Archives of Physical Medicine and Rehabilitation
Volume 90, Issue 7 , Pages 1073-1075, July 2009

On “Impact of Surface Type, Wheelchair Weight, and Axle Position on Wheelchair Propulsion by Novice Older Adults”

  • Stephen Sprigle, PhD, PT

      Affiliations

    • Corresponding Author InformationCorrespondence to Stephen Sprigle, PhD, PT, School of Applied Physiology, Georgia Institute of Technology, 490 10th St, NW, Atlanta, GA 30332-0156

School of Applied Physiology and the Industrial Design Program, Georgia Institute of Technology, Atlanta, GA

Article Outline

Abstract 

Sprigle S. On “Impact of surface type, wheelchair weight, and axle position on wheelchair propulsion by novice older adults.”

The mechanical efficiency of propelling manual wheelchairs is a very important topic. Wheelchair users, clinicians, manufacturers and payers would all benefit from better understanding of mechanical efficiency. However, the measurement of the mechanical efficiency is a nontrivial challenge. Cowan et al deserve a lot of credit for tackling such a difficult problem in their article “Impact of surface type, wheelchair weight, and axle position on wheelchair propulsion by novice older adults.”

The study demonstrated good internal validity in detecting a 4% difference in peak propulsion forces in wheelchairs that differed in mass by 9.1kg. However, the instrumentation used to measure forces altered both the mass and inertia of the wheelchair-occupant system—2 factors that directly affect system energy. This approach, therefore, affects external validity, and the results cannot be extended to infer differences across wheelchair codes. That said, this study adds important information to the body of work into mechanical efficiency of wheelchairs. We now have evidence to suggest that addition of 9kg and an 8cm posterior displacement of axle position adversely affects propulsion biomechanics in an elderly cohort. Improved methodology can lead to mechanical efficiency measurement of different wheelchair models and different wheelchair options.

Key Words: Aged, Biomechanics, Rehabilitation, Wheelchairs

List of Abbreviations: ANOVA, analysis of variance, KE, kinetic energy

 

THIS STUDY TACKLED a very important topic: the mechanical efficiency of manual wheelchairs. The overwhelming majority of research into wheelchair propulsion has focused on biomechanics, targeting the propulsion stroke of the upper extremity. Biomechanics studies are extremely valuable and have resulted in improved clinical knowledge. However, the field has yet to develop a means by which the mechanical efficiency of wheelchair designs and configurations can be measured. The importance of characterizing mechanical efficiency can be stated from 3 perspectives:

1.A large proportion of manual wheelchair users propel independently. These users (and their clinicians) would clearly benefit from objective efficiency measures of different wheelchair designs, configurations, and components. More efficient chairs may permit additional functional capacity as well as reduced stress on the upper extremity.

2.Currently, manufacturers do not have a valid means to inform design of their products. The ability to measure mechanical efficiency would provide invaluable information to manufacturers during the design process.

3.The coding of durable medical equipment by insurance carriers attempts to reflect the primary function of the equipment. In the case of manual wheelchairs, this is ease of propulsion. However, coding is not reflective of propulsion effort and efficiency because no ability to measure these constructs exists.

Cowan and colleagues1 make an important point when stating that the bulk of work on propulsion efficiency uses ergometers, dynamometers, and treadmills. In contrast with studying the biomechanics of propulsion, characterizing the mechanical efficiency of a wheelchair needs to be done over the ground because the wheelchair is endowed with real momentum. This approach has rarely been used because of the challenges associated with monitoring forces and motion as the wheelchair travels over ground. This area of inquiry started years ago, and the first over-ground study of which I am aware was from Roger Glaser, PhD,2 who confirmed that wheelchair propulsion over carpet required more metabolic effort than wheelchair propulsion over tile. This article by Cowan et al extends the work of Glaser2 and others3, 4 as well as the work of their own research team.5

The challenges of attempting to measure mechanical efficiency are illustrated in these prior studies. In fact, this research problem has, in my opinion, the toughest characteristic of a research problem: it sounds easy but is exceedingly hard. The authors of this article deserve a lot of credit for tackling such a difficult problem. To illustrate the complexity, I first present some basic mechanics that govern this type of study and then segue into how this study's results reflect the mechanics of the problem and how results might be applied clinically.

