Arthritis detection using MAchine LearninG and AcceleroMetry (AMALGAM)

  • Research type

    Research Study

  • Full title

    Integrating wearable accelerometer data with machine learning for enhanced osteoarthritis diagnosis – a feasibility study

  • IRAS ID

    343294

  • Contact name

    Andrew McCaskie

  • Contact email

    awm41@cam.ac.uk

  • Sponsor organisation

    Cambridge University Hospital NHS Foundation Trust & University of Cambridge

  • Duration of Study in the UK

    1 years, 6 months, 1 days

  • Research summary

    Osteoarthritis (OA) of the hip and knee is a common condition that can severely affect quality of life by causing pain, stiffness, and difficulty with everyday tasks. Our study investigates whether wearable devices (small sensors worn on the wrist or ankle) can detect differences in walking patterns between people with OA and healthy volunteers. We will enrol around 60 participants, some with OA and others without joint problems, to measure their daily activity and complete a supervised walk test. The sensors will capture data such as step count and gait patterns, providing detailed information on how people move. By applying “machine learning” (computer programmes that find hidden patterns in large datasets) we aim to see if these devices can reliably distinguish OA from healthy joints in a real-world setting.
    In this feasibility study, we will focus on whether the technology is user-friendly and suitable for longer-term use, and we will also explore participants’ comfort and willingness to use these devices. We will not yet combine these movement data with our existing models for predicting OA progression, but successful outcomes here could lead to larger, extended research. Future studies may then explore if wearable-based measurements can improve early diagnosis and guide more personalised interventions, such as tailored rehabilitation programmes or new therapies. Since OA is often managed with medications, physical therapy, and, in severe cases, surgery, discovering subtle changes in walking patterns could help identify people who need additional support sooner. Throughout this process, we will involve patients and the public to ensure our methods align with the needs and preferences of those most affected by OA. Ultimately, by better understanding how movement patterns reflect joint health, we aim to lay the groundwork for improved early detection and treatment, potentially reducing the impact of OA on individuals, families, and healthcare systems.

  • REC name

    London - Bromley Research Ethics Committee

  • REC reference

    25/PR/0675

  • Date of REC Opinion

    16 Jun 2025

  • REC opinion

    Further Information Favourable Opinion