AUTHORS
Kristen Wheeden, PhD, MBA, Ronak Mistry, DO2, Jori May, MD3, Ariela Marshall, MD4, Richard C. Godby, MD5
AFFILIATIONS
1 United Porphyrias Association, Bethesda, Maryland, United States
2 University of Pennsylvania, Philadelphia, Pennsylvania, United States
3 University of Alabama at Birmingham, Birmingham, Alabama, United States
4 University of Minnesota, Minneapolis, Minnesota, United States
5 Mayo Clinic, Rochester, Minnesota, United States
ABSTRACT
Introduction: As artificial intelligence (AI) and digital health tools rapidly enter the healthcare landscape, understanding patient adoption behaviors is critical for optimizing care delivery and improving health literacy. Patients managing complex hematologic conditions may struggle to identify clinicians with expertise and frequently navigate dense clinical information, complex test results, and nuanced treatments. This project aims to evaluate how patients with porphyria and their care partners utilize AI. Furthermore, it seeks to understand patient trust in these digital technologies and assess comfort with clinical care team use of AI.
Methods: A voluntary, anonymous, cross-sectional survey was made available to patients diagnosed with porphyria and their care partners through the United Porphyrias Association for approximately one month (June-July 2026). It was made available through social media platforms and email listservs. Domains assessed included baseline demographic characteristics, educational attainment, clinical treatment status, and baseline health literacy. Descriptive statistical analysis was performed to establish baseline cohort characteristics. Open-ended responses regarding patients' greatest hopes and fears were qualitatively analyzed to capture perspectives.
Results: There were 113 survey responses, predominantly patients with porphyria (n=85, 75.2%) and their care partners (n=25, 22.1%). Responses captured a broad spectrum of the clinical journey, including treatment-naïve individuals (n=31, 27.4%), patients currently on medication (n=39, 34.5%), and clinical trial participants (n=23, 20.4%). The majority of respondents were 35-54 years old (n=61, 53.9%) and held a bachelor’s degree or higher (n=70, 61.9%). Most respondents reported familiarity with AI and endorsed active, self-directed usage of AI (n=85, 75.2%).
AI can function as an asynchronous educational bridge for vulnerable patient populations. Respondents who indicated they "Often" needed assistance understanding complex medical information (N=9) utilized AI to independently interpret diagnostic test results at a substantially higher rate (n=7, 77.8%) compared with those who "Never" required such assistance (N=50) and reported lower rates of AI use for diagnostic interpretation (n=13, 26.0%).
Regarding systems-based practice and clinical workflow optimization, respondents demonstrated readiness for AI integration into clinical care. Patients in the 35-44 year old group (N=30) reported comfort with AI ambient scribes "listening" to clinical encounters to autonomously generate medical notes (n=22, 73.3%), higher than for other age groups. However, there was a divergence between utilization and trust. Even among the most frequent users of AI (N=44), less than half (n=19, 43.2%) explicitly trusted the health information provided by general, non-customized AI models.
Qualitative analysis of open-ended responses revealed that patient fears heavily center on AI "hallucinations," clinical inaccuracies, and data privacy vulnerabilities. Conversely, patient hopes focus on enhanced diagnostic speed, precision in complex clinical scenarios, and the ability of AI to translate dense medical literature into accessible, multimedia-friendly educational formats.
Conclusions: Patients with porphyria and their care partners are actively and independently adopting AI as an asynchronous educational tool, leveraging the technology to translate complex laboratory results and clinical diagnoses. This is particularly prevalent among those facing baseline health literacy barriers, highlighting a critical and unprompted shift in the digital era of healthcare delivery. Concurrently, patients report comfort with use of AI by clinical care teams. However, the persistent and significant gap between high AI utilization and low intrinsic trust underscores an urgent, unmet need. Healthcare institutions and clinicians must move beyond passive observation and actively develop patient-facing guidelines for safe AI interaction, while maintaining care team oversight of clinical AI tools. Modernizing both patient and provider education to include AI literacy is essential for the future of healthcare delivery.
DISCLOSURES
Kristen Wheeden, PhD, MBA: No relevant conflicts of interest to declare.