Abstract
Purpose
To evaluate whether continuous training and education of posture can help children to improve kyphosis.
Method
A smart harness consisting of a tight-fitting harness and a posture sensing system was developed to measure kyphosis and to provide vibratory feedback during daily activities. The posture sensing system consisted of two sensor units and both units contained a 3-axis accelerometer and a 2-axis gyroscope to calculate the orientation. The dimensions and weight of each unit were 55 mm x 35 mm x 15 mm and 25g, respectively. One unit served as a master (placed at the T3 vertebral level) and the second unit served as a slave (placed at the T12 level) and they communicated wirelessly. The master unit calculated the kyphotic angle, similar to the vertebral centroid method but based on the sagital profile, and provided the vibratory feedback. One volunteer wore the unit and performed different postures and activities (walking, sitting, bending and sudden change from sitting to walking) in a gait analysis laboratory. The posture sensing system was sampled at 30Hz and a gait analysis 8-camera system was sampled at 60Hz. The kyphotic angles captured by the smart harness and camera system were compared. After this validation, the system was tested by 5 normal subjects (M, 25 10 years old) 3 hours per day for 4 consecutive days. For the first 2 days there was no feedback and the last 2 days there was feedback. The system took a sample every 30 seconds. When an undesirable posture was detected, the system switched to a fast sample mode at which time the system took ten measurements with a sample rate of 10 Hz for 1 second to further validate the measured kyphotic angle. These 10 measures were averaged to avoid feedback for postures that lasted only for a very short period of time. Posture orientation data was stored in the sensing unit memory and downloaded for outcomes evaluation.
Results
Compared with the gait analysis camera system, the differences in the kyphotic angle during static and dynamic activities were 1.6 1.2, and 3.5 1.9 degrees, respectively. The largest error was 6.8 degrees which occurred during a dramatic change in posture during dynamic activities. The baseline data from the first 2 days (without feedback) showed the kyphotic angle was 48 12 degrees, during which time all subjects were working in front of a computer. The feedback days showed a slight improvement of kyphotic angles from day 1 to day 2, from 45 11 to 42 8 degrees, respectively. There was 12% improvement on day 2 when compared with the baseline data.
Conclusion
This study showed the kyphotic angle could be fairly accurately measured using the smart harness. The kyphotic angle had a slight improvement when feedback was provided, however a longer clinical trial will be required to determine how lasting the effect will be.
