Factors influencing the online health knowledge communication gap among older adults: Evidence from TikTok on chronic diseases

Abstract

Objective

To investigate the multidimensional factors influencing the health knowledge communication gap among older adults in China, within the context of TikTok-based chronic disease content dissemination.

Methods

A sequential mixed-methods design was adopted. Qualitative interviews with 16 older adults were thematically analyzed to identify barriers to digital health communication and to inform questionnaire development. A cross-sectional survey of 407 older adults was then conducted. Psychometric evaluation included expert review, pilot testing, and exploratory factor analysis. K-means clustering identified heterogeneity in communication ability and guided confounder selection. Multiple linear regression assessed the independent contributions of cognitive and technical mastery, income level, attitude and needs, information characteristics, and social support, adjusting for key sociodemographic variables.

Results

Cluster analysis revealed low, moderate, and high-ability groups with systematic differences in age, education, and residence. Regression analysis confirmed that all five dimensions significantly predicted older adults’ health communication abilities. Information characteristics (β = 0.227) and cognitive level and technical mastery (β = 0.221) were the most influential factors, followed by personal attitude and needs (β = 0.211), income level (β = 0.164), and social support (β = 0.125). The integrated model explained 41.5% of the variance in communication ability, with acceptable multicollinearity levels. These patterns empirically support the complementary roles of structural, informational, and relational mechanisms.

Conclusion

The study highlights that bridging the digital health communication gap among older adults requires addressing both structural and cognitive-social factors. By integrating behavioral, ecological, and social dimensions, the findings reveal that digital literacy, content design, and social support are essential for closing digital health communication gaps in aging populations.

Keywords

Digital health communication knowledge gap theory social support theory older adults health knowledge sharing health knowledge communication gap

Introduction

Health knowledge communication is an interdisciplinary and dynamic field gaining significant attention among both health professionals and the general public.^1–3 For policymakers, effective health communication is essential in achieving positive public health policy outcomes and mitigating crises such as pandemics.^4,5 Existing literature suggests that effective health communication can help enhance positive outcomes of public policy.^4–6 For the public, accessible and understandable health knowledge communication directly influences critical decisions related to health management, lifestyle adjustments, and disease prevention, sometimes marking the difference between life and death.^1,7,8 Consequently, public interest in health information has markedly increased, especially in the context of recent global health crises such as the COVID-19 pandemic.

In recent years, the communication gap has led to tense relationships between patients and healthcare professionals, increased disputes and resource waste in medical domain.⁹ After the COVID-19 pandemic, the public is increasingly concerned about health topics. Therefore, the health knowledge communication is no longer limited between patients and healthcare professionals but is extended to the public. Open-ended expression on social media is vital for communication, whereas the effectiveness of communication is significantly diminished when messages are restricted. While it has huge potential for amplifying knowledge, it can equally spread misinformation. The health knowledge communication gap refers to the difference in opinions on health issues between healthcare professionals and the public. They may debate over health or be intent on simply expressing their own views blocking out other distractions.

Besides, the rapid expansion of digital media has reshaped health information dissemination channels, significantly elevating the role of social media platforms in health communication. Social media platforms, particularly short-video platforms like TikTok, have increasingly become popular for health knowledge dissemination. With over one billion monthly active users in China alone and more than 40 million health-related videos,¹⁰ these platforms are instrumental within national strategies such as “Healthy China 2030” and “Internet + Health,” aiming to provide tailored and accessible health services. Despite these advancements, the digital transformation has not been equitable, with older adults often excluded from its full benefits due to substantial barriers such as limited digital skills, sensory impairments, and heightened susceptibility to misinformation.^11–13

This demographic vulnerability becomes increasingly pressing against the backdrop of global demographic shifts. Worldwide, the older adult population is projected to surpass 1.4 billion by 2030, with China alone currently accounting for over 313 million older adults.¹⁰ The aging of the global population magnifies the challenge of bridging digital inequalities and ensuring equitable access to health knowledge, particularly as older populations face a rising burden of non-communicable diseases. Effective self-management of these chronic conditions critically depends on timely, accurate, and actionable health information, increasingly disseminated through digital platforms such as TikTok.

Nevertheless, while healthcare professionals actively employ short-video platforms to communicate health knowledge, the effectiveness of older adults’ engagement with this content remains inadequately understood. Current literature indicates that social media significantly improves health literacy, peer support, and disease self-management.^14,15 However, issues such as content quality variability, the prevalence of entertainment over scientific rigor, and algorithm-driven content personalization further complicate effective health communication on platforms like TikTok.^16–20

Moreover, older adults frequently encounter substantial obstacles to effective digital engagement, exacerbating disparities outlined by the Knowledge Gap Theory, which posits that socioeconomic advantages enable quicker and more beneficial use of new information technologies, thereby widening existing knowledge gaps.²¹ Simultaneously, research highlights how factors such as perceived usefulness, ease of use, and fears of online fraud significantly shape older adults’ attitudes toward digital health technologies. Older adults often encounter substantial barriers, including sensory impairments, limited e-health literacy, and concerns over online fraud. Digital-inclusion research identifies three cascading divides:²² access, skills and use, and outcomes. Chinese surveys find that older adults score lowest on information evaluation and content creation skills.²³ Perceived usefulness and ease of use predict seniors’ intention to adopt mobile health apps.^24,25 Income and rural residency remain robust predictors of reduced Internet use.²⁶ Social-support studies further show that children's “reverse mentoring” significantly boosts elders’ digital confidence and subjective wellbeing.²⁷ A substantial body of work confirms that socio-economic advantage accelerates the uptake and beneficial use of new information technologies, thereby widening knowledge gaps in the general population.^21,28

Existing literature highlights critical barriers in digital health communication among older adults, such as limited digital literacy, sensory impairments, algorithm-driven echo chambers, and insufficiently supportive environments. While several studies underscore the theoretical importance of these barriers, empirical investigations specifically focusing on short-video platforms remain sparse. Furthermore, research on the practical mechanisms facilitating or impeding older adults’ engagement in health knowledge dissemination through social media is notably limited. Besides, current studies predominantly focus on general digital literacy or isolated instances of misinformation without comprehensively addressing the interactive effects of individual, technological, and environmental factors.

Given the identified questions, the primary objective of this research is to systematically examine how individual characteristics, technological factors, and environmental contexts collectively influence older adults’ ability to engage effectively with health knowledge communication on TikTok. The research targets older adults aged 60 and above in China, aiming to provide empirical insights into the health knowledge communication gap within the context of chronic disease management on TikTok.

Literature review

The persistence of stratified access in online knowledge communication

Despite widespread optimism surrounding the democratizing potential of digital technologies, scholars have demonstrated that access to technology does not guarantee equal outcomes in digital participation. Instead, what emerges is a second-level divide: differences in skills, literacy, and usage purposes among various socio-demographic groups.²⁸ Rooted in Knowledge Gap Theory,²¹ this perspective posits that when new information enters a social system, individuals with higher socioeconomic status (SES) and educational attainment tend to absorb and use that information more quickly and effectively than those with fewer resources, leading to widening knowledge gaps over time.

Empirical studies from the Chinese context reveal enduring stratification in digital access and outcomes, particularly across urban-rural and income divides.^29,30 Intervention trials show that even modest smartphone-training workshops can produce double-digit gains in e-health literacy scores.³¹ Moreover, digital inequalities extend beyond access to encompass differences in digital literacy, self-efficacy, and purposeful engagement. At the cognitive level, differences in information processing and critical thinking deepen the online communication gap. Users with low information literacy are less capable of evaluating digital content credibility and relevance, making them more susceptible to echo chambers and misinformation.^20,22 Digital inequalities are also amplified by language, cultural capital, and self-efficacy. Research shows that individuals with limited digital capital are more likely to consume passive and entertainment-oriented content, rather than informative or civic content, thereby limiting their communicative empowerment.³²

In this study, we conceptualize cognitive and technical literacy as older adults’ capacity to understand health information and to use core functions of digital platforms, and income as a proxy for economic resources that shape access to devices, connectivity, and supportive learning opportunities. In the specific context of TikTok, older adults with higher cognitive and technical literacy are more likely to be able to search for relevant health videos, understand and evaluate what they watch, and transform this information into conversations with family members, peers, and health professionals. Likewise, higher and more stable income enables older adults to own and maintain smartphones, purchase data plans, and participate in social activities where digital skills can be acquired and practiced, thereby creating more opportunities for health knowledge exchange. Knowledge Gap Theory suggests that these advantages accumulate over time, such that older adults with more cognitive and economic resources are better positioned to benefit from the influx of health information on short-video platforms, whereas those with fewer resources risk being left further behind.

