Using AI to Boost Genetic Literacy and Health Equity in Reproduction

Published Date : 23-07-2025 

A comparative study explores the potential of AI and large language models (LLMs) to simplify patient education materials (PEMs) in reproductive genetics, enhancing health literacy and equity. 

The integration of artificial intelligence (AI) and large language models (LLMs) in healthcare is gaining traction, particularly in the field of patient education. A recent study from the Department of Clinical Genetics at Maastricht University Medical Centre (MUMC+) and the Department of Genetics and Cell Biology has shown promising results in using AI to simplify patient education materials (PEMs) in reproductive genetics. This study aims to address the complexities of genetic testing and counseling, which often hinder patient understanding and informed decision-making.

The study, led by Marjan Naghdi and colleagues, sought to evaluate the effectiveness of AI and LLMs in simplifying PEMs to advance reproductive genetic literacy and health equity. The researchers hypothesized that AI could help healthcare professionals generate more accessible and understandable materials for patients, thereby reducing the burden of genetic disorders and promoting better health outcomes.

Reproductive genetic testing and counseling are crucial for personalized healthcare, but the complexity of these tests and the way PEMs are designed often limit their uptake. Patients, especially those in vulnerable situations, may either overuse or underuse genetic testing technologies due to a lack of understanding. This study aimed to address this gap by leveraging AI to simplify complex medical information.

The study design was comprehensive, involving a comparative observational approach. The researchers evaluated the capacity of four AI/LLMs—GPT-3.5, GPT-4, Copilot, and Gemini—to simplify 30 PEMs covering six topics in reproductive genetics. These PEMs were sourced from well-recognized platforms such as the World Health Organization (WHO), MedlinePlus, and Johns Hopkins. Each PEM was processed by the four AI/LLMs using a fixed prompt, resulting in 120 simplified outputs.

To assess the effectiveness of the simplified PEMs, the researchers used five validated readability metrics, including the Simple Measure of Gobbledygook (SMOG), to measure improvements in readability. Additionally, a panel of 30 experts in reproductive genetics independently evaluated the clinical accuracy and completeness of the simplified texts.

The results were significant. All four LLMs significantly improved the readability of the PEMs, reducing the text complexity to an average 6th-7th grade reading level. Gemini and Copilot achieved the highest improvements in readability scores, while GPT-4 received the highest expert ratings across all criteria—accuracy (4.1 ± 0.9), completeness (4.2 ± 0.8), and relevance of omissions (4.0 ± 0.9; P < 10-8). These findings highlight the importance of balancing readability with content integrity to ensure that essential medical information is not compromised.

To support the integration of AI-assisted approaches in clinical practice, the researchers developed an open-access graphical user interface (GUI) that provides real-time PEM simplification and readability analysis. This tool is available at https://huggingface.co/spaces/CellularGenomicMedicine/HealthLiteracyEvaluator and can be a valuable resource for healthcare professionals.

However, the study also highlights the need for caution. While AI/LLMs can significantly improve readability, careful evaluation is necessary to ensure that critical information is not omitted. The researchers emphasize the importance of real-world patient feedback to fully assess the potential of these tools in advancing reproductive genetic literacy and promoting health equity.

The study was funded by several grants, including the EVA specialty program (KP111513) of MUMC+, the Horizon-Europe (NESTOR-101120075), the Estonian Research Council (PRG1076), the Horizon-2020 innovation (ERIN-EU952516) grants of the European Commission, the Swedish Research Council (grant no. 2024-02530), and the Novo Nordisk Foundation (grant no. NNF24OC0092384). The authors declare no conflict of interest relevant to this study.

In conclusion, the integration of AI and LLMs into patient education strategies holds great promise for advancing health equity by improving understanding and facilitating informed decision-making. By making complex medical information more accessible, patients can engage more effectively in reproductive genetic testing programs, leading to better health outcomes. 

Frequently Asked Questions (FAQS): 

Q: What is the main goal of using AI in patient education materials?

A: The main goal is to simplify complex medical information to improve patient understanding and facilitate informed decision-making, particularly in reproductive genetic testing and counseling.

Q: Which AI models were used in the study?

A: The study used four AI models: GPT-3.5, GPT-4, Copilot, and Gemini.

Q: How were the PEMs evaluated for clinical accuracy?

A: A panel of 30 experts in reproductive genetics independently scored each simplified PEM for accuracy, completeness, and relevance of omissions.

Q: What is the significance of the open-access GUI developed by the researchers?

A: The GUI provides real-time PEM simplification and readability analysis, supporting healthcare professionals in generating effective and accessible patient education materials.

Q: What are the potential limitations of using AI to simplify PEMs?

A: The potential limitations include the risk of omitting critical information and the need for careful evaluation to ensure clinical accuracy and completeness. 

More Related Topics :