Abstract
Objective: Current methods of assessing the affect a patients’ health has on their daily life are extremely limited. The aim of this work is to develop a sensor based approach to health status measurement in order to objectively measure health status.
Methods: Techniques to generate human behaviour profiles, derived from smartphone accelerometer and gyroscope sensors, are proposed. Experiments, using SVM regression models, are then conducted in order to evaluate the use of the proposed behaviour profiles as a predictor of health status.
Results: Experiments were conducted on data from 171 participants, with an average of 114 hours of data per participant. Regression models were trained and tested on the 10 SF-36 self-ratings. Results showed that the 8 individual SF-36 scales and 2 component scores could be predicted with an average correlation of 0.683 and 0.698 respectively. General Health was predicted with an average correlation of 0.752.
Conclusion: Research shows that the Clinically Important Difference for SF-36 self-ratings are approximately 10 points. Health status prediction errors in this work were 11.7 points on average. While the problem has not been fully solved, this work present a hugely promising direction for health status prediction.
Significance: Using the proposed techniques, health status could be measured using unobtrusive, inexpensive and already available hardware. It could provide a means for clinicians to accurately and objectively assess the daily life benefits of treatments on an individual patient basis.
Methods: Techniques to generate human behaviour profiles, derived from smartphone accelerometer and gyroscope sensors, are proposed. Experiments, using SVM regression models, are then conducted in order to evaluate the use of the proposed behaviour profiles as a predictor of health status.
Results: Experiments were conducted on data from 171 participants, with an average of 114 hours of data per participant. Regression models were trained and tested on the 10 SF-36 self-ratings. Results showed that the 8 individual SF-36 scales and 2 component scores could be predicted with an average correlation of 0.683 and 0.698 respectively. General Health was predicted with an average correlation of 0.752.
Conclusion: Research shows that the Clinically Important Difference for SF-36 self-ratings are approximately 10 points. Health status prediction errors in this work were 11.7 points on average. While the problem has not been fully solved, this work present a hugely promising direction for health status prediction.
Significance: Using the proposed techniques, health status could be measured using unobtrusive, inexpensive and already available hardware. It could provide a means for clinicians to accurately and objectively assess the daily life benefits of treatments on an individual patient basis.
| Original language | English |
|---|---|
| Pages (from-to) | 1750-1760 |
| Number of pages | 11 |
| Journal | IEEE Journal of Biomedical and Health Informatics |
| Volume | 21 |
| Issue number | 6 |
| Early online date | 9 Jan 2017 |
| DOIs | |
| Publication status | Published (in print/issue) - Nov 2017 |
Keywords
- Smartphone
- Health Status
