Accelerometer and Global Positioning System Measurement of Recovery of Community Ambulation Across the First 6 Months After Stroke: An Exploratory Prospective Study

Exploring the relationship of clinical walking tests with 8-months inertial measurement unit (IMU)-based real world mobility tracking in stroke and spinal cord injury survivors

BACKGROUND

Mobility is crucial for participation and quality of life in individuals with sensorimotor impairments, yet scientific evidence on its course in real-world settings is limited. So-called wearables for measuring physical activity might help to overcome this knowledge gap allowing daily measurements of mobility. The aim of the present study is to examine the relationship between clinical walking tests and inertial measurement unit-based mobility tracking in the community setting of stroke and spinal cord injury (SCI) survivors.

METHODS

At a single observational time point, the precision of the activity tracker was evaluated in a standardized parcours in healthy subjects and stroke or SCI survivors (n=57). This was followed by a multicenter observational cohort study (n=116 participants), in which the mobility of stroke and SCI survivors was assessed over 8 months immediately after discharge from acute inpatient rehabilitation. Daily distances covered in the community setting were recorded using the activity tracker. Established walking tests-including the 10-meter walk test (10MWT) and the timed up and go test (TUG)-were conducted at baseline, as well as at 4- and 8-month follow up visits. The relationship between daily distances in the ambulatory setting and 10MWT or TUG performance at discrete study visits (baseline, 4 months (midterm), and 8 months (final) after hospital discharge) was analyzed using regression models.

RESULTS

The precision of the activity tracker in measuring covered distance in a standardized parcours varied by mobility type. The highest precision was achieved in manual wheelchair users (deviation from zero: -1.5±1.03% (p=0.15) while the least favorable precision was observed in participants with SCI and significant walking impairment (-14.6±2% (p<0.001). The widely used 10MWT speed showed a relationship with the ambulatory daily distance. The regression coefficients [m/(1m/s)] were: 874 (95% CI: 578-1171) at baseline (p<0.001), 895 (95% CI: 614-1176) at midterm (p<0.001), and 824 (95% CI: 537-1112) at the final visit (p<0.001). Interestingly, in the category of good walkers with the most favorable walking speeds the daily covered distance unmasked distinct subgroups with shorter and longer daily distances.

CONCLUSIONS

For SCI and stroke survivors, especially medium to fast walkers, activity tracking in real-world settings adds valuable insight beyond clinical walking tests. Clinical studies on rehabilitative interventions for mobility improvement should consider real-life daily distance as a key endpoint.

Fusing GPS, Activity Monitors, and Self-Report to Improve Assessment of Walking Activity and Community Participation After Stroke

BACKGROUND AND PURPOSE

Walking and participation in the community are important goals for people post-stroke (PPS). These constructs are challenging to measure given limitations in current data collection methodologies. The purpose of this study was to (1) develop a data fusion approach that combined data from global positioning system (GPS), activity monitor (AM), and daily trip log to identify walking activity and participation in the community, and (2) to examine the construct validity of the data fusion method.

METHODS

At 60 days post-stroke, PPS wore a GPS and AM and completed a daily trip log for 7 days. Using a combination of a density-based spatial clustering algorithm and geocoding GPS, AM, and daily trip log data were time synched and fused to identify total trips taken outside the home; locations visited per trip; number of steps taken in the home, in the community, at each location visited, and in total. Associations between stroke outcomes and the data fusion metrics were determined to support the construct validity of the data fusion method.

RESULTS

Forty-four PPS took a mean of 2,541 steps/day, of which 56% were in the community, and took a mean of 0.39 trips/day outside the home and visited a mean of 0.42 locations. A social visit was the most common reason for going into the community. There were fair associations between number of trips outside the home and gait speed (GS), r = 0.49, Berg Balance Scale (BBS), r = 0.48, modified Rankin Scale (mRS), r = -0.47, and Stroke Impact Scale participation subscale (SIS-P) (0.45). There were moderate associations between steps taken in the community and GS, r = 0.63, BBS, r = 0.51, mRS, r = -0.61, and SIS-P, r = 0.43.

DISCUSSION AND CONCLUSIONS

Participants did not often access their community. Fusing GPS, AM, and trip log data may provide a comprehensive method to identify walking activity and community participation in PPS.

