UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The effect of mass on the kinematics of steady state wheelchair propulsion in adults and children with… Bednarczyk, Janet H. 1993

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_1993_spring_bednarczyk_janet_h.pdf [ 4.84MB ]
Metadata
JSON: 831-1.0077075.json
JSON-LD: 831-1.0077075-ld.json
RDF/XML (Pretty): 831-1.0077075-rdf.xml
RDF/JSON: 831-1.0077075-rdf.json
Turtle: 831-1.0077075-turtle.txt
N-Triples: 831-1.0077075-rdf-ntriples.txt
Original Record: 831-1.0077075-source.json
Full Text
831-1.0077075-fulltext.txt
Citation
831-1.0077075.ris

Full Text

THE EFFECT OF MASS ON THE KINEMATICS OF STEADY STATE WHEELCHAIR PROPULSION IN ADULTS AND CHILDREN WITH SPINAL CORD INJURY by JANET H . BEDNARCZYK B.S . Chemistry University of Michigan, 1969 Diploma in Physical Therapy Northwestern University, 1973  A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF PHYSICAL EDUCATION in THE FACULTY OF GRADUATE STUDIES School of Physical Education and Recreation We accept this thesis as conforming to the required standard  THE UNIVERSITY OF BRITISH COLUMBIA FEBRUARY, 1993 © Janet H. Bednarczyk, 1993  In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study . I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives . It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission .  (Signature)  Department of  PHYSICAL EDUCATION AND RECREATION  The University of British Columbia Vancouver, Canada  Date  DE-6 (2/88)  April 6 .  1993  ABSTRACT A recent trend in wheelchair design has been the reduction of the mass  of wheelchairs . The purpose of this study was to examine the effect of mass on the kinematics of steady state wheelchair propulsion. The mass of test chairs (9 .3 kg) was manipulated by mass additions (5 and 10 kg) in two, neurologically matched, groups (n=10) of adults and children with spinal cord injury . Three dimensional video analysis was used to determine the movement of upper body angles (elbow, shoulder, trunk, and shoulder abduction) . Statistical analysis were multiple univariate, repeated measures analysis of variance (ANOVA) and analysis of covariance (ANCOVA) with significance set at adjusted p values <0 .05. The average mass and age of the pediatric group was much smaller than the adult group (37.4 kg and 11 .3 years versus 68 .5 kg and 33 .5 years) . The averaged group wheeling velocities were 2 .26 m/sec (pediatric) and 2 .38 m/sec (adult). A two-(groups)-by-four-(conditions) ANOVA of the actual wheeling velocities showed a significant groups effect and a nonsignificant interaction effect . The two groups spent comparable proportions of the wheeling cycle in propulsion (pediatric -= 24 .45 %, adult = 24 .41 %) . A two-(groups)-by-four(conditions) ANCOVA of the % propulsion data showed that both the groups effect and the groups-by-condition interaction effect were not significantly different . A two-(groups)-by-four-(conditions)-by-six-(portion of wheeling cycle, first 25%) ANCOVA of the angular data (with velocity as the covariate) showed significant differences for three (elbow, shoulder, and shoulder abduction) of the four angular parameters and nonsignificant groups-by-conditions effects. These results, based on a test sample of chairs and subjects, indicate that mass additions did not affect the angular kinematics, % propulsion or wheeling velocities of two groups of subjects with spinal cord injury in steady state, short distance, level wheelchair propulsion . The pediatric group did show significant absolute angular differences from the adult group, but the angular changes over time and across experimental conditions were the same in both groups .  ii    TABLE OF CONTENTS ABSTRACT TABLE OF CONTENTS LIST OF TABLES LIST OF FIGURES ACKNOWLEDGEMENTS INTRODUCTION Statement of the problem Model of wheelchair propulsion Hypothesis Experimental design Significance Definition of terms 1. Mass and weight 2. Kinematics 3. Drag (rolling resistance) 4. Normalization and propulsive/recovery portions of the wheeling cycle 5. Ensemble average 6. Coefficient of variation (CV) 7. Subject definitions pediatric and adult 8. Fitting 9. Direct linear transformation (DLT) 10. Meningomyleocele or Spina Bifida Assumptions Delimitations LITERATURE REVIEW Occupant studies Methods of measuring output in wheelchair propulsion Studies involving mass in wheelchair propulsion Study populations Pediatric studies Clinical applications Summary METHODS Subjects Selection Neurological assessment Matching Wheelchairs Selection Fitting Mass addition Laboratory set-up Data collection Data analysis iii  i  ' v vi ix 1 1 2 3 3 4 4 4 4 5 5 6 6 7 7 7 7 7 8 8 8 9 10 12 12 13 13 14 14 14 16 16 16 16 17 17 18 20 22  Angular data Determination of wheeling velocities Drag force calculations Determination of % propulsion Statistical analysis Determination of the coefficient of variation (CV)  RESULTS Subject characteristics Wheeling velocities 0/0 Propulsion data Angular data Comparison of test chairs and subjects own chairs Angular changes under mass addition conditions Differences between adult and pediatric subjects Angle-angle plots for pediatric-adult data Average within and between subjects CVs Coast down accelerations and drag force Summary of results DISCUSSION Timing data Angular data Mass addition data Pediatric vs adult data Variability (CV) data Acceleration and drag data CONCLUSIONS RECOMMENDATIONS AND SUGGESTIONS FOR FUTURE RESEARCH BIBLIOGRAPHY APPENDIX A . Determination of the validity of the time normalizing and averaging software programs APPENDIX B . Determination of the reliability of the determination of angular and velocity data APPENDIX C . Determination of the robustness of the angular kinematics over time APPENDIX D . Individual subject angular data  iv  22 25 25 26 26 27 28 28 29 30 31 31 34 36 40 41 46 50 52 52 52 53 55 56 58 59 60 63 74 76 77 79  LIST OF TABLES Table 1 . Definitions of angles and segments Table 2 . Characteristics of the twenty subjects in the study group Table 3 . Characteristics of the subjects' own wheelchairs Table 4 . Description of the average values for acceleration and correlation coefficients for line fit data Table 5 . Summary of angular ranges of motion in wheelchair propulsion studies Table 6 . Comparison of the 5 trials of time normalized elbow angular data to the averaged file as produced by the averaging software program  v  24 29 47 49 53 75  LIST OF FIGURES Figure 1 . Factors which affect wheelchair propulsion Figure 2 . Link segment model of user-chair system and segment movement over a complete wheeling cycle Figure 3 . Schematic representation of the subject in the wheelchair identifying the marker locations Figure 4 . Average overground wheeling velocities recorded from the wheel markers for the pediatric and adult groups of subjects Figure 5 . Average % propulsion for the two groups under the four wheeling conditions Figure 6 . Comparison of angular patterns (elbow, shoulder, trunk, and shoulder abduction, in degrees) in pediatric subjects (n=10) wheeling in test chairs (Kuschall 3000) and subjects' own chairs Figure 7 . Comparison of angular patterns (elbow, shoulder, trunk, and shoulder abduction, in degrees) in adult subjects (n=10) wheeling in the test chairs (Kuschall 3000) and subjects' own chairs Figure 8 . The elbow angle averaged across the first 25% of the wheeling cycle (in degrees) for the two groups (adult and pediatric, n=10/group) under the four test conditions Figure 9 . The effect of mass additions on the angular patterns (elbow, shoulder, trunk, and shoulder abduction, in degrees) in the pediatric group (n=10) over a normalized wheeling cycle Figure 10 . The effect of mass additions on the angular patterns (elbow, shoulder, trunk, and shoulder abduction angle, in degrees) in the adult group (n=10) over a normalized wheeling cycle Figure 11 . Comparison of pediatric and adult angular data (elbow, shoulder, trunk, and shoulder abduction angle, in degrees) in the subjects' own chairs Figure 12 . Comparison of pediatric and adult angular data (elbow, shoulder, trunk, and shoulder abduction angle, in degrees) in test chairs with no added mass Figure 13 . Comparison of pediatric and adult angular data (elbow, shoulder, trunk, and shoulder abduction angle, in degrees) in test chairs with 5 kg of added mass Figure 14 . Comparison of pediatric and adult angular data (elbow, shoulder, trunk, and shoulder abduction angle, in degrees) in test chairs with 10 kg of added mass Figure 15 . Shoulder angle-elbow angle plots (in degrees) for the pediatric and adult groups Figure 16 . Average within subject CVs for the four angular variables (elbow, shoulder, trunk and shoulder abduction ) in the adult group during the propulsive (PROP) and recovery (REC) phases of wheeling Figure 17 . Average within subject CVs for the four angular variables (elbow, shoulder, trunk, and shoulder abduction )in the pediatric vi  2 5 21 30 31  32  33 34 35 36 37 38 39 40 41  42  group during the propulsive (PROP) and recovery (REC) phases of wheeling Figure 18 . Average between subject CVs over the wheeling cycle for the four angular variables (elbow, shoulder, trunk, and shoulder abduction) for the adult group during the propulsive (PROP) and recovery (REC) phases of wheeling Figure 19 . Average between subject CVs for pediatric group for the four angular variables (elbow, shoulder, trunk, and shoulder abduction ) during the propulsive (PROP) and recovery (REC) phases of wheeling Figure 20 . Summary of average within-(A) and average between-(B) subject average CVs for the angular data over the two groups Figure 21 . Summary of average measured accelerations for the coast down tests for the pediatric (n=10) and adult groups (n=10) Figure 22 . Example of line fit of velocity time curve for coast down test for a single adult subject in a test chair with 10 kg of added mass Figure 23 . Summary of average drag force for the pediatric (n=10) and adult groups (n=10) Figure 24 . The result of time normalization on the same elbow angle file for adult subject #10 . The .3AD file at the left is the raw file and the .ADN file at the right is the time normalized file Figure 25 . Reliability of the angular data (elbow, shoulder, trunk, and shoulder abduction) when the same trial was digitized 5 times consecutively Figure 26 . Comparison of the angular data of two adult subjects collected over a two year time period Figure 27 . Individual angular data (elbow, shoulder, trunk, shoulder abduction angle, and shoulder angle-elbow angle, in degrees) for pediatric subject #01 Figure 28 . Individual angular data (elbow, shoulder, trunk, shoulder abduction angle, and shoulder angle-elbow angle, in degrees) for pediatric subject #02 Figure 29 . Individual angular data (elbow, shoulder, trunk, shoulder abduction angle, and shoulder angle-elbow angle, in degrees) for pediatric subject #03 Figure 30 . Individual angular data (elbow, shoulder, trunk, shoulder abduction angle, and shoulder angle-elbow angle, in degrees) for pediatric subject #04 Figure 31 . Individual angular data (elbow, shoulder, trunk, shoulder abduction angle, and shoulder angle-elbow angle, in degrees) for pediatric subject #05 Figure 32 . Individual angular data (elbow, shoulder, trunk, shoulder abduction angle, and shoulder angle-elbow angle, in degrees) for pediatric subject #06 Figure 33 . Individual angular data (elbow, shoulder, trunk, shoulder abduction angle, and shoulder angle-elbow angle, in degrees) for pediatric subject #07  vii  43  44  45 46 48 49 50 74 76 77 79 80 81 82 83 84 85  Figure 34 . Individual angular data (elbow, shoulder, trunk, shoulder abduction angle, and shoulder angle-elbow angle, in degrees) for pediatric subject #08 Figure 35 . Individual angular data (elbow, shoulder, trunk, shoulder abduction angle, and shoulder angle-elbow angle, in degrees) for pediatric subject #09 Figure 36 . Individual angular data (elbow, shoulder, trunk, shoulder abduction angle, and shoulder angle-elbow angle, in degrees) for pediatric subject #10 Figure 37 . Individual angular data (elbow, shoulder, trunk, shoulder abduction angle, and shoulder angle-elbow angle, in degrees) for adult subject #01 Figure 38 . Individual angular data (elbow, shoulder, trunk, shoulder abduction angle, and shoulder angle-elbow angle, in degrees) for adult subject #02 Figure 39 . Individual angular data (elbow, shoulder, trunk, shoulder abduction angle, and shoulder angle-elbow angle, in degrees) for adult subject #03 Figure 40 . Individual angular data (elbow, shoulder, trunk, shoulder abduction angle, and shoulder angle-elbow angle, in degrees) for adult subject #04 Figure 41 . Individual angular data (elbow, shoulder, trunk, shoulder abduction angle, and