Open research questions in Health Literacy and Information Accessibility

What the literature leaves open

• Schools and community health centers may consider piloting HPM-informed e-booklet– based anemia education as part of existing adolescent health programs while evaluating its feasibility, participant engagement, and knowledge retention. Future research should use controlled study designs, report complete evidence of instrument validity and reliability, preserve continuous outcome measures where appropriate, document intervention fidelity, include participant-level paired transition tables, recruit participants from multiple schools, and evaluate iron-folic acid supplementation adherence, dietary practices, and hemoglobin outcomes over longer follow-up periods. REFERENCES Abu-Baker, N. N., Eyadat, A. M., & Khamaiseh, A. M. (2021). The impact of nutrition education on knowledge, attitude, and practice regarding iron deficiency anemia among female adolescent students in Jordan. Heliyon, 7(2), e06348. https://doi.org/10.1016/j.heliyon.2021.e06348 Cimahi City Health Office. (2023). Anaemia screening report for senior-high-school students in the Melong Asih Public Health Center service area [Unpublished routine programme report]. 2026 Jurnal Kesehatan Budi Luhur | Published by STIKes Budi Luhur 428 Ambarsari et al. (2026): Pre–Post Changes in Anemia Knowledge and Attitudes Following HPM-Informed E-Booklet Education Among Adolescent Girls Imanuna, H. (2022). Penyuluhan anemia gizi besi menggunakan media booklet untuk meningkatkan pengetahuan dan sikap siswi SMAN 7 Malang. Nutriture Journal, 1(1), 1-8. https://doi.org/10.31290/nj.v1i1.3526 M, F. A., Rahma, N. R. S., & Rahayu, B. (2024). The influence of JASA TAMI (Healthy Adolescents Without Anemia) booklet media on adolescent girls in Islamic boarding schools Mizanul Ulum Sanrobone in 2023. Jurnal Kebidanan Kestra, 6(2), 195-204. https://doi.org/10.35451/jkk.v6i2.2134 Ministry of Health of the Republic of Indonesia. (2022, November 16). Remaja bebas anemia: Konsentrasi belajar meningkat, bebas prestasi. Ayo Sehat Kementerian Kesehatan Republik Indonesia. Nikmah, A., Supadi, J., & Setiadi, Y. (2022). The effect of nutrition education with e-booklet media on knowledge and attitude about anemia in adolescent girls at SMP N 1 Gemuh, Kendal Regency. Journal of Nutrition Science, 3(1), 9-16. https://doi.org/10.35308/jns.v3i1.5242 Paramita, F., Rachmawati, W. C., Ekawati, R., Pherenis, H. S. A., Septiani, S. T., & Awaliahmunazila, M. (2024). The effectiveness of nutrition education using e-booklet on knowledge and attitude about anemia among adolescent girls in Turen Malang. Preventia: The Indonesian Journal of Public Health, 9(1), 75-80. https://doi.org/10.17977/um044v9i12024p75-80 Pender, N. J., Murdaugh, C. L., & Parsons, M. A. (2015). Health promotion in nursing practice (7th ed.). Pearson. Puspadiina, N., & Kurniawati, H. F. (2023).

  Pre–Post Changes in Anemia Knowledge and Attitudes Following HPM-Informed E-Booklet Education Among Adolescent Girls · 2026 · DOI

• doi:10.15294/ujph.v0i0.30866 Journal. https://journal.unnes.ac.id/nju/phpj/article/view/18127/9165 Ayu Titisari I, Riyanti E, Nugraha PP. Aplikasi Teori Helath Belief Model pada Partisipasi Wanita Usia Subur (WUS) dalam Pemeriksaan IVA di Kelurahan Kalibanteng Kulon. Jurnal Kesehatan Masyarakat. 2018;6(5):751–9. doi:10.14710/mkmi.11.1.75-85 Artha Sari NPY, Subrata IM. Hubungan Tingkat Pengetahuan Dan Sikap Dengan Perilaku Ibu Rumah Tangga Terkait Pemeriksaan Payudara Klinis (Sadanis) Di Wilayah Kerja Puskesmas Kuta Selatan Saat 2019;4(1):30–6. 15. 16. Vol. XXI No.