VIDEO ABSTRACT AVAILABLE

for more insights from the authors (see the Video, Supplemental Digital Content, available at: http://links.lww.com/JNPT/A529).

Speed- and Endurance-Based Classifications of Community Ambulation Post-Stroke Revisited: The Importance of Location in Walking Performance Measurement

BACKGROUND

Gait speed or 6-minute walk test are frequently used to project community ambulation abilities post-stroke by categorizing individuals as household ambulators, limited, or unlimited community ambulators. However, whether improved clinically-assessed gait outcomes truly translate into enhanced real-world community ambulation remains uncertain.

OBJECTIVE

This cross-sectional study aimed to examine differences in home and community ambulation between established categories of speed- and endurance-based classification systems of community ambulation post-stroke and compare these with healthy controls.

METHODS

Sixty stroke survivors and 18 healthy controls participated. Stroke survivors were categorized into low-speed, medium-speed, or high-speed groups based on speed-based classifications and into low-endurance, medium-endurance, or high-endurance groups based on the endurance-based classification. Home and community steps/day were quantified using Global Positioning System and accelerometer devices over 7 days.

RESULTS

The low-speed groups exhibited fewer home and community steps/day than their medium- and high-speed counterparts (P < .05). The low-endurance group took fewer community steps/day than the high-endurance group (P < .05). Despite vast differences in clinical measures of gait speed and endurance, the medium-speed/endurance groups did not differ in their home and community steps/day from the high-speed/endurance groups, respectively. Stroke survivors took 48% fewer home steps/day and 77% fewer community steps/day than healthy controls.

CONCLUSIONS

Clinical classification systems may only distinguish home ambulators from community ambulators, but not between levels of community ambulation, especially beyond certain thresholds of gait speed and endurance. Clinicians should use caution when predicting community ambulation status through clinical measures, due to the limited translation of these classification systems into the real world.

Geographic intelligence to investigate community participation: a scoping review

PURPOSE

To understand the applicability and unique contributions of global positioning system (GPS) and geographic information system (GIS) technologies to investigate individuals' participation in the community.

METHODS

A scoping review was conducted to summarize studies that employed GPS and/or GIS to assess community participation. Systematic literature searches were performed using nine electronic databases, up to April 2022. Two independent reviewers screened studies for eligibility and extracted data from the selected studies.

RESULTS

The searches retrieved 628 articles, of which 36 met the inclusion criteria. In 31 studies, a GPS unit or GPS-enabled device monitored the community mobility of individuals with different health conditions. Tracking periods varied from five hours to 30 consecutive days. The spatiotemporal parameters obtained from satellite-based data provided information about individuals' presence in the community. Most studies combined GPS with other measuring tools (self-report diaries/questionnaires, qualitative interviews) to capture a broader description of community participation.

CONCLUSIONS

GPS and GIS are viable approaches for advancing research as they provide unique information about community participation not easily captured by other methods. The combination of available methods comprehensively address the physical and social dimensions of this construct.IMPLICATIONS FOR REHABILITATIONGlobal positioning system (GPS) and geographic information system (GIS) technology allows direct and real-time quantification of patterns of individuals' community mobility.The geographic area where individuals participate in daily living activities complements information from self-report diaries/questionnaires.Combining GPS technology with self-report diaries/questionnaires allows a broader description of individuals' community participation, addressing respectively both its physical and social dimensions.Knowledge of individuals' access to and involvement in community locations/situations may help therapists design interventions to improve community participation.

Sympathetic nervous system responses during complex walking tasks and community ambulation post-stroke

