| Summary: | <p>Subacromial shoulder pain is the third most common musculoskeletal complaint
presenting to primary care. Despite the extensive amount of research carried out on the role of
exercise-based rehabilitation, there remains uncertainty around the effectiveness of specific
treatment strategies. Wearable sensors that enable practical monitoring of patient activity and
objective measurement of movement could provide opportunities to address this problem.
Magneto-inertial measurement units (MIMUs) offer clear advantages over existing motion
capture systems, and can facilitate careful measurement of treatment implementation and
response. There remains, however, a distinct lack of both technical and usability validation for
the clinical application of these tools.</p>
<p>The aim of this DPhil thesis was to develop a practical MIMU based motion capture
system and investigate its value in providing meaningful and useful feedback on shoulder
movement and shoulder rehabilitation. This began with the development and testing of a
processing pathway to derive reliable kinematic data from a commercial MIMU sensor under
controlled motion conditions representative of human movement. The sensors were
subsequently incorporated into custom-made wearable accessories, and shoulder motion data
was collected from 70 healthy volunteers and 45 patients with subacromial shoulder pain.
Processing of raw inertial data to clinically meaningful output was performed in MATLAB
using a combination of custom-written and open-access code. Investigations were performed
to evaluate the validity of a range of MIMU derived metrics for the assessment of shoulder
activity and functional performance. Finally, the use of a prototype end-to-end remote
monitoring system, offering near-real time feedback, was studied in 15 patients undergoing a
course of home rehabilitation. System accuracy, usability and perceived impact of the
technology was evaluated.</p>
<p>Metrics derived from an upper arm MIMU-based kinematic data demonstrated good
psychometric properties in the assessment of patients with subacromial shoulder pain.
Movement smoothness demonstrated reliability, known group validity and responsiveness to
treatment. Metrics based on ranges of accelerometer and gyroscope data showed similar
psychometric properties, as well as convergent validity, demonstrated by significant
correlation with the Oxford Shoulder Score. Significant effect sizes for changes of these
metrics following treatment were seen over a 12-week period. Implementing a single sensor
setup, a fully automated processing pathway was validated for independent recognition and
classification of shoulder exercise activity, with up to 97.2% overall accuracy. Quantitative
and qualitative assessment of usability demonstrated that patients were comfortable with this
form of monitoring and many felt these systems could improve their engagement with
treatment.</p>
<p>MIMU based biofeedback systems offer a practical and reliable approach to measuring
treatment compliance and recovery of function in patients with subacromial shoulder pain.
They also have the potential to address factors that influence adherence, through selfmonitoring
and the provision of automated feedback. Through the development of novel data
processing techniques and testing under clinically relevant conditions, this thesis has made
several contributions to this field of study and provides translatable findings for researchers
and clinicians aiming to employ wearable technologies for the assessment and treatment of
musculoskeletal disease.</p>
|
|---|