  Effect of the SADIVA Digital Application on Knowledge, Counseling, and Screening Skills of Health Cadres for Early Breast and Cervical Cancer Detection: A Quasi-Experimental Study · 2026 · DOI

• Although assistive technologies such as screen readers, magnifiers, mobile applications, and AI-based tools have facilitated physical access to information, our findings indicate that technological solutions alone are insufficient to address the broader cognitive and contextual challenges of HL.

  HEALTH LITERACY AMONG INDIVIDUALS WITH VISUAL IMPAIRMENTS IN TURKEY: A CROSS-SECTIONAL EVALUATION AND POLICY RECOMMENDATIONS · 2026 · DOI

• The limitations of this study are that it used a single search engine, searched for a single term, evaluated written texts in a single lan- guage, and did not analyse video content. This aligns with previous evalu- ations of online health information, which have shown that key risk and adverse-effect information is often underreported and that websites meeting established quality benchmarks tend to provide more comprehen- sive content (28).

  Readability and quality of online resources on prolotherapy · 2026 · DOI

• treatment Xie, Y., Seth, I., Hunter-Smith, D. J., Rozen, W. M., Ross, R., & Lee, M. (2023). Aesthetic surgery advice and counseling from artificial intelligence: a rhinoplasty consultation with ChatGPT.

  Evaluating ChatGPT-4 as an Educational Tool for Physical Activity Counseling in Patients with Hyperlipidemia · 2026 · DOI

• immunoglobulin, anti-D corticosteroids, immunoglobulin; is corticosteroid conditional, whereas the intravenous immunoglobulin and anti- D recommendations are strong (1). For children requiring treatment because of non-life-threatening mucosal bleeding or diminished quality of life, scoring also considered short-course treatment principles and avoidance of unnecessary prolonged corticosteroid exposure. in the Some parent-facing diagnostic and counseling questions, including routine ANA testing, bone marrow examination in typical childhood ITP, and MMR vaccination after ITP, are not pediatric management 2019 fully covered recommendations. For relevant ASH 2011 those areas, recommendations carried forward for context in the 2019 guideline document were also considered (1, 13). The term guideline/reference concordance was therefore used to avoid implying that every parent-facing item was scored only against a single ASH recommendation. Evaluation instrument: AI-ITP parent response score The primary evaluation instrument was the AI-ITP Parent Response Score (AI-ITP-PRS), a seven-domain rubric developed specifically for this study to summarize the quality of parent- facing educational responses about childhood ITP. The rubric was created before blinded model scoring. Its content was derived from three sources: (i) the clinical counseling domains represented in the 40-question bank, (ii) ASH 2019 pediatric ITP management principles, carried-forward ASH 2011 diagnostic and vaccination recommendations anticipated LLM- specific failure modes in parent education, including fluent but guideline-discordant advice, unsafe triage language, unsupported remedies, false certainty, and excessive or insufficient reassurance. Content validity was addressed through expert clinical review rather than formal psychometric validation. The initial domain set and anchors were drafted by the study team and reviewed by the three-clinician panel for relevance to pediatric ITP counseling, clarity of the 1–5 anchors, non-redundancy between domains, and coverage of both clinical safety and parent communication.

  Freely accessible large language models for parent education in pediatric immune thrombocytopenia: an expert-rated cross-sectional study of safety, readability, and guideline concordance · 2026 · DOI

• Future research should move beyond content evaluation toward implementation-readiness testing. Multicenter studies should use larger and multilingual parent-question banks, include caregivers with different health-literacy levels, and measure whether AI-generated explanations improve or impair understanding, anxiety, trust calibration, and intended behavior. Comparative studies should test free public models against clinician-supervised, retrieval-augmented, guidelinelocked pediatric hematology systems. They should also examine whether parent-facing AI tools reduce after-visit uncertainty or unintentionally increase unnecessary emergency visits, diagnostic testing, or vaccine hesitancy. Future work should also validate or refine the AI-ITP-PRS itself before it is used as a stand-alone benchmarking instrument, with attention to caregiver input, external rater reproducibility, construct validity, and links between scores and clinically meaningful parent outcomes. after model updates, Longitudinal evaluation is also needed. LLM outputs can interface safety change retrieval-system changes, and prompt-policy modifications, revisions. Pediatric hematology applications should therefore document model versions, undergo periodic re-auditing, include high-risk prompt testing, and use predefined escalation language.

