quantitative and qualitative findings.

Quantitative pathway	Statistical relationship	Qualitative theme	Illustrative quote
DHL → self-efficacy	Direct effect (β = 0.21, p < .05)	Behavioral transformation	“Each time I successfully find health information online, I feel more capable of managing my health problems.” (P11, Male, 67)
DHL → social support	Positive relationship (β = 0.25, p < .05)	Socio-emotional support	“The health knowledge I gain online gives me something valuable to share with friends, which strengthens our connections.” (P8, Female, 64)
Social support → life satisfaction	Positive relationship (β = 0.38, p < .01)	Socio-emotional support	“Having friends to discuss health information with makes me feel less alone in managing health issues, which makes life more satisfying.” (P14, Male, 72)
Life satisfaction → self-efficacy	Positive relationship (β = 0.55, p < .01)	Cognitive-motivational processes	“When I'm satisfied with life, I approach new health challenges with more confidence.” (P20, Female, 69)
Health consciousness moderation	Stronger effects with high health consciousness	Cognitive-motivational processes	“I actively compare different health recommendations before deciding what's most scientific and suitable for me.” (P12, Female, 68) vs. “There's too much information online—I don't know what to believe.” (P27, Male, 70)

Convergence and complementarity in mixed-methods integration

The systematic integration of quantitative and qualitative findings through connecting, building, and merging strategies yielded insights unattainable through either method alone ⁵⁴ :

Quantitative-Qualitative Convergence: Statistical associations between constructs (e.g., DHL-social support correlation r = .26, p < .01) found direct experiential confirmation in participant narratives describing how digital health knowledge facilitated valued social exchanges. This convergence across methods strengthens confidence in the validity of findings (methodological triangulation).

Qualitative Elaboration of Quantitative Patterns: While SEM analyses established the statistical significance of mediation pathways, qualitative data revealed the experiential mechanisms underlying these pathways. For instance, the sequential mediation through social support and life satisfaction (β = 0.05, 95% CI [0.01, 0.09]) might appear modest in magnitude, but qualitative insights revealed this reflects a meaningful multi-step process: digital literacy → social connection → reduced loneliness → positive life evaluation → enhanced confidence. The modest coefficient reflects the cumulative attenuation across multiple steps, each representing psychologically meaningful transitions described by participants.

Direct Pathway (DHL → Self-Efficacy, β = 0.21, p < .05): Qualitative findings illuminated that this direct association operates through mastery experiences. Participant 11 explicitly described the psychological process: “Each time I successfully find health information online, I feel more capable of managing my health problems"—illustrating how successful digital navigation provides performance accomplishments that directly strengthen efficacy beliefs, consistent with social cognitive theory's primary efficacy source. ²⁸

Social Support Mediation (Indirect effect β = 0.10, 95% CI [0.06, 0.14]): The quantitative mediation effect gains explanatory depth through qualitative insights revealing two specific mechanisms. First, Participant 8's narrative illustrated informational support: “The health knowledge I gain online gives me something valuable to share with friends, which strengthens our connections"—showing how Digital Health Literacy facilitates social resource accumulation. Second, Participant 9 described vicarious learning: “In the diabetes forum, seeing others successfully control their blood sugar makes me believe I can do it too"—demonstrating how digitally accessed social support provides observational experiences that enhance efficacy beliefs. This dual mechanism (resource accumulation + vicarious learning) explains the statistically significant mediation pathway.

Life Satisfaction Mediation (Indirect effect β = 0.17, 95% CI [0.12, 0.22]): The substantial mediation through life satisfaction—accounting for 32.08% of total effect—receives mechanistic clarification through qualitative data. Participant 14 described the affective pathway: “Having friends to discuss health information with makes me feel less alone in managing health issues, which makes life more satisfying.” Subsequently, Participant 20 articulated the satisfaction-to-efficacy link: “When I'm satisfied with life, I approach new health challenges with more confidence.” This sequential process aligns with the broaden-and-build theory, with qualitative evidence supporting the proposed mechanism whereby positive evaluative states expand cognitive resources underlying confidence.

Unexpected Qualitative Insights: The qualitative analysis revealed mechanisms not explicitly hypothesized in the quantitative model, including (1) goal-directed feedback loops whereby achievement notifications from health apps provided efficacy-enhancing reinforcement; (2) cognitive overload experiences among moderate health consciousness participants that limited DHL benefits; (3) value-congruence processes strengthening identity integration of digital health competencies. These emergent themes suggest refinements for future theoretical models and measurement approaches.

Explanation of Moderation Magnitude: The stronger path coefficients in the high health consciousness group (DHL → Self-Efficacy: β = 0.32 vs. β = 0.09) gain substantive interpretation through qualitative comparisons revealing fundamentally different information processing strategies, attribution patterns, and emotional engagement between groups—demonstrating that the statistical moderation reflects meaningful psychological differences rather than mere statistical artifact.

