Effectiveness of Human Versus Computer-Based Instructions for Exercise on Physical Activity–Related Health Competence in Patients with Hip Osteoarthritis: Randomized Noninferiority Crossover Trial (Preprint)

Abstract

Hip and knee osteoarthritis is ranked as the 11th highest contributor to global disability. Exercise is a core treatment in osteoarthritis. The model for physical activity–related health competence describes possibilities to empower patients to perform physical exercises in the best possible health-promoting manner while taking into account their own physical condition. Face-to-face supervision is the gold standard for exercise guidance.

The aim of this study was to evaluate whether instruction and guidance via a digital app is not inferior to supervision by a physiotherapist with regard to movement quality, control competence for physical training, and exercise-specific self-efficacy.

Patients with clinically diagnosed hip osteoarthritis were recruited via print advertisements, emails and flyers. The intervention consisted of two identical training sessions with one exercise for mobility, two for strength, and one for balance. One session was guided by a physiotherapist and the other was guided by a fully automated tablet computer-based app. Both interventions took place at a university hospital. Outcomes were assessor-rated movement quality, and self-reported questionnaires on exercise-specific self-efficacy and control competence for physical training. Participants were randomly assigned to one of two treatment sequences. One sequence started with the app in the first session followed by the physiotherapist in the second session after a minimum washout phase of 27 days (AP group) and the other sequence occurred in the reverse order (PA group). Noninferiority was defined as a between-treatment effect (gIG)<0.2 in favor of the physiotherapist-guided training, including the upper confidence interval. Participants, assessors, and the statistician were neither blinded to the treatment nor to the treatment sequence.

A total of 54 participants started the first training session (32 women, 22 men; mean age 62.4, SD 8.2 years). The treatment sequence groups were similar in size (PA: n=26; AP: n=28). Seven subjects did not attend the second training session (PA: n=3; AP: n=4). The app was found to be inferior to the physiotherapist in all outcomes considered, except for movement quality of the mobility exercise (gIG –0.13, 95% CI –0.41-0.16). In contrast to the two strengthening exercises in different positions (supine gIG 0.76, 95% CI 0.39-1.13; table gIG 1.19, 95% CI 0.84-1.55), movement quality of the balance exercise was close to noninferiority (gIG 0.15, 95% CI –0.17-0.48). Exercise-specific self-efficacy showed a strong effect in favor of the physiotherapist (gIG 0.84, 95% CI 0.46-1.22). In terms of control competence for physical training, the app was only slightly inferior to the physiotherapist (gIG 0.18, 95% CI –0.14-0.50).

Despite its inferiority in almost all measures of interest, exercise-specific self-efficacy and control competence for physical training did improve in patients who used the digital app. Movement quality was acceptable for exercises that are easy to conduct and instruct. The digital app opens up possibilities as a supplementary tool to support patients in independent home training for less complex exercises; however, it cannot replace a physiotherapist.

German Clinical Trial Register: DRKS00015759; http://www.drks.de/DRKS00015759