Effects of education level on natural language processing in cardiovascular health communication.
Saved in:
| Title: | Effects of education level on natural language processing in cardiovascular health communication. |
|---|---|
| Authors: | Joseph S; Medical College of Georgia, Augusta University, Augusta, GA, United States., Bhardwaj A; Medical College of Georgia, Augusta University, Augusta, GA, United States., Skariah J; Medical College of Georgia, Augusta University, Augusta, GA, United States., Aggarwal I; Medical College of Georgia, Augusta University, Augusta, GA, United States., Shah V; Case Western Reserve University School of Medicine, Cleveland, OH, United States., Harris RA; Medical College of Georgia, Augusta University, Augusta, GA, United States.; Georgia Prevention Institute, Augusta University, Augusta, GA, United States. |
| Source: | Frontiers in public health [Front Public Health] 2025 Nov 13; Vol. 13, pp. 1688173. Date of Electronic Publication: 2025 Nov 13 (Print Publication: 2025). |
| Publication Type: | Journal Article |
| Language: | English |
| Journal Info: | Publisher: Frontiers Editorial Office Country of Publication: Switzerland NLM ID: 101616579 Publication Model: eCollection Cited Medium: Internet ISSN: 2296-2565 (Electronic) Linking ISSN: 22962565 NLM ISO Abbreviation: Front Public Health Subsets: MEDLINE |
| Imprint Name(s): | Original Publication: Lausanne : Frontiers Editorial Office |
| MeSH Terms: | Cardiovascular Diseases* , Natural Language Processing* , Educational Status* , Health Communication*/methods , Health Literacy*, Humans ; Comprehension |
| Abstract: | Competing Interests: The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. Introduction: Cardiovascular disease (CVD) remains the leading cause of mortality worldwide, underscoring the importance of accessible health communication. Artificial intelligence (AI) tools such as ChatGPT and MediSearch have potential to bridge knowledge gaps, but their effectiveness depends on both accuracy and readability. This study evaluated how natural language processing (NLP) models respond to CVD-related questions across different education levels. Methods: Thirty-five frequently asked questions from reputable sources were reformatted into prompts representing lower secondary, higher secondary, and college graduate levels, and entered into ChatGPT Free (GPT-4o mini), ChatGPT Premium (GPT-4o), and MediSearch (v1.1.4). Readability was assessed using Flesch-Kincaid Ease and Grade Level scores, and response similarity was evaluated with BERT-based cosine similarity. Statistical analyses included ANOVA, Kruskal-Wallis, and Pearson correlation. Results: Readability decreased significantly with increasing education level across all models ( p < 0.001). ChatGPT Free responses were more readable than MediSearch ( p < 0.001), while ChatGPT Free and Premium demonstrated higher similarity to each other than to MediSearch. ChatGPT Premium explained the greatest variance in readability ( r = 0.350; p < 0.001), suggesting stronger adaptability to user education levels compared to ChatGPT Free ( r = 0.530; p < 0.001) and MediSearch ( r = 0.227; p < 0.001). Discussion: These findings indicate that while NLP models adjust readability by education level, output complexity often exceeds average literacy, highlighting the need for refinement to optimize AI-driven patient education. (Copyright © 2025 Joseph, Bhardwaj, Skariah, Aggarwal, Shah and Harris.) |
| References: | Arthrosc Sports Med Rehabil. 2024 Oct 10;7(1):101027. (PMID: 40041831) Resuscitation. 2024 Jan;194:110077. (PMID: 38081504) Turk Kardiyol Dern Ars. 2025 Jan;53(1):35-43. (PMID: 39797456) J Med Internet Res. 2024 Aug 14;26:e55138. (PMID: 39141910) Int J Med Inform. 2024 Oct;190:105562. (PMID: 39059084) Clin Transl Gastroenterol. 2024 Nov 01;15(11):e00765. (PMID: 39212302) Cureus. 2024 May 8;16(5):e59898. (PMID: 38721479) Curr Probl Cardiol. 2024 Nov;49(11):102797. (PMID: 39159709) Cureus. 2024 Jun 4;16(6):e61680. (PMID: 38841294) |
| Contributed Indexing: | Keywords: artificial intelligence; cardiovascular disease; health communication; large language models; natural language processing; patient education; readability |
| Entry Date(s): | Date Created: 20251201 Date Completed: 20251201 Latest Revision: 20251203 |
| Update Code: | 20251203 |
| PubMed Central ID: | PMC12657148 |
| DOI: | 10.3389/fpubh.2025.1688173 |
| PMID: | 41323602 |
| Database: | MEDLINE |
Full Text 
Full Text Finder 
Nájsť tento článok vo Web of Science | Abstract: | Competing Interests: The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.<br />Introduction: Cardiovascular disease (CVD) remains the leading cause of mortality worldwide, underscoring the importance of accessible health communication. Artificial intelligence (AI) tools such as ChatGPT and MediSearch have potential to bridge knowledge gaps, but their effectiveness depends on both accuracy and readability. This study evaluated how natural language processing (NLP) models respond to CVD-related questions across different education levels.<br />Methods: Thirty-five frequently asked questions from reputable sources were reformatted into prompts representing lower secondary, higher secondary, and college graduate levels, and entered into ChatGPT Free (GPT-4o mini), ChatGPT Premium (GPT-4o), and MediSearch (v1.1.4). Readability was assessed using Flesch-Kincaid Ease and Grade Level scores, and response similarity was evaluated with BERT-based cosine similarity. Statistical analyses included ANOVA, Kruskal-Wallis, and Pearson correlation.<br />Results: Readability decreased significantly with increasing education level across all models ( p < 0.001). ChatGPT Free responses were more readable than MediSearch ( p < 0.001), while ChatGPT Free and Premium demonstrated higher similarity to each other than to MediSearch. ChatGPT Premium explained the greatest variance in readability ( r = 0.350; p < 0.001), suggesting stronger adaptability to user education levels compared to ChatGPT Free ( r = 0.530; p < 0.001) and MediSearch ( r = 0.227; p < 0.001).<br />Discussion: These findings indicate that while NLP models adjust readability by education level, output complexity often exceeds average literacy, highlighting the need for refinement to optimize AI-driven patient education.<br /> (Copyright © 2025 Joseph, Bhardwaj, Skariah, Aggarwal, Shah and Harris.) |
|---|---|
| ISSN: | 2296-2565 |
| DOI: | 10.3389/fpubh.2025.1688173 |