A quality assurance study on the accuracy of measuring physical function under current conditions for use of clinical video telehealth

Abstract

Objective: To determine whether conditions for use of clinical video telehealth technology might affect the accuracy of measures of physical function. Design: Repeated measures. Setting: Veterans Administration Medical Center. Participants: Three healthy adult volunteers for a sample size of n=30 independent trials for each of 3 physical function tasks. Interventions: None. Main Outcome Measures: Three tasks capturing differing aspects of physical function: fine-motor coordination (number of finger taps in 30s), gross-motor coordination (number of gait deviations in 10ft [3.05m]), and clinical spatial relations (identifying the proper height for a cane randomly preset ±0-2in [5.1cm] from optimal), with performance simultaneously assessed in person and video recorded. Interrater reliability and criterion validity were determined for the measurement of these 3 tasks scored according to 5 methods: (1) in person (community standard), (2) slow motion review of the video recording (criterion standard), and (3-5) full speed review at 3 Internet bandwidths (64kps, 384kps, and 768kps). Results: Fine-motor coordination - Interrater reliability was variable (r=.43-.81) and criterion validity was poor at 64kps and 384kps, but both were acceptable at 768kps (reliability r=.74, validity β=.81). Gross-motor coordination - Interreliability was variable (range r=.53-.75) and criterion validity was poor at all bandwidths (β=.28-.47). Motionless spatial relations - Excellent reliability (r=.92-.97) and good criterion validity (β=.84-.89) at all the tested bandwidths. Conclusions: Internet bandwidth had differing effects on measurement validity and reliability for the fine-motor task, the gross-motor task, and spatial relations, with results for some tasks at some transmission speeds well below acceptable quality standards and community standards.