Given this theoretical grounding, we hypothesize that:

H1: Older adults with higher information literacy will demonstrate greater capacity to engage in health knowledge communication online.

H2: Higher income levels among older adults will be associated with reduced communication gaps in health knowledge access and sharing.

At the same time, Knowledge Gap Theory also implies that structural characteristics such as education and residence may jointly shape these advantages; we therefore treat education, residence, and related background factors as potential confounders and adjust for them in our empirical analyses.

Platform dynamics and the challenges of health information communication

The transition from general digital inequality to topic-specific communication gaps is particularly pronounced in the health domain. Online health communication involves high stakes: individuals must not only access and understand information but also evaluate its credibility and apply it to complex medical and personal decisions. A significant challenge is the variable quality of health content online, with platforms containing both professional medical advice and widespread misinformation. Studies have shown that even on accounts labeled as “medical,” content may prioritize engagement and virality over accuracy and safety.¹⁹ During the COVID-19 pandemic, health misinformation surged, often targeting or disproportionately affecting vulnerable populations.¹⁸

Health literacy, and specifically e-health literacy, is a central construct in understanding these disparities. Jiang and Estrela demonstrate that digital health literacy levels are strongly associated with individuals’ ability to seek, evaluate, and act on online health content.³³ Low e-health literacy is further linked to poorer chronic disease management, lower adherence to health advice, and increased mortality risks. In many cases, trust emerges as a crucial factor shaping health communication outcomes. Users frequently rely on heuristics such as “verified” accounts, popularity metrics, or visual cues to assess credibility rather than engaging in critical scrutiny.³⁴

From the perspective of Information Ecology Theory, these challenges can be understood as imbalances within an information ecosystem composed of information subjects (users), information objects (content), technological infrastructures (platform algorithms and affordances), and the surrounding socio-cultural context.³⁵ Building on this perspective, we distinguish between older adults’ internal dispositions toward digital health content and the external characteristics of the information environment. Internally, attitudes and needs encompass perceived usefulness, perceived ease of use, motivation to seek health knowledge, and trust in digital media. Prior work based on the Technology Acceptance Model shows that these attitudinal factors strongly predict whether older adults are willing to invest effort in learning new digital tools and integrating them into everyday health practices. Externally, information characteristics pertain to how health content is actually presented on platforms such as TikTok. Information Ecology Theory suggests that effective health communication is most likely to occur when users’ motivations and needs are aligned with an information environment that provides accessible, well-structured, and trustworthy content. In our study context, an older adult who strongly feels the need to understand their own chronic condition, believes that TikTok health videos are useful and manageable, and encounters content that is visually clear and delivered by trusted doctors or experts, is far more likely to watch attentively, remember key points, and talk about them with others. Conversely, weak motivation combined with cluttered, entertainment-driven or untrustworthy content will discourage active communication even if technical access is available.

Building upon these considerations, we hypothesize that:

H3: Older adults with stronger information needs, higher perceived usefulness, and greater trust in digital media will engage more effectively in health knowledge communication.

H4: Accessible, credible, and age-appropriate content formats will positively predict older adults’ health knowledge communication capacity.

Barriers to health knowledge communication among older adults

Older adults encounter a distinct set of challenges in digital health communication that go beyond technological unfamiliarity. Age-related cognitive decline, sensory limitations, and lower confidence in digital environments can collectively undermine their ability to appraise, integrate, and disseminate health knowledge.^36,37 Even when infrastructural access is ensured, structural disadvantages arising from low income, rural residence, and limited education may constrain both the perceived relevance of digital health content and the resources available to act on it.^26,30 Fear of online fraud and privacy breaches further reduces perceived usefulness and willingness to experiment with new health technologies.^38,39

At the same time, digital engagement is embedded within social relationships. Social Support Theory emphasizes that informational, instrumental, and emotional assistance from close ties can buffer technological anxiety and facilitate learning, whereas lack of support or stigmatizing attitudes can deter help-seeking.⁴⁰ Empirical studies in China have shown that mentoring by adult children and grandchildren, in which they teach older family members how to use smartphones, apps, and online services, can reduce technophobia, expand the range of online activities, and enhance subjective wellbeing.^27,41 Conversely, negative peer norms, such as shame associated with asking for help, may impede participation.^42,43 Research on information design further suggests that age-friendly features can significantly improve comprehension among low-literacy viewers,⁴⁴ yet such features are not completely incorporated into health videos on short-video platforms.¹⁶

Taken together, older adults’ health knowledge communication ability is shaped not only by individual skills and platform design, but also by the availability and quality of social support in both online and offline spheres. From a Social Support Theory perspective, informational support, instrumental support, and emotional support all help to reduce anxiety and build confidence in digital environments. In the specific scenario of TikTok, an older adult who can discuss videos with family members, send clips to friends, or receive feedback in online chat groups is more likely to develop a sense of efficacy and belonging in online health communities, which in turn can translate into more frequent and higher-quality conversations about health knowledge. By contrast, older adults who lack such support, or who fear being ridiculed for asking for help, may avoid experimenting with health content altogether, even if they own smartphones and have basic literacy.

Accordingly, we hypothesize that:

H5: Older adults who report stronger informational and emotional social support will exhibit higher levels of health knowledge communication engagement.

Synthesis and conceptual framework

The review above suggests that health knowledge communication gaps among older adults on TikTok arise from the interplay of structural inequalities, platform dynamics, and social–relational contexts. Knowledge Gap Theory highlights how socio-economic resources and educational capital shape older adults’ baseline capacity to access and process health information, suggesting the importance of cognitive level and technical literacy and income, as well as the need to adjust for confounding factors such as education and residence. Information Ecology Theory foregrounds the mutual shaping of users, technologies, and content, drawing attention to older adults’ motivations and needs, their trust in digital media, and the perceived characteristics of health information in a short-video environment. Social Support Theory situates digital engagement within networks of family and peers, emphasizing that interpersonal support and interaction climates can either facilitate or hinder older adults’ ability to experiment with, appropriate, and share health knowledge.

Building on these complementary perspectives, we develop a conceptual framework in which older adults’ health knowledge communication ability is jointly shaped by four interrelated dimensions of the information ecology, each of which is operationalized by one or more constructs in our empirical model. The information literacy dimension captures the human-capital mechanism emphasized in Knowledge Gap Theory: older adults with higher cognitive and technical literacy are better equipped to interpret and use health content on TikTok. The information person dimension reflects the resource and motivational position of the individual user. It includes economic resources, represented by income, which determine access to devices, data and learning opportunities, as well as personal attitudes and needs, which capture perceived usefulness, motivation and trust in digital media. The information quality dimension corresponds to the properties of the information objects circulating in the TikTok ecosystem; here focus on perceived clarity, accessibility and credibility of health content, consistent with Information Ecology Theory's emphasis on the “fit” between users’ capacities and the affordances of information. Finally, the information environment dimension represents the social interaction climate: the extent to which older adults are embedded in supportive online and offline networks that encourage exploration, discussion, and sharing of health knowledge, as articulated by Social Support Theory.

Knowledge Gap Theory primarily informs the information literacy and economic aspects of the information person dimension, explaining why older adults with more education, resources and skills are structurally advantaged in a rapidly digitalizing health information environment. Information Ecology Theory links the information person, information quality, and information environment dimensions by emphasizing that effective communication depends on the alignment between users’ needs and motivations, the design and credibility of content, and the broader communicative context in which platforms like TikTok operate. Social Support Theory, in turn, complements these perspectives by highlighting that even when structural resources and platform affordances are present, the absence of supportive social relationships may still prevent older adults from converting opportunities into actual health knowledge communication behavior.

By integrating all three perspectives, our framework treats the health knowledge communication gap among older adults as an emergent outcome of who the users are, what kind of content and affordances they encounter, and with whom they navigate these environments. On this theoretical basis, we specify five parallel hypotheses that correspond to the five focal constructs in our model. H1 and H2 derive primarily from Knowledge Gap Theory and posit that higher cognitive and technical literacy and higher income will be associated with greater health knowledge communication ability. H3 and H4 are grounded in Information Ecology Theory and propose that stronger information needs and more positive attitudes toward digital media, together with clearer and more trustworthy information characteristics, will facilitate health knowledge communication. H5 is informed by Social Support Theory and states that more supportive social interaction climates will be associated with greater engagement in health knowledge communication. In our empirical analyses, we test these hypotheses simultaneously in a multivariable framework, while controlling for key structural confounders (education, residence, age, and occupation) in order to estimate the independent contribution of each dimension. Figure 1 summarizes this conceptual model, showing how the information literacy, information person, information quality, and information environment dimensions jointly shape older adults’ health knowledge acquisition and sharing on TikTok and, ultimately, the health knowledge communication gap.

