Matthias Tietsch

Advisors

Dr. Felix Kluge, Martin Ullrich (M.Sc.), Prof. Dr. Björn Eskofier

Duration
10/2019 – 03/2020

Abstract
Gait speed is thought to be a proxy for fitness, quality of life and ultimately survival in older adults [1], further Studenski et al. argue that a change in gait speed of 0.1 m/sec is medically meaningful. However, measuring gait speed in short clinical tests may not be representative for a subject’s habitual gait. Therefore, it is desirable to measure habitual gait speed in the subject’s real world (e.g. home and outside, often termed “free-living”) rather than in clinical walk tests. On the other hand, measuring real-world gait speed with inertial sensor units (IMU) is not straight forward as complex signals need to be filtered and converted into gait speed estimates, eventually with reliable step detection as a prerequisite. Transformation of the IMU signal for interpretation and/or modelling of gait speed is easier when the sensor is worn on the body’s center of mass e.g. as a belt buckle [2] compared to positions such as the wrist with more dynamic and unpredictable movements (arm ambulation vs hands in the pocket vs non-walking movements of the arms). Transformation of 3-axial acceleration is complex and may require filtering and analytical steps to derive signals suitable for gait speed calculations or modelling (see [3] for a specific analytical approach and [4] for a tutorial).

In addition to technical and analytical challenges human factors such as wearing comfort and appeal influence the subject’s compliance wearing the sensor, as well as robustness and ease of handling, and some of these factors may be population dependent. The recommended wearing position can impact compliance, e.g. the wrist or waist is more comfortable than the ankle or foot which would be best for measuring gait [5], and a good compromise of signal periodicity, stability and wearing comfort might be a sensor position around the waist [6].

Continuous monitoring of mobility and walking speed in particular are becoming important end points in clinical studies to demonstrate functional impact of therapies in many indications such as in the neuroscience (e.g. PD [7]) or muscular skeletal disease fields.

In a recent interventional study from Novartis with a muscle growing agent in a frail population physical performance end points (e.g. clinical walk tests) were measured. In addition, real-world walking and gait speed as functional end points were monitored with a center of mass worn accelerometer together with an algorithm to derive steps and gait speed via supervised modelling (unpublished).

In current and upcoming studies for which mobility might be monitored as a functional end point Novartis wants to use IMUs worn at any positions around the waist rather than requiring the patient to wear the sensor at the center of mass. To achieve comparable accuracy of step and speed estimation from these alternative positions our existing algorithm needs to be tested on different positions around the waist, and a new algorithm may need to be developed (potentially based on the principals of our existing method) to cope with additional noise and position-induced asymmetries.

References:

1. Studenski, S et al. “Gait speed and survival inolder adults”. JAMA. 2011 Jan 5;305(1):50-8. doi: 10.1001/jama.2010.1923. PubMedPMID: 21205966; PubMed Central PMCID: PMC3080184.
2. Schimpl M, Lederer C, Daumer M. “Development and validation of a new method tomeasure walking speed in free-living environments using the actibelt® platform.”PLoS One. 2011;6(8):e23080. doi: 10.1371/journal.pone.0023080. Epub 2011 Aug 5.PubMed PMID: 21850254; PubMed Central PMCID: PMC3151278.
3.  Sabatini AM, Ligorio G, Mannini A. “Fourier-based integration of quasi-periodicgait accelerations for drift-free displacement estimation using inertial sensors”.Biomed Eng Online. 2015 Nov 23;14:106. doi: 10.1186/s12938-015-0103-8. PubMedPMID: 26597696; PubMed Central PMCID: PMC4657361.
4.  Del Din S, Hickey A, Ladha C et al. “Instrumented gait assessment with a single wearable: an introductory tutorial”l [version 1; referees: 1 approved, 1 approved with reservations]. F1000Research 2016, 5:2323
   (https://doi.org/10.12688/f1000research.9591.1)
5. Simpson LA, et al. “Capturing step counts at slow walking speeds in older adults:comparison of ankle and waist placement of measuring device”. J Rehabil Med. 2015Oct 5;47(9):830-5. doi: 10.2340/16501977-1993. PubMed PMID: 26181670.
6. Rosenberger, M et al. “Estimating activity and sedentary behavior from an accelerometer on the hip or wrist”. Medicine and science in sports and exercise, 2013; 45(5), 964-75.
7. Del Din S, Godfrey A, Mazzà C, Lord S, Rochester L. “Free-living monitoring ofParkinson’s disease: Lessons from the field”. Mov Disord. 2016 Sep;31(9):1293-313.doi: 10.1002/mds.26718. Epub 2016 Jul 25. Review. PubMed PMID: 27452964.