4 The Short Physical Performance Battery does not correlate with daily life gait quality and quantity | 61 of the other foot and 3) tandem stance with one foot placed in front of the other, with the heel of the one foot touching the toe of the other foot. In the Four Meter Walk Test the participants were asked to walk a distance of four meters at a comfortable speed. The time to walk this distance is the outcome measure, and to minimize burden, a single walk test has been shown reliable and valid (200, 201). During the Five Times Sit to Stand test the time was recorded a participant needed to stand up and sit down from a chair five times as quickly as possible, with timing stopped after the fifth rise. Participants were instructed not to use walking aids during the SPPB assessment unless it was necessary for safety reasons. The total score of the SPPB is a sum of the scores of the Three Stage Balance test, Four Meter Walk test and Five Times Sit to Stand test. Each test results in a score between 0 (least favourable) to 4 (most favourable), thus the SPPB has a range from 0-12. Daily life gait quality and quantity measures To determine gait quality and quantity, participants received a tri-axial inertial sensor (DynaPort MoveMonitor, McRoberts, The Hague, The Netherlands). The inertial sensor was attached to a three-centimetre wide, tight elastic belt and was worn on the back of their trunk at belt height. Participants were instructed to wear the inertial sensor for seven consecutive days and to remove the sensor only during water activities such as bathing, showering and swimming. Participants were included in the analysis when they wore the inertial sensor for at least two days for 18 hours per day. To remove data collected during participants’ transportation back to their home environment after the measurements, we excluded the first six hours and the last six hours from the analysis. The sensor measures accelerations at a frequency of 100 Hz in the vertical (VT), mediolateral (ML) and anteroposterior (AP) direction and has a range of +/− 60 m/s2. For every participant, locomotion episodes that lasted for at least ten seconds were selected from the acceleration signal using the manufacturer’s algorithm (202). Many of the gait characteristics we calculate have a dependence on data series length. By dividing these bouts in epochs of 10 seconds, this effect is negated. All subsequent analyses happened with our own custom Matlab codes (MATLAB R2021b Mathworks, Natrick, MA, USA; and https://github.com/VU-HMS/Gait-Analysis) (203). First, the locomotion episodes were divided in epochs of ten seconds and gait characteristics were calculated for each of these epochs to obtain detailed information about gait dynamics (204). Then, gait quality characteristics were estimated from the trunk accelerometry data. A vast number of gait quality and quantity measures can be calculated from these data, and analysing all possible measures could lead to type-1 errors and a potential source of bias (205). Therefore, a literature search was done in PubMed to find relevant gait quality and quantity measures that are correlated to clinical outcome measures by using the following search string: ((acceleromet*[tiab]) OR (inertial sensor*[tiab]) OR (DMO[tiab])) AND ((functional status[tiab]) OR (SPPB[tiab]) OR (fall*[tiab]) OR (clinical outcome*[tiab])) AND ((pls[tiab]) OR (regression[tiab])). Based on this literature search, it was decided to include the following
RkJQdWJsaXNoZXIy MjY0ODMw