shoulder angle-elbow angle, in degrees) for adult subject #05 Figure 42 . Individual angular data (elbow, shoulder, trunk, shoulder abduction angle, and shoulder angle-elbow angle, in degrees) for adult subject #06 Figure 43 . Individual angular data (elbow, shoulder, trunk, shoulder abduction angle, and shoulder angle-elbow angle, in degrees) for adult subject #07 Figure 44 . Individual angular data (elbow, shoulder, trunk, shoulder abduction angle, and shoulder angle-elbow angle, in degrees) for adult subject #08 Figure 45 . Individual angular data (elbow, shoulder, trunk, shoulder abduction angle, and shoulder angle-elbow angle, in degrees) for adult subject #09 Figure 46 . Individual angular data (elbow, shoulder, trunk, shoulder abduction angle, and shoulder angle-elbow angle, in degrees) for adult subject #10  viii  86 87 88 89 90 91 92 93 94 95 96 97 98  ACKNOWLEDGEMENTS This project would not have been completed without the support of a number of people whom I would like to thank . First, my thanks to the B .C. Medical Services Foundation who saw fit to support this project and made it possible for this dream to become a reality . Also to Mr. Marco Ferrara of Motion 2000 Inc., whose personal commitment to this project made it possible for new test chairs to be available at the right place and at the right time for data collection to occur . My thanks to Mr. Don Alder, who stepped in with short notice, to train and guide me in the subtleties of wheelchair adjustments . Next, my thanks to the clinical staff at the Spina Bifida Clinic at B .C . Children's Hospital and in particular to the Director of the Clinic, Dr. W. Arnold, and the Coordinator Bev Irwin . The warmth and kindness of these people made the process of pediatric subject recruitment not only possible but a delight . I also would like to acknowledge the support of Mr . Norm Haw of the B .C . Paraplegic Association who assisted in the recruitment of the adult subjects. I received invaluable support from the staff of the School of Physical Education and Recreation at U .B.C. A number of members of the support staff of the School of Physical Education and Recreation made it possible to bring people and equipment together in order for the project to proceed . The Director of the School, Dr . R. Schutz was instrumental in obtaining much needed, wheelchair accessible space for data collection . As well, he provided guidance in the statistical analysis of the data along with Mr . Han Joo Eom who supplied the expertise to teach me how to use BMDP . Also, Dr . K. Coutts was a valuable resource in the analysis of drag force . The successful conclusion of this project was only possible due to the long term supervision of Dr. D. Sanderson. I would also like to acknowledge the support from the clinical and administrative staff of the B .C. Rehabilitation Society which enabled me to pursue graduate studies while maintaining a continuous clinical link . In particular, Mr . E . Desjardins, Mr . M. Ewan and Mrs . M . Bryson provided patient, consistent, sympathetic support which was much appreciated. Finally, I would like to thank my husband and two sons whose unconditional support over the past three years (and many dinners of fast food, without me) made everything possible .  ix  INTRODUCTION Wheelchair propulsion is a repetitive, cyclical movement which is the product of a user-machine interaction . Most user-machine systems, such as occur in cycling or rowing, are of relatively short duration and are often recreational . The situation is entirely different for the person with a disability who is dependent for their mobility on a wheeled, human-powered machine . In this case, the interaction with the machine is permanent . The wheelchair becomes an extension of the person propelling it . Yet very little is known about the factors in the user-chair interaction that may affect wheeling performance (Van der Woude, Veeger & Rozendal, 1986). The number of persons in the United States who currently depend on a wheelchair for their mobility is reported to be more than 750,000 (Majaess, Kirby & Stolarz,1991) . Comparable Canadian estimates are limited . It is known that there are a total of 101,870 individuals in Canada who are living at home and who rely upon a wheelchair as a primary means of mobility (Statistics Canada, 1986) . Many of these disabled people have close to normal life expectancies (Treishman, 1987) . Therefore, the issue of effective user-machine interface will become even more critical as the mean age of our population rises and the numbers of persons with disability, who are dependent on wheelchairs, increases. For a child, who is born with disabilities which make mobility by way of the lower extremities impossible, the interaction with a machine is a life-long experience . There are very few studies that have involved children with disabilities, and none in the literature that have examined the kinematics of wheelchair propulsion by pediatric users . The consequence of the absence of such data means that clinicians have no objective data by which they can appropriately prescribe wheelchairs for children with disability. Statement of the problem A recent trend in wheelchair design has been the reduction of the mass of wheelchairs (Parziale, 1991 ; Ragnarsson, 1990) . This design trend is especially important for the pediatric wheelchair user because pediatric chairs, though smaller in size, have as much mass as comparable adult chairs (Robbins, 1991) . Because the pediatric user often has a low mass, the chair mass constitutes a large proportion of the system mass . There is little 1  experimental evidence to support the belief that low mass chairs are preferable, and no published data on wheelchair propulsion by pediatric subjects . The drag force, or the inertial force of the wheelchair that the user must overcome, is almost the same in pediatric and adult wheelchairs (Jarvis & Rolfe, 1982) . Thus children, who are smaller and not as strong as adults, are likely faced with overcoming larger net forces than the adult wheelchair user. Model of wheelchair propulsion Current models of wheelchair propulsion describe it as a complex activity with many factors which can affect the user-machine output (Cooper, 1990 ; Van Der Woude, Veeger & Rozendal, 1986, 1989) . The critical factors in wheelchair propulsion fall into three basic categories : 1) those factors which affect the occupant and his/her ability to work ; 2) those factors which affect the wheelchair and its stability/resistance to movement ; and, 3) those factors which involve the interaction, or interface, between the chair and user (Figure 1) . The effect of changes in any, or all, of these factors can be observed in the resultant kinematic or metabolic output of the user-chair system .  Figure 1 . Factors which affect wheelchair propulsion Kinematic measurements have been shown to be important descriptors of wheelchair performance (Bednarczyk & Sanderson, 1991 ; Gaines, Zomlefer & Zhao, 1984 ; Ridgeway, Pape & Wilderson, 1989 ; Sanderson & Sommer, 2  1985 ; Van der Woude, 1989) . Kinematics provides a systematic, non-invasive description of technique and timing in wheelchair propulsion which is essential to our basic understanding of this movement pattern (Van Der Woude et al ., 1986) . The key features of wheelchair propulsion, as measured kinematically, have been shown to be the angular displacements of the upper extremities, and the timing of the propulsive and recovery phases of the wheeling cycle (Masse, Lamontagne & O'Riain, 1992 ; Rodgers, Gayle, Figoni, Kobayashi, Glaser & Gupta, 1992 ; Sanderson & Sommer, 1985). Hypothesis The purpose of this study was to examine the effect of mass on the kinematics of steady state wheelchair propulsion in adults and children with spinal cord injury . The hypothesis was that mass additions (in the 10 kg range) would affect the kinematics of wheelchair propulsion by changing both the timing and angular characteristics of wheelchair propulsion. Experimental design The experiment to test this hypothesis was a 2 x 4 (groups-by-chair mass) mixed factorial design with repeated measures on the second factor . The dependent variables were : 1) the angular kinematics of wheeling (shoulder, elbow, shoulder abduction and trunk) at a steady wheeling speed (2 m/sec) ; 2) the timing of wheeling (% propulsion) ; 3) the actual wheeling velocities ; and, 4) the average coefficients of variation (CVs) of the angular data . The factors were users (two groups ; adults and children with spinal cord injury) and chair mass (repeated measures ; subject's chair, test chair alone, test chair +5 kg and test chair +10 kg .). The experimental design was sensitive to the interface of user and chair and reflected the reality of the situation facing the clinician and user in wheelchair prescription . Identical, new, low mass test chairs (9 .3 kg) were provided in a variety of sizes to match the users . Factors such as residual neurological ability, wheeling speed, and uniform addition of mass were controlled so that the effect of mass on propulsion kinematics could be determined . Adults with traumatic paraplegia were selected because they often have a large mass and there is existing research information on this population. Children with Spina Bifida (Meningomyelocele) were selected as the second group because they often have a small mass and they represent a group of  3  users with paraplegia resulting from congenital defects in the spinal cord who are available in sufficient numbers for research purposes. Significance The results of this study will benefit clinicians and users in understanding the influence that mass might have in overground, steady state wheelchair propulsion . It has application to the process of appropriate wheelchair selection for adult and pediatric users with paraplegia . The presence of significant changes in wheeling style in response to chair mass might support clinicians and users in their selection of low mass chairs . However, nonsignificant mass effects in the short distance, level wheeling condition selected in this study does not necessarily mean that there are not significant long distance, or long term, effects. Additionally, this study describes, in detail, the kinematic wheeling style of children with disability . Children have been assumed to be miniature versions of adult wheelchair users with respect to wheelchair propulsion without any research basis for this assumption . The demonstration of similar kinematics and variability of wheeling in children to adults would permit the inference of research and clinical experience from the adult to the pediatric disabled population . Furthermore, the kinematic description of wheeling in children with Spina Bifida provides an important first step in documenting and understanding some of the conditions which this growing group of children with disabilities are experiencing . This will provide basic information for pediatric clinicians as well as children with Spina Bifida and their families. Definition of terms 1. Mass and weight Weight and mass are different, though related, entities . Mass is a quantity of matter or the amount of inertia that an object has ; the units of measure are kilograms . Weight is defined as the gravitational force acting upon an object ; the units of measure are Newtons . Thus, the mass of an object is independent of position in space while the weight of an object will change depending upon the object's position relative to the centre of the earth. 2. Kinematics Kinematics is defined as the movement of a body (or body parts) in space with respect to time and a known reference system . The analysis of human movement is based on the assumption that the movement can be represented 4  by body segments which are assumed to be rigid, uniform rods with the centre of mass of the segment located at a single, fixed point on this rod . In the following analysis of wheelchair propulsion, the link segment model and the movement of these segments over a complete wheeling cycle for the user-chair system would be described as shown in Figure 2.  Figure 2 . Link segment model of user-chair system and segment movement over a complete wheeling cycle. Kinematic measurement involves the direct measurement of markers placed on the body which define the link segments . From these direct measurements over time, linear and angular displacements, velocities and accelerations of the body segments over time -re calculated. 