Demographics and baseline characteristics of the 30 patients are presented in table 1.
Walking distances and walking speeds are listed in table 2.
Outside walking versus estimated walking distance
A moderate reliability was found between these outcomes (ICC=0.62). The median patients'
estimated MWD was 400m, which is significantly higher than measured during outside walking (median=309m, P=.02) (table 2). The mean difference found in the Bland-Altman plot in MWD between both outcomes was 122m (95% limits of agreement: -677.4m – 433.6m). A substantial and significant divergent scatter was reflected by the Bland-Altman plot (figure 1) indicating that the agreement between outside walking and estimated distances decreased when patients judged that their walking distance was longer than it actually was. This variability between outcomes was confirmed by a 55% COV value.
Outside walking versus Gardner Skinner Treadmill Protocol (GSP)
Again a moderate reliability was found between these outcomes (ICC=0.67). No significant differences were found in median MWD as measured during outside walking (309m) and GSP (250m, P=.70, table 2). The mean difference in MWD as shown in the Bland Altman plot between outside walking and the GSP was 35m (95% limits of agreement: -328.5m – 399.0m) (figure 2). A 42% COV was found, dropping to 19% after removing two extremes (figure 2).
Outside walking versus Non Graded Treadmill Protocol (NGTP)
A moderate reliability was found between these outcomes (ICC=0.46). No significant differences in median MWD were found comparing outside walking (309m) with NGTP (317m, P=.20, table 2). A -122m mean difference (95% limits of agreement: -920.5 – 676.4) was calculated and depicted in the Bland Altman plot (figure 3). A significant negative scatter trend was seen (figure 3) indicating that the agreement between outside walking and the NGTP outcomes was lower in patients with larger walking distances. An 89% COV variability was found.
WIQ scores and correlations.
The WIQ distance score correlated well with values of MWD as estimated by the patient or as measured during outside walking or following GSP or NGTP testing (table 3). In contrast, the WIQ total score correlated moderately with these distances.
Walking speed during outside walking was significantly faster than during the GSP and NGTP (medians: outside walking=3.8 km/h, GSP=3.2 km/h, NGTP=2.8 km/h; P<.001, table 2). Outside walking speed correlated moderately with the WIQ speed score (r = 0.62, p<.001).
Outside walking measuring methods.
Median MWD as measured with the iPhone (309m) were not different compared to values obtained with a measuring wheel (290m, P=.064). A strong correlation was observed between these values (ICC=0.997).