Challenges and opportunities of the implementation of machine learning in burn clinical care: A systematic review

Background and aims: Machine learning as one of the components of artificial intelligence isknown to play an effective role in increasing the accuracy of diagnosis, sensitivity and clinicalfeatures for predicting complications related to burns such as acute kidney injury. However, thewidespread adoption of machine learning has challenged the development of clinically relevantmodels due to the lack of expert programming in the medical community and access to healthdata. The aim of this study was to assess the challenges and opportunities of the implementationof machine learning in burn clinical care.Method: The protocol of this systematic review followed the PRISMA guideline. An extensivesearch was carried in online databases including PubMed, Web of Science, Scopus, Google Scholar,and ProQuest with the keywords such as “Machine Learning”, “Transfer Learning”, “ArtificialIntelligence”, “Burn”, “Burn Unit”, “care”, “clinical care”, from the earliest records up to October۲۰, ۲۰۲۲. Also, all English-language studies related to the purpose of the present study were included.Letters to the editor, opinions, conference abstracts, intervention, reviews were excludedfrom this study. The appraisal tool for cross-sectional studies (AXIS tool) was used to assess thequality of included studies. All stages of search and quality evaluation of articles were conductedby two researchers, independently.Results: A total of ۶ out of ۲۱۵ studies were included in the study. The challenges of machinelearning in burn clinical care were including lack of complete expertise in using machine learning,lack of expert programming and inappropriate access to health data. On the other hand, thesechallenges lead to the creation of opportunities such as the development of automatic machinelearning platforms to facilitate clinical studies to fully understand the high capabilities of artificialintelligence in the health field, especially burn clinical care, extensive education of healthcareworkers, especially doctors, and improving the standard of burn clinical care.Conclusion: The most challenges of machine learning in burn clinical care were including lackof complete expertise in using machine learning, lack of expert programming and inappropriateaccess to health data. In general, the use of machine learning, despite the challenges raised, canavoid heavy costs for the patient, family and medical systems by accurately and early diagnosisof complications when care resources are limited and expensive. Therefore, these challenges arestill a major concern for the implementation of machine learning in burn clinical care. Hence,more research is needed to address the challenges of using machine learning for burn clinical care.

کلیدواژه ها:

machine learning ، artificial intelligence ، burns