JMIR Publications

ABSTRACT

Background:

Hip osteoarthritis (OA) affects patients’ function during activities of daily living (ADL); however, objective means of measuring this function are lacking. To characterize daily living function, the activities performed during daily life need first to be accurately identified. Human activity recognition (HAR) models based on wearable sensor measurements, such as Inertial Measurement Units (IMU), have previously been implemented in individuals with various functional impairments, but not yet in people with end-stage hip OA. In addition, the balance between HAR accuracy and wearability, crucial for patients’ acceptance, has not yet been thoroughly explored. Finally, to assess deficits of pathological populations, the comparison with asymptomatic controls is necessary.

Objective:

This study aimed to evaluate the minimal accurate IMU configuration, from 8 to 1 IMU, in patients with hip OA and asymptomatic controls.

Methods:

We collected data from 21 patients and 10 controls. Participants completed a tour of ADL in the hospital vicinity while equipped with eight IMUs. Activities focused on mobility and included gait on flat ground, gait up and down a ramp, stairs ascent/descent, turns, sitting up/down, and static sitting. Bi-directional Gated recurrent unit (Bi-GRU) models were trained to classify ADL based on 3D accelerations and angular velocities of multiple IMU configurations.

Results:

Model accuracy was assessed on test sets of patient and control data using Cohen’s Kappa, obtaining values of 0.95 for 8-IMUs, 0.93 for 4-IMUs, 0.90 for 2-IMUs and 0.79 using 1-IMU. The accuracy was higher for controls than for patients. The best 2-IMU configurations were two shanks and two feet, and the best 1-IMU configuration was the shank.

Conclusions:

The reduction of IMU number from 8 to 2 showed a minimal decrease in Kappa, indicating that minimal IMU setups could be as accurate as larger setups. Regarding gait detection, a single IMU was sufficient to reach very high accuracy (Kappa≥0.90). The single pelvic IMU showed the lowest accuracy for gait while still reaching high accuracy (Kappa>0.65). This study provides proof-of-concept that an accurate and minimal HAR system could be developed using IMU data from people with hip OA.