Amna A Othman, Kendall A Flaharty, Suzanna E Ledgister Hanchard, Ping Hu, Dat Duong, Rebekah L Waikel, Benjamin D Solomon
{"title":"Assessing large language model performance related to aging in genetic conditions.","authors":"Amna A Othman, Kendall A Flaharty, Suzanna E Ledgister Hanchard, Ping Hu, Dat Duong, Rebekah L Waikel, Benjamin D Solomon","doi":"10.1038/s41514-025-00226-z","DOIUrl":null,"url":null,"abstract":"<p><p>Most genetic conditions are described in pediatric populations, leaving a gap in understanding their clinical progression and management in adulthood. Motivated by other applications of large language models (LLMs), we evaluated whether Llama-2-70b-chat (70b) and GPT-3.5 (GPT) could generate plausible medical vignettes, patient-geneticist dialogues and management plans for a hypothetical child and adult patients across 282 genetic conditions (selected by prevalence and categorized based on age-related characteristics). Results showed that LLMs provided appropriate age-based responses in both child and adult outputs based on Correctness and Completeness scores graded by clinicians. Sub-analysis of metabolic conditions including those typically presents neonatally with crisis also showed age-appropriate LLM responses. However 70b and GPT obtained low Correctness and Completeness scores at producing plausible management plans (55-66% for 70b and a wider range, 50-90%, for GPT). This suggests that LLMs still have some limitations in clinical applications.</p>","PeriodicalId":94160,"journal":{"name":"npj aging","volume":"11 1","pages":"33"},"PeriodicalIF":4.1000,"publicationDate":"2025-05-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12049513/pdf/","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"npj aging","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1038/s41514-025-00226-z","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q2","JCRName":"GERIATRICS & GERONTOLOGY","Score":null,"Total":0}
引用次数: 0
Abstract
Most genetic conditions are described in pediatric populations, leaving a gap in understanding their clinical progression and management in adulthood. Motivated by other applications of large language models (LLMs), we evaluated whether Llama-2-70b-chat (70b) and GPT-3.5 (GPT) could generate plausible medical vignettes, patient-geneticist dialogues and management plans for a hypothetical child and adult patients across 282 genetic conditions (selected by prevalence and categorized based on age-related characteristics). Results showed that LLMs provided appropriate age-based responses in both child and adult outputs based on Correctness and Completeness scores graded by clinicians. Sub-analysis of metabolic conditions including those typically presents neonatally with crisis also showed age-appropriate LLM responses. However 70b and GPT obtained low Correctness and Completeness scores at producing plausible management plans (55-66% for 70b and a wider range, 50-90%, for GPT). This suggests that LLMs still have some limitations in clinical applications.