Simplifying healthcare communication: Evaluating AI-driven plain language editing of informed consent forms

This repository contains the supplementary materials for the study "Simplifying healthcare communication: Evaluating AI-driven plain language editing of informed consent forms", presented at the [AI4PL Workshop of the MT Summit Conference, 2025].

The study investigates how generative AI can be used to simplify cancer-related informed consent forms (ICFs) using Plain Language (PL) strategies. Two prompt engineering approaches were tested—Simple AI Edit and Complex AI Edit—and their output was evaluated using standard readability metrics.

The materials provided allow full replication of the study and support further research on readability, health literacy, and AI for accessible communication.

Contents:
📁 Corpus - Informed Consent Forms.zip:

• Original ICFs (in TXT format)

• AI-edited versions using the Simple AI Edit and Complex AI Edit strategies

📄 data_analysis.xlsx:

• Readability scores (Flesch Reading Ease, Gunning Fog Index, and SMOG Index) for each version of each ICF

📓 AI4PL_Paper_in_MT_summit.ipynb:

• Google Colab / Jupyter Notebook to reproduce the readability evaluation and generate visualisations

• The notebook is accessible here: https://colab.research.google.com/drive/1YMVUk7yDbTMjea3eCseUacp6EWI5MbHn?usp=sharing 

How to cite:
Briva-Iglesias, V., & Peñuelas Gil, I. (2025). Simplifying healthcare communication: Evaluating AI-driven plain language editing of informed consent forms.  MT Summit 2025.