Stroke survivors frequently report increased perceived challenge of walking (PCW) in complex environments, restricting their daily ambulation. PCW is conventionally measured through subjective questionnaires or, more recently, through objective quantification of sympathetic nervous system activity during walking tasks. However, how these measurements of PCW reflect daily walking activity post-stroke is unknown. We aimed to compare the subjective and objective assessments of PCW in predicting home and community ambulation. In 29 participants post-stroke, we measured PCW subjectively with the Activities-specific Balance Confidence (ABC) Scale and objectively through electrodermal activity, quantified by change in skin conductance levels (SCL) and skin conductance responses (SCR) between outdoor-complex and indoor-steady-state walking. High-PCW participants were categorized into high-change SCL (ΔSCL ≥ 1.7 μs), high-change SCR (ΔSCR ≥ 0.2 μs) and low ABC (ABC < 72%) groups, while low-PCW participants were categorized into low-change SCL (ΔSCL < 1.7 μs), low-change SCR (ΔSCR < 0.2 μs) and high-ABC (ABC ≥ 72%) groups. Number and location of daily steps were quantified with accelerometry and Global Positioning System devices. Compared to low-change SCL group, the high-change SCL group took fewer steps in home and community (p = 0.04). Neither ABC nor SCR groups differed in home or community steps/day. Objective measurement of PCW via electrodermal sensing more accurately represents home and community ambulation compared to the subjective questionnaire.

Risk factors and ambulatory outcome in ischemic stroke patients with pre-stroke depression

OBJECTIVE

It is well known that post-stroke depression might be a negative factor for stroke recovery, however there is limited evidence to establish the link between pre-stroke depression and stroke outcome such motor recovery. The objective is to determine clinical risk factors in ischemic stroke patients with pre-stroke depression that are associated functional ambulatory outcome.

METHODS

Data from acute ischemic patients from a regional stroke registry were collected for consecutive recombinant tissue plasminogen activator(rtPA)-treated acute ischemic stroke patients between January 2010 and June 2016. Logistic regression model was used to predict risk factors that served as predictive variables, while the increase or reduce odds of improvement in ambulatory outcome was considered as the primary outcome. Multicollinearity and possible interactions among the independent variables were analyzed using the variance inflation factor.

RESULTS

A total of 1446 patients were eligible for recombinant tissue plasminogen activator (rtPA) and 596 of these patients received rtPA. Of the 596 ischemic stroke patients, 286 patients presented with recent pre-stroke depression, 310 had no pre-stroke depression. Carotid artery stenosis (OR = 11.577, 95% CI, 1.281-104.636, P = 0.029) and peripheral vascular disease (OR = 18.040, 95% CI, 2.956-110.086, P = 0.002) were more likely to be associated with increase odds of improvement in ambulation in patients with no pre-stroke depression treated with rtPA, while antihypertensive medications (OR = 0.192, 95% CI, 0.035-1.067, P = 0.050),previous TIA (OR = 0.177, 95% CI, 0.038-0.818, P = 0.027), and congestive heart failure (OR = 0. 0.160, 95% CI, 0.030-0.846, P = 0.031) were associated with reduced odds of improvement in ambulation. In addition, carotid artery stenosis (OR = 0.078, 95% CI, 0.10-0.614, P = 0.015, congestive heart failure (OR = 0.217, 95% CI, 0.318-0.402, P = 0.030), previous TIA (OR = 0.444, 95% CI, 0.517-0.971, P = 0.012), higher NIHSS scores ((OR = 0.887, 95% CI, 0.830-0.948, P < 0.001), and antihypertensive medications (OR = 0.810, 95% CI, 0.401-0.529, P = 0.019) were associated with the reduced odd of improvement in ambulation in an ischemic stroke population with pre-stroke depression treated with rtPA.

CONCLUSION

Our findings indicate that more risk factors were associated with the decreased odds of an improvement in ambulation following thrombolytic therapy in an ischemic stroke population with pre-stroke depression when compared with those without pre-stroke depression. This finding maybe helpful in the development of management strategies to increase the use of thrombolytic therapy for pre-stroke depressed ischemic stroke to increased their eligibility for rtPA.

High-intensity treadmill training and self-management for stroke patients undergoing rehabilitation: a feasibility study

BACKGROUND

Physical activity undertaken by stroke survivors is generally low. This trial investigated the feasibility of delivering a high-intensity treadmill and self-management program to people with stroke undergoing inpatient rehabilitation and determine whether physical activity, walking ability and cardiorespiratory fitness could be increased.

METHOD

A phase I, single-group, pre-post intervention study was conducted with stroke survivors undergoing inpatient rehabilitation who could walk. Participants undertook a high-intensity treadmill and self-management program for up to 30 min, three times a week for 8 weeks under the supervision of their usual physiotherapist. Feasibility was determined by examining compliance, satisfaction and adverse events. Clinical outcomes were amount of physical activity, walking ability, and cardiorespiratory fitness collected pre-training (week 0), post-training (week 8), and at follow-up (week 26).

