Mapping the use of artificial intelligence for skin injury assessment and care in hospitalized patients: A scoping review

Ângelo Antônio  Paulino Martins Zanetti; Denise  Desconsi; Rafaella Manhoni Lima De  Miranda; Luisa Brolacci  Lana; Lucas Daniel Del Rosso  Calache; Silvana Andréa Molina  Lima; Clarita Terra Rodrigues  Serafim

doi:10.18502/npt.v13i1.20595

Ângelo Antônio Paulino Martins Zanetti Department of Nursing, Medical School, São Paulo State University, Botucatu, São Paulo, Brazil
Denise Desconsi Department of Nursing, Medical School, São Paulo State University, Botucatu, São Paulo, Brazil
Rafaella Manhoni Lima De Miranda Department of Nursing, Medical School, São Paulo State University, Botucatu, São Paulo, Brazil
Luisa Brolacci Lana Department of Aeronautical Engineering, School of Engineering, São Paulo State University, São João da Boa Vista, Brazil
Lucas Daniel Del Rosso Calache Department of Aeronautical Engineering, School of Engineering, São Paulo State University, São João da Boa Vista, Brazil
Silvana Andréa Molina Lima Department of Nursing, Medical School, São Paulo State University, Botucatu, São Paulo, Brazil
Clarita Terra Rodrigues Serafim Department of Nursing, Medical School, São Paulo State University, Botucatu, São Paulo, Brazil

DOI: https://doi.org/10.18502/npt.v13i1.20595

Keywords: skin; wounds and injuries; artificial intelligence; machine learning; digital health

Abstract

Background & Aim: Skin injuries are frequent hospital complications, and the role of artificial intelligence in management remains unclear. This review aimed to identify, map, and analyze the evidence on the use of artificial intelligence in the assessment, monitoring, and management of skin injuries in hospitalized patients worldwide.

Methods & Materials: A scoping review was conducted following the Joanna Briggs Institute guidance and the PRISMA Extension for Scoping Reviews (PRISMA-ScR). Searches were carried out in Embase, PubMed, Scopus, CINAHL, Cochrane Library, Web of Science, SciELO, BVS, LILACS, and the CAPES thesis and dissertation catalog. Eligible sources included primary studies, technical notes, dissertations, and theses. All references were organized in EndNote Web and transferred to Rayyan to support duplicate removal and facilitate screening by reviewers.

Results: The search resulted in the identification of 1,240 studies, of which eight were included and published in English. Most studies are technological development studies with samples ranging from 10 to 5,729 images or participants. Studies have shown that artificial intelligence techniques applied to pressure injuries, including Convolutional Neural Networks, Random Forest, Support Vector Machine, and Extreme Gradient Boosting, improve detection, measurement, classification, risk prediction, and clinical decision support, potentially reducing workload and enhancing care safety.

Conclusion: The application of artificial intelligence in the domain of skin injuries revealed a variety of uses. However, it was predominantly focused on the specific context of pressure injuries in hospitalized individuals. Consequently, a noticeable gap in the literature was identified regarding alternative categories of injuries affecting this population segment.