An AI model using gait data before and after hip replacement could help predict which patients will walk close to normal and who may need extra rehabilitation.
The tool analyses gait biomechanics, the study of how people walk, using data collected before and after operations in people with hip osteoarthritis, caused by cartilage wear in the hip joint.
The study examined 109 patients with unilateral hip osteoarthritis before total hip replacement. Of these, 63 were reassessed after surgery, while 56 healthy people acted as a control group.
The analysis, developed by researchers at Karlsruhe Institute of Technology and Universitätsmedizin Frankfurt, identified three patient groups with different patterns of change in how they walked.
Some improved markedly and walked almost normally after surgery, while others still showed clear deviations from the control group.
Three-dimensional joint angle and joint loading data were gathered through musculoskeletal modelling. Certain measurements, including hip angles and joint loads, were especially useful in assigning patients to the right group. The groups also differed in age, height, weight, walking speed and osteoarthritis severity.
Dr Bernd J. Stetter, who leads a musculoskeletal health and technology research group at Karlsruhe Institute of Technology, said: “Our model makes it possible to predict who will benefit especially well from an operation and who will need additional intensive therapy afterwards.
“Since the algorithms are explainable and transparent, we expect the model to enjoy a high level of clinical acceptance.”
The researchers said the model could help clinicians make better-informed decisions, set more realistic expectations with patients and tailor rehabilitation after surgery.
While the model was trained for artificial hip joints, the team said the approach could potentially be adapted to other joints and conditions.
Around 200,000 people in Germany received artificial hip joints in 2024, making it one of the country’s most common orthopaedic procedures.
