Reproducibility and Acceptability of Short Physical Function Tests Scores Obtained via Virtual versus Face-to-Face Assessments

Abstract

Gait speed (GS), 30-second chair stand test (CST), timed up and go test (TUG), and step test (ST) have been validated for face-to-face assessment (F2F) but not for virtual testing (VT). This study examined reproducibility and acceptability of GS, CST, TUG, and ST for VT. Three student physical therapists administered GS, CST, TUG, and ST to 52 healthy participants in F2F and VT formats. Participants' experience and acceptability of VT were captured via a survey. Reproducibility of these measures for VT was assessed using intraclass correlation coefficient (ICC), where ICC > 0.75 and > 0.90 indicated good and excellent reliability. Paired <i>t</i>-tests examined whether the scores differed in VT versus F2F. Interview responses were coded for convergence of information for acceptability of VT and audiovisual communication during VT. Scores for GS (ICC = 0.83), CST (ICC = 0.87), TUG (ICC = 0.95), and ST (ICC = 0.93) obtained in VT were reproducible with F2F. However, <i>t</i>-tests revealed that the VT scores were statistically different (<i>P</i> < 0.0001) for all measures, with exception of GS (<i>P</i> = 0.14), where performance was superior in F2F testing. Thematic analysis suggested high level of acceptability with VT but unanimous preference for F2F assessment. Results indicate that scores for GS, CST, TUG, and ST obtained via VT are reproducible with F2F, albeit superior performance was observed in F2F. Participants found VT acceptable but indicated preference for F2F. We caution clinicians that the established norms for F2F for these measures should not be used and different normative values for these measures may be required for VT.