Table 4.

Correlation coefficients comparing average daily vigorous, moderate, moderate-to-vigorous, and light physical activity and steps measured by the Fitbit One, ActiGraph GT3X+, Omron pedometer, and a physical activity diary among 23 men with localized prostate cancer.

Measuring devicea,b	Fitbit One versus ActiGraph GT3X+		Fitbit One versus Diaryc		Fitbit One versus Omron pedometer		ActiGraph GT3X+ versus Diary		ActiGraph GT3X+ versus Omron pedometer
No. of men	22		20		21		20		21
Activity Category	r	P value	r	P value	r	P value	r	P value	r	P value
MVPA	.85	<.001	.84	<.001	-	-	.68	.001	-	-
Vigorous activity	.65	.001	.47	.04	-	-	.66	.001	-	-
Moderate activity	.70	<.001	.38	.10	-	-	.57	.009	-	-
Light activity	.72	<.001	-	-	-	-	-	-	-	-
Steps	.94	<.001	-	-	.67	.001	-	-	.72	<.001

^aPearson correlations were used for measures that were normally distributed: steps, light activity, and moderate activity. Spearman correlations were used for skewed measures: vigorous activity.

^bCorrelation coefficients were not calculated if one of the devices did not measure the activity category of interest. The diary did not measure light activity or steps. The pedometer does not measure light, moderate, or vigorous physical activity.

^bOne individual reported high levels of tennis, which was classified as a vigorous activity in the diary data based on the standard MET value for tennis (7.3 METs) but was classified as a moderate activity for this individual by the Fitbit. The correlations for vigorous and moderate physical activity as assessed by the Fitbit One and diary excluding this individual were r=.61, P=.006 and r=.61, P=.005, respectively. We also reclassified tennis as a moderate activity in the diary data, and the correlations between the Fitbit One and the diary for vigorous and moderate activity were similarly improved (r=.58, P=.008 and r=.73, P<.001).