Table 5. Triangulating the validity of Fitbit-derived PA metrics.
| Step count | Time in PA | Time in MVPA | |
|---|---|---|---|
| Did criterion measures agree with each other? | |||
| Scripted tasks | + | ++ | ++ |
| Free living, Epoch level | -- | ++ | +/++ |
| Free living, Daily level | -- | + | + |
| Free living, Average level | - | - | ++ |
| Consistency across severity strata* | + | + | - |
| Did criterion measures correlate with each other? | |||
| Scripted tasks | ++ | na | na |
| Free living, Epoch level | - | na | na |
| Free living, Daily level | ++ | ++ | ++ |
| Free living, Average level | ++ | ++ | ++ |
| Consistency across severity strata* | + | + | - |
| Did Fitbit agree with criterion measures? | |||
| Scripted tasks | + | ++ | -- |
| Free living, Epoch level | -- | +/++ | -/+ |
| Free living, Daily level | -/+ | --/- | + |
| Free living, Average level | + | -/+ | +/++ |
| Consistency across severity strata* | - | - | - |
| Did Fitbit meet or exceed the agreement exhibited by criterion measures? | + | -/+ | - |
| Did Fitbit correlate with criterion measures? | |||
| Scripted tasks | ++ | na | na |
| Free living, Epoch level | - | na | na |
| Free living, Daily level | ++ | +/++ | + |
| Free living, Average level | ++ | +/++ | +/++ |
| Consistency across severity strata* | -/+ | - | - |
| Did Fitbit meet or exceed the correlations exhibited by criterion measures? | + | -/+ | - |
| Did Fitbit relate to clinical outcomes? | |||
| Did Fitbit exhibit the expected correlations with clinical measures? | + | + | + |
| Did Fitbit-derived PA metrics differ across known groups? | + | - | + |
| Did Fitbit meet or exceed the relationships exhibited by criterion measures? | + | + | + |
| Validity of Fitbit-derived PA metrics | |||
| Can criterion measures be considered equivalent? | - | - | - |
| Can Fitbit and Actigraph be considered equivalent? | - | - | - |
| Did Fitbit exhibit evidence of construct validity? | + | + | + |
++: Excellent agreement or strong correlations (0.75–1.0)
+: Fair to good agreement or moderate correlations (0.4–0.75)
-: Poor agreement, weak correlations (0.2–0.4)
—: Very weak or complete lack of agreement or correlation (<0.2)
+/-: Evidence was mixed; Binary yes/no responses are indicated with + or -, respectively
*Consistency across severity strata describes whether the trends observed during scripted tasks and at the epoch, daily, and average levels were consistent across subgroups with mild, moderate, and severe MS. A positive rating in this category does not necessarily mean that measures correlated or agreed with each other.