The overall propulsion efficiency of a wheelchair is based on a combination of factors. From a mechanical perspective, 2 factors greatly influence manual wheelchair movement: friction and inertia. These factors are manifest as rolling resistance, bearing friction, overall system (chair + occupant) mass, and mass distribution, to name a few. Note that overall system mass is a factor. While wheelchair mass obviously influences this quantity, the mass of the occupant is the dominant contributor. For example, the mass of an 80kg (176lb) person sitting in an 11kg (25lb) chair differs less than 3% from that person sitting in a 13.6kg (30lb) chair.

Studies that seek to compare wheelchair designs (such as Cowan et al1) attempt to use propulsion effort (system input) and the resulting wheelchair motion (system output) to characterize mechanical efficiency. In fundamental mechanical terms, mechanical efficiency can be described as the input work required per unit distance traveled. A primary goal of the wheelchair user is to impart KE to the wheelchair. At any instant in time, the KE is closely approximated by the summation of terms that contribute to the KE of linear travel, yaw (turning), and rotation of the drive wheels and casters.

If one were to consider only straight motion (as used by Cowan et al1), system inertia caused by a turning (yaw) wheelchair and differences in left and right side wheels are eliminated, and the basic equation becomes the following:

where KE is the total kinetic energy, m is the mass of the occupied wheelchair, v is the speed of the occupied wheelchair, ID is the moment of inertia of the drive wheel about its axle, ωD is the rotation rate of the right and left drive wheels, IC is the moment of inertia of the caster wheel about its axle, and ωC is the rotation rate of the right and left caster wheels. The neglected terms are primarily the caster yaw and the KE of whole-system yaw (turning), pitch (wheelie motion), and roll (sideways tipping motion), which are expected to be small during straight motion.

Cowan et al1 used a SmartWheel to measure the variables reflective of mechanical efficiency. System output (and, therefore, KE) was represented by chair velocity, and the work supplied by the subjects was represented by forces at the handrim (peak resultant and tangential forces). Two additional variables, push frequency and stroke length, were reported as additional measures of system input.

As one considers the described study with respect to the dynamic environment, a few things are noted. The use of SmartWheels added about 7kg mass to the wheelchair-occupant system and, just as importantly, altered the inertia of the drive wheels and overall system. Therefore, within the simplified KE equation, the use of SmartWheels directly impacted the mass of the occupied wheelchair (m) and the moment of inertia of the drive wheels (ID).

When assessing this or any article, one considers both internal and external validity. The authors did a nice job controlling for risks of internal validity. For instance, they recruited a cohort of non–wheelchair users over a defined age range and randomized their independent variables (weight, axle location, surface). This high internal validity is exhibited by the effect sizes compared with the actual differences in measurements. For example, the peak resultant force varied 4% or less than across the 9.1kg mass difference, but the respective distributions were significantly different for each surface. Distinguishing small differences requires good internal validity.

The challenge when assessing this article's results lies in addressing external validity. Specifically, 2 factors must be considered: (1) how well the results generalize to elder wheelchair users and (2) how well the results generalize to commercial wheelchairs. These 2 factors are affected by the use of an ambulatory cohort and by the tested wheelchairs' masses and inertial properties.

Cowan et al1 recruited elderly subjects who did not use a wheelchair. Recruiting non–wheelchair users for a propulsion study may seem odd, but valid reasons exist. For one, more ambulatory people can be recruited than elderly wheelchair users. In any experimental design, adequate sample size is important. Recruiting a sufficient number of elderly wheelchair users would have been difficult for at least 2 reasons: (1) not all elderly wheelchair users propel with only their upper extremities—rather, many propel with their lower extremities or use both upper and lower extremities—and (2) the protocol was lengthy enough that subject fatigue is a natural concern, with the thought being that elderly wheelchair users are more likely to experience fatigue than healthy elderly subjects. Despite the good reasons to recruit non–wheelchair users, one must adequately state the limitations of this choice. In general, 2 issues present: (1) whether ambulatory elderly subjects propel the same as elderly wheelchair users, and (2) whether novice users propel the same as experienced users because they are still learning. Propulsion does differ between experienced and nonexperienced users,6 and researchers from Delft Institute of Technology have looked at both immediate and longer-term changes over time.7, 8 Cowan et al had subjects propel for a few minutes before the trial, so immediate learning was probably accommodated. When investigating learning over a 3-week period, timing variables—such as push frequency—changed with learning, but other variables, including force application, did not change with learning.8 The good news for this study is that the measured force variables would not necessarily change if the subjects underwent additional practice, so their force results are a valid measure of their propulsion. How well the results would generalize to elderly wheelchair users is unknown and stands as a fair assessment of study limitations.