RESULTS

Forty stroke survivors participated, completing 10 (SD 6) sessions, 94% at the specified training intensity, with high satisfaction and no adverse events related to the intervention. At week 8, participants completed 2749 steps/day (95% CI 933 to 4564) more physical activity than at week 0. Walking distance increased by 110 m (95% CI 23 to 196), walking speed by 0.24 m/s (95% CI 0.05 to 0.42), and VO2 peak by 0.29 ml/kg/min (95% CI 0.03 to 0.56). At week 26, increases in physical activity, walking distance and speed, and cardiorespiratory fitness were maintained.

CONCLUSIONS

A high-intensity treadmill training program embedded within a self-management approach during inpatient rehabilitation appears feasible and potentially may offer sustained improvements in physical activity, walking ability, fitness, and quality of life. A randomised trial is warranted.

TRIAL REGISTRATION

This feasibility study was registered with the Australian New Zealand Clinical Trials Registry ( ACTRN12613000764730 ).

Goal attainment in mobility after acute rehabilitation of mobility-restricting paralysis syndromes with regard to the ambulatory therapeutic level of participation NeuroMoves : A German national multicenter observational cohort study

Background

A central goal of rehabilitation in patients with paralysis syndromes after stroke or spinal cord injury (SCI) is to restore independent mobility as a pedestrian or wheelchair user. However, after acute rehabilitation, the mobility frequently deteriorates in the ambulatory setting, despite the delivery of rehabilitative interventions such as physical therapy or the prescription of assistive devices. The aim of the NeuroMoves study is to identify factors that are associated with changes of mobility in the ambulatory setting after acute inpatient rehabilitation, with a particular focus on participation according to the ICF (International Classification of Functioning, Disability and Health).

Methods

The NeuroMoves study is intended as a national multicenter observational cohort study with 9 clinical sites in Germany. A total of 500 patients with mobility-restricting paralysis syndromes (i.e. stroke or SCI) are to be recruited during acute inpatient rehabilitation prior to discharge to the ambulatory setting. Patients will have 8 months of follow-up in the ambulatory setting. Three study visits at the clinical sites (baseline, midterm, and final) are planned at 4-months intervals. The baseline visit is scheduled at the end of the acute inpatient rehabilitation. During the visits, demographical data, neurological, functional, quality of life, and implementation measures will be assessed.

At baseline, each study participant receives an activity tracker (sensor for recording ambulatory mobility) along with a tablet computer for home use over the 8 months study duration. While mounted, the activity tracker records mobility data from which the daily distance covered by walking or wheelchair use can be calculated. Customized applications on the tablet computer remind the study participants to answer structured questionnaires about their health condition and treatment goals for physical therapy. Using the study participants’ tablet, therapists will be asked to answer structured questionnaires concerning treatment goals and therapeutic measures they have applied. The primary analysis concerns the association between mobility (daily distance covered) and the degree of participation-oriented rehab interventions. Further exploratory analyses are planned.

Discussion

The findings could inform healthcare decision-making regarding ambulatory care in Germany focusing on mobility-promoting interventions for patients with mobility-restricting paralysis syndromes.

Study registration

German Clinical Trials Register, DRKS-ID: DRKS00020487 (18.02.2020).

Supplementary Information

The online version contains supplementary material available at 10.1186/s12883-021-02167-y.

Examining the relationship between community mobility and participation using GPS and self-report data