3. Drag (rolling resistance) Drag is a force that resists the rolling movement of the wheelchair . It is dependent on the mass distribution over the wheels, wheel radius, acceleration, and total mass of the user and chair (Bennedik, Engel, & Hildebrandt, 1978) . It is influenced by the frictional coefficients of the chair (in particular the wheel bearings) and flutter of the front caster wheels (Brubaker, 1981) as well as the frictional coefficients of the rolling surface (Wolfe, Waters, & Hislop, 1977). 4. Normalization and propulsive/recovery portions of the wheeling cycle Wheelchair propulsion is a repetitive action involving common units of propulsion comprised of the components of push and recovery . However, the timing of these components can vary from individual to individual (Sanderson & Sommer, 1985 ; Van Der Woude, Hendrich & Veeger, 1988) . In order to 5  compare cycles of differing durations, all wheeling cycles were normalized to the same unit of time and expressed as a percentage of that normalized cycle. In all cases, 0% corresponds to the point of first hand contact on the wheel (grab) and 100% corresponds to the subsequent grab . Release, or the point of "hand off" the wheel occurs somewhere within the wheeling cycle (usually at about 25% of the cycle) . The percentage of the wheeling cycle in which force is being applied to the wheel by the hand is termed the propulsive portion of the cycle and recovery is used to describe that portion of the cycle when the hand is not in contact with the wheel and no force is being applied by the user. 5. Ensemble average Ensemble average is the average pattern of propulsion for either several subjects or the average of several trials for the same subject . In this study, two types of averaging were performed . In the first case, five consecutive trials for the same subject were expressed as a within subject ensemble average . In the second case, the within subject average for all of the subjects in each group (n=10) was further averaged to produce between subject ensemble averages. 6. Coefficient of variation (CV) The CV is defined as the standard deviation of a group of scores divided by the mean score multiplied by 100 (Dixon & Massey, 1969) . Within-and between-subject CVs are indicators of the degree of the variability of the wheeling movement. There is no information on how variable the wheeling movement is kinematically although there is some indication, in small subject populations (Sanderson & Sommer, 1985 ; Su, Chou, Lu, & Lai, 1991), that it may be quite variable . Because there is so little information on the pediatric wheelchair user, there is also little information on the variability of the pediatric wheelchair user relative to the adult user . Therefore, the CV was felt to be important to any comparison of the kinematics of wheeling in the two groups. One would expect that the variability of the angular data would be less during the propulsive phase because there are limited options as to how the user can come in contact with the wheel . In contrast, during the recovery phase, as has been demonstrated by Sanderson & Sommer (1985) and Higgs (1986), there are a variety of movement patterns that are available to the user . Thus in this study average CVs were computed over both the recovery and propulsive phases of the wheeling cycle .  6  7. Subject definitions : pediatric and adult For the purposes of this study, the term pediatric was used to denote a person between 8 and 14 years of age (inclusive) and the term adult was used to describe a person between 22 and 52 years of age (inclusive). 8. Fitting Fitting refers to the process of adjusting the portions of a wheelchair (seat height, inclination, position, wheel angle, footrest height) to user comfort and preference. In some wheelchair designs, only the footrest height is adjustable. In others, there are many adjustments that can be made . The relationship of chair fit to wheelchair performance is unclear . For this, reason chair fitting was kept to a minimum in this study. 9. Direct linear transformation (DLT) The process of direct linear transformation computes three dimensional data coordinates from two sets of two dimensional images . It uses an equation containing 11 coefficients and a least squares fit, that translates and scales the data according to coordinate points previously taken from a calibration frame. DLT has the advantage of not requiring specific camera placements, special lenses or camera leveling procedures in order to obtain a high degree of accuracy in the analysis of movement (Abdel-Aziz & Karara, 1971 ; Cheetham & Sheirman, 1988). 10. Meningomyleocele or S ina Bifida Meningomyelocele or Spina Biflda is a defect that occurs during neural tube closure in the fetus (21-26 days post conception), resulting in an abnormal development of the bones and coverings of the spinal cord . The causes of the defect are unknown but are believed to be both genetic and environmental. Spina bifida is the most common birth defect in British Columbia (affecting approximately 1 in 1,000 newborns) with a range of disabilities, many of which are permanent . Assumptions Air or wind resistance can affect drag (Kyle et at ., 1978) . This effect is dependent on the relative velocities of the user-machine system and the air around them . The assumption was made that at the velocities at which short distance, indoor wheeling occurs (0 .5-2 .5 m/sec) that drag effects due to air resistance were negligible . This assumption was based on studies which  7  demonstrated minimal wind resistance in indoor wheeling situations (Hedrick, Wang, Moeinzadel & Adrian,1990 ; Van Der Woude et al ., 1986). Delimitations The results of this study will reflect the populations of adults with traumatic paraplegia and children with Spina Bifida . It will not be possible to infer the results of this study to other disabled populations such as children with cerebral palsy or adults with head injury . However, it is hoped that these selected populations will reveal the presence of underlying principles relevant to wheelchair performance which may later be shown to be consistent with wheeling style in other subject populations. Because the wheelchair serves many functions, this study will not be able to provide information about the user-chair interaction in situations other than level, steady state wheeling . One function of a wheelchair, most certainly, is as a means of moving the user from place to place . A certain portion of this mobility involves initiating movement from a standstill and another portion involves maintaining movement . This study focuses on the maintenance of mobility once the user-chair system is in motion . Another important function of the wheelchair, aside from mobility, is to serve as a stable base from which the user transfers from chair to car or chair to bed, for example . The wheelchair must also be able to fit into, and be lifted into the user's car . All of these factors are important in the final prescription decision . This study can only provide information about the forward, short distance, overground mobility of the disabled person using a manual wheelchair under controlled conditions . It is also unable to address the demands of long term wheeling under a variety of irregular wheeling conditions . LITERATURE REVIEW Most research designs of wheelchair propulsion have attempted to control the many factors in the process by selecting single components for study . Thus, the focus has been either on the occupant or the chair with few studies appreciating the importance of the interaction of the two. Occupant studies There has been considerable work done on the wheelchair in isolation from the user (Brubaker, 1986 ; Brubaker & McClay, 1984, 1985 ; Collins & Kauzlarich, 1988 ; Kauzlarich, 1986 ; Nitz & Bullock, 1983 ; Peizer, Wright, & Freiberger, 1964 ; Reger & Hobson, 1983) . Some important information was 8  obtained regarding prototype wheelchair design which has influenced the wheelchair manufacturing industry, but even the authors admit that it is the interface between the user and the chair which is the critical determinant of performance (Brubaker, 1990). Other studies have focused on the occupant in isolation and have attempted to measure the effect of training on conditioning of disabled subjects (Davis, Plyley, & Shepard, 1991 ; Dreisinger & Londeree, 1982 ; Gangelhoff, Cordain, Tucker, & Sockler,1988 ; Stotts, 1986 ; Zwiren & Bar-Or, 1975) . These studies disregard possible affects of chair design and fitting . Only the user is considered when discussing the effect of training on user-chair output. Methods of measuring output in wheelchair propulsion In addition to the question of considering the interaction of the occupant to the chair, there is another problem of how to measure output in this complex, interrelated system . There have been two basic approaches : metabolic studies and kinematic studies . Some researchers have chosen to measure the metabolic parameters of the occupant of the wheelchair . The intent was to identify the physiological responses of the wheeler regardless of the chair characteristics (the user in isolation approach) . In order to use metabolic measures however, the experimental design requires continuous wheeling of at least 4 minutes duration . Thus researchers such as Van der Woude et al. (1986), Van der Woude, Veeger, Rozendal & Sargent (1989), Voight & Bahn (1969), Whiting, Dreissinger, & Hayden (1984), Lakomy, Campbell, & Williams (1987), Franklin, Swantek, Grais, Johnstone, & Gordon (1990), Reid, Lawrie, Hunter, & Warren (1990), and Gass & Camp (1979) used treadmill wheeling by elite athletes, who were fit enough to wheel long distances. Thus, it is not clear how the results of such studies relate to the "average", non-athletic user in ordinary wheeling situations. Recently Cooper (1992) has combined metabolic, kinetic, kinematic and EMG analysis of wheelchair propulsion in order to obtain a "complete energy profile" of the wheelchair-user interaction . However, this type of research methodology continues to be applied exclusively to the elite athletic population of wheelchair users. It is not clear what the relationship is between everyday, overground wheeling and long distance, treadmill, or ergometer, wheeling . Van IngenSchenau (1980) has examined overground versus treadmill running in able bodied athletes and found them to be similar . A recent study by Sanderson & 9  Bednarczyk (1991) has shown that the kinematics of wheelchair propulsion by 18 subjects with spinal cord injury is not the same in overground as compared to ergometer wheeling . A different study by Cooper (1992) has shown that the kinematics of wheeling is the same overground and on an ergometer if the inertial properties of the ergometer are closely matched to the overground wheeling situation . Therefore, overground wheeling was selected in this study in an attempt to imitate wheeling conditions most commonly experienced by the wheelchair user. Heart rate responses of the users were used as outcome measures in a few studies (Anderson, Brattgaard, 1970 ; Brattgaard & Severinsson, 1976; Engel & Holdebrandt, 1974 ; Staros, 1981) . This method offers an inexpensive, low technology approach to evaluation of the wheeling movement . However, heart rate responses are too slow and too global to permit insight into the factors which may be critical in routine, short distance wheeling. Kinematic measurements have been shown to be important descriptors of wheelchair performance (Bednarczyk & Sanderson, 1991 ; Gaines et al ., 1984 ; Ridgeway et al ., 1989 ; Sanderson & Sommer, 1985 ; Van der Woude, 1989) . However, one issue in the kinematic assessment of wheelchair propulsion has been the question of whether two (2D) or three (3D) dimensional analysis is more appropriate . Two dimensional analysis assumes that the movement to be analyzed occurs in a single plane . The third dimension is required to capture out of plane movements . Previous studies (Sanderson & Sommer, 1985 ; Ridgeway et al ., 1988 ; Van Der Woude et al .,1986) have described the wheeling movement in two dimensions . However, several research teams (Rodgers et al ., 1992 ; Sanderson & Bednarczyk, 1991 ; Su et al ., 1991 ; Wang, Xu, Hedrick, Adrian, & Morse, 1989) have demonstrated that there are significant out of plane movements of the arms of subjects in wheeling . Therefore, 3D analysis of the wheeling movement was selected in this study. Studies involving mass in wheelchair propulsion There have been some studies which have examined the effect of mass in wheelchairs (Kauzlarich, 1986 ; Kirby, Kumbhare, & MacLeod, 1989 ; Lemaire, Lamontagne, Barclay, & John, 1991 ; Loane & Kirby, 1985) . These studies have examined only the effects of adding mass to the rear of a static system . The outcome measure used was the instability (tipping point) of the user-chair 10  combination . No mention is made in these studies of the effect of mass in a moving system. Another approach has been to compare performance of different types of wheelchairs (Hilbers & White, 1987 ; Van der Woude et al . 