Related Experiment Video
Updated: May 10, 2025
06:45
Development of a Benchtop Model for Evaluating the Compatibility of Wound Dressing Materials with Negative Pressure Wound Therapy Systems
Published on: May 2, 2025
94
Artificial Intelligence Versus Human Systematic Literature Review Into Negative-pressure Wound Therapy in Plastic
Augustine J Deering1, Payden A Harrah1, Melinda Lue2
1From the Long School of Medicine, University of Texas Health Science Center San Antonio, San Antonio, TX.
Plastic and Reconstructive Surgery. Global Open
|April 21, 2025
Summary
Artificial intelligence (AI) shows promise for systematic reviews but struggles with evidence quality assessment. Human investigators generally assigned higher evidence quality scores than AI in this study on negative-pressure wound therapy (NPWT).
Area of Science:
- Medical Informatics
- Evidence-Based Medicine
- Wound Care
Background:
- Artificial intelligence (AI) has significant potential to aid physicians in evidence-based medicine.
- Systematic reviews are crucial for synthesizing evidence, especially in complex fields like negative-pressure wound therapy (NPWT).
- Challenges in trial design for NPWT have historically made high-level evidence synthesis difficult.
Purpose of the Study:
- To compare the ability of AI and human investigators to perform a systematic literature review.
- To evaluate the efficacy of NPWT and AI's capacity for assessing evidence quality in this domain.
Main Methods:
- A systematic literature search was performed across PubMed, SCOPUS, and CINAHL.
- Articles were screened using Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) guidelines.
- The Grading of Recommendations, Assessment, Development, and Evaluation (GRADE) criteria were used by both AI and humans to assess evidence quality.
Main Results:
- Eighteen studies involving 3131 patients were reviewed.
- NPWT was associated with shorter hospital stays in 5 of 7 studies, improved infection rates in 8 of 14 studies, and reduced wound closure time in 9 of 12 studies.
- AI consistently assigned lower quality of evidence scores compared to human reviewers.
Conclusions:
- AI is a valuable tool for systematic reviews but currently has limitations in accurately determining evidence quality.
- AI's lower scores may indicate reduced bias, but confounders in NPWT research limit high-level evidence.
- Further development is needed for AI to reliably assess evidence quality in complex medical fields.