Figure 1.

Conceptual model which proposes that cognitive/technical literacy, income resources, personal attitudes and needs, social interaction climate, and knowledge affordances jointly shape older adults’ health-knowledge acquisition and sharing behavior.

Methods

Study design

This study employed a sequential mixed-methods approach, integrating qualitative and quantitative techniques to comprehensively explore the factors influencing the health knowledge communication gap among older adults using TikTok. Initially, qualitative data were gathered via semi-structured interviews to gain deep insights into older adults’ experiences, perceptions, and challenges related to online health communication. The qualitative findings informed the subsequent development of a structured questionnaire designed to quantitatively test the hypothesized relationships among individual, technological, and environmental factors.

This study was conducted in accordance with the Declaration of Helsinki and approved by the Medical Ethics Committee of Tongji Medical College Huazhong University of Science and Technology (Approval No. 2023S193). Informed consent was obtained from all participants, who were informed about this research and its objectives. All participants were told that the research data would be used only following the study objectives.

Participants and sampling

Qualitative phase

As Table 1 shows, 16 participants aged 60 or above were recruited through community contacts and referrals from local community workers. Inclusion criteria were: (1) aged ≥60 years; (2) residence in the local community for at least six months; (3) basic exposure to smartphones or short-video platforms, either through personal use or regular exposure in the household; and (4) sufficient cognitive and sensory capacity to participate in a 30–45 minutes interview. Older adults with severe cognitive impairment, serious psychiatric disorders, or severe hearing or visual problems that prevented meaningful communication were excluded.

Table 1.

Demographic characteristics of interviewees.

ID	Gender	Age (years)	Education	Occupation	Health status
A	M	60	High school	Worker	Just Good
B	F	63	Middle school	Worker	Fair
C	M	67	Middle school	Farmer	Poor
D	F	71	Illiterate	Farmer	Fair
E	M	75	Semi-literate	Worker	Fair
F	F	64	Middle school	Employee	Just Good
G	F	68	Primary school	Farmer	Fair
H	F	70	Middle school	Farmer	Good
I	M	62	High school	Self-employed	Good
J	M	65	Primary school	Farmer	Just Good
K	F	62	High school	Self-employed	Just Good
L	M	62	Middle school	Worker	Fair
M	F	69	High school	Employee	Poor
N	F	68	Middle school	Worker	Fair
O	M	73	Illiterate	Farmer	Poor
P	F	70	Primary school	Farmer	Good

Interviews were held in local community centers or homes, lasting 30–45 minutes. Open-ended questions were asked including (1) attitudes toward discussing health topics online, (2) experiences using short video apps for health information, (3) perceived challenges in understanding or sharing health knowledge, and (4) influences of family, neighbors, or media on their health information use. Interviews were audio-recorded and thematically analyzed. The full interview guide is provided in Supplemental material 1.

Data collection and analysis were conducted iteratively. After every few interviews, preliminary themes were reviewed and compared across cases. Data saturation was defined as the point at which no substantially new themes emerged in at least two consecutive interviews. In this study, saturation was reached by the 14th interview; two additional interviews were conducted to confirm that no new themes appeared, resulting in a final sample of 16 participants. This sample size is consistent with qualitative studies focusing on specific topics in aging and digital health, where 12–20 interviews are often sufficient to achieve thematic saturation.^45,46

Qualitative analysis of the interview transcripts revealed several recurrent themes. Many older adults exhibited technological resistance and preferred traditional healthcare channels. Some participants noted that seniors “do not want to learn unfamiliar channels” and still regard county hospitals and offline clinics as the norm, reflecting a conservative mindset and reluctance to adopt new digital media. Respondents also reported limited information channels: those with lower digital skills confined their online behavior to simple actions such as scrolling or sharing short videos and often encountered repetitive, fragmented content, describing an algorithm-driven information cocoons. Participants further described poor social feedback: during the COVID-19 pandemic, many seniors were exposed to rumors, and attempts to share health information with younger relatives were frequently ignored, discouraging further information exchange. Finally, interviewees were vulnerable to misinformation; without strong media literacy, many could not distinguish credible from false health advice, and even when they suspected “miracle cures,” they remained uncertain about their validity. In all, entrenched attitudes, narrow media use, social isolation, and low information literacy converge to widen the health knowledge communication gap among older adults.

Quantitative phase and sample size

A cross-sectional survey was conducted in multiple urban and rural communities across southeastern Shandong Province. Convenience sampling was adopted due to resource constraints, while ensuring coverage across different geographic and socioeconomic contexts. Older adults were eligible if they: (1) were aged 60 years or older; (2) had lived in the study communities for at least six months; and (3) had basic experience with smartphones or short-video platforms, either as users or through regular exposure in the household. Individuals with severe cognitive impairment, serious psychiatric disorders, or sensory limitations that prevented completion of the questionnaire were excluded. Trained researchers approached eligible older adults in community centers, village gatherings, and local events. For low-literacy respondents, questionnaires were administered face-to-face using standardized neutral probes to avoid interviewer bias.

The minimum sample size for the multiple linear regression analysis was estimated using Green’s rule of thumb for regression models,⁴⁷ and the equation was:

N = 50 + 8 m

Where m is the number of predictors. With five core predictors (X₁–X₅), a minimum of 90 participants was required. To allow for additional sociodemographic covariates and potential non-response, we aimed to recruit at least 200 participants. Ultimately, 407 valid questionnaires were obtained, exceeding the minimum requirement. A post-hoc power analysis indicated that, with N = 407, α = 0.05, and five core predictors, the study had approximately 0.95 power to detect small to medium-sized effects (Cohen's f² ≈ 0.05) in the overall regression model, suggesting adequate statistical power for the main analyses.

Measures and questionnaire design

Based on qualitative findings and theoretical frameworks including Knowledge Gap Theory, Information Ecology Theory, and Social Support Theory, a structured questionnaire was developed. The questionnaire was developed by adapting items from previously validated instruments, but the full instrument has not been previously validated as a whole. Therefore, expert review, pilot testing, and psychometric analyses were conducted to ensure appropriateness for the target population. The full questionnaire, including all items and response options, is provided in Supplemental material 2.

Item development and content validation

Initial item pools were generated based on qualitative themes and relevant theoretical propositions. To establish content validity, three experts in health communication and aging research independently evaluated each item for relevance, clarity, and cultural appropriateness. Items receiving inconsistent ratings were revised for wording or specificity. This step ensured that the final instrument adequately captured the conceptual domains implied by the three theoretical frameworks.

A pilot test involving 20 older adults from a non-study community was conducted to assess clarity, readability, and response burden. Participants reported no major comprehension difficulties, though minor wording modifications were made to improve accessibility for individuals with low literacy. Feedback from the pilot confirmed that the number of items and response format were suitable for the target population.

Measures

The questionnaire consisted of several major components. First, demographic information was collected, including gender, age, occupation, education, and income range. Second, health communication ability (Y) was measured using three items on a 5-point Likert scale that assessed older adults’ willingness and perceived ease of discussing and sharing health topics, with higher scores indicating stronger communication ability. In addition, five constructs aligned with the conceptual framework were assessed as independent variables (X₁–X₅). Cognitive level andtechnical mastery (X₁) was measured through three items evaluating respondents’ ability to distinguish authentic from false health information, comprehend health knowledge disseminated by doctors and experts, and use basic communicative functions on short-video platforms. Income level (X₂) was assessed through a single item capturing the respondent's monthly income range. Attitudes and needs (X₃) were measured through three items reflecting motivation, perceived usefulness, and personal or family health-related drivers of engagement with online health information.

Social support (X₄) was assessed through three items reflecting the perceived convenience and openness of the online communication environment, the degree to which online interactions felt natural or relaxed, and the quality of interpersonal relationships with family members, friends, and acquaintances who may facilitate or constrain health knowledge exchange. Information characteristics (X₅), conceptualized according to Information Ecology Theory, referred to respondents’ perceptions of the clarity, accessibility, and trustworthiness of online health content. This construct was operationalized through three items capturing the clarity and simplicity of different presentation formats (such as pictures, text, and videos), the extent to which respondents used multiple channels or approaches to acquire health knowledge, and the perceived credibility of health information delivered by doctors and experts on short-video platforms.

The final investigation tools are shown in Table 2. All Likert items were scored from 1 (completely disagree) to 5 (completely agree). Higher values indicate greater literacy, more positive attitudes, stronger social support, or higher perceived information quality. We hypothesized that higher health communication ability, reflecting a reduced knowledge gap. Content validity was assessed by a panel of three experts in health communication and aging research, who evaluated each item for relevance, clarity, and cultural appropriateness. Minor wording revisions were made based on their feedback. The draft questionnaire was then pilot-tested with 20 older adults from a community not included in the main survey to assess clarity, comprehensibility, and response time; participants’ feedback led to further minor adjustments to item wording and layout.

