Step-Count Accuracy of Three Motion Sensors for Older and Frail Medical Inpatients

Publication date: Available online 22 September 2016
Source:Archives of Physical Medicine and Rehabilitation
Author(s): Ruth McCullagh, Christina Dillon, Ann Marie O'Connell, N.Frances Horgan, Suzanne Timmons
ObjectivesTo measure the step-count accuracy of an ankle-worn accelerometer, a thigh-worn accelerometer and one pedometer in older and frail inpatients.DesignCross-sectional design study.SettingResearch room within a hospital.ParticipantsConvenience sample of inpatients aged ≥65 years, able to walk 20 metres unassisted, with or without a walking-aid.Intervention: Patients completed a 40-minute programme of predetermined tasks while wearing the three motion sensors simultaneously. Video-recording of the procedure provided the criterion measurement of step-count.Main Outcome MeasuresMean percentage (%) errors were calculated for all tasks, slow versus fast walkers, independent versus walking-aid-users, and over shorter versus longer distances. The Intra-class Correlation was calculated and accuracy was visually displayed by Bland-Altman plots.ResultsThirty-two patients (78.1 ±7.8 years) completed the study. Fifteen were female and 17 used walking-aids. Their median speed was 0.46 m/sec (interquartile range, IQR 0.36-0.66). The ankle-worn accelerometer overestimated steps (median 1% error, IQR -3 to 13). The other motion sensors underestimated steps (40% error (IQR -51 to -35) and 38% (IQR -93 to -27), respectively). The ankle-worn accelerometer proved more accurate over longer distances (3% error, IQR 0 to 9), than shorter distances (10%, IQR -23 to 9).ConclusionsThe ankle-worn accelerometer gave the most accurate step-count measurement and was most accurate over longer distances. Neither of the other motion sensors had acceptable margins of error.



from Rehabilitation via xlomafota13 on Inoreader http://ift.tt/2d4HaEX
via IFTTT