RESNA 27th International Annual Confence
A ModelBased Criterion for Assessing Appropriateness of Wheelchair Setup
Wheelchair setup, and specifically, hub position can affect pushrim biomechanics, and may contribute upper extremity injury. The purpose of this study was to develop a clinical tool to allow clinicians to easily determine optimal axle position. We developed a model to estimate the theoretical distance of the front caster from the floor, WheelD, during a wheelie. In our sample of 35 persons with paraplegia, WheelD ranged from 1.49” to 8.30”. There was an average of 0.34” difference between experimental wheelie angle, and our theoretical angle. Individuals with the hub more anterior to the shoulder, thus better positioned, were those with a WheelD between 1.49 and 3.58”. Clinicians may use this range as an estimation for quality of wheelchair setup of manual wheelchair users with paraplegia.
Wheelchair biomechanics, assistive technology, wheelie,
Wheelchair configuration can affect several aspects of pushrim biomechanics including: push frequency, mechanical efficiency, rolling resistance, and the rate of rise of propulsive forces (1). Boninger et al (2000) investigated the propulsion biomechanics of 40 individuals with a spinal cord injury (SCI) using both kinetic and kinematic analysis. In their study, they found a significant correlation between the horizontal position of the wheel hub (xpos) to the rate of rise of pushrim forces, as well as a correlation between the horizontal and vertical position (ypos) of the wheel hub to the push angle. In order to optimize propulsion biomechanics, these investigators suggest that the axle of the wheelchair should be set in the most anterior position possible without compromising stability and that the distance between the shoulder and the hub of the wheel should be minimized to reduce the risk for upper extremity injury. In the clinical environment, these adjustments are timeconsuming and cumbersome, potentially involving several trialanderror attempts for a highquality setup. A final check for wheelchair setup quality would save time and assure appropriate fit.
The purpose of this study was to develop a guideline, based on a mathematical model, which gives clinicians a relatively simple and objective estimation of how well a client is positioned in their wheelchair as it relates to axle position.
A mathematical model incorporating user characteristics and wheelchair setup parameters was developed. Model inputs: The model utilized available data from Boninger et al (1). Data were taken from 35 wheelchair users with a spinal cord injury at the 3 rd thoracic level and below and included the wheelchair user's height, wheelchair seat plate to floor distance, horizontal distance of the center of the rear wheel to the center of the caster (L), wheelchair/rider system weight distribution (F1F4), xpos and ypos. To calculate the weight distribution, the wheelchair user sitting in their own wheelchair was placed on four force plates, one for each of the rear wheels, and one for each of the front wheels (Fig 1). F1 and F2 represent the weight distributed through the right and left wheels, respectively, and F3 and F4 represent the weight distributed through the right and left casters, respectively.
Calculations: The Xposition of the center of mass of the wheelchair/ wheelchair user system (CM x ) was calculated using weight distribution data for Equation 1.
ΣM_{F1x}=0; CM_{x}= [(F_{3} + F_{4}) L]/mg.
The subject's Yposition center of mass (Ind y ) in sitting (90º trunk flexion, 90º knee flexion, with the arms fully adducted, and the elbows flexed to 90º), was estimated using 50 th percentile anthropometic tables of males (2,3). According to DuvalBeaupre and Robain (1991), the center of mass of individuals with paraplegia in sitting shifts upward by 3% of the individual's height, and this adjustment was made for our data. The Yposition of the total center of mass of the individual in sitting was calculated as superior to the level of the greater trochanter by a distance of 11.3% of the subject's total height, and from kinematic data, this relative position was determined with respect to the wheelchair hub.
Based on inhouse experimentation, the yposition of the center of mass of the wheelchair was estimated at 5cm above the level of the hub. Because of the relatively small percentage of weight contribution of the wheelchair compared to the wheelchair/ rider system, this estimation will lead to minimal errors. The combined Yposition of the center of mass of the wheelchair/ rider system was then calculated (CM y ).
Knowing the X and Yposition of the center of mass of the wheelchair/ rider system, we calculated the angle needed to tip the wheelchair in order for the subject to perform a wheelchair wheelie. A wheelie is defined as lifting the front wheels of a wheelchair off the floor and balancing this position while the user is in the chair. To find the angle of tilt necessary to reach equilibrium for a wheelie position, we first defined a coordinate system (X, Y) with an origin at the wheelchair hub (Fig 2). In this way, the x and yposition of the center of mass will be directly related to the horizontal and vertical adjustment of the wheelchair axle.
The vector, ‘r', locates the center of mass of the wheelchair/ rider system, with î and j unit vectors. Theta is defined as the angle needed to tip the wheelchair such that the center of mass of the wheelchair/ rider system is directly above the pivot point (pp). The theoretical wheelie angle, theta is given by Equation 2.
? = cos 1 ((r·j)/ r)
Equation 3 gives the theoretical height of the front caster from the floor, WheelD, when in a wheelie position.
WheelD= L*sin(?)
Since we considered the subjects from Boninger et al (2000) to be a good indication of the normal distribution of wheelchair configuration, we separated our results into quartiles using SPSS software (version 11.5).
Model Validation: In order to ensure our approximations did not contribute a large amount of error to our model, we compared our theoretical wheelie angle to actual wheelie inclination measures as reported by Tharakeshwarappa et al (in review). Only subjects with paraplegia who had performed a wheelie were included in the validation analysis.
Calculations: For our subjects, the theoretical distance of the front caster to the floor as they perform a wheelie would have ranged from 1.49 to 8.30”. Quartile distributions are displayed in Table 2.