  Freely accessible large language models for parent education in pediatric immune thrombocytopenia: an expert-rated cross-sectional study of safety, readability, and guideline concordance · 2026 · DOI

• This study provides novel insight into public awareness and expectations of the CAPQ services and is among the first to assess both awareness and user perceptions in the general population. Inclusion of public perspectives beyond clini- cians directly involved in poisoning events is a key strength. Nonetheless, the findings should be interpreted in light of several limitations. The sample size was adequate for descriptive analyses but limited statistical power and the precision of subgroup estimates, which remain explora- tory. Voluntary participation may have introduced selection bias, with possible underrepresentation of remote regions and overrepresentation of healthcare workers despite tar- geted outreach efforts. Nevertheless, because the healthcare worker category was broadly defined, prior CAPQ awareness or use did not necessarily occur within the context of direct patient care. Prior CAPQ use was also similarly distributed between healthcare and non-healthcare workers, suggesting that satisfaction evaluations were not confined to healthcare professionals and thereby limiting potential bias related to professional background. Information bias, including recall and social desirability bias, remains possible, although con- fidentiality and anonymity were emphasized to mitigate this risk.

  Public awareness and satisfaction with the centre antipoison du quebec (CAPQ): a population-based survey · 2026 · DOI

• The findings of this study should be interpreted in light of several limitations. First, the data were collected from individuals residing in a single province (Yozgat) using a convenience sampling method. This limits the direct gen- eralizability of the results to different geographical, cultur- al, and socioeconomic contexts. Since patterns of digital technology use and attitudes toward artificial intelligence may vary across regions, the findings should be evaluated within the specific context in which the data were collect- ed. Future research conducted with larger and multi-cen- ter samples would help strengthen the external validity of the findings. In addition, the intention to adopt ChatGPT-generated health information was measured using a single-item in- dicator, consistent with prior research (19). Single-item measures have been shown to be valid and predictive for clear and unidimensional constructs such as behav- ioral intention (21; 30). Among their key advantages are the reduction of participant burden, shorter administra- tion time, and improved data quality, particularly in field- based survey research. The literature also suggests that narrowly scoped multi-item scales may be perceived as repetitive or fatiguing by respondents, potentially com- promising response quality (30). From this perspective, when the construct being measured is conceptually lim- ited in scope, a single-item measure may provide a more direct and parsimonious assessment.

  The Role of E-Health Literacy and Trust in the Adoption Intention of ChatGPT-Generated Health Information · 2026 · DOI

• This study had several limitations. First, only readability scores were assessed, and the scientific accuracy, quality, and timeliness of the websites were not evaluated. Second, the search terms, such as ‘autism,’ ‘atypical autism,’ ‘Asperger syndrome,’ ‘pervasive developmental disorder,’ and included medical terminology. The frequent use of these technical terms may have inflated readability scores and exaggerated the difficulty of the texts. Third, the evaluation did not consider the overall quality of the websites, and only the top 50 sites per search term that met the inclusion criteria were analyzed. Fourth, the study was limited to materials published in English, so that no conclusions can be drawn about resources in other languages. Fifth, only the Google search engine was used, and materials from other search engines or social media platforms were not included. In addition, although readability tools provide objective metrics based solely on text, this study did not evaluate the potential impact of graphics, visuals, videos, interactive features, or page layout on readability. Finally, the research was conducted within a specific timeframe. Since website content can change over time, future studies should adopt a more dynamic design that accounts for long-term variations.

  Readability of Online Patient Education Materials on Autism: An Invisible Barrier to Information Access · 2026 · DOI

• Despite the significant increase in demand for health information on adult ADHD, there is a substantial gap in research evaluating the information sources on adult ADHD.