Higher levels of community mobility have been shown to be associated with better physical health, mental health, and quality of life. The ability to move about one's community is also expected to facilitate community participation, which is an aspect of health functioning. This study uses Global Positioning Systems (GPS) technology to track various dimensions of community mobility, such as destinations, time outside the home, and distance traveled, and examine the relationship between these variables and community participation in a sample of individuals with serious mental illnesses (SMI). This population was selected because they are known to have diminished health functioning in terms of their community participation, and the goal is to explore the extent to which mobility limitations may account for this. A total of 103 individuals with serious mental illnesses were recruited from mental health agencies and consented to having their mobility tracked using GPS for 13 days and answering questions about their community-based activities. Greater amount of participation was associated with having more destinations and spending more time out of the house, but not with traveling larger distances and having a greater activity space. None of the mobility variables were related to the number of important participation areas or sufficiency of participation. The findings support the hypothesis that greater mobility is related to more participation, although satisfaction with the degree to which one participates does not appear to be impacted, suggesting that other factors need to be accounted for. Health policymakers and providers should pay attention to community mobility as a factor that affects health outcomes such as participation, in individuals with serious mental illnesses, and other populations. In particular, attending to access to personal transport, public transportation, and other mobility options appears to be important, as well as interventions aimed at encouraging greater community mobility.

Stepping After Stroke: Walking Characteristics in People With Chronic Stroke Differ on the Basis of Walking Speed, Walking Endurance, and Daily Steps

BACKGROUND

What contributes to free-living walking after stroke is poorly understood. Studying the characteristics of walking may provide further details that guide interventions.

OBJECTIVE

The objectives of this study were to examine how the walking characteristics of bouts per day, median steps per bout, maximum steps per bout, and time spent walking differ in individuals with various walking speeds, walking endurance, and daily steps and to identify cutoffs for differentiating ambulators who were active versus inactive.

DESIGN

This study involved a cross-sectional analysis of data from the Locomotor Experience Applied Post-Stroke trial.

METHODS

Participants were categorized by walking speed, walking endurance (via the 6-minute walk test), and daily steps (via 2 consecutive days of objective activity monitoring). Differences in walking characteristics were assessed. Linear regression determined which characteristics predicted daily step counts. Receiver operating characteristic curves and areas under the curve were used to determine which variable was most accurate in classifying individuals who were active (≥5500 daily steps).

RESULTS

This study included 252 participants with chronic stroke. Regardless of categorization by walking speed, walking endurance, or daily steps, household ambulators had significantly fewer bouts per day, steps per bout, and maximum steps per bout and spent less time walking compared with community ambulators. The areas under the curve for maximum steps per bout and bouts per day were 0.91 (95% confidence interval = 0.88 to 0.95) and 0.83 (95% confidence interval = 0.78 to 0.88), respectively, with cutoffs of 648 steps and 53 bouts being used to differentiate active and inactive ambulation.

LIMITATIONS

Activity monitoring occurred for only 2 days.

CONCLUSIONS

Walking characteristics differed based on walking speed, walking endurance, and daily steps. Differences in daily steps between household and community ambulators were largely due to shorter and fewer walking bouts. Assessing and targeting walking bouts may prove useful for increasing stepping after stroke.

Factors which Influence Risk Taking and the Evolution of Social-Identity in Stroke Narratives: A Thematic Synthesis

Background: The disruption of a stroke can impact an individual’s sense of social identity. A comprehensive review is required to understand the factors and processes that influence changes in social identity following a stroke. Aims: To undertake a review of literature to discover a process of social identity evolution for people with stroke and identify the factors which influence it. Methods: A meta-ethnographic approach to review was undertaken and a subtle realist viewpoint was assumed. Studies were included if they documented experiences and perceptions relating to stroke. Eight electronic databases were searched from January 2009 until January 2019. Quality assessment and synthesis techniques were applied. Findings: Out of the 18 papers included, a total of 251 (141/251, 56% male, 109/251, 43% female, 1/251, 0.4% undisclosed) individuals were included within the synthesis. The evolution of social-identity model was developed and identified with five key stages to represent a process that individuals with stroke can experience. Factors which influence the process were identified and direct implications for clinical practice are given. Conclusion: This review has highlighted the major themes within the evolution of social identity and management strategies for risk taking to achieve a desired future. Further research is required to consider how these findings may be tested in clinical practice.

Which impairments, activity limitations and personal factors at hospital discharge predict walking activity across the first 6 months poststroke?