1986) . The results did show differences between wheelchairs propelled by small numbers of a mix of able bodied and disabled subjects over a long distance, level surface . The two test chairs in these studies might possibly have had different masses, but that is unclear. The results of the Hilbers study were not supported when disabled subjects were used in short distance, level, and incline wheeling conditions (Bednarczyk & Sanderson, 1991) . In this case, no differences were found in propulsion performance in a variety of test chairs. Peizer, Wright & Freiberger (1964) examined differences in mechanical and biomechanical factors in a light weight (29 lb) and conventional weight (49 lb) wheelchair. Unfortunately most of their analysis was confined to the chair alone and the performance data (no kinematics) was limited to data collected from three subjects with lumbar level paraplegia . In the almost 30 years since this study was done the mass of wheelchairs has changed considerably . At this point in time, 29 lbs would hardly be considered a "low mass" wheelchair. Parziale (1991) studied the effect of standard and lightweight chairs in a group of 26 people with spinal cord injury . The subjects were asked to wheel at two paces (sprint and at a slower pace) for a two distances (400 and 820 feet). Speed was neither monitored nor controlled . The subjects were asked to wheel as fast as they could for the short course and to wheel as far as they could in a given time period (4 minutes) for the long course . No differences were found in heart rate, blood pressure or respiratory rate in the two chair conditions or the two wheeling distances . However, differences were found between the subjects with paraplegia and those with quadriplegia, irrespective of chairs or distance. Unfortunately, the reader is given no information as to the neurological status of the subjects (complete or incomplete spinal cord injury, or thoracic versus lumbar paraplegia), so the reported differences could have been due to the selected subject populations . In addition, there was no attempt to describe the way in which the subjects performed the required wheeling tasks (no kinematic description was employed).  11  Study populations Another problem with some studies in the literature has been the selection of the study population . Many of the studies on wheelchair propulsion have used able bodied subjects (Cerguilini, Figura, Marchetti, & Ricci, 1981; Glaser, Ginger, & Laubach,1977 ; Van Der Woude et al ., 1986) . Able bodied subjects, unlike individuals with disability, are able to use their lower extremities for balance and stability . Brown, Knowlton, Hamill, Schneider, & Hetzler (1990) demonstrated significant physiological and biomechanical differences between wheelchair dependent and able bodied subjects when propelling a wheelchair on an ergometer . They concluded "these differences are important considerations to the interpretation of data in wheelchair studies" . Veeger, Lute, Roeleveld, & Van Der Woude (1991) demonstrated significant differences between able bodied trained and untrained wheelchair users . In addition, persons with disability have been shown (Duval-Beaupere & Robain, 1991) to have a different distribution of body mass than their able bodied counterparts. Thus, the conclusions from experimental studies involving able bodied subjects is suspect when applied to the disabled population. Much of the published work on wheelchair propulsion, which has used subjects with disability, has selected primarily for athletic populations of persons with paraplegia (Cooper, 1989, 1990,1991 ; Coutts & Strogyn, 1987 ; Glaser, Sawka, Young, & Suprasayad, 1980 ; Hardison & Israel, 1987 ; Stoboy, Rich, & Lee, 1971 ; Tupling, Davis, & Pierrynowski, 1986 ; Veeger, Vander Woude, Drexhage, & Koperdraat, 1989 ; Wicks, Olderidge, Cameron, & Jones, 1983). This has created a false impression that all wheelchair users are fit, athletic, and young . While it is true that elite, athletic members of the disabled community require research support for their endeavors, information from studies on this subject population cannot be applied to the majority of wheelchair users. Pediatric studies There are very few studies on disabled groups other than persons with paraplegia (Bostom & Bates, 1987 ; Hunter, 1987 ; Knowlton, Fitzgerald, & Sedlock, 1981 ; Lehman, Warren, Halar, Stonegridge, & Delateur,1974 ; Warren, 1990) . There are even fewer studies which have included children with disabilities (Findley & Agre, 1988 ; Franks, Palisano, & Darbee, 1991 ; Jarvis & Rolfe, 1982 ; Luna-Reyes & Reyes, 1988 ; O'Connell, Barnhart, & Parks, 1992). Jarvis & Rolfe (1982) used a machine which simulated wheelchair propulsion to 12  measure the power output of the first propulsive cycle in fifty-four children (twenty-two of whom had Spina Bifida and thirty-two of whom were able bodied) . Luna-Reyes & Reyes (1988) looked at the energy expenditure for ambulation in children with disability by crutch walking (with and without orthoses) and wheelchair propulsion . Though the subject pool consisted of sixteen children with Poliomyelitis (as well as forty-one able bodied children), only eight children were examined in wheelchair propulsion . No mention is made of the types of wheelchairs used or fittings of the wheelchairs to the subjects in the study . Findley & Agre (1988) did similar metabolic studies on a small number of adolescents with Spina Bifida . Franks et al . (1991) compared the energy cost of walking and wheeling in three children with Spina Bifida and found that the high metabolic cost of walking in these children had a negative effect on some aspects of school performance . There are apparently no published studies which have examined the three dimensional kinematics of wheelchair propulsion in children. Clinical applications In addition to the aforementioned shortcomings in the literature on wheelchair propulsion, there is a gap between the existing research information and application of this information to wheelchair prescription in the clinical setting (Fahland & Grendahl, 1967 ; Harburn & Spaulding, 1986 ; McLaurin, 1990 ; Rozendal, Rebroech, & Van der Woude, 1990 ; Shapcott, Heinrich, Brubaker, & Fergruson-Pell, 1986) . This gap may be due to an inappropriate experimental protocol (use of able bodied subjects, use of long distance wheeling on ergometers or treadmills) or to poor communication between clinicians, users, and researchers. Summary There are few studies which have examined the effect of mass in wheelchair propulsion and none of these have used biomechanical measurement methods to describe wheeling . There are apparently no studies in the literature which have described the biomechanics of wheelchair propulsion in children with disability .  13  METHODS Subjects Selection Eleven adult subjects were recruited from within the population of spinal cord injured adults residing in the Vancouver area . Many of these subjects already had volunteered as participants in pilot studies . Others were referred to the study by counsellors of the B .C. Canadian Paraplegic Association. Complete data collection occurred on all of the adult subjects . Data from one of the adult subjects was not used in this study because it was not possible to find an adequate neurological match for this subject to a pediatric subject. Inclusion criteria for the adult group were: 1) a traumatic spinal cord injury with resultant complete or incomplete neurological deficit between T6 and L2 (inclusive); 2) the absence of other concomitant injuries such as head injuries which made the subjects unable to provide informed consent; 3) dependency on a manual wheelchair for the majority of daily activities; 4) the dependence on a wheelchair for more than one year, (i .e., at least one year post injury) ; and, 5) age from 22 to 52 years of age (inclusive). Paraplegic subjects were selected because there is a considerable existing body of knowledge on wheelchair propulsion in this group . They were asked to give informed consent on forms previously approved by the U .B .C. Ethics Review Committee. The pediatric subjects were recruited through the Meningomyelocele Clinics held at the B .C. Children's Hospital (BCCH) . After receiving permission from the Clinical Research Review Committee, the Director of the Meningomyleoce!e Clinic and the Vice-President in charge of Medical Services, the author attended 10 different clinics between January 30 and April 9, 1992. Over this period of time, a total of 32 children (out of the approximately 200 children with Meningomeyelocele who were present at those clinics) were assessed for possible inclusion into the study . The charts of all prospective subjects were reviewed (after receiving permission from the Director of Medical Records at BCCH) in order to determine if all of the inclusion criteria had been met . From these 32 children . 21 were found to meet all of the inclusion criteria. 14  Complete data collection occurred in the Biomechanics Laboratory at the University of British Columbia on 18 children . From this group of 18, 10 were found to meet the matching criteria of this study and their data is presented in this document. Inclusion criteria for the pediatric group were: 1) the presence of spinal cord injury as a result of neural tube defect at birth (Spina Bifida) . This included subjects with motor and sensory involvement between T6 and L2 (inclusive); 2) age between 8 and 14 years, inclusive ; and, 3) dependence upon a wheelchair for the majority of mobility needs for at least one year . Because many children with Spina Bifida walk with long leg braces, the inclusion criteria of being primarily dependent on the use of a wheelchair for mobility was difficult to rigidly apply to this group . However, every attempt was made to insure that the pediatric users were experienced in wheelchair propulsion. Excluded from the study were subjects with upper extremity involvement or cognitive deficits resulting from uncontrolled hydrocephalus or other concomitant birth defects . Each prospective pediatric subject was also evaluated for wheeling ability and their preferred wheeling speed was recorded at the clinic . Those subjects who were unable to propel their wheelchairs at speeds of more than 1 .5 m/sec were excluded from the study . A wrist x-ray was obtained through the Meningomyelocele clinic for each prospective pediatric subject in order to determine the bone age of each pediatric subject because some children with Meningomyelocele also have growth defects. Consent was provided by the child's parent (or guardian) for participation in the study along with the assent of the child according to forms previously approved by the U .B .C. Ethical Review Committee and the Research Review Committee of BCCH. Children with Spina Bifida or Meningomyelocele were selected because they represent a large group of children with spinal cord injury . Fortunately the numbers of children with traumatic spinal cord injury is low . However, this makes them an unsuitable sample for research purposes . The numbers of children with Spina Bifida is much larger . From within this group it was possible to select children with relatively uncomplicated paraplegia . While technically the neural tube defect which results in Spina Bifida is a multisystem disorder, 15  functionally there are many children who present with an uncomplicated paraplegia . These were the subjects who were selected for this study. Neurological assessment All subjects were assessed by the author, a trained physiotherapist, according to the American Spinal Injury Association (ASIA) scale . This scale has been shown to be rapid, sensitive, and accurate (Bednarczyk & Sanderson, 1992) in classifying disability . Each subject was given an ASIA score between 0 and 100 based on selected manual muscle tests . This assessment occurred in conjunction with screening at the Meningomyelocele clinic held at BCCH for the pediatric subjects . The adult subjects were assessed upon arrival at the laboratory. Matching Subjects were matched according to their ASIA scores . There was a need for neurological matching in order to ensure equivalency in propulsion ability in the two subject populations . Adult and pediatric subjects were considered matched if their ASIA score fell within 4 ASIA points of each other. Subjects were recruited, assessed, and matched until 10 pairs of subjects were obtained . Wheelchairs Selection The majority of wheelchairs in the marketplace range in mass from 10-20 kg with pediatric wheelchairs having the same mass as adult wheelchairs (Robbins, 1991) . The Kuschall Champion 3000 was selected as the test wheelchair from amongst the many available in the marketplace because it is a low mass chair (9 .3 kg) which is available in identical styles but in a range of sizes for both adult and pediatric users. The rationale for the addition of 5 and 10 kg to the test chair was that it replicated the range of chair mass available commercially and thus reflected the reality faced by the clinician and wheelchair user. New chairs for the study were supplied by Motion 2000 Inc . Once the recruitment of subjects had been completed, information about the type and dimension of chairs which each subject was using at that time were forwarded to Mr. Marco Ferrara (President of Motion 2000 ) who provided the best possible match between the test chairs and the users' own chairs . A total of seven new Kuschall Champion 3000 wheelchairs were provided for use in the study. 