Table 2.

Scale construction and theoretical mapping.

Dimension	Sub-dimension	Question item	Source
Health communication ability		1. I am willing to discuss health topics with others and acquire health knowledge	⁴⁸
		2. I am willing to take the initiative to promote and popularize health knowledge among people around me
		3. When communicating health knowledge with others, I can usually communicate smoothly and talk freely
Income level	Monthly income	4. Your monthly income is (sources include but are not limited to children's support, agricultural income, and government subsidies)	⁴⁹
Information literacy	Cognitive level and technical mastery	5. In daily life, I can distinguish the authenticity of health knowledge and judge its practicality	^50–53
		6. I can understand the relevant health and medical knowledge disseminated by doctors and experts in daily life
		7. On short-video platforms, I can use information technology to send text, pictures and videos, and I am familiar with the relevant functions
Information Person	Usage attitude and personal needs	8. I often follow or search for health knowledge on Tiktok to understand my own condition	^54,55 ⁵⁶
		9. I spend a lot of time acquiring health knowledge
		10. One of the major reasons why I pay attention to online health knowledge exchanges is the health issues of my friends and family around me
Information environment	Social support	11. The simple and convenient online environment has created a good channel for me to communicate health knowledge with others	^57,58
		12. Online conversations give me a natural and relaxed atmosphere
		13. I get along very well with my friends around me and the people I know well
Information quality	Information characteristics	14. The more straightforward and clear the form of presentation of knowledge (pictures, text, videos) is, the easier it is for me to understand	⁵⁹ ⁶⁰
		15. I will use multiple approaches to acquire and understand health knowledge
		16. Compared with offline learning, I prefer to trust the health knowledge imparted by doctors and experts on short-video platforms

Data analysis methods

The qualitative analysis followed an inductive, thematic approach aimed at identifying recurrent patterns in older adults’ experiences with online health communication. Rather than applying line-by-line coding or formal grounded-theory procedures, the research team conducted repeated close readings of the interview transcripts to summarize key ideas and compare emerging themes across participants, which is common in exploratory mixed-methods studies where the goal is to inform instrument development rather than to generate a formal theory. Two members of the research team independently reviewed all transcripts, highlighted meaningful segments, and documented preliminary thematic categories. These initial codes were discussed in group meetings and refined into a set of higher-order themes that captured the most salient challenges and perceptions expressed by the participants. Data collection and analysis proceeded iteratively. After approximately 12 interviews, no substantively new themes emerged; between the 12th and 14th interviews, content was largely repetitive. Two additional interviews were conducted to confirm this pattern, and again no new themes were identified, indicating thematic saturation in line with recommendations for studies of relatively homogeneous older populations.

Quantitative data were analyzed using SPSS 27. Reliability of the scales was assessed using Cronbach's alpha coefficients. Internal consistency for the multi-item constructs ranged from 0.726 to 0.885. Construct validity was examined using exploratory factor analysis (EFA) with principal component extraction and varimax rotation. The Kaiser–Meyer–Olkin (KMO) measure of sampling adequacy was 0.806, and Bartlett's test of sphericity was significant (p < 0.001), indicating that the data were suitable for factor analysis. The factor solution was broadly consistent with the hypothesized structure, supporting the construct validity of the measurement instrument.

Descriptive statistics were used to summarize sociodemographic characteristics and scale scores. To identify subgroups of older adults with distinct levels and profiles of health communication ability, we first standardized the composite scores for the key constructs and then performed k-means clustering analysis. Candidate solutions with k = 2 to k = 5 clusters were compared using the elbow method, average silhouette coefficients, and the substantive interpretability of cluster profiles.

Beyond subgroup description, the clustering procedure also served to inform the selection of potential confounders for the regression models. Sociodemographic variables that exhibited systematic and statistically significant differences across clusters, together with factors identified in prior literature as structural determinants of digital engagement among older adults, were treated as covariates.

Group differences in key variables across the three clusters were examined using Pearson's chi-square tests for categorical variables and Kruskal–Wallis tests for non-normally distributed continuous variables. Finally, multiple linear regression analysis was conducted to estimate the independent effects of cognitive level and technical mastery, income level, attitudes and needs, social support, and Information characteristics (X₁–X₅) on health communication ability (Y), while adjusting for the identified sociodemographic covariates. Model diagnostics included inspection of residual plots and assessment of multicollinearity using variance inflation factors (VIF), all of which were below 2, indicating no problematic collinearity among predictors.

Results

Questionnaire reliability and validity testing

Internal consistency was high for every multi-item scale, as Table 3 shows. Cronbach's alpha coefficients were 0.818 for the three-item communication ability scale, 0.830 for the cognitive level and technical mastery scale (X₁), 0.791 for attitude and needs (X₃), 0.840 for social support (X₄), and 0.726 for information characteristics (X₅). All alphas exceeded the conventional 0.70 threshold, indicating good reliability of the measures. Removing any single item did not substantially increase α, implying all items were important for the scale. Corrected item-total correlations for all items exceeded 0.40.

Table 3.

Reliability statistics (Cronbach's α) for scales.

Scale dimension	Number of items	Cronbach's α
Health communication ability	3	0.818
Cognitive level and technical mastery	3	0.830
Attitude and needs	3	0.791
Social support	3	0.840
Infomation characteristics	3	0.726

As Table 4 shows, EFA was performed on the independent variable items to evaluate construct validity. The KMO measure was 0.806 (above 0.6) and Bartlett's test of sphericity was significant (p < 0.001), indicating the data were suitable for factor extraction. Unrotated eigenvalues indicated a five-factor solution consistent with the theorized structure, explaining 76.35% of total variance. All communalities exceeded 0.40, and rotated factor loadings showed clear grouping consistent with the hypothesized constructs.

Table 4.

Result of validity test.

	Factor loading
Scale item	Factor 1	Factor 2	Factor 3	Factor 4	Factor 5	Communality
Item 4	0.080	0.025	0.015	−0.010	0.983	0.973
Item 5	0.843	0.036	0.096	0.020	0.081	0.728
Item 6	0.855	0.114	0.141	0.122	−0.021	0.779
Item 7	0.803	0.072	0.158	0.144	0.061	0.699
Item 8	0.112	0.069	0.854	0.073	0.027	0.752
Item 9	0.115	0.027	0.742	0.425	0.060	0.749
Item 10	0.241	0.173	0.787	0.079	−0.052	0.717
Item 11	0.041	0.870	0.100	0.047	0.153	0.794
Item 12	0.117	0.846	0.125	0.039	−0.062	0.750
Item 13	0.056	0.846	0.027	0.210	−0.049	0.766
Item 14	0.616	0.063	0.155	0.512	−0.039	0.672
Item 15	0.126	0.023	0.359	0.810	0.072	0.807
Item 16	0.205	0.396	0.044	0.728	−0.097	0.740
Eigen Value(Unrotated)	4.447	2.000	1.513	1.067	0.899	-
% of Variance(Unrotated)	34.206%	15.387%	11.636%	8.207%	6.917%	-
Cumulative % of Variance(Unrotated)	34.206%	49.593%	61.229%	69.436%	76.353%	-
Eigen value(Rotated)	2.632	2.404	2.135	1.725	1.030	-
% of Variance(Rotated)	20.248%	18.495%	16.421%	13.269%	7.920%	-
Cumulative % of Variance(Rotated)	20.248%	38.743%	55.164%	68.433%	76.353%	-
KMO	0.806	-
Bartlett's Test of Sphericity	2154.670	-
df	78	-
P value	<0.001	-

Note: Boldface indicates factor loadings > 0.40.

We further conducted PCA on the antecedent items to reduce dimensionality and confirm factor structure. As Table 5 shows, the analysis indicated that three principal components should be retained. These three components explained 41.93%, 20.04%, and 16.73% of variance, respectively. Weighting these by their eigenvalues yielded contributions of 53.28%, 25.46%, and 21.25%, with a cumulative explained variance of 78.70%.

Table 5.

Total variance explained.

	Eigen			PCA
Factor	Eigen value	Variance (%)	Cumulative of variance (%)	Eigen value	Variance (%)	Cumulative of variance (%)
1	2.097	41.934	41.934	2.097	41.934	41.934
2	1.002	20.038	61.972	1.002	20.038	61.972
3	0.836	16.726	78.698	0.836	16.726	78.698
4	0.640	12.804	91.502	-	-	-
5	0.425	8.498	100.000	-	-	-

As Table 6 shows, all items loaded strongly (>0.4) on one of these components, indicating clear factor grouping, affirming that the survey items cluster as expected and that the PCA model is robust.