This integrated interpretation demonstrates how quantitative pathways identified through SEM correspond to lived experiences described by participants, with qualitative data revealing the cognitive, emotional, and social processes through which statistical relationships manifest in daily life. Particularly valuable was the qualitative insight into how information processing strategies, social validation experiences, and goal-directed feedback loops facilitate the translation of digital competencies into psychological resources in the developmental context of later life.

Direct association between DHL and self-efficacy

Our results confirm a significant direct association between DHL and self-efficacy, supporting the application of social cognitive theory in digital health contexts. This pattern aligns with Park et al., ¹⁴ who identified electronic health literacy as a robust predictor of self-efficacy among community-dwelling older adults. In contrast to general internet use, health-specific digital competencies—such as verifying medical information or navigating appointment systems—provide older adults with mastery experiences. These successful interactions constitute performance accomplishments, which Bandura identified as the most influential source of efficacy beliefs.

Qualitative findings further reinforced this interpretation through the theme of “behavioral transformation.” Participants reported that being able to monitor vital signs independently or cross-check health advice reduced their dependence on others. These experiences also changed how they viewed their ability to manage health challenges. Together, the quantitative and qualitative results suggest that the psychological impact of DHL arises from an enhanced sense of control over one's health management trajectory rather than from functional skills alone.

Mediating roles of social support and life satisfaction

Beyond the direct effect, the mediation analysis clarifies the socio-affective pathways that link DHL to self-efficacy. The finding that social support mediates this relationship is consistent with socioemotional selectivity theory, which emphasizes that older adults prioritize emotionally meaningful social interactions. ¹³ In the digital era, DHL becomes a key tool for sustaining and expanding these connections. This result is compatible with Song et al., ³ who showed that overcoming technological barriers is crucial for reducing social isolation.

Qualitative data echoed this mechanism in the theme of “socio-emotional support,” where participants described how digital platforms facilitated health information sharing within family and peer networks. Exchanging digital health information also worked as a form of social currency. It strengthened intergenerational ties and peer support and, in turn, increased psychological confidence. In this context, DHL does more than enable information acquisition. It also reshapes how older adults participate in relational networks that support health management and everyday coping.

Life satisfaction also emerged as a mediator in the relationship between DHL and self-efficacy. While previous research has documented bidirectional associations between life satisfaction and self-efficacy,^35,36 the present study demonstrates that life satisfaction can function as a pathway through which digital competencies and social resources contribute to enhanced efficacy beliefs. This finding is consistent with the broaden-and-build theory, which proposes that positive emotional states expand cognitive and behavioral repertoires while helping individuals build durable personal resources. ¹⁷ This process may be especially salient in later life, when positive emotional experiences increasingly guide goal setting and prioritization. ³² Conceptually, this perspective differs from models that position efficacy as an antecedent of well-being and instead emphasizes how positive life evaluations create a psychological context in which confidence can develop. However, because the data are cross-sectional, longitudinal studies are needed to test the temporal ordering of these relationships and to examine potential reciprocal influences between life satisfaction and self-efficacy over time.

Sequential mediation pathway

The sequential mediation pathway from DHL through social support and life satisfaction to self-efficacy represents an important theoretical contribution. It specifies a multi-step process through which technological competencies are converted into psychological resources in later life, extending previous research on digital engagement among older adults.^12,13 The mixed-methods analysis also showed how this statistical pathway appeared in everyday life. DHL made it easier for older adults to connect with others by sharing health information and joining online communities. These social resources reduced isolation and fostered a sense of belonging. Higher life satisfaction then supported more positive emotional states and broader thinking. Together, these changes strengthened confidence in managing future challenges. ⁸

The modest size of this indirect effect (9.43% of the total effect) reflects cumulative attenuation across multiple steps rather than limited theoretical importance. Each step marks a substantive transition from digital engagement to perceived social support, from social connection to positive life evaluation, and from life evaluation to self-efficacy. This pathway highlights that digital interventions may be most effective when they are designed not only to improve individual skills but also to shape the relational and emotional environments in which these skills are used.

Moderating role of health consciousness

The moderation analysis further refines this picture by identifying health consciousness as a critical boundary condition. Associations between DHL and psychological outcomes were significantly stronger among participants with high health consciousness. This pattern is consistent with the Integrative Model of eHealth Use, which highlights the role of motivational factors in determining how efficiently individuals convert resources into health outcomes. ¹⁶ Participants with high health consciousness engaged in what Liu et al. ¹⁶ describe as “health-promoting lifestyles,” actively using digital resources to pursue specific health goals.