16  These chairs had a considerable range in seat dimensions from twelve by twelve inches to fifteen by eighteen inches . Two inch foam cushions were used on all test chairs. Some of the factors which affect drag in a wheelchair are the wheeling surface (Wolfe et al ., 1977), the tire pressure and material, the bearings and size of the main wheels and casters, and wheel shimmy (Brubaker et al ., 1986; Cooper, 1990 ; Kauzlarich, Bruning, & Thacker,1984 ; Kauzlarich & Thacker,1985) . Identical wheel types, pushrim sizes, and wheel camber settings (4 degrees) were used in all test chairs . Other factors relate to the position of the user relative to the center of mass of the system . All of these factors were maintained in as constant a state as possible throughout the experiment . This was accomplished by frequent inspection and maintenance of the test chairs by the author under the supervision of a designated representative of the manufacturer (Mr . Don Alder from Sunrise Medical Inc .). All tire pressures were maintained at 100 psi (according to the manufacturers specifications). Fitting Each subject was fitted into a test chair of the proper basic dimensions for their body size . Because seat height and seat position have been shown to affect wheelchair propulsion (Reger & Hobson, 1983 ; Van der Woude, 1990; Veeger, Van der Woude, & Rozendal, 1989 ; Walsh, Marchiori & Steadward, 1986), the fitting aspects of the test chairs were kept as constant as possible throughout the experiment . Seat height was maintained as constant as subject safety would permit . Seat position was held constant such that the users' hip joint centres were 5 cm above and 5 cm behind the wheel centre . This position was selected because it is the position recommended by the manufacturer. Foot rest height was adjusted to provide for subject comfort and safety. Mass addition The experimental design involved adding mass to the test chairs. Because pilot studies and others (Kirby, Atkinson, & MacKay, 1989 ; Lemaire et al., 1991 ; Loane & Kirby, 1985, 1986) have shown that the drag of the usermachine combination changes depending on where the mass is added, the two mass additions were placed such that they were equally distributed over the contact surface area of the seat of the chair . This was accomplished by use of several straps which supported either a 5 kg or 10 kg mass just beneath the 17  chair seat . The correct number of strips of cold roll steel were sewn into a fabric sleeve in order to comprise the needed mass which was determined by measurement on a floor scale . The fabric permitted the steel bars to be expanded or contracted in order to evenly cover the surface area of the different seat sizes in the different test chairs. Laboratory set-up A temporary laboratory was set up in a large wheelchair accessible gymnasium . Two parallel wheeling runways were established with identical dimensions by use of long strips of smooth canvas placed on the wooden gymnasium floor . One runway served as the practise runway and the second as the test runway . Two gen-locked Panasonic digital video cameras (model WV-D5100) were positioned at an approximate distance of four meters and at 45 degrees to the test runway (at 90 degrees to each other) . The wheeling movement was recorded at 60 Hertz (Hz) on super VHS quality videotapes. Lights were placed just above each camera lense to illuminate the markers and the aperture on the video camera lens was reduced (SES set at 1/500) in order to darken the image producing the maximum contrast between the light reflective markers and the background. A volume of five meters (length) by two meters (width) by one and a half meters (height) was calibrated by construction of a metal rectangular prism of these dimensions placed in the center of the test runway . Fishing line was hung from eleven points placed around the prism from which a total of thirty-three styrofoam spheres were suspended . The spheres were distributed over the entire volume of the rectangular prism . Weights were attached to the bottom of each line and placed in large containers of water to minimize oscillations in the position of the spheres . The position of these thirty-three spheres was determined by direct measurement and trigonometric calculation based on a single sphere being given the designated coordinates in space of 0,0,0 mm and all the others being defined relative to this origin . The prism was then videotaped and the position of the thirty-three spheres was digitized from both camera angles . After direct linear transformation, the three dimensional coordinates of the thirty-three spheres was determined . The net residual mean square error between the actual position of the spheres and the video determined position of the spheres was found to be 7 .7 mm . Once the calibration of the wheeling space had been completed, the rectangular prism 18  was removed from the test runway and the video cameras were maintained in the same position throughout the remainder of the data collection phase of the study . Two photoelectric cells (Armaco, model R401) were positioned at the beginning of the calibrated section of the runway . When the subject passed through the beam between the photocells, an event light was triggered which was visible in both cameras . This event light was used to synchronize the wheeling movement in the two video cameras . There was sufficient space on the test runway for the subjects to complete three to four wheeling cycles before they entered into the calibrated volume . They also had sufficient space to complete several more cycles at the end of the calibrated volume . In this manner, the physical dimensions of the test runway permitted steady state wheeling to occur. The events of the hand contact (grab) on the wheel rim and hand off (release) was determined by the use of a hand switch which responded to pressure by producing an electrical signal . A number of switches of varying sizes were made in order to provide the maximal comfort and minimal distraction for the subjects . It was important to have several very small switches because the hands of some of the subjects were very tiny and it would have been conceivable that a large switch would have altered wheeling style in those subjects . The switch was linked by way of a comparator to two small light emitting diodes (LEDs) which were placed within view of both cameras . In this way, the moment of contact with the wheel (and release from the wheel) was determined with the accuracy of the sampling rate of the Peak System (60 Hz). The mass of each subject and each wheelchair (as well as chair parts) was determined by use of a floor scale (Toledo, model 8132) with a surface area measuring one by one meters which was rented for this study . The subjects could be easily and safely positioned on the scale . The scale was calibrated to within 0 .1 kg. The subjects' mass was determined by subtracting the mass of chair alone from the mass of the user plus chair mass . A variety of wrenches and tools was available to be used to remove the castor wheels for weighing. Tire pressures were measured and adjusted with a standard bicycle pump with a pressure gauge. A padded treatment table of adjustable height was positioned in the laboratory so that the subject could sit comfortably while his/her wheel chair 19  was being weighed . In addition to the test equipment, a VCR and monitor with tapes from local disabled sporting associations was made available for the parents and friends of the subjects to view during the testing . Included in these tapes were videotapes demonstrating various wheelchair skills in a variety of environments . The VCR was also used to permit each subject to view their own wheeling style as recorded by the video cameras at the end of the testing session. A play area was also provided for subjects and siblings to use. Data collection Each subject was assessed in the laboratory after consent, neurological testing, matching, and wheelchair fitting had been completed . Each subject was given as much time as necessary to become comfortable in the test chair and to become familiar with the test space and equipment . The subjects were asked to propel their test wheelchairs across the practise runway at a constant speed of 2 m/sec . This speed was selected based on data from prior studies with adults (Bednarczyk & Sanderson, 1991 ; Van Der Woude et al ., 1986 ; ). There are few published reports on wheelchair propulsion in children . Jarvis et al ., (1982) found that disabled children, though weaker than their able bodied counterparts, were able to comfortably propel a simulated wheelchair at 2 .5 m/sec . After consultation with pediatric clinicians as well as screening of the children at the Meningomyeleocele Clinic at BCCH, it was evident that the children were able to comfortably wheel at 2 m/sec . None of the subjects who were tested indicated either verbally or through their wheeling behavior that they had any difficulty with this wheeling speed . The subjects had a number of trials on the practise runway until they felt comfortable with the test wheeling speed . Reflective markers were placed over the right side of the subjects' neck (spinous process of C7), joint centres of the shoulder, elbow, wrist, hip, knee, and ankle as well as the right tire and wheel centre . The placement of markers is illustrated schematically in Figure 3 .  20  Figure 3 . Schematic representation of the subject in the wheelchair identifying the marker locations. The thumb switch was taped to the subjects right thumb and the cabling which went to the comparator and LEDs was taped along the length of the subjects right arm . The cable was supported by the author during all trials in order to ensure subject safety and to eliminate any extra drag on the system. The subjects were given feedback as to their wheeling speed by the author who used a stopwatch and a measured distance on the floor to compute approximate wheeling speeds . In pilot studies, a cycle computer had been used to give the subjects feedback on their wheeling speeds . However, it was found that the pediatric subjects were unable to concentrate on simultaneous wheeling and watching the digital readout on the cycle computer . Thus, in this study, all subjects were given verbal feedback from the author by way of the stop watch times and encouraged to either maintain, increase or decrease their wheeling speed. After warm ups and practise on the first runway, the subjects were then escorted to the test runway and the purpose and function of the photoelectric cells and the video cameras explained . Each subject was given a minimum of one trial on the test runway before data collection began . Each condition (no added mass, + 5 kg, and +10 kg) was presented in a previously determined 21  counterbalanced order . The exception was the "own chair" condition which was always presented as the last condition in order to evaluate if there were any differences in wheeling style at the end of the experimental session relative to the beginning of the session which might have indicated fatigue . The subjects were asked to complete five pushing trials for each condition. Drag force has been measured in a variety of ways. Van der Woude, De Groot, Hollander, Van-Ingen Schenau, & Rozendal (1986) defined drag as equal to the rolling resistance, internal friction, and the effect of gravitational force . In their studies, drag force was measured by towing the user and chair on a motor driven treadmill at a variety of slopes of the treadmill . Another method of measuring drag was described by Coutts (1990) . In this case, the usermachine system was accelerated to a given velocity and