BACKGROUND
The availability of low cost ubiquitous wearable sensors has enabled researchers, in recent years, to collect a large volume of data in various domains including healthcare. The goal has been to harness wearables to further investigate human activity, physiology and functional patterns. As such, on-body sensors have been primarily used in healthcare domain to help predict adverse outcomes such as hospitalizations or fall, thereby enabling clinicians to develop better intervention guidelines and personalized models of care to prevent harmful outcomes.
In the previous studies [9,10] and the patent application [11], we introduced a generic framework (Sensing At-Risk Population) that draws on the classification of human movements using a 3-axial accelerometer and extraction of indoor localization using BLE beacons, in concert. This work is to address the longitudinal analyses of a particular cohort using the introduced framework in a skilled nursing facility.
OBJECTIVE
(a) To observe longitudinal changes of physical activity and indoor localization features of rehabilitation-dwelling patients, (b) to assess if such changes can be used at early stages during the rehabilitation period to discriminate between patients that will be re-hospitalized versus the ones that will be discharged to a community setting and (c) to investigate if the sensor based longitudinal changes can imitate patients changes captured by therapist assessments over the course of rehabilitation.
METHODS
Pearson correlation was used to compare occupational therapy (OT) and physical therapy (PT) assessments with sensor-based features. Generalized Linear Mixed Model was used to find associations between functional measures with sensor based features.
RESULTS
Energy intensity at therapy room was positively associated with transfer general (β=0.22;SE=0.08;p<.05). Similarly, sitting energy intensity showed positive association with transfer general (β=0.16;SE=0.07;p<.05). Laying down energy intensity was negatively associated with hygiene grooming (β=-0.27;SE=0.14;p<.05). The interaction of sitting energy intensity with time (β=-0.13;SE=.06;p<.05) was associated with toileting general. Dressing lower body was strongly correlated with overall energy intensity (r = 0.66), standing energy intensity (r = 0.61), and laying down energy intensity (r = 0.72) on the first clinical assessment session.
CONCLUSIONS
This study demonstrates that a combination of indoor localization and physical activity tracking produces a series of features, a subset of which can provide crucial information on the storyline of daily and longitudinal activity patterns of rehabilitation-dwelling patients.
[1]
Majid Sarrafzadeh,et al.
A Combination of Indoor Localization and Wearable Sensor–Based Physical Activity Recognition to Assess Older Patients Undergoing Subacute Rehabilitation: Baseline Study Results
,
2019,
JMIR mHealth and uHealth.
[2]
Majid Sarrafzadeh,et al.
Low-cost indoor health monitoring system
,
2016,
2016 IEEE 13th International Conference on Wearable and Implantable Body Sensor Networks (BSN).
[3]
S. Intille,et al.
Estimating activity and sedentary behavior from an accelerometer on the hip or wrist.
,
2013,
Medicine and science in sports and exercise.
[4]
Alireza Atri,et al.
An Overview of Longitudinal Data Analysis Methods for Neurological Research
,
2011,
Dementia and Geriatric Cognitive Disorders Extra.
[5]
Der-Chiang Li,et al.
A learning method for the class imbalance problem with medical data sets
,
2010,
Comput. Biol. Medicine.
[6]
R. Newton,et al.
Ability of Functional Independence Measure to accurately predict functional outcome of stroke-specific population: systematic review.
,
2010,
Journal of rehabilitation research and development.
[7]
Joseph G. Ibrahim,et al.
Missing data methods in longitudinal studies: a review
,
2009
.
[8]
N. Breslow,et al.
Approximate inference in generalized linear mixed models
,
1993
.