Modelling and artificial intelligence using sensor data to personalise rehabilitation following joint replacement

In 2017, over 218,000 people in the UK had a hip (THR) or knee (TKR) replacement. Provided there are no other underlying conditions, these patients should return to normal activities. However, published studies show that fewer than 50% joint replacement patients regain a normal healthy walk (gait) 1 year post-op. Gait deficiencies have been linked to osteoarthritis in other joints, poor mobility and reduced independence in Activities of Daily Living (ADLs) resulting a lower Quality of Life (QoL). This project will use a medical device plus automated exercises (vGym) to help improve the rehabilitation phase following joint replacement, with the goal to correct gait abnormalities and hence improve a patient's mobility and QoL and reduce healthcare costs.

GaitSmart plus vGym, an innovative cloud based, smart sensor system, will be used to determine hip and knee replacement patients' gait kinematics in the outpatient clinic and provide exercises. In this project this will be linked to an artificial intelligence (AI) machine learning system, to optimise the personalised rehabilitation programme.

Trials will take place on unilateral hip and knee replacement patients from Norwich Hospital. Patients in the intervention group will receive personalised exercise programmes at each appointment; 6, 9, 12 and 15 weeks post-op, based on their GaitSmart data. A control group will follow the Standard of Care (SoC).

Evidence of clinical efficacy will be determined by comparing digital gait kinematics data, speed, PROMS and QoL data from patients following the new care pathway and SoC at the start and end of the intervention period. The economic benefit to the NHS will be determined by comparing all patient outcome data for both groups and predicting future costs based on gait deficiencies relative to changes in QoL.

For this project, DML will develop artificial intelligence (AI) so the personalised exercise programmes are produced automatically. The GaitSmart test and automated exercise programme will be an integral part of the patient's rehabilitation and delivered by Healthcare Assistants.

GaitSmart has already completed a GaitSmart intervention study in to the NHS for older patients who have fallen and are under the care of a community hospital. These patients received four GS sessions and an automated GaitSmart personalised exercise programme and the clinical and health economic data presented to MPs. This approach is clinically effective and produces a positive ROI. The learnings will be applied to this study.