Research Spotlight 07: Health Communication and Policy

Background: Lesbian, gay, and bisexual (LGB) adults in the U.S. often encounter systemic barriers in healthcare (e.g., discrimination, marginalization), which can contribute to lower trust in medical providers and the health care system. Poor patient-clinician communication (especially dismissive attitudes) may exacerbate these difficulties and discourage LGB individuals from seeking preventive care. While medical mistrust among LGB individuals has received increasing attention, limited research has explored how patient-centered communication relates to trust in both individual clinicians and the broader healthcare system. This study addresses that gap by examining these associations among a national sample of LGB adults.

Purpose: Using a subsample of LGB adults (n=418; weighted n=16,887,874) from the 2024 Health Information National Trends Survey (HINTS 7), we tested the pre-registered hypothesis that higher levels of patient-centered communication would be associated with a) higher trust in doctors and b) higher trust in the healthcare system.

Method: Multivariable linear regression models were conducted to test the study hypotheses. A composite score of 7 patient-centered communication items (e.g., “How often did health professionals help you deal with feelings of uncertainty about your health or health care?”) was entered as the predictor, and trust in doctors (“how much would you trust information about cancer from a doctor?”; 1-4 scale) as well as trust in the healthcare system (“how much do you trust the health care system (for example, hospitals, pharmacies, and other organizations involved in health care?”; 1-4 scale) were entered as the outcome variable in separate models, controlling for age, sex, insurance, and sexual orientation subgroup.

Results: Patient-centered communication was positively associated with greater trust in the healthcare system (b=0.44, SE=0.13, 95% CI [0.17,0.72]), but not significantly associated with trust in doctors (b=0.27, SE=0.15, 95% CI [-0.03, 0.58]). Individuals aged 65-74 (compared to 18-34) reported higher trust in the healthcare system (b=0.45, SE=0.17, 95% CI [0.11, 0.80]).

Conclusions: Among LGB adults, patient-centered communication was positively associated with trust in the healthcare system but not with trust in individual doctors. Future research and clinician training efforts should explore how inclusive, affirming communication may support trust-building across multiple levels of care.

Presenter - Adrian Casiano, Loyola Marymount University
Co-Author - Whitney Brymwitt, BA, MA, Loyola Marymount University
Co-Author - Timothy Williamson, PhD, MPH, Loyola Marymount University

The 2025 Myanmar earthquake unfolded amid conflict, repression, and widespread internet shutdowns, offering a lens on health communication under “polycrisis.” Most research on crisis communication has not fully examined the complex dynamics of how health information circulates when political, infrastructural, and digital disruptions converge. This gap is especially salient amid rapid advances in tools such as satellite internet and mesh networks, whose use in volatile settings remains poorly understood. We conducted semi-structured interviews with 24 stakeholders (humanitarian workers, health professionals, media actors, and community volunteers) involved in disseminating health information. Each interview used participatory mind-mapping to co-construct diffusion pathways, surfacing intermediaries, trust dynamics, and adaptations often missed in conventional qualitative work. Reflexive thematic analysis integrated transcripts, notes, and maps into a system-level account of information flows. Dissemination spanned three phases: assessment, planning, and delivery. Assessments relied on existing community networks; credibility rose when aid preceded data collection. Planning emphasized speed using pre-approved or adapted materials tailored to scarce resources. Delivery was strongly community-based: aid distribution paired with WASH and health education; mobile clinics combined treatment with health talks; local coordination mechanisms enhanced trust and efficiency. Barriers cut across censorship and surveillance, displacement, misinformation, and power and transport failures. Stakeholders used VPNs and encrypted apps to bypass censorship, leveraged Starlink satellite internet to sustain coordination during shutdowns, and piloted mesh-style networks; while radio, printed flyers, and religious gatherings remained indispensable for disconnected or wary communities. Digital tools enabled rapid coordination and updates but posed drawbacks: surveillance risks, constant platform switching, and exclusion of less-connected, lower-literacy groups who relied on trusted in-person intermediaries. Analyzing this case shows how health information systems adapt through hybrid digital and face-to-face approaches, pairing communication with tangible aid and rooting trust in local actors. As polycrises worldwide grow more complex and dangerous, resilience will hinge on balancing technological innovation with equity, trust, and contextual adaptation in health communication.

