Activity monitor placed at the non-paretic ankle is accurate in measuring step counts during community walking in post-stroke individuals : a validation study
Article [Accepted Manuscript]
Is part ofPM&R ; vol. 11, no. 9, pp. 963-971.
Background: Different environmental factors may affect the accuracy of step-count activity monitors (AM). However, the validationconditions for AM accuracy largely differ from ecological environments.Objectives: To assess and compare the accuracy of AM in counting steps among poststroke individuals: during different locomotortasks, with AM placed at the nonparetic ankle or hip, and when walking in a laboratory or inside a mall.Design: Validation study.Settings: Laboratory and community settings.Participants: Twenty persons with chronic hemiparesis, independent walkers.Methods: First session: participants performed level walking (6-minute walk test [6MWT]), ramps, and stairs in the laboratory withAM placed at the nonparetic ankle and hip. Second session: participants walked a mall circuit, including the three tasks, with AMplaced at the nonparetic ankle. The sessions were video recorded.Main Outcome Measurements: Absolute difference between the steps counted by AM and the steps viewed on the video recordings(errors, %); occurrence of errors greater than 10%.Results: Median errors were similar for the 6MWT (0.86 [0.22, 7.70]%), ramps (2.17 [0.89, 9.61]%), and stairs (8.33 [2.65, 19.22]%)with AM at the ankle. Step-count error was lower when AM was placed at the ankle (8.33 [ 2.65, 19.22]%) than at the hip (9.26[3.25, 42.63]%, P = .03). The greatest errors were observed among the slowest participants (≤0.4 m/s) on ramps and stairs, whereassome faster participants (>1 m/s) experienced the greatest error during the 6MWT. Median error was slightly increased in the mallcircuit (2.67 [0.61, 12.54]%) compared with the 6MWT (0.50 [0.24, 6.79]%, P = .04), with more participants showing errors >10% dur-ing the circuit (7 vs 2, P = .05).Conclusions: Step counts are accurately measured with AM placed at the nonparetic ankle in laboratory and community settings.Accuracy can be altered by stairs and ramps among the slowest walkers and by prolonged walking tasks among faster walkers.Level of Evidence: III .