than allowed to coast when only the drag force was being applied to the system . Drag, measured in this way, was defined for a user-chair system as the product of the total mass, acceleration and two inertial factors for the front and rear wheels (Coutts, 1991). The method described by Coutts was selected for use in the present study in which the user-chair system was passively accelerated to a starting speed of approximately 2 m/sec and then allowed to accelerate with only a drag force being applied . Two coast down tests were recorded at the end of each condition . This allowed a short rest period for the subjects because they were not required to push their chairs during this portion of the data collection. The combination of wheeling and coast down tests meant that a minimum of twenty-eight trials were recorded for each subject. The author rejected wheeling trials on the basis of poor contact with the thumb switch or inadequate wheeling speed and additional trials were completed until five acceptable wheeling trials and two acceptable coast down trials had been recorded for each condition . The entire period of time the subject was in the test laboratory for data collection was between 60 and 75 minutes . Data collection occurred once for all subjects except for two subjects (one adult and one pediatric subject) for whom two sets of data were collected due to technical difficulties . Data analysis Angular data A single wheeling cycle (from grab to the subsequent grab) was selected from each trial for analysis. Sufficient distance was available for the subjects to 22  complete several wheeling cycles before entering the calibrated space . The first complete cycle after the subject entered the calibrated space was selected for analysis . The positions of the light reflective markers were digitized from both camera angles and converted to real spatial units by use of an algorithm called the Direct Linear Transformation (Peak Performance Technologies software) . In this process, the host computer combined the digitized data coordinates from the two-dimensional data from each camera view and produced a single three-dimensional coordinate data file . During this process the three-dimensional coordinate data were smoothed with a fourth order, lowpass Butterworth digital filter at 6 Hz . The manufacturer recommends filtering at 6-10 Hz for most normal human movement patterns . Filtering at 6 Hz was found, through experimentation, to provide the best signal to noise ratio for the wheeling data . The process of digitization of the more than 1,400 trials from both camera angles in this study was done manually with the assistance of the automatic tracking software program in the Peak System . In those cases where markers were temporarily obscured from view by body part movements, interpolation (available with the Peak System software) was used to estimate the position of the markers . Once the three-dimensional coordinates of the markers had been generated, the displacement data were then differentiated with respect to time within the Peak software programs to produce the linear velocity data files. The angular data were then computed based on link segment modeling of the three dimensional marker coordinates based on the following definitions of angles . The upper arm segment was defined by the elbow and shoulder markers, and the lower arm segment by the elbow and wrist markers . The trunk segment was defined by the shoulder and hip markers . The elbow angle was defined as the angle between the upper and lower arm segments . Full elbow flexion corresponded to 0 degrees and full elbow extension to 180 degrees. The shoulder angle was defined as the angle between the upper arm segment and the trunk segment . The position of the upper arm segment parallel to the trunk was defined as 0 degrees and the position of the shoulder in extension behind the trunk was defined as 90 degrees. The trunk angle was a projected angle between the trunk segment and a vertical plane drawn through the hip marker . The position of the trunk segment parallel to the vertical was defined as 0 degrees and the position of full forward 23  flexion of the trunk as 90 degrees . The shoulder abduction angle was also a projected angle and was defined by the projection of the upper arm segment on a sagittal plane drawn through the shoulder marker . The position of shoulder adduction with the upper arm parallel to the vertical was defined as 0 degrees and full abduction as 90 degrees . The angular definitions and ranges of motion of the defined angles in this study are summarized in Table 1. Table 1 . Definitions of angles and segments Segments  Markers which define segments  Upper arm Lower arm Trunk  elbow - shoulder elbow - wrist shoulder - hip  Angles  Segments which define angles  Elbow Shoulder Trunk Abduction  upper - lower arm 0° = full flexion upper arm - trunk 0° = arm at side trunk - vertical plane 0° = trunk upright upper arm-sagittal 0° = arm at side plane  Ranges  of Motion 180°=full extension 90°=full extension 90°= full flexion 90° = full abduction  After the various body joint angles had been computed, and the angular velocity and acceleration files had been generated, each file was time normalized using software developed in the Biomechanics Laboratory at U .B .C. The time normalized angular files were then within-subject ensemble averaged (by another laboratory program) for each condition . A minimum of four trials was used in all of the within subject averages . Both software programs (the normalizing and the averaging program) were validated to determine their accuracy (details are in Appendix A) . After checking for errors on the Peak System by viewing each of the five trials in a given condition through the graphics portion of the program (and redigitizing trials as necessary), all of the files were exported to a Macintosh computer for between-subject ensemble averaging of the two groups (adult and pediatric) and graphing (using Microsoft Excel and Kaledeigraph programs) . 24    The reliability of the angular data was determined by digitizing the same trial five times consecutively . The accuracy (as defined by the standard deviation) was found to be under two degrees for each of the four angular parameters (See appendix B for details). Determination of wheeling velocities The three-dimensional linear velocity files of the wheel marker which had been generated on the Peak System by differentiation of the three-dimensional linear displacement files, were used to determine the actual average resultant velocity of the wheel marker for each of the trials . This average velocity of the wheel marker was then used to determine average wheeling velocities for each subject, for each condition . Ensemble averages of the average wheeling velocities for each subject were created in order to determine the average wheeling velocities of the two subject groups (adults and children). Drag force calculations Drag force was determined from average resultant linear velocities of the axle marker recorded during coast down tests under each condition . After all of the digitization for the angular data had been completed, the project setup file in the Peak System was altered to record only the movement of the axle for the coast down tests . A single trial was digitized for the coast down conditions of own chair, test chair with no added mass, and test chair with 10 kg of added mass. In each case, eighty frames of the coast down trial were digitized from each of the two cameras and the resultant three-dimensional coordinates of the two markers were determined . Then the linear velocity files for the markers were exported from the Peak System to the Macintosh computer where the line fitting program in Kaleidagraph was used to determine the value for the acceleration by calculating the best fit of the slope of the velocity-time curve . In each case r2 values were used to determine the acceptability of the line fit to the coast down data file . This value for average acceleration was used, along with measured values for chair mass and inertial properties, to determine the drag force . The following formula for drag force was used in this calculation: F= Ma + 2 lr(ar/rr)/rr + 2 lc(ac/rc)/rc (Equation 1) where: F = the force of drag M = the mass of the user/chair system a = linear acceleration of the system 25  Ir= the moment of inertia about the rear wheel lc= the moment of inertia about the caster wheel rw= the radius of the rear wheels rc= the radius of the caster wheels ar= the angular acceleration of the rear wheel and ar= ai/rr ac= the angular acceleration of the 'caster wheel and ac= ai/rc rr= the radius of the rear wheel (in m) and al= the resultant linear acceleration of the rear wheel which was measured. The linear acceleration of the front wheel was assumed to be equal to that of the rear wheel. The moment of inertia of the caster and rear wheel was estimated according to the method described by Coutts (1991) where: If = 0.8 (d02 + d ; 2) and do is the outside diameter of the wheel (in m) and di is the inside diameter of the wheel (in m). Determination of % propulsion The timing of wheel contact and release was used from the events (grab and release) recorded from the LEDs on the videotapes (generated by the thumb switch) . The percent propulsion was determined by taking the percentage of a raw wheeling cycle (defined as one grab to the next consecutive grab) in which the hand was in contact with the wheel (grab to release) . This was determined for all five trials for each subject and each condition . Averages across subjects and across conditions were then computed. Statistical analysis, The analysis of variance (ANOVA) of the data was performed using BMDP Biomedical Computer program (Dixon, 1990) . The angular data was time normalized into twenty-one intervals of 5% increments in which 0% was defined by the first grab and 100% of the cycle by the subsequent grab . A two(groups)-by-six-(propulsive phase of wheeling cycle)-by-four-(conditions) repeated measures ANCOVA of the angular and % propulsion data, with velocity as the covariate, was performed . An ANCOVA was used in order to correct for the observed differences in wheeling velocities of the two groups 26    because other teams (Masse & Montagne, 1989 ; Van der Woude, Veeger, Rozendal, Van Ingen Schenau, Rooth, & Van Nierop, 1988) had shown changes in kinematics as a result of changes in wheeling speed . Multivariate ANCOVAs could have been used but univariate ANCOVAs were selected instead because the ratio of sample size and the number of dependent variable and levels of the repeated measure was less than optimal (Huberty & Morris, 1989). The first 25% of the wheeling cycle (six increments in the normalized wheeling cycle) was selected in the angular data analyses in order to reduce the number of repeated points in time and because it represents the portion of the wheeling cycle where force is being applied (which is the portion of the movement of greatest clinical interest). For the other data sets (average angular CVs, and wheeling velocity), a two-(groups)-by-four-(condition) repeated measures ANOVA was used . In all cases, the Greenhouse-Geisser adjusted p values were used in order to compensate for any violations of assumptions of sphericity. Significance was defined by adjusted p values < 0 .05. Where significance was determined, posthoc Scheffe tests were employed. Determination of the coefficient of variation (CV) The determination of average CVs was felt to be important in order to compare the variability of the angular parameters in the adult and pediatric group . In this study, the CV was determined for the angular data by determining the value of the standard deviation (s .d.) divided by the mean value of the twenty-one five percent intervals over the time-normalized wheeling cycle . The equation for determination of the CV is shown below. CV = (s .d ./ average) X 100  (Equation 2)  Once the CV had been calculated for each of the twenty-one points in the time normalized wheeling cycle, average CVs were computed across the propulsive phase (first 25%) and recovery phase (last 75%) of the wheeling cycle . The within-subject CVs were computed by determining the CV of the five trials per condition per time point for each subject . Average CVs were computed across the subjects within each group and across the propulsive and recovery portions of the wheeling cycle . This process was repeated for each of the four angular  27  variables . Thus, the determination of each average within-subject CV involved manipulation of 1,600 data files . RESULTS Subject characteristics The mass of the two groups differed considerably with the mean pediatric mass being 37 .41 ± 9 .95 kg and the adult mass being almost double that at 68 .46 ± 8 .67 kg . The mean age of the ten pediatric subjects was 11 .3 ± 2.2 years and the mean age of the ten adult subjects was three times that at 33 .5 ± 8.9 years . Pediatric and adult subjects were paired and matched to within 4 ASIA points of one another . The subjects were closely matched neurologically as evidenced by the mean ASIA scores in the two groups (pediatric = 56 .1 ± 3 .48, adults = 55 .2 ±4 .73) . There were 3 female and 7 male subjects in each of the two groups . The characteristics of the twenty subjects in the study are summarized in Table 2. Bone scans for nine of the ten pediatric subjects were obtained through physician referral while the subjects were at the Meningomyelocele Clinic in order to determine if any of the pediatric subjects had growth defects . The radiology reports of these scans indicated that the bone ages were within one standard deviation of expected normal values for the stated chronological age of the child . These bone scans indicated that none of the pediatric subjects in this study demonstrated the presence of concomitant growth defects.  