The qualitative theme of “cognitive-motivational processes” helped explain these results. Participants with higher health consciousness described more systematic ways of processing information and reported deliberately integrating digital advice into their daily routines. In contrast, participants with lower health consciousness tended to consume information more passively. They were less likely to act on what they had learned, which limited the translation of digital skills into psychological efficacy. These findings suggest that DHL provides the basic capability for self-efficacy, whereas health consciousness determines how strongly this capability is used.

Cultural context and generalizability

The cultural specificity of the Chinese context introduces important boundary conditions for interpreting these findings. First, the prominence of the social support pathway may be especially salient in a collectivistic value system that emphasizes interdependence and social harmony. In contexts where interpersonal relationships are central to self-concept, digital health engagement may yield distinctive psychological benefits through social connection. Whether the socially mediated pathway identified here operates with the same strength in more individualistic cultures, where autonomy is emphasized, remains an open empirical question. In such settings, direct pathways from competence to efficacy might play a larger role.

Second, traditional Confucian values of filial piety are likely to shape technology-use patterns, as adult children often act as primary facilitators of their parents’ digital engagement, as illustrated by Participant 23's reliance on their daughter. ²⁴ This configuration may not generalize to contexts where older adults prioritize independence or rely more heavily on formal training and peer networks. Finally, China's digital ecosystem—dominated by integrated super-apps such as WeChat—and its rapidly digitizing healthcare system create a distinctive environment for digital health engagement. Future cross-cultural studies are needed to separate general psychological mechanisms (such as the sources of efficacy described in social cognitive theory) from culture-specific patterns.

Practical implications for digital health interventions

The findings of this study have practical implications for policymakers, healthcare practitioners, and technology developers seeking to narrow the digital divide in aging populations. The results suggest that fostering self-efficacy requires a shift from interventions that focus solely on functional skills to a broader, resource-oriented empowerment approach.

Given that social support significantly mediates the relationship between DHL and self-efficacy, interventions should move beyond solitary instruction. Community-based programs can establish “digital health circles”—peer-learning groups in which older adults learn collaboratively. ⁸ Such programs activate the social support pathway from the outset, transforming learning into a social bonding experience. In light of the importance of intergenerational relationships in the Chinese context, educational initiatives should also involve family members as “digital health partners,” equipping them with strategies to support older relatives while avoiding over-dependence. ¹³ Because this study relied on perceived DHL, future intervention trials should also include objective performance-based assessments. This will help check whether gains in self-reported confidence translate into actual competence in core digital health tasks.

The moderating role of health consciousness indicates that a “one-size-fits-all” approach is unlikely to be effective. Interventions should be tailored to different levels of health consciousness. For older adults with lower health consciousness, the primary barrier is often motivational. Programs for this group should first highlight the practical value of digital tools, such as easier appointment booking or medication refills, rather than focusing on abstract long-term health benefits. Simplified interfaces and immediate positive feedback can help reduce cognitive load and prevent disengagement. ¹⁶ For older adults with higher health consciousness, the focus should shift toward “critical digital literacy.” Because these individuals actively seek information, they may be particularly vulnerable to misinformation. Training should therefore stress source verification, critical appraisal of online content, and the ability to distinguish evidence-based advice from commercial advertising. ¹⁴

Because life satisfaction operates as a cognitive and affective expansion mechanism, digital health platforms should incorporate principles from positive psychology. User interfaces can also provide small emotional rewards, such as visualizing health trends, celebrating adherence milestones, or acknowledging incremental progress. These features may help start positive emotional cycles. ⁶ By recognizing users’ achievements and making progress visible, digital technologies can support positive self-evaluations and, over time, stronger self-efficacy.

Limitations and future research directions

Although this study provides robust support for the proposed integrated model, several limitations should be acknowledged. First, the reliance on self-reported measures of DHL, particularly the eHEALS, ⁵⁵ means that the study captures perceived confidence rather than objective performance. As recent critiques have noted, ⁵ self-reports may be affected by the Dunning–Kruger effect: some older adults may overestimate their competence, while others underestimate their skills because of low technology confidence. Our findings indicate that perceived capability is an important predictor of psychological self-efficacy, but future research should integrate performance-based assessments (e.g., simulated tasks such as booking appointments or locating reliable information) to examine how perceived and actual competencies align. Second, the cross-sectional design precludes causal inference. Longitudinal studies using cross-lagged panel models are needed to disentangle the reciprocal dynamics between DHL and self-efficacy over time.

Third, sampling constraints limit the generalizability of the findings. The quantitative sample was drawn primarily from community centers in established neighborhoods, which may underrepresent older adults who are homebound, lack internet access, or live in remote rural areas. These groups are likely to encounter different barriers from those faced by participants in the current study. Future research should use stratified sampling strategies to include these hard-to-reach populations so that digital health interventions can be designed to meet the needs of the most vulnerable segments of the aging population. ⁸