The other major external validity issue concerns the tested wheelchairs. The authors acknowledged increased mass but did not completely relate these changes to their potential influences. First, the study compared chairs with masses of 18.1 and 27.2kg (40 and 60lb). The authors' intent was to reflect the quantitative difference in weights of standard versus ultra lightweight chairs. The tested wheelchairs also reflected a greater contribution to overall system mass (occupant + wheelchair) and differences in moments of inertia compared with the 11.3 and 18.1kg chairs they represented. Based on the equation presented, we know these mechanical differences will have an impact, but the impact is unknown. The generalizability of the results depends, in large part, on understanding the collective impact.

The experimental design included 3 factors (surface type, wheelchair mass, axle position) and reported 5 outcome variables (peak resultant force, peak tangential force, average velocity, push frequency, stroke length). The analysis used a 2-way ANOVA (mass × axle position) coupled with a separate 1-way ANOVA for surface type. This obviated the ability to consider interactions with surface type. The authors make the statement, “These data suggest an interaction between axle position and surface type.” This interaction may, in fact, exist, but the chosen analysis did not address it.

Analysis of the main factors resulted in some confirmatory and new findings. The finding that outcome variables change across surfaces confirms prior work on wheelchair rolling resistance9, 10 and the early study by Glaser et al,2 but contradicts that of Koontz et al.5 The finding that a more forward axle position enhances kinetics and kinematic propulsion variables confirms earlier work with dynamometers (see examples11, 12, 13, 14). The finding that propulsion variables differ when pushing an 18.1kg chair versus a 27.2kg chair contradicts that of Bednarczyk and Sanderson,4 who compared wheelchairs with masses of 9.1, 14.0, and 19.0kg (20, 31, 42lb) and found no differences in propulsion kinematics. Collectively, these results are important and illustrate the need for additional study. This need is highlighted by the authors' desire to relate combinations of variables.

The authors discuss and propose a relationship between self-selected velocity and resultant propulsion force. This relationship attempts to relate an input to an output of the system as a means to reflect the mechanical efficiency of the wheelchair. The methodology was not configured to quantify mechanical efficiency but, given the variables collected in this study, it is a clever way to address efficiency. Because the data are available to quantify this relationship, using a regression between velocity and push frequency (or a related approach) might have led to an interesting conclusion. Currently, a hypothesis is generated using group means.

The authors make a very large inference when discussing the interaction between weight and axle position, stating, “A lack of interaction between wheelchair weight and axle position suggests a decrease in wheelchair weight will reduce peak resultant force regardless of axle position. Clinicians can maximize reductions in peak resultant force by securing the lightest possible wheelchair and shifting the axle position as far forward as tolerated by their client.” Unfortunately, the lack of an interaction is not the same as proof, and we cannot generalize an effect at a higher resolution than the tested variables. The study found that a 9.1kg increase had an adverse effect on propulsion kinetics and kinematics. That is an important finding. Our goal now should be to determine whether smaller weight differences have an impact, and what type of users gain a biomechanical benefit from certain weight thresholds.

We must also be very cautious in referencing wheelchair codes, although it is naturally tempting to do so. The coding of wheelchairs uses weight as a differentiating factor, but thresholds were not based on any measure of mechanical efficiency. This illustrates a fundamental flaw in coding and highlights the need for coding that actually reflects performance. K0001 (standard) chairs, by definition, have a mass greater than 16.3kg (36lb); K0003 and K0004 (lightweight) chairs must have a mass less than 16.3 and 15.4kg (36 and 34lb), respectively; and K0005 (ultra lightweight) wheelchairs must have a mass less than 13.6kg (30lb). Note that these weight thresholds are without front rigging. Front riggings have a mass of about 2 kg. Some ultra lightweight wheelchairs exhibit fixed footrests that may or may not be included in the overall reported weight. Therefore, some complete K0001 wheelchairs have a mass of about 18.1kg (40lb) or about 4½kg greater than a K0005 (ultra lightweight wheelchair). The Ti Lite wheelchair in this study had a mass of 11.3kg, which is probably 6.8kg less than some K0001 wheelchairs. Therefore, this study did not compare a K0001 to a K0005 wheelchair by mass or mass differential. The study used a more liberal differential of 9kg. Although this difference certainly exists across certain wheelchair models, the coding definitions do not dictate such a difference.