28  Table 2 .	 Characteristics of the twenty subjects in the study group Pediatric Subject # 01  02 03 04 05 06 07 08 09 10 Adult Subject # 01  02 03 04 05 06 07 08 09 10  MASS (kq)  ASIA SCORE  AGE (yrs)  SEX  34.6 43 .4 32 .8 54 .5 43 .4 45 .8 32 .4 23 .9 40 .4 22 .9  50 56 58 63 54 54 54 56 58 58  14 14 13 12 12  M M M M F M M F M F  84 .7 70 .5 65 .6 74 .7 63 .9 69 .2 61 .1 76 .0 65 .2 53 .7  50 56 58 66 52 54 50 54 58 54  30  10 11  8 11  8 52 32 27 32 22 42 41 26 31  M M F M M M F M M F  Wheeling velocities The actual wheeling velocities determined from the linear velocity files based on wheel markers are shown in Figure 4 for the two groups wheeling over the four conditions. The subjects were asked to wheel at an overground velocity of 2 .0 m/sec. It is apparent that all of the subjects were able to achieve the nominal wheeling speed over all of the test conditions . In all cases, the actual wheeling velocities were slightly higher than the target wheeling velocity of 2 .0 m/sec. The averaged group wheeling velocities were 2 .26 ± 0 .39 m/sec for the pediatric group and 2 .38 ± 0 .31 m/sec for the adult group wheeling in their own chairs . A two-(groups)-by-four-(conditions) repeated measures ANOVA of these data showed a significant groups effect (F (i,i8 )= 4.87, p=0 .04), a nonsignificant conditions effect (F(3 , 54)= 2 .36, p=0 .12), and a nonsignificant groups-bycondition effect (F( 3,54)= 2.27, p=0 .13). These data indicate that while the  29    pediatric group were wheeling at velocities significantly lower than the adult group on average, that the two groups responded in a similar fashion in terms  4 PEDIATRIC ADULT 3  OWN  0 kg 5 kg CONDITION  10 kg  Figure 4 . Average overground wheeling velocities recorded from the wheel markers for the pediatric and adult groups of subjects (10 subjects/group) . Each value is the average of at least 4 trials.  of wheeling velocities to the test conditions and that the velocities were similar under all four test conditions. 0/0Propulsion data, The percentage of the the wheeling cycle that the subjects spent in contact with the wheel (% propulsion) is shown in Figure 5 for both groups under the four wheeling conditions . The two groups spent comparable proportions of the wheeling cycle in propulsion (pediatric group = 24 .45 ± 7.29%, adult group = 24 .41 ± 7 .61%) in the own chair condition A repeated measures two-(groups)-by-four-(conditions) ANCOVA of these data, with wheeling velocity as the covariate, showed that neither the main nor interaction effects were significant (group F(1 , 18)=0 .00, p=0 .99 ; condition F(3 ,54)=0.66, p=0 .54, and group-by-conditions interaction F(3,54)=0 .37, p=0 .72). Interpretation of these results were that the portion of the time spent in the propulsive and recovery phases was similar in both groups and that both  30    groups showed no alteration in the relative timing of grab and release in response to the mass additions.  PEDIATRIC  ADULT GROUP Figure 5 . Average % propulsion for the two groups under the four wheeling conditions . Each value is the average of five trials for 10 subjects in each group .  Angular data Comparison of test chairs and subjects own chairs The first angular data which are presented are the comparison for the two groups wheeling in their own chairs and the unloaded test chairs . This is important in order to demonstrate that there was not an effect on the angular kinematics caused by the test chairs themselves . The data for the four angular parameters (shoulder, elbow, trunk, and shoulder abduction) are shown in Figure 6 . It is evident from visual inspection of these graphs, that for the pediatric subjects the differences in the angular patterns of wheeling between the test chairs and the subject's own chairs are small in all of the four angular parameters .  31    PEDIATRIC SUBJECTS  J W  I 0  25  50 % CYCLE  75  100  I  0  25  50 % CYCLE  75  100  0  25  50 %CYCLE  75  100  15-  t z  TEST CHAIR -- -- OWN CHAIR  10  0  X W  J LL  5  0  1  0  25  50 %CYCLE  75  100  Figure 6 . Comparison of angular patterns (elbow, shoulder, trunk, and shoulder abduction, in degrees) in pediatric subjects (n=10) wheeling in test chairs (Kuschall 3000) and subjects' own chairs.  Similar angular data are presented in Figure 7 for the adult group . For the adult group, the differences between the wheeling angular kinematics in the subjects' own chairs and the unloaded test chairs were also small . From visual inspection of the graphical data, it appears that the adult group demonstrated more trunk flexion in the test chairs than in their own chairs . This also appears to be more trunk movement generally in the adult group than seen in the previous graphs for the pediatric group . However, it should be noted that the variability in the trunk angular data was large.  32    ADULT SUBJECTS  .S   0  r 0  25 '  50 %CYCLE  75  100  0  i 25  TEST CHAIR OWN CHAIR  50 % CYCLE  t 75  t 100  80 ¶601 40-  ~ .~  z  i 01 0  TEST CHAIR OWN CHAIR  220 0  i 25  — 50 %CYCLE  1— 75  a  -~  00  100  25  50 %CYCLE  75  100  Figure 7 . Comparison of angular patterns (elbow, shoulder, trunk, and shoulder abduction, in degrees) in adult subjects (n=10) wheeling in the test chairs (Kuschall 3000) and subjects' own chairs.  The results of the ANCOVA of these data, using velocity as the covariate, were that the main effects for these analyses were significantly different and will be discussed in more detail in the following sections . The two by four, groupsby-conditions effect was found to be nonsignificant for three of the four angular variables (shoulder : F(3,54)= 1 .25, p=0.30 ; trunk F(3,54)= 1 .17, p=0.91 ; shoulder abduction : F(3,54) = 1 .33, p=0 .28) over the propulsive phase (first 25%) of the wheeling cycle . Only the elbow angle showed a significant groups-bycondition effect (F(3,54)= 4.87, p=0.01) . A post hoc Scheffe's test indicated that it was the own chair condition for the pediatric group that differed from the others for this variable . These results are summarized in Figure 8 where the elbow 33    angular data is averaged across the first 25% of the wheeling cycle and across subjects in each group for the four conditions .  130   ADULT PEDIATRIC  120 110 z 0  L.--  ui N100 z 0 w  a  w >  x  90  a  80 0 kg 5 kg 10 kg OWN  Figure 8 . The elbow angle averaged across the first 25% of the wheeling cycle (in degrees) for the two groups (adult and pediatric, n=10/group) under the four test conditions.  The post hoc Scheffe's test indicates that the significant differences found in the elbow angle ANCOVA were due to the fact that the pediatric group showed, on average, more elbow extension in their own chairs than in the test chairs whereas the adults had a slight tendency to show less elbow extension in their own chairs. Angular changes under mass addition conditions The data in Figure 9 demonstrate the effect of the two mass additions (5, and 10 kg) on the angular kinematic data ensemble averaged across the ten subjects in the pediatric group . It is evident that the angular data shows very little change in any of the four angular parameters between the mass addition conditions .  34    160 —  0  PEDIATRIC GROUP  25  50 % CYCLE  75  100  r 25  0  50 % CYCLE  75  100  15-  + 0 kg Mass + 5 kg Mass +10 Kg Mass  10 5  00  25  50 % CYCLE  75  100  0  25  50 %CYCLE  75  100  Figure 9 . The effect of mass additions on the angular patterns (elbow, shoulder, trunk, and shoulder abduction, in degrees) in the pediatric group (n=10) over a normalized wheeling cycle.  As noted earlier, the only significant groups-by-condition effect was for the elbow angle and that was due to the differences in the own chair condition and not due to any significant changes in response to the mass addition conditions . Similar data is presented in Figure 10 for the adult group . Once again it is evident from inspection of the angular data that the mass additions had little effect on the four angular kinematic variables.  35    -+ 0 kg Mass  + 5 kg Mass --- +10 kg Mass  0  25  50 % CYCLE  75  100  15-  10-  z 5  xw  + 0 kg Mass + 5 kg Mass --- +10 kg Mass   LL 0 0  25  50 % CYCLE  75  100  Figure 10 . The effect of mass additions on the angular patterns (elbow, shoulder, trunk, and shoulder abduction angle, in degrees) in the adult group (n=10) over a normalized wheeling cycle.  Differences between adult and pediatric subjects Despite the fact that differences were not found in either group as a result of the mass additions, it was apparent that the basic wheeling parameters of the two groups (adult and pediatric) were different . The two groups are compared with respect to the four angular parameters in Figure 11 in the own chair condition . The elbow angular data appear to be very similar during the propulsive phase between the two groups . The pediatric group shows more shoulder extension (minimum pediatric group 33 .8 ° versus adult group 23 .2 °) than the adult group and more shoulder abduction (maximum pediatric group 65.6 ° versus adult group 56 .3 °) than the adult group . The average trunk angular data for the two groups appears to be quite similar. 36    -PEDIATRIC  ADULT 0 25  50 % CYCLE  75  100  15  0  0 z  25  50 % CYCLE  75  100  80  a 10 z 0 X  J I  i.  0  -PEDIATRIC  ADULT  so  02  po D wa  m U  40  Q  5  0 0  25  50 CYCLE  75  100  Figure 11 . Comparison of pediatric and adult angular data (elbow, shoulder, trunk, and shoulder abduction angle, in degrees) in the subjects' own chairs . Each group contains the mean from 10 subjects.  The average angular data for the two groups and four angular parameters in the test chair condition are shown in Figure 12 . The pediatric group demonstrate, on average, more shoulder extension (maximum pediatric group 66 .6 ° versus adult group 60 .3 °) and less elbow extension than the adult group (maximum pediatric group 138 .0 0 , adult group 144 .7 °) throughout the propulsive phase of wheeling . The pediatric group also demonstrate less trunk flexion (pediatric maximum = 5 .52 °, adult maximum = 8 .47 0) and less shoulder abduction (pediatric maximum = 55 .95 °, adult maximum = 66 .48 °) than the adult group . However, the pattern of change of all of these angular variables over time was similar in both groups .  37    TEST CHAIR + 0 kg MASS  0  25  50 % CYCLE  75  80,  100  25  50 % CYCLE  75  15  PEDIATRIC ADULT  z  5 w LL  0 0  25  50 %CYCLE  75  100 % CYCLE  Figure 12 . Comparison of pediatric and adult angular data (elbow, shoulder, trunk, and shoulder abduction angle, in degrees) in test chairs with no added mass . Each group contains the mean from 10 subjects.  A repeated measures ANCOVA, with velocity as the covariate, of these angular parameters showed a significant groups effect for three of the four variables (elbow : F(1 , 18)= 12 .27, p=0 .003 ; shoulder F(1 ,18)= 16 .46, p=0.0007; shoulder abduction F(1,18)= 20 .58, p=0 .0003) . Only the trunk angular data showed a nonsignificant groups effect (F(1 ,1 8)= 0 .40, p=0 .53) which was perhaps due to the small absolute value and large measurement error in this parameter. These data indicate that there were significant differences in the angular kinematics between the two groups . What is of interest, is that the groups-by-time-by-condition effects were nonsignificant for all of the angular variables (elbow : F(15 ,270)= 0 .60, p=0.62 ; shoulder F(15,270)= 0 .85, p=0 .50; trunk F(15,270)= 0.88, p=0 .58; and shoulder abduction F(15,270)= 0 .83, p=0 .50). 