Author - Hein Thu, MPH, Digital Transformations for Health Lab
Author - Myat Min Thu, Department of Computer Engineering, Chiang Mai University
Presenter - Shahmir Ali, PhD, National University of Singapore

Background: Multicancer detection (MCD) screening tools can detect multiple cancers simultaneously and hold promise for addressing cancer screening disparities. Message framing research suggests that selective use of gain-framed (emphasizing benefits) and loss-framed (emphasizing costs) messages could alter receptivity to MCD screening. However, responses may differ according to race and gender, including when overlayed with culturally targeted messaging. Objective: This study examined effects of gain versus loss-framed messaging, with and without culturally targeted communication, on African Americans' and White Americans' receptivity to MCD testing. Methods: African American (female n=70, male n=47) and White American (female n=52, male n=73) adults 45-75 years old with no prior MCD use were recruited from Henry Ford Health to complete an online educational module about MCD testing. Participants were randomly assigned to view a gain-framed or loss-framed message. Half of African Americans received an additional culturally targeted message emphasizing personal control. Receptivity was measured using Theory of Planned Behavior constructs. Results: Between-race analyses revealed a three-way interaction (ethnicity × gender × message frame) for TPB attitudes (F(1,177)=3.71, p=.056). Among African American men, loss-framed messages produced higher receptivity (M=6.54, SD=0.90) than gain-framed messages (M=5.59, SD=1.54). Among White American men, gain-framed messages produced higher receptivity (M=6.33, SD=0.80) than loss-framed messages (M=5.95, SD=1.03). Effects were less pronounced among women. Within-race analyses among African Americans revealed a significant three-way interaction (gender × frame × culturally targeted messaging) for perceived behavioral control (F(1,109)=3.69, p=.057). Among African American men, culturally targeted messaging enhanced loss-framed effectiveness (M=5.77, SD=1.12 vs. M=5.55, SD=1.26) while reducing gain-framed effectiveness (M=5.57, SD=1.24 vs. M=5.95, SD=1.00). Effects were less pronounced among African American women. Conclusion: Loss-framed messaging appears more persuasive for African Americans, while gain-framed messaging remains persuasive for White Americans. Culturally targeted messaging may help mitigate negative receptivity among African American men, suggesting important gender differences. Findings highlight the need to consider race and gender when designing health messages about emerging screening technologies.

Author - Todd Lucas, PhD, Michigan State University
Co-Author - Christine Neslund-Dudas, PhD, Henry Ford Health
Co-Author - Anurag Dawadi, MPH, Michigan State University
Co-Author - Punith Shetty, Henry Ford Health
Co-Author - Olivia Aspiras, PhD, Michigan State University

BACKGROUND: Alcohol increases cancer risk, but population awareness of this association has historically been low. Recent communication strategies have attempted to increase awareness. Understanding how knowledge of the alcohol use-cancer risk association has changed over time is necessary to assess ongoing health education and communication needs to increase public awareness.

METHODS: Between 2017 and 2024, five cycles of the Health Information National Trends Survey, which uses complex survey sampling techniques for a nationally representative sample, included a question about the link between alcohol use and cancer risk. Using complex survey analysis techniques in STATA 19.5, we examined the proportion of accurate (belief that alcohol increases cancer risk), inaccurate (belief that alcohol has no effect or decreases cancer risk), and uncertain (“don’t know”) responses. We also examined the relation between educational attainment and accurate responses in each survey year.

RESULTS: The proportion of the US adult population reporting accurate knowledge of the link between alcohol use and cancer risk generally increased over time. In 2017, 26% of adults believed that alcohol use increased cancer risk, whereas in 2024 37% reported an accurate belief. At the same time, the proportion of adults reporting uncertainty also generally increased, from 36% in 2017 to 54% in 2024. Higher educational attainment was associated with accurate knowledge in each survey year. Consistent with the knowledge gap phenomenon -- that those with higher education are quicker to be exposed to and acquire new health knowledge -- the relation became generally stronger across survey years, from OR = 1.05 in 2017 to OR = 1.29 in 2024.

DISCUSSION: The observed changes in beliefs about the relation between alcohol consumption and cancer risk demonstrate a trend towards greater accurate knowledge over time. However, even in 2024 only about 1/3 of adults had accurate knowledge, highlighting the need for ongoing messaging. Moreover, the strengthening relation of educational attainment to risk knowledge highlights the important need for focused health education and health communication to eliminate the education-related knowledge gap in awareness of the cancer risk associated with alcohol use.

Presenter - Marc Kiviniemi, PhD, MBA, CPH, FSBM, University of Kentucky
Co-Author - Eleanor Hudd, PhD, University of Kentucky
Co-Author - Amy McQueen, PhD, FSBM, Washington University, School of Medicine
Co-Author - Heather Orom, PhD, University at Buffalo
Co-Author - Anne Ray, PhD, MEd, University of Kentucky
Co-Author - Erika Waters, PhD, MPH, FSBM, Washington University School of Medicine
Co-Author - Jennifer Hay, PhD, FSBM, Memorial Sloan Kettering Cancer Center