Table 6.

Factor loading (rotated).

Name	Factor loading (Rotated)
Name	Factor 1	Factor 2	Factor 3	Communality
x₁	0.730	0.106	−0.336	0.657
x₂	0.142	0.979	0.095	0.988
x₃	0.737	−0.074	−0.189	0.584
x₄	0.545	−0.101	0.819	0.978
x₅	0.839	−0.128	−0.089	0.728

Note: Boldface indicates factor loadings > 0.40.

Cluster of health communication ability

We applied k-means clustering to the scores on the three principal components derived from the health communication ability items (Questions 1–3). The three-cluster solution was selected based on the elbow criterion, average silhouette widths, and the interpretability of the resulting profiles. The algorithm converged after four iterations, and cluster sizes were reasonably balanced. As Table 7 shows, the cluster centers indicate distinct profiles on the principal component dimensions.

Table 7.

Result of cluster.

Item	Initial cluster center			Final cluster center
	cluster_1	cluster_2	cluster_3	cluster_1	cluster_2	cluster_3
Y1	−0.24938	−1.92700	1.42825	0.00481	−1.07444	0.85259
Y2	−1.08071	−0.17567	1.63441	−0.22367	−0.94718	0.94824
Y3	0.66824	−1.99818	1.55705	0.04204	−1.10209	0.84252

As Table 8 shows, the clusters included 132 (32.43%), 122 (29.98%), and 153 (37.59%) participants, respectively. The clusters were well-balanced in size, with none dominating, suggesting meaningful segmentation.

Table 8.

Descriptive statistics of clustering categories.

		Cluster		Error
Category	N(%)	Mean square	df	Mean square	df	F	p-value
Zscore: y1	132(32.43)	126.030	2	0.381	404	330.753	<0.001
Zscore: y2	122(29.98)	126.813	2	0.377	404	336.230	<0.001
Zscore: y3	153(37.59)	128.510	2	0.369	404	348.490	<0.001

The three clusters derived from the k-means analysis reflect substantively distinct behavioral profiles in older adults’ engagement with health knowledge communication. Cluster 2, representing the low-ability group, is characterized by limited digital skills, weak information needs, low perceived information quality, and minimal social support. Members of this group exhibit limited confidence in using TikTok for health-related purposes and demonstrate difficulty evaluating content credibility that align closely with Knowledge Gap Theory, which posits that individuals with fewer cognitive and socioeconomic resources are less able to benefit from new information environments. These findings also mirror qualitative accounts describing confusion, avoidance, and mistrust toward digital health content.

Cluster 1, the moderate-ability group, demonstrates intermediate scores across all constructs, suggesting partial engagement with digital health information. This profile reflects transitional users who possess sufficient digital literacy to navigate TikTok health videos but lack either strong motivation or consistent social reinforcement to sustain active communication. Their mixed patterns of engagement indicate that relatively modest improvements in content design, motivation, or social support may facilitate upward mobility in communication capacity.

Cluster 3, the high-ability group, exhibits the strongest performance across all theoretically grounded dimensions, including digital literacy, perceived usefulness, information needs, and social support. Members of this cluster actively seek out, comprehend, and disseminate health information on TikTok. This profile is consistent with Information Ecology Theory, which emphasizes the alignment between individual capacities, content affordances, and supportive communication environments. Qualitative evidence further supports this interpretation: participants resembling this cluster reported confidence in navigating TikTok, proactively discussing videos with others, and effectively integrating online content into their daily health practices.

Taken together, these cluster profiles, when viewed alongside the qualitative findings, illustrating that older adults experience digital health communication environments in systematically differentiated ways shaped by structural resources, cognitive capacities, motivational orientations, and social interaction contexts.

Descriptive characteristics of respondents

As Table 9 shows, the 407 respondents were 57.2% male. The largest age cohort was 64–67 years old, while only a small fraction were over 70. Educationally, the sample was skewed low: over 75% of respondents had education at or below middle school, reflecting the rural context. By occupation, more than half were engaged in primary (farming) or secondary (factory/construction) industries, and a significant minority were self-employed, consistent with the region's employment structure. Residence type was moderately balanced (45.7%). Most constructs had mean scores around 3.2–3.3 on the 5-point scale, indicating moderate levels on average. Variability (SD) was relatively low except for income (X₂) and social support (X₄), whereas attitude and needs (X₃) and information characteristics (X₅) were less dispersed (SD < 1.00). The distributions of most variables were approximately symmetric (skewness near 0) with kurtosis values close to 0, suggesting no extreme departures from normality in summary statistics.

Table 9.

Demographic characteristics of survey sample (N = 407).

Variable	N(%)/Mean ± SD
Gender (Male)	232(57.2)
Age
60–63 years	95(23.3)
64–67 years	154(37.8)
68–71 years	85(20.9)
72–75 years	50(12.3)
75 years older	23(5.7)
Education
Illiterate	145(35.6)
Primary school	163(40.0)
Junior high	69(17.0)
High school/Tech school	30(7.4)
Residence(Rural)	186(45.70)
Occupation
Farmer	107(26.3)
Worker	118(29.0)
Self-employed	89(21.9)
Unemployed	70(17.2)
Retired	23(5.6)
Monthly income
<¥500	67(16.5)
¥500–1500	111(27.3)
¥1500–2500	90(22.1)
¥2500–3500	115(28.3)
>¥3500	24(5.9)
Cognitive level and technical mastery	2.799 ± 1.188
Attitude and needs	3.328 ± 0.982
Social support	3.242 ± 1.046
Information characteristics	3.241 ± 0.919
Health communication ability	3.247 ± 0.977

Group differences across clusters

The results shown as Table 10, the clustering results that empirically support the integrated conceptual framework. Patterns in SES and literacy are consistent with Knowledge Gap Theory, environmental differences align with Information Ecology Theory, and gradients in interpersonal interaction correspond to Social Support Theory.

Table 10.

Group differences across clusters.

	Category (Medium IQR/Percentage)			Kruskal–Wallis H/ Pearson's chi-square(df)	p
Item	Low-ability group (n = 122)	Moderate-ability group (n = 132)	High-ability group (n = 153)	Kruskal–Wallis H/ Pearson's chi-square(df)	p
Age
60–63 years	18(14.75)	36(27.27)	41(26.80)	19.94(8)	0.011
64–67 years	44(36.07)	51(38.64)	59(38.56)
68–71 years	25(20.49)	26(19.70)	34(22.22)
72–75 years	25(20.49)	10(7.58)	15(9.80)
75 years older	10(8.20)	9(6.82)	4(2.61)
Education
Illiterate	63(51.6)	45(34.1)	37(24.2)	37.71(6)	<0.001
Primary school	47(38.5)	54(40.9)	62(40.5)
Junior high	12(9.8)	24(18.2)	33(21.6)
High school/Tech school	0(0.0)	9(6.8)	21(13.7)
Occupation
Farmer	33(27.05)	37(28.03)	37(24.18)	5.546(8)	0.698
Worker	32(26.23)	43(32.58)	43(28.10)
Self-employed	26(21.31)	22(16.67)	41(26.80)
Unemployed	23(18.85)	22(16.67)	25(16.34)
Retired	8(6.56)	8(6.06)	7(4.58)
Residence
Rural	69(56.56)	62(46.97)	55(35.95)	11.745(2)	0.003
Urban	53(43.44)	70(53.03)	98(64.05)
Gender
Male	69(56.56)	74(56.06)	89(58.17)	0.143(2)	0.931
Female	53(43.44)	58(43.94)	64(41.83)
Cognitive level and technical mastery	2.667(2.0,3.3)	3.333(2.7,4.0)	4.000(3.0,4.3)	73.119	<0.001
Income Level	2.000(2.0,3.0)	3.000(2.0,4.0)	4.000(2.0,4.0)	20.589	<0.001
Attitudes and needs	2.667(2.0,3.4)	3.667(2.3,4.0)	4.000(3.3,4.3)	62.144	<0.001
Social support	3.000(2.0,4.0)	3.000(2.0,4.0)	3.667(3.0,4.3)	32.344	0.003
Information characteristics	2.667(2.3,3.3)	3.000(2.3,3.7)	4.000(3.0,4.3)	83.082	<0.001

Significant sociodemographic and construct-level differences were observed across the three communication-ability clusters. Age differed notably, with the low-ability group containing a higher proportion of the oldest adults, whereas younger seniors predominated in the high-ability group (χ² = 19.94, p = 0.011). Education showed a strong gradient (χ² = 37.71, p < 0.001), over half of the low-ability group were illiterate, while the high-ability group had the greatest share of individuals with junior-high education or above. Residence also varied significantly (χ² = 11.745, p = 0.003), with rural residents concentrated in the low-ability group and urban residents in the high-ability group. Gender and occupation did not differ significantly.

