<< Back to the abstract archive
Educating Artificial Intelligence in Plastic Surgery: A Systematic Review on the Status of Demographic Reporting in Plastic Surgery AI Research
Alexis Henderson, MPH; Nicolás M. Kass, BA ;Onyedi Moses, MBBS; Tobi J. Somorin BS; Angel M. Dixon, BA; Janina Kueper, MD; Ashley Rogers, MD; Jesse A. Goldstein, MD, FAAP, FACS
University of Pittsburgh Medical Center, Children's Hospital of Pittsburgh
2025-01-09
Presenter: Alexis Henderson
Affidavit:
The entirety of this submission is the original work of the research team.
Director Name: Jesse A. Goldstein
Author Category: Medical Student
Presentation Category: Clinical
Abstract Category: General Reconstruction
Background: Artificial Intelligence (AI) is becoming increasingly integrated into plastic surgery research and clinical practice. Effective implementation requires comprehensive demographic data to ensure models address diverse patient populations, promoting equitable, personalized care. This study evaluates the current landscape of demographic reporting in plastic surgery AI literature.
Methods: A systematic search of plastic surgery literature from the last 10 years was performed using the PubMed database using PRISMA guidelines. The search was executed with the following keywords: ((Artificial Intelligence) OR (Machine learning)) AND ((Plastic Surgery) OR (Reconstructive Surgery)). Exclusion criteria included articles with non-human subjects, cadaveric studies, unretrievable full-texts, non-English language, letters, commentaries, and reviews.
Results: Initial results yielded a total of 1,193 articles published from 2018-2024. Of these, 146 articles met the inclusion criteria. 50.7% of the included articles were published from 2023-2024. Amongst all included articles, age was the most reported demographic variable (62.3%), and SES/ Education was the least reported variable (2.7%). 18.5% of plastic surgery AI literature included the reporting of race/ethnicity in their demographic variables.
Conclusions: Race and ethnicity are heavily underreported in plastic surgery AI research. The advent of AI requires due diligence from healthcare professionals to prevent bias in models, as unchecked algorithms may worsen health disparities between groups. We recommend disclosing demographic data used in clinical AI and machine learning models to ensure the technology serves diverse populations equitably. Transparent data reporting enables plastic surgeons to improve prediction accuracy and treatment for all patient groups, fostering more inclusive and fair medical practices.
