<oai_dc:dc xmlns:oai_dc="http://www.openarchives.org/OAI/2.0/oai_dc/" xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="http://www.openarchives.org/OAI/2.0/oai_dc/ http://www.openarchives.org/OAI/2.0/oai_dc.xsd"><dc:title>Balancing Readability and Quality: Using Artificial Intelligence as a Tool to Improve Patient Education in Neurodevelopmental Genetics</dc:title><dc:creator>Ratnasamy, Maria </dc:creator><dc:subject>Neurodevelopmental</dc:subject><dc:subject>Genetics</dc:subject><dc:coverage>Biochemistry and Molecular Biology</dc:coverage><dc:relation>B S</dc:relation><dc:description>Declining patient health education and literacy present a growing public health crisis within the United States. With low health literacy directly correlated to poorer socioeconomic outcomes, it is critical that patients have access to accurate, understandable patient education materials (PEMs); however, many continue to be written at a reading level above suggested national standards. To combat this information gap, studies have begun to evaluate the ability of large language models (LLMs) to generate PEMs with improved readability for lay audiences. While many studies have shown that LLMs can effectively improve readability scores, few have commented on the quality and factual accuracy of rewritten PEMs. Additionally, few studies have analyzed LLMs’ ability to improve readability scores of educational materials focused on neurodevelopmental genetics. In this study, three major LLMs (ChatGPT, Gemini, and Claude) were tested on their ability to improve the readability of various neurodevelopmental genetics abstracts. With standardized prompting, abstracts were rewritten by each LLM and subsequently evaluated for improved readability and linguistic quality using the Flesch-Kincaid Grade Level and an established quality rubric, respectively. Results showed that all tested LLMs produced high-quality rewrites of improved readability, with Gemini producing the greatest proportion of high-quality abstracts that fell within target readability scores tailored to a lay audience. This study ultimately establishes a foundation for the use of LLMs in the development of neurodevelopmental genetics PEMs, paving the way for greater patient autonomy and improved healthcare outcomes.  </dc:description><dc:contributor>Santhosh Girirajan, Thesis Supervisor</dc:contributor><dc:contributor>Lorraine C Santy, Thesis Honors Advisor</dc:contributor><dc:rights>open_access</dc:rights><dc:date>2026-03-18T20:47:16Z</dc:date><dc:identifier>https://honors.libraries.psu.edu/catalog/9946mar7083</dc:identifier></oai_dc:dc>