  Too Complex, Too Unreliable: Gaps in the Quality of Online Health Information for Adults with ADHD · 2026 · DOI

• This web-based analysis may be affected by geographic and language biases due to search engine personalization, which may limit the generalizability of the findings. Acknowledgements This scientific research paper was derived from a research grant funded by Taibah University, Madinah, Kingdom of Saudi Arabia. All authors have read and agreed to the published ver- sion of the manuscript. 1 3Journal of Maxillofacial and Oral Surgery Funding No fund. Data Availability The raw data supporting the conclusions of this arti- cle will be made available by the authors on request.

  Oro-Antral Communication: Web-Based Health Information Quality Study · 2026 · DOI

• While this study provides empirical evidence for artificial intelli- gence-assisted health education text generation, it has five key limita- tions. First, the sample size is constrained (total n = 100), covering only five mainstream large language models (LLMs) and five health education topic categories. This narrow scope may not fully capture the characteristics of other models (e.g., the GPT series, iFlytek Spark) or specialized topics such as rare diseases and genetic disor- ders. Second, the assessment relies solely on objective indicators, without integrating patients’ subjective experience data (e.g., text comprehension scores, operational implementation rates) or clinical evaluations by medical professionals. This makes it challenging to comprehensively reflect the actual application effectiveness of the generated texts. Third, the study did not account for the impact of patients’ individual differences (e.g., age, educational background, health literacy level) on text acceptance, which means the pertinence of the proposed optimization strategies requires further enhance- ment. Fourth, model outputs were collected at a single time point (November 25, 2025). LLMs are frequently updated, which may cause temporal shifts in response quality and readability; future studies should include longitudinal assessments. Fifth, this study remains largely descriptive in design and did not experimentally manipulate prompt strategies. Although standardized prompts ensured fair cross- model comparison, real-world clinical applications often use guided, constrained, or role-based prompts. Future research should conduct controlled experiments to compare how different prompting approaches affect the quality and readability of AI-generated health education materials. Sixth, due to the lack of validated and widely accepted readability indices for Chinese medical texts, all questions and responses required dual translation (Chinese-to-English and English-to-Chinese). Although the entire translation process was completed by a professional obstetrics and gynecology specialist to ensure consistency, potential linguistic differences caused by transla- tion may still exist and should be considered when interpreting the results. Nevertheless, all outputs were processed in an identical manner, thus preserving full cross-model comparability. Seventh, all models were evaluated using English prompts and outputs. Doubao, Deep Seek, and Wenxin Yiyan are Chinese-preferred large language models primarily optimized for Chinese understanding and genera- tion. Their performance on English medical questions may be inher- ently disadvantaged compared with internationally developed models such as GPT-5 and Gemini, which could introduce systematic differ- ences unrelated to actual clinical quality. This language-origin disad- vantage should be considered when interpreting between-group comparisons.

  Comparative evaluation of the quality, reliability, and readability of five large language model responses to frequently asked questions on gestational hypertension · 2026 · DOI

• This study has the following limitations. First, the survey population was recruited solely from patients with coronary heart disease in a tertiary Grade A hospital in Guangxi, China. This may limit the generalizability of the findings to patients in other regions or those treated in primary and secondary healthcare settings. Consequently, the results may lack broad representativeness. Second, data were collected within 24–48 h of hospitalization. Although patients were clinically stable at the time of assessment, this early stage may still have influenced self-reported responses. Third, all data were collected via self-report questionnaires, which may introduce common method biases, including social desirability and recall biases. Fourth, although the classification accuracy of the latent profile analysis was relatively high (entropy >0.8), classification uncertainty was not adjusted for in the regression model. Future studies may adopt more appropriate bias correction methods to improve estimation precision. Finally, this study adopted a cross-sectional, one-off survey design and therefore could not examine causal relationships among the variables. In the future, multicenter, large-sample longitudinal studies should be considered to further clarify the potential causal relationships and pathways of influence among digital health literacy, social support, health beliefs, and their associated factors in patients with coronary heart disease.