Purpose: To determine which impairments, activity limitations and personal factors at hospital discharge poststroke predict volume, frequency, and intensity of walking activity 1, 3, and 6 months later.Materials and Methods: Prospective longitudinal observational study. Thirty-six people with stroke (71 SD 14 years, 69% male) were recruited at hospital discharge and predictors including fatigue, mood, executive function, walking speed, walking endurance, age, prestroke activity, self-efficacy, and perceived stroke recovery and health were collected. At 1, 3, and 6 months follow-up, participants wore an ActivPAL™ accelerometer to collect measures of walking activity.Results: At 1 month, walking endurance predicted all walking activity (R² > 0.29, p < 0.01). At 3 months, walking endurance and prestroke activity predicted activity volume and intensity (R² = 0.46-0.61, p < 0.001), and prestroke activity predicted activity frequency (R² = 0.31, p = 0.004). At 6 months, age-predicted activity volume and frequency (R² = 0.34-0.35, p < 0.003), while prestroke activity, discharge walking endurance, and executive function together predicted activity intensity (R² = 0.79, p < 0.001).Conclusion: Walking endurance contributes to walking activity outcomes across the first 6 months following hospital discharge poststroke. After 1 month of discharge, factors other than poststroke changes also contribute to activity outcomes, and should be considered when targeting poststroke physical activity.Implications for rehabilitationWalking endurance should be addressed during stroke rehabilitation as higher scores are linked to more walking activity in the first month after discharge.Prestroke factors such as low prestroke activity levels and older age predict reduced walking activity after stroke, so approaches to address barriers these factors may pose are needed in people with stroke.Physical activity interventions should be tailored to the individual, their environment, and context, and take into consideration prestroke factors.

Functional Outcome Measures of Recombinant Tissue Plasminogen Activator-Treated Stroke Patients in the Telestroke Technology

The efficiency of telestroke programs in improving the rates of recombinant tissue plasminogen activator (rtPA) in stroke patients has been reported. Previous studies have reported favorable treatment outcomes with the use of telestroke programs to improve the use of rtPA, but functional outcomes are not fully understood. This study investigated the effect of telestroke technology in the administration of rtPA and related functional outcomes associated with baseline clinical variables. Retrospective data of a telestroke registry were analyzed. Univariate analysis was used to compare demographic and clinical variables in the rtPA group and the no rtPA group and between the improved functional ambulation group and the no improvement group. A stepwise binary logistic regression identified factors associated with improved functional outcome in the total telestroke population and in the subset of the telestroke population who received rtPA. In adjusted analysis and elimination of any multicollinearity for patients who received rtPA in the telestroke setting, obesity (odds ratio [OR] = 2.138, 95% confidence interval [CI], 1.164-3.928, P < .05), higher systolic blood pressure at the time of presentation (OR = 1.015, 95% CI, 1.003-1.027, P < .05), and baseline high-density lipoprotein at the time of admission (OR = 1.032, 95% CI, 1.005-1.059, P < .05) were associated with improved functional outcomes. Increasing age (OR = 0.940, 95% CI, 0.916-0.965, P < .0001) and higher calculated National Institutes of Health Stroke Scale (OR = 0.903, 95% CI, 0.869-0.937) were associated with a poorer outcome in rtPA-treated patients. Telestroke technology improves functional outcomes at spoke stations where neurological expertise is unavailable. Further studies are necessary to determine how telestroke technology can be optimized, especially to improve contraindications and increase eligibility for thrombolysis therapy.

Sleep Duration, Sedentary Behavior, Physical Activity, and Quality of Life after Inpatient Stroke Rehabilitation

OBJECTIVE

The aim of this study was to describe accelerometer-derived sleep duration, sedentary behavior, physical activity, and quality of life and their association with demographic and clinical factors within the first month after inpatient stroke rehabilitation.

MATERIALS AND METHODS

Thirty people with stroke (mean ± standard deviation, age: 63.8 ± 12.3 years, time since stroke: 3.6 ± 1.1 months) wore an activPAL3 Micro accelerometer (PAL Technologies, Glasgow, Scotland) continuously for 7 days to measure whole-day activity behavior. The Stroke Impact Scale and the Functional Independence Measure were used to assess quality of life and function, respectively.

