Has Artificial Intelligence Improved Management Outcomes of Hand Fractures in the Tropics? A Systematic Review and Meta Analysis
Kelechi Uzodinma Imediegwu *
Department of Orthopaedic Surgery, National Orthopaedic Hospital, Enugu, Nigeria and Division of Surgery and Interventional Science, University College London, UCL, UK.
Ebuka L. Anyamene
College of Medicine, University of Nigeria, UNN, Nigeria.
Makata C. Amarachi
College of Medicine, Enugu State University of Technology, ESUT, Nigeria.
Chika A. Nmaju
College of Medicine, University of Nigeria, UNN, Nigeria.
Chukwuemeka V. Mokwe
College of Medicine, Enugu State University of Technology, ESUT, Nigeria.
Iloabuchi C. Collins
College of Medicine, Enugu State University of Technology, ESUT, Nigeria.
Kenechukwu D. Nwafor
College of Medicine, University of Nigeria, UNN, Nigeria.
Chidiogo I. Nwafor
College of Medicine, University of Nigeria, UNN, Nigeria.
Chukwuemeka P. Ikemsinachi
College of Medicine, University of Nigeria, UNN, Nigeria.
Kosisochukwu E. Udeogu
College of Medicine, University of Nigeria, UNN, Nigeria.
Chioma F. Idu
College of Medicine, University of Nigeria, UNN, Nigeria.
Chimbuchi E. Ozoemena
College of Medicine, University of Nigeria, UNN, Nigeria.
Joanne C. Akwaeke
Ansett Diagnostic Hospital, Lagos, Nigeria.
Grace N. Kaluokoro
College of Medicine, University of Nigeria, UNN, Nigeria.
Sylvester S. Eze
College of Medicine, University of Nigeria, UNN, Nigeria.
Ndoh E. Eunice
College of Medicine, University of Nigeria, UNN, Nigeria.
*Author to whom correspondence should be addressed.
Abstract
Introduction: The application of Artificial intelligence (AI) and Machine learning (ML) in hand surgery has rapidly become a powerful and emerging field with a promising patient-care reputation. This study aims to compare the diagnostic performance of AI with that of an experienced hand surgeon in terms of accuracy, specificity, and sensitivity in detecting hand injuries
Methods: A systematic literature search of PubMed, Web of Science, Scopus, and Google Scholar databases was conducted.
Results: The initial search yielded 9,317 articles. Only nine (9) articles were included in the final review with a 45,778-sample size after the application of the inclusion/exclusion criteria.
The result indicated no statistical significance of the application of AI and ML in hand surgery as compared to experienced surgeons in terms of accuracy (OR=0.947, 95% CI 0.552- 1.624; p= 0.844), sensitivity (OR=0.678, 95% CI 0.354- 1.289; p= 0.234) and specificity (OR=1.260, 95% CI 0.525- 3.022; p= 0.604).
Conclusion: There was no significant difference in accuracy, sensitivity, and specificity in terms of diagnosis and clinical application between an expert surgeon and AI or ML in hand surgery. Thus, Artificial intelligence can also be utilized in diagnosing hand surgery in the tropics, however since artificial intelligence is an evolving field, further calibration will be required for its improved diagnostic value. We recommend that more doctors in the tropics are better equipped with AI techniques to be used as an adjunct in clinical practice.
Keywords: Artificial intelligence, Machine learning, Hand fracture