  Latent profile and associated factors of digital health literacy among patients with coronary heart disease: a cross-sectional study · 2026 · DOI

• This study employed a transparent, systematically applied search strategy with a structured CHS content checklist informed by expert toxicologists and supported by standardized reviewer training. It also demonstrated substantial interrater agreement (κ = 0.78). Additionally, the mixed-methods approach to reviewer ratings and impressions provided a nuanced understanding of how CHS is portrayed on YouTube. At the same time, several limitations should be considered when interpreting these findings. The author- developed checklist and usefulness scale, while supported by expert input and high kappa values, were study-specific and have not undergone formal psychometric validation. Classification thresholds were set a priori and deliberately conservative, and different cutoffs could alter the distribution of categories. YouTube search results are dynamic and influenced by personalization factors that cannot be fully controlled. Despite using standardized search terms and logged-out browsing, the dataset represents a time- limited snapshot (April-September 2023) and may reflect selection and platform algorithm biases. The analysis was restricted to English-language content on a single platform, limiting generalizability to other languages, regions, or platforms such as TikTok or Instagram. The modest sample size (n = 97) is adequate for descriptive analysis but constrains statistical power for detecting small associations or subgroup analyses. Our analyses are therefore primarily descriptive, supplemented by simple correlation tests, and are not intended to support complex inferential modeling. Additionally, there were a few YouTube policy changes that affected the content included in this analysis. 2026 Dean et al. Cureus 18(5): e109113. DOI 10.7759/cureus.109113 9 of 12 YouTube disabled the “dislike” reaction in November 2021, which affected 27 videos (26%) of our dataset. Further, a Community Guidelines update in August 2023, during the data extraction period, provided a framework for prevention, treatment, and disinformation categorization for reporting violations [36]. While our dataset was not actively affected by these changes, future content analyses will likely be influenced by these updates. Judgments of usefulness and misleading content involve some degree of subjectivity despite structured tools, reviewer training, and consensus procedures. This may contribute to misclassification at the margins and should be considered when interpreting the exact proportions of videos in each category. Finally, the study characterizes content but does not assess how viewers interpret or act on videos, precluding causal inferences about clinical outcomes. Our recommendations to clinicians and educators should therefore be understood as reasoned implications of the observed content landscape rather than conclusions about measured behavioral effects.

  Assessing the Quality of YouTube Videos About Cannabinoid Hyperemesis Syndrome · 2026 · DOI

• This study provides a theoretically grounded model examining the relationships between e-health literacy, physical literacy, and healthy life- style beliefs, thereby contributing to the existing literature. In particular, this study contributes by empirically demonstrating the mediating role of physical literacy in the relationship between e-health literacy and nutri- tion beliefs, physical activity beliefs, and health beliefs. However, the framework presented should be considered as a partial representation of the relationships among health literacy and health belief constructs, rather than a comprehensive model. Another important limitation of this study is the use of a convenience sampling method. This sampling approach may introduce selection bias, as participants were recruited from acces- sible schools and volunteered to participate, which may limit the repre- sentativeness of the sample. Although the sample size (N = 816) can be considered adequate for statistical analysis, it may not be sufficient to generalize the findings to broader populations. Therefore, the findings should be interpreted with caution in terms of external validity. In addi- tion, since the participant group consisted only of high school students, the generalizability of the findings is limited to similar age groups and contexts. Furthermore, the cross-sectional design of the study restricts the ability to draw conclusions about the directionality of relationships among variables. Finally, the study focused on a limited set of variables, and other potential mediating or moderating factors were not included.

  The mediating role of E-health literacy and physical literacy in developing nutrition, physical activity, and health beliefs · 2026 · DOI

• This study used a random forest model, an advanced machine learning technique, to identify key determinants of health information literacy (HIL) among patients with COPD. Compared with traditional statistical methods, this approach provided a more data-driven and comprehensive analysis by ranking variables based on their relative impor- tance. Additionally, LASSO regression was employed to refine variable selection, enhancing the robustness and reli- ability of the findings. These methodological advancements provide essential insights into health literacy among patients with COPD and lay the foundation for developing targeted interventions to improve HIL in this population. This study also has several limitations. First, the cross- sectional design restricts causal inference and limits the ability to examine temporal relationships between HIL and self-management. Second, although a comprehensive questionnaire was used, differences in reference standards across countries may affect comparability and generalizabil- ity. Third, the sample may not fully represent the broader COPD population, particularly individuals with diverse demographic and cultural backgrounds. Fourth, self- reported HIL and self-management measures may intro- duce reporting bias and affect response accuracy. Finally, although random forest and LASSO regression are suitable for high-dimensional data, correlations among predictors may influence variable-importance estimates, and external validation is recommended to assess robustness. These con- siderations should be taken into account when interpret- ing the findings, which would benefit from confirmation in future longitudinal, multicenter studies.

