Use of an activity tracker as a measurement tool in a knee osteoarthritis clinical trial (active-oa trial)

Abstract

Purpose: In osteoarthritis (OA) clinical trials, reliable and responsive outcome measures to document improvements in function are limited. Current measures predominately rely on self-reported function tests, but they are poorly responsive, subject to misreporting and the placebo response. Functional measures such as observed timed-up-and-go test (TUG) and six-minute walk test (6MWT) are advocated, but their relevance may be limited in the context of a patient’s real-world functional capabilities. Using physical activity tracking devices/wearables as a way to measure improvements in function in day-to-day life is potentially an attractive and easily available alternative to conventional OA trial outcome measures. Consumer-level activity monitors have shown high validity for measurement of steps when compared with research-grade devices. The aim of this study is to assess if the use of an activity tracker in an OA clinical trial is a responsive measurement tool. Secondary objectives include the assessment of the feasibility of an activity tracker use and validity of an activity tracker in an OA clinical trial. Methods: This was a prospective cohort study nested in the Platelet-rich Plasma as a Symptom and Disease-Modifying Treatment for Knee Osteoarthritis - the RESTORE trial (ACTRN12617000853347). Consecutive participants recruited to the RESTORE trial were invited to join this study. Participants ≥ 50 years old with painful knee OA and Kellgren and Lawrence (KL) grade 2-3 were included. Participants were asked to wear an activity tracker (Fitbit Flex (2), Fitbit Inc, San Francisco, CA, USA) for seven consecutive days at the time of baseline assessment (week 0, before their injection therapies), and at the 2-month follow-up assessment. The primary outcome was mean steps per day calculated from a seven-day wear of an activity tracker. To determine the responsiveness of the activity tracker, the standardised response mean (SRM) was calculated for mean steps per day between baseline and 2 months; dividing the mean change scores by the standard deviation of the mean change scores. To investigate validity, correlation coefficient (r) was used to examine the relationship between change in the mean steps per day between baseline and 2 months and participant-perceived global ratings of change (global improvement in function) at 2 months. The relationship of change in mean steps per day and change in secondary outcomes between baseline and 2 months was examined: global improvement (pain and overall improvement), 6MWT, TUG, numeric rating scale (NRS) for pain, knee osteoarthritis outcome score (KOOS), and the Western Ontario and McMaster Universities Osteoarthritis index (WOMAC). Correlation coefficients were estimated for continuous variables. The activity tracker wear was deemed feasible if >70% of patients comply with wearing the trackers at least six out of seven days for each measurement period. Results: 65 participants were recruited (mean age = 61.3 years; 46% female; mean BMI 28.7 kg/m2; 41.5% KLG 2 and 58.5% KLG 3). 8 participants dropped out or had activity tracker misuse/failure at the 2-month follow-up. The SRM was -0.02 (95% CI -0.38 to 0.35) for mean steps per day between baseline and 2 months. The proportion of participants who wore the trackers for six out of seven days was 98% at baseline and 100% at 2-months follow-up. The correlation with change in mean steps per day at baseline and 2 months and outcome measures are detailed in table 1. Conclusions: The use of an activity tracker in a real-world OA clinical trial demonstrated small correlations between change in steps at baseline and 2 months follow-up with global improvement overall and WOMAC function. The use of a tracker is a feasible option with demonstration of good compliance amongst the trial participants. This study has not shown significant responsiveness or validity of the activity trackers in comparison with other regularly utilised outcome measures in OA clinical trials. However, the main challenge of this study is the lack of a gold standard outcome measure to validate against. Due to the complex nature of pain, perceived quality of life, and the interplay between these factors and measured function, a lack of correlation does not necessarily represent a failed validation in this context.