Wearable Sensors Reliable for Home Gait Assessment in Knee OA

Sara Freeman

April 26, 2022

Remote gait assessment in people with knee osteoarthritis using wearable sensors appears reliable but yields results slightly different from those achieved in the laboratory, researchers from Boston University have found.

As reported at the OARSI 2022 World Congress, there was "good to excellent reliability" in repeated measures collected by patients at home while being instructed via video teleconferencing.

Agreement was "moderate to excellent" when the findings were compared with those recorded in the lab, Michael J. Rose of Boston University reported at the congress, sponsored by the Osteoarthritis Research Society International.

"People walked faster and stood up faster in the lab," Rose said. "Later we found that the difference in gait speed was statistically significant between the lab and home environment."

This has been suggested previously and implies that data collected at home may have "greater ecological validity," he observed.

Accelerated Adoption of Telehealth

Assessing how well someone walks or can stand from a seated position are well known and important assessments in knee OA, but these have but have traditionally only been done in large and expensive gait labs, Rose said.

Wearable technologies, such as the ones used in the study he presented, could help move these assessments out into the community. This is particularly timely considering the increased adoption of telehealth practices during the COVID-19 pandemic.

To look at the reliability measurements obtained via wearable sensors versus lab assessments, Rose and associates set up a substudy within a larger ongoing, single-arm trial looking at the use of digital assessments to measure the efficacy of an exercise intervention in reducing knee pain and improving knee function.

For inclusion in the main trial (n = 60), and hence the substudy (n = 20), participants had to have physician-diagnosed knee OA, be 50 years of age or older, have a body mass index of 40 kg/m2or lower, be able to walk at for a least 20 minutes, and have a score of three or higher on the Knee Injury and Osteoarthritis Outcome Score pain subscale for weight-bearing items.

Acceptance of In-Lab Versus Home Testing

The substudy participants (mean age, 70.5 years) all underwent in-person lab visits in which a wearable sensor was placed on each foot and one around the lower back and the participant asked to perform walking and chair stand tests. The latter involved standing from a seated position as quickly as possible without using the arms five times, while the former involved walking 28 meters in two laps of a 7-meter path defined by two cones. These tests were repeated twice.

Participants were then given the equipment to repeat these tests at home; this included the three sensors, a tablet computer, and chair and cones. The home assessments were conducted via video conferencing, with the researchers reminding how to place the sensors correctly. The walking and chair stand tests were then each performed four times: Twice in a row and then a 15-minute rest period before being performed twice in a row again.

The researchers collected participants' feedback about the process on questionnaires and Likert scales that showed an overall positive experience for the remote home visit, with the median rating being "very likely" to participate in another home visit and that the time commitment required was "very manageable."

Good Correlation Found

To determine the correlation and the test-retest reliability of the data obtained during the repeated home tasks, Rose and collaborators used Pearson's correlation R2 and the intra-class correlation coefficients (ICC).

ICCs for various gait and chair stand variables obtained with the sensors were between 0.85 and 0.96 for the test-retest reliability during the remote home visit, and R2 ranged between 0.81 and 0.95. Variables include stance, cadence (steps per minute), step duration and length, speed, and chair stand duration.

With regard to the agreement between the home versus lab results, ICCs ranged between 0.63 and 0.9.

"There were some logistical and technological challenges with the approach," Rose conceded. "Despite written and verbal instructions, 2 of the 20 participants ended up having gait data that was unusable in the home visit."

Another limitation is that the study population, while "representative," contained a higher number of individuals than the general population who identified as being White (95%) and female (85%), and 90% had a college degree.

"Individuals typically representative of an OA population were generally accepting and willing to participate in remote visits showing the feasibility of our approach," Rose said.

"We need to determine the responsiveness of gait and chair stand outcomes from wearable sensors at home to change over time."

The study was sponsored by Boston University with funding from Pfizer and Eli Lilly. The researchers used the OPAL inertial sensor (APDM Wearable Technologies) in the study. Rose made no personal disclosures. Four of his collaborators were employees of Pfizer and one is an employee of Eli Lilly & Company, all with stock or stock options.

This article originally appeared on MDedge.com, part of the Medscape Professional Network.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.