  Identification of key influencing factors of health information literacy in COPD patients: a cross-sectional study using a random forest model · 2026 · DOI

• This study has several strengths. The fact that it was conducted in a socioeconomically disadvantaged urban area and focused on an under-researched and vulnerable population is a significant advan- tage. The relatively large sample size (N = 441) and the assessment of health literacy regarding climate change using a valid and reliable measurement tool enhance the reliability of the findings. However, there are also some limitations. Due to the cross-sectional design, causal relationships between variables cannot be determined. Macro- level determinants (industrial activities, energy consumption, the role of institutional actors) were excluded; examining these factors would require different datasets and methods. Since the study focuses solely on the Marmara Region and Istanbul, generalizations to the whole of Turkey are limited. Collecting data via telephone interviews may have limited the depth of responses and potentially created a bias toward non-response. Additionally, the fact that the study was conducted in a single district of Istanbul limits the generalizability of the findings to other regions. Finally, the Behavior subscale, consisting of two items, was excluded from the analyses due to its low internal consistency (Cronbach’s α = 0.26).

  Health literacy related to climate change: a communıty-based cross-sectional study · 2026 · DOI

• a dual “affective-cognitive” mechanism. youth sleep quality frequency under affects via Specifically, Hypothesis H1 is supported, there is a significant negative correlation between health information exposure frequency under algorithmic recommendation and youth sleep quality. This finding corroborates cognitive activation theory (17), indicating that under the “information seeking out users” paradigm, high-frequency, passive health information flow itself can constitute cognitive load. This directly disrupts young people’s mental tranquility and is correlated with the decline of sleep quality (6). Second, Hypotheses H2 and H3 hold true, this confirms the independent mediating effects of information anxiety and selfefficacy. The pathway H2 indicates that the health information exposure exhibits a negative correlation with youth sleep quality by triggering information anxiety. This pathway clearly illustrates the “affective pathway” from information overload to emotional distress, and ultimately to physiological sleep disorders. This perfectly aligns with Harvey’s (17) cognitive activation theory, which posits that anxiety triggered by information before bedtime leads to cognitive activation, thereby inhibiting the natural relaxation process required for sleep (18, 19). The H3 pathway indicates that the health information exposure frequency exhibits a negative correlation with youth sleep quality by weakening self-efficacy. This demonstrates the mechanism of the “cognitive pathway.” According to Social Cognitive Theory (4), when health information is overly complex or confusing, it can weaken young people’s confidence in their ability to manage their own health (selfefficacy) (24, 25). Individuals with lower self-efficacy are more likely to abandon efforts when facing sleep difficulties, leading to poorer sleep quality (27, 28). The core contribution of this study is to verify hypothesis H4, that is, information anxiety and self-efficacy constitute a continuous “emotion cognition” chain mediation path. This chain mechanism organically integrates cognitive activation theory (17) and social cognitive theory (4), and provides a more refined and dynamic perspective to explain how the algorithmic recommendation environment systematically affects sleep health of young people. Specifically, high-frequency exposure to algorithm-recommended health information is first positively associated with the level of information anxiety (an emotional response). Then, according to the attention control theory (32), this emotional response occupies substantial cognitive resources and is negatively associated with an individual’s self-efficacy, ultimately demonstrating a significant negative correlation with young people’s sleep quality. The discovery of the complete mediation effect of this path is more theoretically illuminating. One explanation for this finding is that health information contact itself is basically harmless to sleep quality.

  The impact of health information exposure frequency on youth sleep quality under algorithmic recommendation: a serial mediation role of information anxiety and self-efficacy · 2026 · DOI

• algorithmic self-efficacy (H3a), which in turn is negatively exposure diminishes correlated with youth sleep quality (H3b).

