Comprehensive Summary
This study, presented by Demir et al., examines the performance of ChatGPT-4o and Gemini Pro to generate PRISMA-A checklist compliant abstracts in order to improve high-level clinical evidence reporting. The authors processed 162 full-text articles published in Q1-ranked orthodontic journals with ChatGPT-4o and Gemini Pro using a PRISMA-A Checklist-aligned structured prompt, and the outputs were scored using a tailored Overall Quality Score (OQS) on a scale from 0 to 22. As a result, both LLMs yielded satisfactory OQS in the abstracts they generated, but ChatGPT-4o consistently achieved higher scores than Gemini Pro. ChatGPT-4o produced more complete and sufficient abstracts, achieving a mean OQS of 21.67 (SD 0.58) versus 21.00 (SD 0.71) for Gemini Pro. To conclude, the authors emphasize the importance of thorough review and verification of all generated content to ensure accuracy, scientific validity, and contextual appropriateness.
Outcomes and Implications
With the demonstrated potential of ChatGPT-4o and Gemini Pro to efficiently generate PRISMA-A-compliant abstracts from systematic reviews and meta-analyses, this provides a practical tool to streamline abstract writing and improve the circulation of high-level clinical evidence. Therefore, this research provides a foundation for AI-assisted tools that supports researchers in generating structured and standard-compliant scientific outputs. Future research should focus on expanding the application of LLMs to other areas of dentistry and broader medical fields outside of orthodontics.