Across all five theoretical constructs, the three clusters were clearly differentiated. Median scores increased monotonically from low to high-ability groups for cognitive level and technical mastery (X₁), income level (X₂), attitude and needs (X₃), social support (X₄), and information characteristics (X₅). In all, communication ability reflects a cumulative advantage: individuals who are younger, more educated, urban-residing, and socially supported also report stronger cognitive, motivational, and content-related capacities.

Multiple regression analysis

Table 11 presents the results of the multiple linear regression analysis examining the predictors of older adults’ health communication ability. After adjusting for sociodemographic covariates, all five theoretical constructs remained significant and positive predictors of communication ability. Cognitive and technical mastery (X₁) (β = 0.221, p < 0.001), income level (X₂) (β = 0.164, p < 0.001), attitude and needs (X₃) (β = 0.211, p < 0.001), social support (X₄) (β = 0.125, p = 0.003), and information characteristics (X₅) (β = 0.227, p < 0.001) were all significantly associated with higher communication ability.

Table 11.

Result of multiple regression analysis.

Variable	Beta	T	Lower 95% CI	Upper 95% CI	p	VIF	Tolerance
Cognitive level and technical mastery	0.221	4.760	0.130	0.312	<0.001	1.45	0.690
Income level	0.164	3.680	0.077	0.252	<0.001	1.34	0.749
Attitudes and needs	0.211	4.630	0.121	0.300	<0.001	1.39	0.720
Social support	0.125	2.960	0.042	0.208	0.003	1.77	0.838
Information characteristics	0.227	4.420	0.126	0.328	<0.001	1.30	0.566
Residence (Ref = Rural)	0.072	0.820	−0.101	0.245	0.415	1.30	0.772
Age (Ref=60–63)
64–67 years old	0.017	0.170	−0.184	0.219	0.866	1.66	0.601
68–71 years old	−0.025	−0.220	−0.255	0.204	0.830	1.51	0.661
72–75 years old	−0.253	−1.820	−0.528	0.021	0.070	1.41	0.709
75 years older	−0.377	−2.060	−0.736	−0.018	0.040	1.19	0.838
Gender (Ref = Male)	0.265	3.020	0.092	0.437	0.003	1.27	0.789
Education (Ref = Illiterate)
Primary school	0.042	0.430	−0.151	0.235	0.670	1.55	0.644
Junior high	−0.004	−0.030	−0.282	0.273	0.975	1.88	0.531
High school/Tech school	0.219	1.160	−0.154	0.592	0.248	1.65	0.606
R Square	0.415
Adjusted R Square	0.395
F	F(14,392) = 19.89, P < 0.001

Among the covariates, gender showed a significant effect (β = 0.265, p = 0.003), with women reporting higher communication ability than men. Age demonstrated a graded negative association, with adults aged 75 years older exhibiting significantly lower ability relative to the 60–63 group (β = –0.377, p = 0.040). Other age categories did not differ significantly. Residence and education level were not statistically significant in the adjusted model.

The overall model explained 41.5% of the variance in communication ability (adjusted R² = 0.395), representing a substantial improvement over earlier models that did not include confounding adjustment. Multicollinearity diagnostics indicated acceptable levels.

Together, the findings support all five hypotheses (H1–H5), as Figure 2 shows: older adults who have higher digital skills and health information literacy, greater economic resources, stronger personal motivation to seek health knowledge, more supportive social networks, and more favorable information features exhibit significantly higher health communication ability (and thus a smaller communication gap).

Figure 2.

Hypothses v

Discussion

This study provides an examination of the factors influencing the health knowledge communication gap among older adults in the context of TikTok. The findings affirm that several individual, social, and technological factors significantly affect older adults’ ability to acquire, assimilate, and disseminate health information. Key predictors identified include cognitive and technical literacy, income level, personal attitudes and needs, social support, and information characteristics of health content.

Theoretical contributions of an integrated framework

A major contribution of this study is the way it integrates Knowledge Gap Theory, Information Ecology Theory, and Social Support Theory to explain health knowledge communication in older adults. Individually, each theory offers a partial explanation; jointly, they provide a comprehensive framework that underpins our conceptual model. Knowledge Gap Theory postulates that as information in society increases, individuals of higher SES tend to acquire information faster than those of lower SES, widening knowledge disparities.²¹ In our model, income level represents this socioeconomic dimension. Although income was measured with a single item, it served as an indicator of the material and educational resources that older adults can leverage to obtain and communicate health information. Knowledge Gap Theory also informed our consideration of content accessibility because more complex information could exacerbate gaps between individuals with different literacy levels. By incorporating Knowledge Gap Theory, we recognize structural inequalities in knowledge communication and the need for strategies to narrow these gaps.

While Knowledge Gap Theory highlights structural differences, Information Ecology Theory offers a holistic view of the information environment in which older adults operate.⁶¹ In our study, cognitive level and technical mastery reflects the information subject's capacity, while attitude and needs and information characteristics reflect how individuals interact with the information technology and content. Social support captures aspects of the information environment, including the ease-of-use and social ambience of online platforms, as well as the presence of a supportive social milieu. By applying Information Ecology Theory, we treat older adults’ health knowledge communication as an emergent outcome of interactions between people, digital platforms, content characteristics, and social contexts. Our findings indeed support this systems perspective: older adults with high health knowledge communication ability tended to be those who not only had personal skills and motivation, but also operated in a conducive environment and embraced the affordances of technology. This aligns with the idea that effective information behavior requires alignment between the user, the technology, and the surrounding environment.

Social Support Theory further complements the model by focusing on the interpersonal dimension.⁴⁰ In the context of our study, social support manifests in several ways. First, it appears as a motivator: some older adults reported that the health concerns of family and friends drive them to seek out health knowledge online. This resonates with prior findings that family health needs can spur older individuals to adopt health technologies.⁶² Social support also appears as a facilitator of communication: participants who indicated they have harmonious relationships and a relaxed, natural atmosphere for conversation were more likely to discuss and disseminate health information. This reflects the classic understanding that social networks provide channels for information exchange and encouragement for engagement.⁶³ By incorporating Social Support Theory, our model recognizes that older adults do not make decisions about sharing health knowledge in a vacuum; rather, these decisions are embedded in social relationships.

The combined theoretical lens suggests that an older person's capacity to communicate health knowledge on TikTok is jointly determined by their personal resources, their technological and informational ecosystem, and their social connections. Each theory maps onto specific dimensions of our model, and together they address the multifaceted nature of the phenomenon. This theoretical integration is one of the study's strengths: it moves beyond single-theory explanations and demonstrates that older adults’ health knowledge communication is a product of individual capacity, content and technology, and social context interacting. In doing so, our findings contribute to each theory. For Knowledge Gap Theory, we provide contemporary evidence from a social-media context, showing that digital engagement and content format can moderate traditional SES gaps. For Information Ecology, we illustrate its applicability to personal health information behavior on a popular platform (TikTok), highlighting how a user-friendly environment and technology use can empower older users. For Social Support Theory, we reaffirm that even in online health activities, real-world social support and interpersonal motivations remain pivotal.

Practical implications

The findings of this study suggest several practical directions for enhancing health knowledge communication among older adults in digital environments. Strengthening cognitive and technical literacy should be prioritized through community-based digital training programs that teach basic smartphone operations, information verification, and short-video platform navigation.⁶⁴ Attitudinal and need-based components are equally important; peer-led workshops and interactive sessions with healthcare professionals may increase older adults’ confidence and perceived value of engaging with online health information.⁶⁵ The effects of information characteristics and social support further underscore the need for age-inclusive communication ecosystems, where public health agencies and platform designers collaborate to produce clear, trustworthy, and visually accessible content. Intergenerational support may also amplify the benefits of such interventions.⁶⁶ In addition, income-related disparities highlight the need for policy efforts to improve digital access through subsidized data plans and better rural connectivity. Given that the low-ability cluster was disproportionately rural, less educated, and digitally inexperienced, targeted outreach is essential to prevent further widening of digital and health knowledge gaps.⁶⁷

Limitation

While this study provides meaningful insights into older adults’ health knowledge communication on TikTok, several limitations should be acknowledged. First, the use of self-report measures for both predictors and outcomes raises the possibility of common method bias, as collecting data from the same respondents at a single time point may inflate correlations due to consistency motives or respondent fatigue. Although procedural remedies were applied, this bias cannot be fully ruled out. Potential social desirability bias may have influenced responses to items measuring willingness or perceived ease of sharing health information. Older respondents may have overstated altruistic or pro-social intentions, contributing to inflation in the outcome and related predictors. The absence of a social desirability scale limited our ability to correct for this bias.