  The impact of health information exposure frequency on youth sleep quality under algorithmic recommendation: a serial mediation role of information anxiety and self-efficacy · 2026 · DOI

• This study has several limitations. First, all data were self- reported and may be influenced by recall and social desir- ability bias, particularly in estimating the frequency of You- Tube use and reliance before clinical procedures. Second, the sample was limited to Malaysian undergraduate dental stu- dents, which may restrict generalizability to other countries or educational systems. Third, the survey was distributed via student networks and social media, so the total number of students who received the invitation could not be deter- mined, and a precise response rate was unavailable; self- selection bias is therefore possible. Fourth, the study did not directly assess the quality, accuracy, or guideline concordance of the YouTube videos students used, nor did it systematically distinguish between institutional or professional association channels and non-institutional content. Finally, we did not include objective measures of clinical performance, so the impact of YouTube use on actual procedural competence could not be evaluated. Future research should address these limitations by combining learner surveys with video quality assessment and performance-based outcomes.

  YouTube As a Learning Tool for Clinical Procedures: A National Survey of Malaysian Dental Students · 2026 · DOI

• This study has several limitations. The patients were spread across a wide age range and, accordingly, a wide developmental range. This is not comparable to age dif- ferences among adult patients. The children’s develop- ment inevitably and consciously influences the results of the study. Further studies should examine specific age groups to determine causal relationships. The FL ques- tionnaire in our study was designed from scratch. As nearly 75% showed a high score, we conclude that the pic- tural food questions were too easy. The questionnaire is therefore not a suitable tool in this form and needs to be adapted. HLS-EU-Q16’s ceiling effect and self-report (vs. demonstrated skills) nature limit the granularity, particu- larly in samples with higher results; the small inadequate HL group (n = 20) in our study reduced power for subtle effects. However, continuous collinearity checks con- firmed robustness of null findings, but the results might favor detection of developmental patterns over subtle HL effects. Collinearity with education/age was minimal (Supplementary 12). Clinical outcomes were reported incompletely, as they constituted a secondary compo- nent of the emergency medicine specialists’ retrospective case reviews. Consequently, we did not evaluate delays in diagnosis, diagnostic errors, or adverse events. Although these endpoints are clinically important, they fall outside our primary research question. Our analysis focused on comparing subjective condition severity assessments across age and HL levels including their role for severe outcomes. Our data therefore cannot directly establish the causal effect of HL on emergency care outcomes. All subjective evaluations are inherently prone to bias. In our study, professional assessments were conducted by a dedicated expert team, which minimized biases such as confirmation bias and anchoring bias, although these biases cannot be eliminated. During retrospective case assessment, specialists operated independently and were blinded to contextual variables including arrival time and triage category. To ensure feasibility and comprehen- sive coverage, we enrolled only patients presenting from 10 AM to 8 PM. This sampling strategy may have intro- duced inclusion bias, as emergencies more common at night such as febrile convulsions were underrepresented. Other aspects, such as reduced resources or increased stress levels at night, may also have had an influence. Nevertheless, we consider the sample largely representa- tive: it encompasses a large time period and captures the full spectrum of clinically relevant conditions. Addition- ally, the demographic characteristics of our participants including gender, age, HL level, education, and timing of PED visit are consistent with earlier epidemiological surveys. Due to the non-inclusion of critically ill patients requiring immediate intervention and those who arrived by EMS, particularly severe cases are underrepresented overall. However, in order to maintain the feasibility of the study during routine operations, no other type of inclusion was possible. Ideally, subjective assessments of condition severity would have been obtained prior to triage to minimize procedural bias and unconscious influence from clini- cal staff. This was not feasible within the constraints of routine emergency department operations. However, the triage process in our PED does not routinely include dis- tinct feedback and the ESI category is not communicated by default. Steinke et al. BMC Emergency Medicine (2026) 26:99 Conclusion Parental HL did not affect accuracy or discrepancies regarding condition severity evaluations. Underestima- tion signaled higher risk, as can be expected. Children’s self-assessment of illness and, from around early school age on, anxiety predicted severe outcomes properly while pain did not. In multivariable models, nurses out- performed physicians, patients, and parents in predict- ing severe outcome. Children’s FL did not correlate with parental HL or severity. Age-appropriate use of childrens’ self-assessments could further support risk stratification in the PED.

