P241: A Human+AI Process for Generating Fully-Annotated, Plain-Language Trial Summaries from Source Materials in Under 24 Hours
Poster Presenter
Walter Bender
Chief Scientific Officer
Sorcero, Inc. United States
Objectives
Timely dissemination of plain-language trial summaries is both an economic and ethical imperative[1,2]; consequently we have developed a bespoke Human+AI collaborative process engineered for the rapid generation of scientifically robust, regulatory-compliant, patient-accessible PLTS.
Method
The input is a clinical study report. The process combines a generative-AI framework and industry guidelines to create a source-verified output. An annotated draft plain-language trial summary (PLTS) is provided to medical writers for revision before medical, legal, and regulatory review.
Results
The deployed process combined a RAG framework[3], which grounds all AI-generated output in verified source documents, with a proprietary definitive AI scoring rubric to ensure data fidelity, elimination of hallucinations, and adherence to industry and regulatory guidelines, including (EU) No 536/2014. A core component of the process was a mandatory review by a subject-matter expert (SME) that ensures scientific integrity before the draft is delivered to the study team. The process has been applied repeatedly (N=7) to deliver a fully formed draft PLTS with source-traced data points, full tables, and design explanations for figures. These high-quality drafts were generated in 8 to 24 hours. The medical writing team reported that trust was enhanced because the process reproduces results consistently and includes source traceability, immutable audit trails, transparency and explainability, promotional-language checks, reading-level metrics, and benchmarking against existing processes, which takes weeks rather than days.
References:
1. Regulation (EU) No 536/2014 of the European Parliament and of the Council of 16 April 2014 on clinical trials in medicinal products for human use and repealing Directive 2001/20/EC. Off J Eur Union. 2014;158
2. Penlington M, Silverman H, et al. Plain Language Summaries of Clinical Trial Results: A Preliminary Study to Assess Availability of Easy-to-Understand Summaries and Approaches to Improving Public Engagement. Pharmaceut Med. 2020;34(6):401-406.
3. Abo El-Enen, M., Saad, S. & Nazmy, T. A survey on retrieval-augmentation generation (RAG) models for healthcare applications. Neural Comput & Applic 37, 28191–28267 (2025).
4. Meppelink CS, Smit EG, et al. Health Literacy and Online Health Information Processing: Unraveling the Underlying Mechanisms. J Health Commun. 2016;21(sup2):109-120.
5. Lee H, Jin S, Henning-Smith C, et al. Role of Health Literacy in Health-Related Information-Seeking Behavior Online: Cross-section
Conclusion
This capability directly addresses the challenge of timely dissemination of clinical study results to study participants by accelerating the high-quality generation of these critical publications[4,5]. To formally quantify its impact on time-to-publish and quality, a rigorous comparative validation against traditional human-only and generic LLM methods is currently underway, with initial results expected in early 2026.
