A practical guide to the implementation of AI in orthopaedic research—Part 5: Data management

Abstract

While the magnitude and types of data available to orthopaedic researchers are steadily growing, standardized and efficient data management workflows for orthopaedic research using artificial intelligence (AI) are currently lacking. This work introduces essential principles and best practices for planning, collecting, storing, processing, labelling and governing data in AI-based orthopaedic research. The various domains of available data quality guidelines for medical AI research are reviewed and discussed in terms of their adaptability to orthopaedic research datasets. In addition, future areas of improvement, such as registry development, the potential of synthetic data and gradual transition to continuous data streams for AI applications, are outlined. Level of Evidence: Level V.