RESULTS

Sleep duration ranged from 6.6 to 11.6 hours/day. Fifteen participants engaged in long sleep greater than 9 hours/day. Participants spent 74.8% of waking hours in sedentary behavior, 17.9% standing, and 7.3% stepping. Of stepping time, only a median of 1.1 (interquartile range: .3-5.8) minutes were spent walking at a moderate-to-vigorous intensity (≥100 steps/minute). The time spent sedentary, the stepping time, and the number of steps differed significantly by the hemiparetic side (P < .05), but not by sex or the type of stroke. There were moderate to strong correlations between the stepping time and the number of steps with gait speed (Spearman r = .49 and .61 respectively, P < .01). Correlations between accelerometer-derived variables and age, time since stroke, and cognition were not significant.

CONCLUSIONS

People with stroke sleep for longer than the normal duration, spend about three quarters of their waking hours in sedentary behaviors, and engage in minimal walking following stroke rehabilitation. Our findings provide a rationale for the development of behavior change strategies after stroke.

Activity Recognition for Persons With Stroke Using Mobile Phone Technology: Toward Improved Performance in a Home Setting

BACKGROUND

Smartphones contain sensors that measure movement-related data, making them promising tools for monitoring physical activity after a stroke. Activity recognition (AR) systems are typically trained on movement data from healthy individuals collected in a laboratory setting. However, movement patterns change after a stroke (eg, gait impairment), and activities may be performed differently at home than in a lab. Thus, it is important to validate AR for gait-impaired stroke patients in a home setting for accurate clinical predictions.

OBJECTIVE

In this study, we sought to evaluate AR performance in a home setting for individuals who had suffered a stroke, by using different sets of training activities. Specifically, we compared AR performance for persons with stroke while varying the origin of training data, based on either population (healthy persons or persons with stoke) or environment (laboratory or home setting).

METHODS

Thirty individuals with stroke and fifteen healthy subjects performed a series of mobility-related activities, either in a laboratory or at home, while wearing a smartphone. A custom-built app collected signals from the phone's accelerometer, gyroscope, and barometer sensors, and subjects self-labeled the mobility activities. We trained a random forest AR model using either healthy or stroke activity data. Primary measures of AR performance were (1) the mean recall of activities and (2) the misclassification of stationary and ambulatory activities.

RESULTS

A classifier trained on stroke activity data performed better than one trained on healthy activity data, improving average recall from 53% to 75%. The healthy-trained classifier performance declined with gait impairment severity, more often misclassifying ambulatory activities as stationary ones. The classifier trained on in-lab activities had a lower average recall for at-home activities (56%) than for in-lab activities collected on a different day (77%).

CONCLUSIONS

Stroke-based training data is needed for high quality AR among gait-impaired individuals with stroke. Additionally, AR systems for home and community monitoring would likely benefit from including at-home activities in the training data.

Using a Real-Time Location System for Assessment of Patient Ambulation in a Hospital Setting

OBJECTIVE

To assess the feasibility of using an infrared-based Real-Time Location System (RTLS) for measuring patient ambulation in a 2-minute walk test (2MWT) by comparing the distance walked and the Johns Hopkins Highest Level of Mobility (JH-HLM) score to clinician observation as a criterion standard.

DESIGN

Criterion standard validation study.

SETTING

Inpatient, university hospital.

PARTICIPANTS

Patients (N=25) in an adult neuroscience/brain rescue unit.

INTERVENTIONS

Not applicable.

MAIN OUTCOME MEASURES

RTLS and clinician-reported ambulation distance in feet, and JH-HLM score on an 8-point ordinal scale.

RESULTS

The RTLS ambulation distance for the 25 patients in the 2MWT was between 68 and 516ft. The mean difference between clinician-reported and RTLS ambulation distance was 8.4±11.7ft (2.7%±4.6%). The correlation between clinician-reported and RTLS ambulation distance was 97.9% (P<.01). The clinician-reported ambulation distance for 2 patients was +100ft and -99ft compared with the RTLS distance, implying clinician error in counting the number of laps (98ft). The correlation between the RTLS distance and clinician-reported distance excluding these 2 patients is 99.8% (P<.01). The accuracy of the RTLS for assessment of JH-HLM score for all 25 patients was 96%. The average patient speed obtained from RTLS data varied between 0.4 and 3.0mph.

CONCLUSIONS

The RTLS is able to accurately measure patient ambulation and calculate JH-HLM for a 2MWT when compared with clinician observation as the criterion standard.