Quartile 1 
Quartile 2 
Quartile 3 
Quartile 4 

WheelD (inches) 
1.493.58 
3.595.12 
5.135.98 
>5.98 
Mean % weight (lbs) in the front of the wheelchair, (F3+F4) 
16.0+/4.77 
25.7+/5.94 
33.3+/3.76 
33.6+/4.87 
When compared to experimental wheelie data, our theoretical values for a wheelie angle when compared to experimental values were within an average of 0.34”. Table 2 shows a comparison of the theoretical WheelD to the actual WheelD.

Subject 1 
Subject 2 

Theoretical WheelD (inches) 
2.20 
3.88 
Actual WheelD 
2.03 
3.36 
Authors recommend setting the axle of the wheelchair in the most anterior position possible without compromising stability to optimize wheelchair propulsion efficiency (1). In this study, we have developed a simple, objective measure that incorporates this recommendation. The 1 st quartile results from our model represent those individuals who are positioned with a greater percentage of weight posterior to the hub, which relates to fore and aft axle position. Based on this, and the findings from Boninger et al (2000) relating axle position to forces, they are likely to be in the best position for wheelchair propulsion. We propose that asking the client to perform a wheelie could become an integral part of wheelchair assessment. Based on the range we have specified for distance of the front caster to the floor during a wheelie, if the individual is able to perform a wheelie, a clinician will be able to quickly and easily approximate fore and aft axle position. On the other hand, if the individual is unable to perform a wheelie, the clinician can tip the chair while providing support. In addition, inability to perform a wheelie may need to be addressed.
This model is limited by the fact that it assumes there are no postural changes of the subject sitting in the wheelchair as they perform a wheelie, so trunk movement could affect results. However, our validation calculations make an initial demonstration that the assumptions required in this model do not significantly affect the results. Furthermore, our model is specific to individuals with a spinal cord injury, and would not necessarily apply to other manual wheelchair users. Future should attempt to further verify these results and use different populations.
This study was supported by the Department of Veterans Affairs and the VA Department of Education (Grant # B3057T), and by NIDDR (Grant # CDF 84:133A)
Fabrisia Ambrosio, MPT , MS ;
Human Engineering Research Laboratories,
VA
Pittsburgh Healthcare System,
7180 Highland Drive , Building 4,
2 nd Floor
East;
Pittsburgh , PA 15206 .
Ph: (412) 3654850;
Email: Faa7@pitt.edu