Second, several constructs exhibited limited measurement validity. Income relied on a single categorical item and therefore lacked reliability and granularity, and components of the social support were represented by broad single items that may not capture the full complexity of interpersonal and contextual support. Moreover, certain items adapted from prior studies have not been extensively validated among older Chinese users of short-video platforms. These constraints likely contributed to measurement error and may have attenuated or obscured more nuanced relationships.

Third, the cross-sectional design prevents causal inference. Although interpreted through theoretical frameworks, associations between digital literacy, motivation, social context, and communication ability may be bidirectional; longitudinal or experimental designs are needed to clarify directionality and mediating processes.

Fourth, although structural equation modeling (SEM) could theoretically offer a more comprehensive test of the conceptual framework, preliminary analyses indicated model instability and poor fit—likely attributable to limited sample size, correlated constructs, and partial reliance on single-item measures. Regression models therefore offered a more parsimonious analytic approach. Nevertheless, the conceptual model may still omit relevant mediators or moderators (e.g. health status, prior digital experience, platform trust), which could partially explain the modest R² of the final model. Future research incorporating additional theoretically grounded covariates may reveal mechanisms not captured here.

Fifth, sampling bias must be considered. Participants were relatively digitally engaged older adults willing to complete an online or interviewer-assisted survey, which means the study may not represent digitally excluded seniors who face more severe structural barriers. The findings therefore reflect patterns among active or semi-active social media users rather than the broader older adult population in China.

Finally, the selection of covariates was necessarily constrained. Although we included demographic variables and residence type based on theoretical justification and empirical cluster differentiation, we were unable to measure other potentially influential confounders such as general health status, cognitive functioning, chronic disease burden, digital access infrastructure, or prior exposure to digital training. The omission of these variables may have limited the detection of more complex mediating or moderating pathways. Similarly, questionnaire design constraints may have constrained the ability to uncover deeper structural or ecological mechanisms predicted by the theoretical framework.

Future research should build on these findings through designs capable of establishing temporal and causal mechanisms, such as longitudinal studies and experimental or quasi-experimental interventions that track how digital literacy, motivational factors, and social support shape health communication behaviors over time. More comprehensive measurement tools, including multi-item scales for SES, information evaluation skills, and digital experience are needed to strengthen construct validity and reduce measurement error. Future studies should also address potential sampling bias by intentionally recruiting less digitally active or digitally excluded older adults, and should examine how infrastructural inequalities, platform algorithms, and community-level resources shape the evolving information ecology. Finally, the use of SEM or multilevel modeling with larger and more diverse samples could clarify the interrelationships among the theoretical constructs and uncover the mediating or moderating pathways underlying health knowledge communication in later life.

Conclusion

Older adults’ health knowledge communication on TikTok is shaped by intersecting individual, informational, and contextual factors. Cognitive level and technical mastery, attitudes and needs, information characteristics, social support, and income level jointly and positively predicted health communication ability, while cluster analysis revealed distinct low, moderate, and high-ability profiles across age, education, and residence. These findings extend knowledge gap, information ecology, and social support perspectives to short-video environments and highlight leverage points for age-friendly digital health interventions targeting vulnerable subgroups.

Supplemental Material

sj-docx-1-dhj-10.1177_20552076261415942 - Supplemental material for Factors influencing the online health knowledge communication gap among older adults: Evidence from TikTok on chronic diseases

Supplemental material, sj-docx-1-dhj-10.1177_20552076261415942 for Factors influencing the online health knowledge communication gap among older adults: Evidence from TikTok on chronic diseases by Ni Cheng, Heng Dong and Yang Yang in DIGITAL HEALTH

Supplemental Material

sj-docx-2-dhj-10.1177_20552076261415942 - Supplemental material for Factors influencing the online health knowledge communication gap among older adults: Evidence from TikTok on chronic diseases

Supplemental material, sj-docx-2-dhj-10.1177_20552076261415942 for Factors influencing the online health knowledge communication gap among older adults: Evidence from TikTok on chronic diseases by Ni Cheng, Heng Dong and Yang Yang in DIGITAL HEALTH

Footnotes

ORCID iD

Heng Dong

Author contributions

Ni Cheng: conceptualization, supervision, validation, writing—reviewing. Heng Dong: conceptualization, writing—original draft preparation, visualization. Yang Yang: writing—review & editing.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the National Social Science Foundation of China (grant number 23BTQ070).

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Supplemental material

Supplemental material for this article is available online.

References

Sharma

Nahak

Kanozia

. Identifying the gaps of health communication research: a systematic review. J Content Community Commun 2019; 10: 123–133.

Nishiuchi

Taguri

Ishikawa

. Using a marginal structural model to design a theory-based mass media campaign. PLOS One 2016; 11: e0158328.

Sanders

. Triple P – positive parenting program: a population approach to promoting competent parenting. Aust E-j Adv Ment Health 2003 Jan; 2: 127–143.

Sandell

Sebar

Harris

. Framing risk: communication messages in the Australian and Swedish print media surrounding the 2009 H1N1 pandemic. Scand J Public Health 2013 Dec; 41: 860–865.

Spitale

Merten

Jafflin

, et al. A novel risk and crisis communication platform to bridge the gap between policy makers and the public in the context of the COVID-19 crisis (PubliCo): protocol for a mixed methods study. JMIR Res Protoc 2021 Nov 1; 10: e33653.

Lee

Basnyat

. From press release to news: mapping the framing of the 2009 H1N1 a influenza pandemic. Health Commun 2013; 28: 119–132.

Schmidt

. Communication gap: the disconnect between what scientists say and what the public hears. Environ Health Perspect 2009 Dec; 117: A548–A551.

Gajović

Svalastog

. When communicating health-related knowledge, beware of the black holes of the knowledge landscapes geography. Croat Med J 2016 Oct; 57: 504–509.

Chen

Walter

Wei

JCC

. Using ChatGPT-like solutions to bridge the communication gap between patients with rheumatoid arthritis and health care professionals. JMIR Med Educ 2024 Feb 27; 10: e48989.

10.

China Internet Network Information Center. The 55th Statistical Report on China’s Internet Development [Internet]. China Internet Network Information Center; 2025 Jan p. 68. Available from: https://cnnic.cn/NMediaFile/2025/0428/MAIN17458061595875K4FP1NEUO.pdf.

11.

Duan

Chen

Wang

, et al. Digital exclusion and cognitive function in elderly populations in developing countries: insights derived from 2 longitudinal cohort studies. J Med Internet Res 2024 Nov 15; 26: e56636.

12.

Kebede

Ozolins

Holst

, et al. Digital engagement of older adults: scoping review. J Med Internet Res 2022 Dec 7; 24: e40192.

13.

Morgan

Marston

Maddock

, et al. Adjusting technology for an ageing population: co-producing technological solutions with older adults. Innov Aging 2023 Dec 21; 7: 1170–1171.

14.

Elhariry

Malhotra

Goyal

, et al. A SIMBA CoMICs initiative to cocreating and disseminating evidence-based, peer-reviewed short videos on social media: mixed methods prospective study. JMIR Med Educ 2024 Oct 30; 10: e52924.

15.

MacKinnon

Kia

Lacombe-Duncan

. Examining TikTok’s potential for community-engaged digital knowledge mobilization with equity-seeking groups. J Med Internet Res 2021 Dec 9; 23: e30315.

16.

Tan

Liu

Shi

, et al. Information quality of videos related to Helicobacter pylori infection on TikTok: cross-sectional study. Helicobacter 2024; 29: e13029.

17.

Yang

. Why people use TikTok (douyin) and how their purchase intentions are affected by social media influencers in China: a uses and gratifications and parasocial relationship perspective. J Interact Advert 2021; 21: 297–305.

18.

Adebesin

Smuts

Mawela

, et al. The role of social media in health misinformation and disinformation during the COVID-19 pandemic: bibliometric analysis. JMIR Infodemiol 2023 Sept 20; 3: e48620.

19.

Liu

, et al. Analysis and detection of health-related misinformation on Chinese social media. IEEE Access 2019; 7: 154480–9.

20.

Fan

Yuan

, et al. Recognizing fake information through a developed feature scheme: a user study of health misinformation on social media in China. Inf Process Manage 2022 Jan 1; 59: 102769.

21.

Tichenor

Donohue

Olien

. Mass media flow and differential growth in knowledge. [cited 2025 May 22]; Available from: https://dx.doi.org/10.1086/267786.

