Tsang, K.
Hiremath, S.
Ding, D.
Project title: 
Interactive Exercise Technologies and Exercise Physiology for People with Disabilities - Predicting Energy Expenditure in Manual Wheelchair Users with Spinal Cord Injury
Project Number: 
Tool type: 
Tool class: 
Engineering tool
Disability targeted: 
Spinal cord injury
Study target: 
Manual wheelchair users
Study purpose or goal: 
To provide a more accurate estimate of energy expenditure in manual wheelchair users. This will help users to monitor their own physical activity, which in turn promote behavioral changes and develop/maintain a healthier lifestyle.
Who administers this tool?: 
Ease of use: 
Sensitive issues: 
The age, weight and height of the users are required
Are any approvals required?: 
No approvals are required
How is it administered?: 
In order to use this application, you should have access to the raw data collected by the SenseWear armband typically through the SenseWear Professional Software. You should save the raw data in excel file before running analysis using our app.
Development background: 
The use of micro-electromechanical systems in the form of activity monitors to track daily physical activity (PA) levels has become very popular, especially in ambulatory population. Studies have shown that activity monitors are accurate in tracking energy expenditure; these devices also showed potentials in encouraging users to develop/keep a healthier lifestyle by promoting regular PA. However, manual wheelchair users (MWUs) are not benefited from this technology; most of the devices such as the SenseWear are designed to track lower limb movements, which is ineffective for MWUs who rely on their upper limbs as a primary means of mobility for daily tasks. The app is developed to help users to accurately track their PA while using the SenseWear.
Development methodology: 
Our research team has developed two types of custom energy expenditure models, i.e. the general and the activity-specific models. The general model can be used when you do not have a log of activities performed by the wheelchair user. The activity-specific model can be used for four types of activities including resting while seated, light-weight deskwork (e.g., computer operation, reading etc.), wheelchair propulsion, and arm-ergometry exercise. Both models were developed using data collected from 45 manual wheelchair users with SCI. Participants were asked to wear a SenseWear armband and a portable metabolic cart (gas analyzer) while performing the four activities mentioned above in the laboratory setting.
Can this development process be used elsewhere?: 
Specifications met?: 
The energy expenditure models (or algorithms) had been tested for accuracy. It is about 35% more accurate than the SenseWear
Has the design been tested for safety and reliability?: 
The energy expenditure models (or algorithms) had been tested for reliability
Can this tool be used for other purposes/populations?: 
The tool is only validated in manual wheelchair users with spinal cord injury
Data analysis: 
Is complete
Both the general and activity-specific models were developed and tested among manual wheelchair users with SCI. We do not know how the models will perform when they are applied to manual wheelchair users with other diagnoses.
It is not applicable to commercial activity monitors other than the SenseWear Armband.
The two custom models with another group of 45 manual wheelchair users with spinal cord injury who performed a wide range of lifestyle and sporting based activities in both laboratory and home/community settings. Overall, the general model has an error of -2.8±26.1%, while the activity-specific models have MSE of -4.8±25.4% when compared to the portable metabolic cart readings. Both the general model and the activity-specific models showed strong agreement when compared to the gold standard, ICC(3,1) were 0.862 and 0.831 respectively.
The two custom models, general and activity-specific, used in the app are accurate and reliable in estimating energy expenditure in manual wheelchair users with spinal cord injury.
It's specific to manual wheelchair users with spinal cord injury
Research methods: 
The data estimated by the tool and the data collected by the metabolic cart (gold standard) were compared. The mean signed percent error (MSE) was computed to investigate the accuracy of the tool.
In addition, the intra-class correlation coefficient (ICC) of two-way mixed single measures with absolute agreement was calculated to evaluate the agreement between the data estimated by our tool and by the metabolic cart.
Impact of these findings on the field: 
This tool provides a more accurate estimation of energy expenditure in manual wheelchair users which may facilitate other related research. We will continue improving this tool to promote a healthier lifestyle among manual wheelchair users; and hopefully it can be used as a clinical tool to help clinicians inform any physical activity treatments or exercise programs prescribed.
Peer review status: 
The results were reported in a Journal of Rehabilitation Research & Development (JRRD) manuscript titled "Evaluation of Custom Energy Expenditure Modes for the SenseWear Armband in Manual Wheelchair Users". The manuscript is currently under peer review.
Who uses the collected data?: 
Manual wheelchair users who want to develop a healthier lifestyle. Clinicians may help them achieve the goal using the data.
Is this tool available free of charge?: 
Tool contact: 
Dan Ding, PhD,
KaLai Tsang