Quantifying At-Home Physiotherapy Participation: SPARS vs Self-Reported Diaries.

Area of Science:

• Biomedical Engineering
• Rehabilitation Science
• Digital Health

Background:

• At-home physiotherapy adherence is crucial for rehabilitation success.
• Traditional self-report diaries lack objectivity and can suffer from recall bias.
• Wearable technology offers a potential solution for objective monitoring of patient exercise participation.

Observation:

• This study compared smartwatch-derived data (accelerometer/gyroscope) analyzed by a convolutional neural network against patient self-report diaries.
• Data was collected from 53 patients with rotator cuff pathology during the initial two weeks of a 12-week physiotherapy program.
• The system utilized patient-specific in-clinic data for training the machine learning algorithm.

Findings:

• A high agreement (ICC=0.72) was observed between smartwatch data and diary entries for exercise participation.
• The machine learning algorithm achieved an AUROC of 0.99 for identifying exercise periods.
• Patient diaries reported more exercise sessions (0.96 additional days on average) than recorded by the smartwatch system.

Implications:

• Smartwatch-based monitoring provides an accurate and objective alternative to traditional self-report diaries for at-home physiotherapy.
• Discrepancies may highlight technology limitations or over-reporting in diaries.
• Physical therapy monitoring technology shows promise for long-term assessment as diary adherence declines.