Abstract

Aim: Digital delivery of genetic healthcare services is increasingly used to expand access to genetic counseling/testing for individuals at risk of hereditary cancer syndromes. We synthesized evidence on their effects on (i)completion of pre-test counseling, (ii) uptake of genetic testing, (iii) knowledge.
Methods: The meta-analysis was performed using RevMan 5.4.1 (The Nordic Cochrane Center, Copenhagen, Denmark). Following PRISMA, we searched PubMed, EBSCO, Embase, Web of Science, PsycINFO, and Scopus (April–May 2025). Reviewers independently screened, extracted, and assessed risk of bias of trials. Random-effects meta-analyses estimated odds ratios (ORs) for categorical and standardized mean differences (SMDs) or mean differences for continuous outcomes; heterogeneity was quantified with I². Protocol registered in PROSPERO (CRD420251042666).
Results: Seventeen studies (13 RCT, 3 quasi-experimental, 1 cohort) including 6,714 participants (3,443 intervention; 3,271 control) published between 2001–2025 were analyzed. Pre-test counseling completion: Digital interventions significantly increased completion across four studies (OR 2.07, 95% CI 1.22–3.51; Z = 2.70; p = 0.0007). Genetic testing uptake: No overall difference was observed between digital and in-person care across seven studies (OR 1.30, 95% CI 0.80–2.10; p = 0.29; I² = 82%). Telehealth showed a nonsignificant trend toward higher uptake (OR 1.96, 95% CI 0.51–7.53; I² = 91%). Knowledge outcomes: Telehealth versus in-person counseling yielded comparable results (SMD −0.09, 95% CI −0.40 to 0.23; p = 0.59). Other digital formats suggested greater gains in some settings (SMD 1.19, 95% CI 0.01–2.37; p = 0.05), though pooled effects were not statistically significant and heterogeneity was substantial. Risk of bias was generally low or rated as “some concerns” in randomized trials; nonrandomized studies were of moderate quality.
Conclusions: Digital genetic counseling is generally non-inferior to in-person care for knowledge and psychosocial outcomes, and shows roughly equivalent effectiveness for genetic test uptake. Importantly, digital approaches doubled the odds of completing pre-test counseling. The greatest benefits appear linked to implementation features that reduce access barriers. To refine best-practice, hybrid models of care, future research should prioritize standardized outcome measures and direct head-to-head comparisons of telehealth, web-based, and AI/chatbot modalities.

Author - Maide Nur Tumkaya Tumkaya, RN, PhD, , Koc University
Author - Seyma Inciser Pasalak, RN. PhD, Independent Researcher
Presenter - Memnun Seven, PhD, RN, University of Massachusetts Amherst

Sustained adherence to annual lung cancer screening extends life expectancy, yet fewer than 20% of eligible individuals are screened. Shared Decision Making (SDM) plays a critical role in supporting adherence. Fewer than 10% of eligible patients report SDM participation, often describing it as inadequate. We developed LungScreenSDM, an AI-powered conversational agent that engages patients in key SDM stages and facilitates informed decisions with providers.
Methods: We utilized a retrieval-augmented generation model to generate context-aware responses based on a pre-assembled knowledge base comprising 601 open-source PubMed articles and other trusted sources (e.g. NIH). The JSON-formatted knowledge base embedding was generated using the Hugging Face BAAI BGE v 1.5 embedding model and transformed into a searchable index with Meta’s Facebook AI Similarity Search. The indexed knowledge base and instructional prompt were used to initialize an interaction with ChatGPT-4o, which was logged in the Langfuse observability platform to track performance, accuracy, and the occurrence of hallucinations. Prompt engineering ensured alignment with SDM stages, an empathetic tone, and 6th-grade readability. Evaluation was conducted in Langfuse using a large language model as a judge. AI was used for clarity and grammar.
Results: LungScreenSDM guided patients through the key SDM stages: introducing the screening decision, eliciting preferences, explaining risks and benefits, and concluding with a conversation summary, along with questions to ask a provider. Additional functionalities included access to source materials, recommended follow-up questions during the dialogue, and an eligibility check for screening, with consistent prompts to consult a healthcare provider. Across 113 test interactions, hallucination was observed in fewer than 3% of responses, and no toxic language was detected. The LungScreenSDM achieved a factual accuracy score of 0.96 and a helpfulness score of 0.8 (scale 0-1). Precision and context relevance were comparatively lower, reflecting the intentional design choice to nudge users through SDM stages. To address this, a structured human review will be implemented in the future.
Conclusion: Implemented in a HIPAA-compliant environment, LungScreenSDM will be offered to individuals who may be eligible for screening before they meet their healthcare provider. Currently, we plan to expand the program's reach to increase the number of annual screenings.

Presenter - Ilona Fridman, Hackensack Meridian Health & Georgetown University
Co-Author - Adil Alaoui, MS, Georgetown University
Co-Author - Subha Jamal, MS, Georgetown University
Author - Samantha Zocher, MS, Georgetown University
Author - Matthew McCoy, PhD, Georgetown University
Author - Camelia Bencheqroun, MS, Georgetown University
Author - Jia Li Dong, MS, Georgetown University
Author - Amina Eman Shorish, BS, Georgetown University
Author - Nawar Shara, PhD, MedStar Health Research Institute
Author - Lisa Carter-Bawa, PhD, MPH, FAAN, FSBM, Center for Discovery & Innovation at Hackensack Meridian Health