Practical Artificial Intelligence: Realistic Ways It Can Help Orthopaedic Surgeons and the Challenges It Will Face.

Abstract

Artificial intelligence can improve various orthopaedic subspecialties in the next 5 to 10 years. There are several image recognition applications particularly in orthopedic trauma and orthopedic spine. Specifically, convolutional neural networks have been shown to work well for making diagnoses and recreating more advanced imaging form radiographs. There are many applications of artificial intelligence with predictive in total joint arthroplasty, particularly with shared decision making. And there are many day-to-day applications that can be improved with natural language processing, particularly administrative tasks. This includes several applications in billing and charting. When investigating the landscape of artificial intelligence in healthcare, there are many barriers to their adoption. This includes overcoming bias, incorporating new applications into clinical workflow, regulatory approval, and billing.