In closing, we learn from over-ground studies such as that described by Cowan et al1. We now have evidence to suggest that addition of 9kg and an 8cm posterior displacement of axle position adversely affects propulsion biomechanics of an elderly cohort. This adds to our body of knowledge, but this study also illustrates the need for tests that can relate wheelchair design to wheelchair mechanical efficiency. Only then can we determine which users would benefit from which types of wheelchair and establish a coding and coverage policy that reflects performance. We need to develop tests that can monitor the mechanical efficiency of wheelchairs as they move in manners that reflect everyday use. This will require us to monitor efficiency while starting, stopping, and turning in addition to straight, steady-state motion. Only then will everyday mobility be accurately reflected and the energy losses caused by inertial changes be adequately captured.

Back to Article Outline

References 

  1. Cowan RE, Nash MS, Collinger JL, Koontz AM, Boninger ML. Impact of surface type, wheelchair weight, and axle position on wheelchair propulsion by novice older adults. Arch Phys Med Rehabil. 2009;90:1076–1083
  2. Glaser RM SM, Woodrow BK, Suryaprasad AG. Energy cost and cardiopulmonary responses for wheelchair locomotion and walking on tile and carpet. Paraplegia. 1981;19:220–226
  3. Beekman CE, Miller-Porter L, Schoneberger M. Energy cost of propulsion in standard and ultralight wheelchairs in people with spinal cord injuries. Phys Ther. 1999;79:146–158
  4. Bednarczyk JH, Sanderson DJ. Limitations of kinematics in the assessment of wheelchair propulsion in adults and children with spinal cord injury. Phys Ther. 1995;75:281–289
  5. Koontz AM CR, Boninger ML, Yang Y, Impink BG, van der Woude LH. A kinetic analysis of manual wheelchair propulsion during start-up on select indoor and outdoor surfaces. J Rehabil Res Dev. 2005;42:447–458
  6. Patterson P, Draper S. A neural net representation of experienced and nonexperienced users during manual wheelchair propulsion. J Rehabil Res Dev. 1998;35:43–51
  7. de Groot S, Veeger HE, Hollander AP, van der Woude LH. Adaptations in physiology and propulsion techniques during the initial phase of learning manual wheelchair propulsion. Am J Phys Med Rehabil. 2003;82:504–510
  8. de Groot S, Veeger DH, Hollander AP, van der Woude LH. Wheelchair propulsion technique and mechanical efficiency after 3 wk of practice. Med Sci Sports Exerc. 2002;34:756–766
  9. Kauzlarich JJ. Wheelchair rolling resistance and tire design. In:  van der Woude JHV editors. Biomedical aspects of manual wheelchair propulsion. Amsterdam: IOS Press; 1999;p. 158–171
  10. Brauer RL, Voegtle PJ. Wheelchair rolling resistance and tire design. In: Proceedings of the 1981 Biomechanics Symposium, Joint ASME/ASCE Mechanics Conference. University of Colorado: American Society of Mechanical Engineers; 1981;p. 267–270
  11. Mulroy SJ, Newsam CJ, Gutierrez DD, et al. Effect of fore-aft seat position on shoulder demands during wheelchair propulsion, part 1: a kinetic analysis. J Spinal Cord Med. 2005;28:214–221
  12. Hughes CJ, Weimar WH, Sheth PN, Brubaker CE. Biomechanics of wheelchair propulsion as a function of seat position and user-to-chair interface. Arch Phys Med Rehabil. 1992;73:263–269
  13. Gutierrez DD, Mulroy SJ, Newsam CJ, Gronley JK, Perry J. Effect of fore-aft seat position on shoulder demands during wheelchair propulsion, part 2: an electromyographic analysis. J Spinal Cord Med. 2005;28:222–229
  14. Boninger ML, Baldwin M, Cooper RA, Koontz A, Chan L. Manual wheelchair pushrim biomechanics and axle position. Arch Phys Med Rehabil. 2000;81:608–613

 No commercial party having a direct financial interest in the results of the research supporting this article has or will confer a benefit on the author or on any organization with which the author is associated.

 Reprints are not available from the author.

PII: S0003-9993(09)00279-2

doi:10.1016/j.apmr.2009.04.002

Refers to article:

  • Impact of Surface Type, Wheelchair Weight, and Axle Position on Wheelchair Propulsion by Novice Older Adults

    Rachel E. Cowan, Mark S. Nash, Jennifer L. Collinger, Alicia M. Koontz, Michael L. Boninger
    Archives of Physical Medicine and Rehabilitation July 2009 (Vol. 90, Issue 7, Pages 1076-1083)

Archives of Physical Medicine and Rehabilitation
Volume 90, Issue 7 , Pages 1073-1075, July 2009

Article Tools