22.

Colom

. The digital divide: by jan van dijk, cambridge, polity press, 2020, 208 pp., £17.99 (paperback), ISBN: 978-1-509-534456. Inf, Commun Soc. 2020 Sept 18;23(11):1706–8.

23.

Liu

Sun

, et al. The digital divide and cognitive disparities among older adults: community-based cohort study in China. J Med Internet Res 2024 Nov 27; 26: e59684.

24.

Sun

Jiang

Zhou

. Associations between perceived usefulness and willingness to use smart healthcare devices among Chinese older adults: the multiple mediating effect of technology interactivity and technology anxiety. Digit Health 2024 May 27; 10: 20552076241254194.

25.

Hou

Wang

. How to improve older adults’ trust and intentions to use virtual health agents: an extended technology acceptance model. Humanit Soc Sci Commun 2024 Dec 18; 11: 1–11.

26.

Huang

Quan

. Narrowing the digital divide: the growth and distributional effect of internet use on income in rural China. China Econ Rev 2025 June 1; 91: 102387.

27.

Sheng

Wang

Zhu

, et al. An exploratory study of digital reverse mentoring based on grounded theory from the youth’s perspective. Document Inform Knowl 2023; 40: 52–61.

28.

Van Deursen

Helsper

. Collateral benefits of internet use: explaining the diverse outcomes of engaging with the Internet. New Media Soc 2018 July; 20: 2333–2351.

29.

Hong

Zhou

Fang

, et al. The digital divide and health disparities in China: evidence from a national survey and policy implications. J Med Internet Res 2017 Sept 11; 19: e7786.

30.

. Internet use and income gaps between rural and urban residents in China. J Asia Pac Econ 2024; 29: 789–809.

31.

Barbati

Maranesi

Giammarchi

, et al. Effectiveness of eHealth literacy interventions: a systematic review and meta-analysis of experimental studies. BMC Public Health 2025 Jan 23; 25: 288.

32.

Romanelli

Kim

Fredriksen-Goldsen

. Bridging access gaps: examining use of digital health technologies among sexual and gender minority older adults. Innov Aging. 2024; 8: 671–672.

33.

Jiang

Wang

Leng

, et al. The level of electronic health literacy among older adults: a systematic review and meta-analysis. Arch Public Health 2024 Nov 7; 82: 204.

34.

Zhang

. Trust in social media brands and perceived media values: a survey study in China. Comput Hum Behav 2022 Feb 1; 127: 107024.

35.

Horton

. Information ecology. J Syst Manag 1978; 29: 32–36.

36.

Thomas

Menon

Boruff

, et al. Applications of social constructivist learning theories in knowledge translation for healthcare professionals: a scoping review. Implement Sci 2014 May 6; 9: 54.

37.

Berkowsky

Czaja

. 2 - challenges associated with online health information seeking among older adults. In: Pak

McLaughlin

(eds) Aging, technology and health [internet]. San Diego: Academic Press, 2018 [cited 2025 June 9], pp.31–48. Available from: https://www.sciencedirect.com/science/article/pii/B9780128112724000026.

38.

Guo

Zhang

Sun

. The privacy-personalization paradox in mHealth services acceptance of different age groups. Electron, Commer, Res, Appl 2016; 16: 55–65.

39.

Luo

. Older people’s intention to use medical apps during the COVID-19 pandemic in China: an application of the unified theory of acceptance and use of technology (UTAUT) model and the technology of acceptance model (TAM). Ageing Soc 2024 July; 44: 1515–1532.

40.

Leahy-Warren

. Social support theory. In: Theories guiding nursing research and practice: making nursing knowledge development explicit. New York, NY, US: Springer Publishing Company, 2014, pp.85–101.

41.

Chan

Chen

. Personal and other factors affecting acceptance of smartphone technology by older Chinese adults. Appl Ergon 2016; 54: 62–71.

42.

Pan

Jordan-Marsh

. Internet use intention and adoption among Chinese older adults: from the expanded technology acceptance model perspective. Comput, Hum, Behav, 2010; 26: 1111–1119.

43.

Zhang

Liu

Deng

, et al. Knowledge sharing motivations in online health communities: a comparative study of health professionals and normal users. Comput Human Behav 2017 Oct; 75: 797–810.

44.

Zhou

Miao

, et al. Effects of font style and font color in news text on user cognitive load in intelligent user interfaces. IEEE Access 2022; 10: 10719–10730.

45.

Guest

Bunce

Johnson

. How many interviews are enough?: an experiment with data saturation and variability. Field Methods 2006 Feb 1; 18: 59–82.

46.

Saunders

Sim

Kingstone

, et al. Saturation in qualitative research: exploring its conceptualization and operationalization. Qual Quant 2018 July 1; 52: 1893–1907.

47.

Austin

Steyerberg

. The number of subjects per variable required in linear regression analyses. J Clin Epidemiol 2015 June 1; 68: 627–636.

48.

da Silva

Mosquera

Soares

. Factors influencing knowledge sharing among IT geographically dispersed teams. Technol Forecasting Social Change 2022 Jan 1; 174: 121299.

49.

Gan

Jia

. Income differences, media use and knowledge acquisition: a study on the influence mechanism of knowledge gap among rural elders: logistics-regression based moderating test and analysis. Journalism Mass Commun 2021; 03: 54–62.

50.

Wang

Zhang

. Behavior patterns and influencing factors: health information acquisition behavior of Chinese senior adults on WeChat. Heliyon 2023 June 1; 9: e16431.

51.

Jia

Bao

Wang

, et al. Factors influencing health information acquisition behavior from a social capital theory perspective: a study of zhihu users. Front Public Health 2025; 13: 1501879.

52.

Kuang

. Exploring influence factors of WeChat users’ health information sharing behavior: based on an integrated model of TPB, UGT and SCT. Int J Hum–Comput Interact 2021 Aug 9; 37: 1243–1255.

53.

Jia

. Users’ health information sharing behavior in social media: an integrated model. Humanit Soc Sci Commun 2024 Dec 4; 11: 1647.

54.

Mirzaei

Aslani

Luca

, et al. Predictors of health information–seeking behavior: systematic literature review and network analysis. J Med Internet Res 2021 Jul 2; 23(7): e21680.

55.

Xiao

Sharman

Rao

, et al. Factors influencing online health information search: an empirical analysis of a national cancer-related survey. Decis Support Syst 2014 Jan 1; 57: 417–427.

56.

Bhattacherjee

. Understanding information systems continuance: an expectation-confirmation model. Mis Q 2001 Sept 1; 25: 351–370.

57.

Deng

Zhang

, et al. Seeking and using intention of health information from doctors in social media: the effect of doctor-consumer interaction. Int J Med Inf 2018 July 1; 115: 106–113.

58.

Soroya

Farooq

Mahmood

, et al. From information seeking to information avoidance: understanding the health information behavior during a global health crisis. Inf Process Manage 2021 Mar 1; 58: 102440.

59.

Jiabei

Zhaohua

Tailai

. Study on the influencing factors of network health information search behavior of professional women and the regulating effect of social support. Libr Inf Sci 2020 Dec 5; 64: 53–62.

60.

Wanjiao

Zigang

. Analysis of impacting factors of health information seeking behavior of Chinese urban residents. J Mod Inform 2019; 39: 77–85.

61.

Davenport

Prusak

. Information Ecology: Mastering the Information and Knowledge Environment. New York: Oxford University Press, 1997.

62.

Chang

Mai

Zhang

, et al. Media use behavior mediates the association between family health and intention to use Mobile health devices among older adults: cross-sectional study. J Med Internet Res 2024 Feb 19; 26: e50012.

63.

Kim

Kreps

Shin

. The role of social support and social networks in health information–seeking behavior among Korean Americans: a qualitative study. Int J Equity Health 2015 Apr 28; 14: 40.

64.

Ngiam

NHW

Yee

Teo

, et al. Building digital literacy in older adults of low socioeconomic Status in Singapore (project wire up): nonrandomized controlled trial. J Med Internet Res 2022 Dec 2; 24: e40341.

65.

Lms

Abah

, et al. Digital Literacy Training for Low-Income Older Adults Through Undergraduate Community-Engaged Learning: Single-Group Pretest-Posttest Study. JMIR Aging 2024 May 14; 7: e51675.

66.

Leedahl

Wilber

Choi

. The promise of digital inclusion for older adults: building bridges for underserved older adults in two states. Innov in Aging. 2023 Dec 21; 7: 523.

67.

Finco

Mir

Gresham

, et al. Ethical considerations of digital health technology in older adult care. Lancet Healthy Longevity 2024 Jan 1; 5: e12–e13.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.03 MB

0.02 MB