  Greater parental health literacy does not improve assessment of their child’s emergency condition severity · 2026 · DOI

• This study has several limitations. First, although we conducted a cross-platform comparison of diabetes education texts generated by 10 LLMs using a multi-metric framework assessing quality, readability, and actionability, the evaluation was limited to English-language outputs. The findings may therefore not be generalisable to patient-education materials produced in other languages or applied in different cultural and health-literacy contexts.

  Challenges of using generative artificial intelligence for diabetes patient education: a cross-platform analysis of the quality, readability, and actionability of text generated by large language models · 2026 · DOI

• Purpose To gain a deeper understanding of users’ health information adoption and to promote the effectiveness of health information spread in the context of online limited information, this paper aims to examine how the information-motivation-behavioural (IMB) skills model can be used to organize online health information by experimenting how different IMB elements (information, motivation and behavioural skills) affect users’ intention to adopt health information.

  Tell me what to do first: how behavioural skills drive users’ intention to adopt online health information · 2024 · DOI

• BACKGROUND: The internet has become a common source of health information; however, little is known about online health information-seeking behaviour (HISB) among patients in low- and middle-income countries (LMICs).

  Online health information-seeking behaviour of patients attending a primary care clinic in Malaysia: a cross-sectional study · 2021 · DOI

Most-cited papers in Health Literacy and Information Accessibility

• Empirical Studies Assessing the Quality of Health Information for Consumers on the World Wide Web · JAMA · 2002 · 1,457 citations
• Association of Health Literacy With Diabetes Outcomes · JAMA · 2002 · 1,375 citations
• Health Literacy: Report of the Council on Scientific Affairs · JAMA · 1999 · 1,047 citations
• Health Information on the Internet · JAMA · 2001 · 911 citations
• Understanding and Responding to Health Literacy as a Social Determinant of Health · Annual Review of Public Health · 2020 · 904 citations
• Health Literacy and Child Health Outcomes: A Systematic Review of the Literature · PEDIATRICS · 2009 · 542 citations
• eLearning: A review of Internet-based continuing medical education · Journal of Continuing Education in the Health Professions · 2004 · 293 citations
• Quality of health information for consumers on the web: A systematic review of indicators, criteria, tools, and evaluation results · Journal of the Association for Information Science and Technology · 2015 · 265 citations
• Health Literacy and Women's Reproductive Health: A Systematic Review · Journal of Women s Health · 2016 · 225 citations
• Electronic health literacy and health-related outcomes among older adults: A systematic review · Preventive Medicine · 2022 · 170 citations

Most recent work

• Differences in online health information source utilization and a mediation model of eHealth literacy, health behaviors, physical well-being, and quality of life among older adults in Israel · Educational Gerontology · 2026
• Challenges of using generative artificial intelligence for diabetes patient education: a cross-platform analysis of the quality, readability, and actionability of text generated by large language models · Frontiers in Public Health · 2026
• Defining and measuring food literacy in late childhood and adolescence: A scoping review · Public Health Nutrition · 2026
• Evaluating ChatGPT-4 as an Educational Tool for Physical Activity Counseling in Patients with Hyperlipidemia · Balıkesır Health Sciences Journal · 2026
• Identification of key influencing factors of health information literacy in COPD patients: a cross-sectional study using a random forest model · BMC Pulmonary Medicine · 2026
• Greater parental health literacy does not improve assessment of their child’s emergency condition severity · BMC Emergency Medicine · 2026
• Scientific quality of colonoscopy-related YouTube videos in the Turkish language · Medicine · 2026
• Content, quality, and reliability of gout-related videos on TikTok and Bilibili: A cross-sectional study · Medicine · 2026
• Development and validation of an educational video for newly initiating peritoneal dialysis patients: from perioperative care to home-based management · Frontiers in Medicine · 2026
• YouTube As a Learning Tool for Clinical Procedures: A National Survey of Malaysian Dental Students · European Journal of General Dentistry · 2026

Related topics in Health Professions