Gait analysis using a tri-axial accelerometer : reliability and difference between elderly persons according to fall risk.

Research output: Chapter in Book/Report/Conference proceedingConference paperResearch


Background: Instrumental gait analysis often requires heavy equipment such as a force platform, thus lowering its accessibility for frail elderly patients. An alternative approach is to evaluate the gait characteristics in these patients using a tri-axial accelerometer. Although an increasing body of literature exists describing several outcome parameters that can be obtained using this technology little is known about the reliability of these evaluations, especially in geriatric patients presenting increased risk for falling. Aim of the study: To investigate the reliability of gait analysis outcomes using a tri-axial accelerometer in elderly persons with and without increased fall risk, and to evaluate the difference in gait characteristics between both groups. Methods: 31 geriatric patients (11 male, 20 female, mean age 81±5 years) with increased fall risk (documented as a recent fall history and/or Timed Get-Up-and-Go test >15 seconds and/or Tinetti-test score ?24/28), 40 healthy elderly persons without fall risk (20 male, 20 female, mean age 79±5 years) and three assessors participate in the study. For each participant three gait evaluations (each containing 2 walks of 18 meters) were performed by 2 assessors in a random order. During each gait analysis, a tri-axial accelerometer (DynaPort MiniMod, MacRoberts Inc., The Hague, The Netherlands; data acquisition at 100Hz) was positioned in the middle of the Os Sacrum with an elastic belt around the pelvis. Gait speed, step time, step time asymmetry (difference between left and right leg step time, divided by the mean step time of both legs), cranio-caudal (CC) and medio-lateral (ML) step & stride regularity (autocorrelation of respectively CC and ML accelerations of subsequent steps & strides) were calculated from the accelerometer data. Intra -and inter-observer reliability were analyzed per subgroup and gender for all outcome parameters (for single walk and mean of 2 walks) using Intra-Class Correlation coefficients (ICC; model 3,1). Reliability was defined as: ICCResults: For each subgroup inter -and intra-observer reliability of gait speed was good to excellent (based on single as well as average of 2 walks). For the other outcome parameters, the average values of two walks showed good to excellent reliability (except poor reliability for CC step regularity in the healthy female and CC stride regularity in healthy male & female). When based on a single walk, reliability was markedly lower (poor to very good). Subjects presenting increased fall risk showed slower gait speed and worse CC & ML step and stride regularity compared to those without fall risk (all p
Correspondence: FRIA, VUB, Laarbeeklaan 103, B-1090 Brussels, Belgium.
Original languageEnglish
Title of host publication32nd Wintermeeting Belgische Vereniging voor Gerontologie en Geriatrie, Oostende
Publication statusPublished - 2009


  • elderly
  • accelerometer
  • reliability
  • fall

Fingerprint Dive into the research topics of 'Gait analysis using a tri-axial accelerometer : reliability and difference between elderly persons according to fall risk.'. Together they form a unique fingerprint.

Cite this