Murdoch University Research Repository

Welcome to the Murdoch University Research Repository

The Murdoch University Research Repository is an open access digital collection of research
created by Murdoch University staff, researchers and postgraduate students.

Learn more

Accelerometer based free-living data: Does macro gait behaviour differ between fallers and non-fallers with and without Parkinson's disease?

Del Din, S., Godfrey, A., Galna, B.ORCID: 0000-0002-5890-1894, Dockx, K., Pelosin, E., Reelick, M., Mirelman, A., Hausdorff, J. and Rochester, L. (2015) Accelerometer based free-living data: Does macro gait behaviour differ between fallers and non-fallers with and without Parkinson's disease? In: 2015 International Society for Posture and Gait Research (ISPGR) World Congress, 28 June - 2 July, 2015, Seville, Spain.



BACKGROUND AND AIM: Gait impairment and falls are frequent among older adults and people with Parkinson's disease (PD), and may lead to loss of functional independence and poor quality of life. Current approaches for evaluating falls risk are based on self-report or testing at a given time point and therefore may be suboptimal. Continuous monitoring of gait is emerging as a powerful tool to assess motor impairment and falls risk in real life using accelerometer-based technology, potentially providing an accurate and objective measure of risk [1]. Macro level gait behaviours (e.g., volume, pattern, and variability of walking bouts) are sensitive to PD pathology [2], however, there are conflicting reports about their association with falls risk. The aim of this study was to explore the association between physical activity (PA) and falls history by analysing whether macro level gait behaviour differs between fallers and non-fallers with and without PD using 7 day accelerometer-based free-living data.

METHODS: 227 fallers (F: 106 elderly, 121 PD; age: 76±6 yrs, and 72±6 yrs, respectively) enrolled in the V-TIME study [3], who fell twice or more in the 6 months prior to assessment, together with 65 participants without a history of falls (NF: 50 elderly, 15 PD, age: 65±9 yrs, 70±7 yrs, respectively) enrolled into ICICLE-GAIT [2] were tested. Data were recorded continuously for 7 days with a tri-axial accelerometer (Axivity AX3, UK, 100Hz, ±8g) placed on the low back (L5). Macro level outcomes (MLO) representing the volume (% walking time, number of steps, mean bout length), pattern (alpha (α)), and variability (S2) of free-living activity were extracted in MATLAB® (R2012a) [2]. General linear modelling examined the effect of fall history (F vs NF) and pathology (PD vs elderly) on MLO, controlling for age, sex and BMI.

RESULTS: Although the % walking time and number of steps was not related to fall history, F tended to walk in shorter bouts (p=.004) and had a less variable walking pattern (lower S2, p=.019) compared to NF. PD spent less time walking (p=.002), took fewer steps (p=.002), and accumulated proportionally more steps in shorter bouts (higher α) compared to the elderly (p=.006), regardless of falls history. There were no interactions between pathology and falls history for any of the outcomes.

CONCLUSIONS: Our results showed that there is an association between falls history and PA. Volume-based MLO, pattern and variability of the walking bouts derived from free-living accelerometer-based data are independently associated with a history of falls and PD. These results support the use of a single accelerometer-based sensor to assess falls risk in free living settings, however, future work is needed to confirm if MLO can predict falls, potentially guiding clinical decision making.

REFERENCES: [1] Lord S et al., Mov Disord, 2013; 28(11):1534-43 [2] Lord S et al., J Neurol, 2013; 260(12):2964-72 [3] Mirelman A et al., BMC Neurol, 2013;13:15

Item Type: Conference Item
Conference Website:
Item Control Page Item Control Page