38     These data indicate that while there were significant differences between the two groups, they both responded in a similar fashion to the test chair and added mass conditions (Figures 13 and 14).  160-  TEST CHAIR + 5 kg MASS  0'  PEDIATRIC - . ADULT I  0  25  I  i  50 % CYCLE  75  100  0  25  0 Z  15-  PEDIATRIC ADULT  75  -  Z 10-  50 % CYCLE  s0-  H V  • p  m f U 2  5  c  .0 /  20PEDIATRIC  ADULT  0 0  I  0  25  50 % CYCLE  1  I  75  100  0 0  25  50 % CYCLE  75  100  Figure 13 . Comparison of pediatric and adult angular data (elbow, shoulder, trunk, and shoulder abduction angle, in degrees) in test chairs with 5 kg of added mass . Each group contains the mean from 10 subjects.  It is interesting that the flattened shape of the elbow angle curve in the recovery phase for the adult group is maintained throughout all of the test conditions . The pediatric group maintained more elbow extension throughout the propulsive phase and more elbow flexion throughout the recovery phase of wheeling than the adult group . The pattern of more shoulder extension in the pediatric group as opposed to the adult group is maintained throughout all of the mass addition conditions . The pediatric group show less shoulder abduction than the adult group in the 0 kg and 5 kg of added mass conditions . It  39     160-  TEST CHAIR + 10 kg MASS  8060, z ~yg40-  .~  •  sN Z xw  I  0  25  50 % CYCLE  1  75  20-  PEDIATRIC ---°- ADULT  1 100  00  25  50 % CYCLE  75  100  100  0  25  50 %CYCLE  75  100  I  15W  z	4 -1  -PEDIATRIC .ADULT  10-  Y =  z2  - -  f a 5,. 00  25  50 % CYCLE  -  75  Figure 14 . Comparison of pediatric and adult angular data (elbow, shoulder, trunk, and shoulder abduction angle, in degrees) in test chairs with 10 kg of added mass . Each group contains the mean from 10 subjects. is only in the 10 kg of added mass condition that the adult group show less abduction and the angular patterns of the two groups become more similar. Angle-angle plots for pediatric-adult data The differences between the pediatric and adult group angular data are summarized in Figure 15 where the shoulder angle is plotted against the elbow angle . This angle-angle plot over a time normalized wheeling cycle demonstrates the relationship of these two linked joints . The propulsive phase of the wheeling cycle is the lower portion of the curve and the recovery phase the upper portion . The propulsive phase of the wheeling cycle shows a similar pattern of elbow extension in both groups . However, the pediatric group show more shoulder extension than the adult group especially towards the end of the propulsive phase . During the recovery phase the adult group shows a rapid  40    PEDIATRIC  z 0120co  Z W  F  ' 100Propulsion 80 20  40 60 SHOULDER ANGLE  i 80 EXTENSION -  20 40 SHOULDER ANGLE  60 EXTENSION  80  -f'  Figure 15 . Shoulder angle-elbow angle plots (in degrees) for the pediatric and adult groups Each plot contains the mean from 10 subjects.  reduction in elbow extension while the pediatric group maintain more elbow extension throughout the recovery phase . This pattern of elbow extension gives the pediatric plot the appearance of a more rounded, open shape than the adult plot . The area enclosed by these two plots was measured and determined to be 598 .7 degrees2 for the adult group and 901 .6 degrees 2 for the pediatric group . In other words, the difference in the average pediatric wheeling angular style resulted in an enclosed area 1 .5 times greater than for the adult group. Average within and between subjects CVs The average within-subject CVs for the angular data for the adult group are shown in Figure 16 . It is apparent that the average within-subject CV for the elbow angular data is low (average mean CV across all three conditions 5 .24 ± 0 .48 %) and that there are no differences between the variability of the angular data during the recovery phase as opposed to the propulsive phase of wheeling . The CVs for the shoulder angular data are greater than in the elbow angular data (average mean CV across all three conditions 9 .93 ± 1 .61 %) . The within-subject CV for the shoulder abduction angle is 6 .34 ± 0 .90 % averaged across the three conditions . The within-subject CV for the trunk angular data for the adult group is much higher than the CV calculated for the other three angular parameters (55 .54 ± 7 .72 %) . It is not surprising that the CVs for the trunk angular data are high because the CV is inversely proportional to the 41    >W  UJ WZ 0 4 WO W  am  PROP  PROP  REC  PROP  REC  PORTION OF CYCLE  PORTION OF CYCLE  REC  PROP  REC  PORTION OF CYCLE  PORTION OF CYCLE  Figure 16 . Average within subject CVs for the four angular variables (elbow, shoulder, trunk and shoulder abduction ) in the adult group during the propulsive (PROP) and recovery (REC) phases of wheeling.  mean values and the mean values for the trunk angular data are very low. Therefore the CV shows more variability with a mean of small magnitude . It is also evident that there is very little change in the CVs in response to the mass addition conditions . The overall angular CV for the adult group, averaged over the shoulder, elbow, shoulder abduction angular data and over both the propulsive and recovery phases of the wheeling cycle was 7 .2%. Similar data are shown in Figure 17 for the pediatric group . The average within-subject CVs for the pediatric group for the elbow angular data are similar in the two groups (average mean CV across all three conditions pediatric 7 .09 ± 0 .87 %, adult 5 .24 ± 0 .48 %) . The shoulder angular data are once again similar in the two groups (average mean CV across all three conditions pediatric 11 .45  42    ± 1 .50 %, adult 9 .93 ± 1 .61 %) . However the values for the shoulder abduction angle average CVs were much higher in the pediatric group than the adult  PROP  REC  PROP  PORTION OF CYCLE  PROP  REC  PORTION OF CYCLE  REC  PROP  PORTION OF CYCLE  REC  PORTION OF CYCLE  Figure 17 . Average within subject CVs for the four angular variables (elbow, shoulder, trunk, and shoulder abduction )in the pediatric group during the propulsive (PROP) and recovery (REC) phases of wheeling .  group (average mean CV across all three conditions pediatric 16 .87 ± 3 .36 %, adult 6 .34 ± 0 .90 %) . The within-subject CVs for the trunk angular data are, once again, much higher (average mean CV across all three conditions pediatric 59 .44 ± 8 .54 %, adult 55.54 ± 7 .72 %) reflecting the low mean value of the trunk data . As noted with the adult data, the differences between the propulsive and recovery phases of wheeling were small as well as the differences in response to the wheeling conditions . When the CV was averaged over the shoulder, elbow, and shoulder abduction data (both propulsive and recovery phases) for the pediatric group, the CV was 11 .8%. This overall average CV for the adult group was 7 .2%. The differences in overall average CVs between the two groups is most likely due to the larger 43     CVs for the shoulder abduction data in the pediatric group . An ANOVA of the within-subject CV data was performed in order to determine if the differences in variability noted between groups and over conditions were significant . The analysis of the average CV data indicated a significant groups effect for the average trunk angular CVs (F(1 , 18)= 4 .56, p= 0 .05) and shoulder abduction angular CVs (F(1,18)= 12.59, p=0 .002) angles . The groups-by-condition and groups-by portion of wheeling cycle were non-significant in all cases . The ANOVA of the elbow and shoulder CVs were non-significant in all cases . All other interaction effects were found to be nonsignificant. The following data demonstrate the between-subject CVs for the angular data in the adult group (Figure 18) . The between-subject CVs are generally expected to be much higher than the within-subject CVs .  PROP  REC  PROP  PORTION OF CYCLE  REC  PORTION OF CYCLE  • 60W  w 2Q 50-  0 • 40>0 3000  D  + 0 kg +5kg + 10 kg  wO  wa Q  e  2010-  O  PROP  REC  y  PORTION OF CYCLE  PROP  REC  PORTION OF CYCLE  Figure 18 . Average between subject CVs over the wheeling cycle for the four angular variables (elbow, shoulder, trunk, and shoulder abduction) for the adult group during the propulsive (PROP) and recovery (REC) phases of wheeling .  44    The average between-subjects CVs for the elbow angular data were 10.26 ± 2 .47 for the adult group averaged over the three test conditions . These CVs were higher than the corresponding within subject CVs previously noted. The average between-subject CVs for the shoulder angular data were 21 .23 ± 3 .06 % for the adult group and are about double the within-subjects CVs for this variable . The average CVs for the shoulder abduction angular data were 20 .25 ± 3.84 % . The average CVs for the trunk angular data were higher than those for the other variables (as seen in the within subject CVs) with an average over the three conditions of 95 .38 ± 4 .65 %. The average between-subject CVs for the angular data in the pediatric group are shown in Figure 19 . As with the adult group, the average CVs for  r  g u- w > w< 43 wm 4W  50  ® + 0 kg + 5 kg 0 + 10 kg  +0 kg p + 5kg 0 +10 kg  40 30  4w 20 10  : : : :•': coawao-o-p  PROP  O >a 2  20 10  REC  PROP  REC  PORTION OF CYCLE  PORTION OF CYCLE  ye 60  200  0 50 oa  0  PROP  PROP  REC  REC  PORTION OF CYCLE  PORTION OF CYCLE  Figure 19 . Average between subject CVs for pediatric group for the four angular variables (elbow, shoulder, trunk, and shoulder abduction ) during the propulsive (PROP) and recovery (REC) phases of wheeling. the elbow angular data were similar to the adult average CVs (average across the three conditions pediatric 11 .03 ± 2 .05 %, adult 10 .26 ± 5 .75) and are 45    higher than the within-subject CVs previously described . The average CVs for the shoulder angles were 21 .01 ± 2 .55 % for the pediatric group (as compared to 21 .23 ± 3 .06 % for the adult group) and were unchanged over the portion of the wheeling cycle . The average CVs for the shoulder abduction angle between-subjects in the pediatric group were 27 .89 ± 3 .29 % (average over the three conditions) which was somewhat higher than the average betweensubject CVs for the adult group (20 .25 ± 3 .84 %) . Finally, the average CVs for the trunk angular data were 56 .56 ± 4 .17 °/O in the pediatric group which is considerably less than the between-subject CVs for the adult group (95 .38 ± 4.65 %) . The grand average of the average elbow, shoulder, and shoulder abduction CVs between-subjects was 16 .9% for the pediatric group and 16 .6% for the adult group . These results are summarized in Figure 20. BETWEEN SUBJECTS CV  WITHIN SUBJECTS CV  125 100 ® p  75  ELBOW SHOULDER SHD ABD TRUNK  50 25  ADULT GROUP  ADULT  PEDIATRIC  PEDIATRIC GROUP  Figure 20 . Summary of average within-(A) and average between-(B) subject average CVs for the angular data over the two groups.  These data demonstrate that the two groups showed similar variability in the angular data with the pediatric group showing more variability within- and between-subjects in shoulder abduction than the adult group did . Also, the pediatric group showed less between-subject variability in trunk angles than the adult group. Coast down accelerations and drag force An attempt was made to measure the drag force of the user-chair system in order to further understand the possible effects of mass additions . This was done by way of coast down tests . The characteristics of the test chairs were 46    uniform in that while they had a range of seat sizes the total mass of the chair was 9 .3 kg, the rear wheel mass was 2 .0 kg and caster wheel mass was 0 .2 kg. The radius of the rear wheel was 0 .30 m . Two inch foam cushions were used in all test chairs. The measured mass of the pediatric subjects' own chairs was 12 .5 - 23.3 kg and the adult subjects' chairs was 11 .9 - 19.7 kg . There were no differences in the average mass of the subjects' own chairs between the pediatric group (14 .4 kg) and the adult group (14 .3 kg) . Individual chair characteristics (both mass and dimensions) were used in the determination of the angular accelerations for the coast down tests and are summarized in Table 3. Table 3 . Characteristics of the subjects' own wheelchairs. ( A= Adult, P=Pediatric)  Subj #  CHAIR TYPE  Chair MASS (kg)  REAR Wheel Mass (kg)  Caster Wheel MASS (kg)  P 01 P 02 P 03 P 04 P 05 P 06 P07 P 08 P 09 P 10 A 01 A 02 A 03 A 04 A 05 A 06 A 07 A 08 A 09 A 10  QUICKIE II QUICKIE Ii QUICKIE GP KUSCHALL QUICKIE QUICKIE II QUICKIE II QUICKIE GP QUICKIE HP SCOTTS QUICKIE GP QUADRA QUICKIE II SHADOW SHADOW QUICKIE II QUICKIE GP KUSCHALL QUICKIE 3 QUICKIE GP  21 .3 19 .2 12 .5 12 .5 23 .3 12 .8 15.0 13.5 13.3 12 .7 19 .7 16 .1 14 .7 11 .9 12 .1 14 .0 12.4 15.8 15 .2 11 .3  5 .0 5 .0 5 .0 2 .4 5 .0 1 .8 2 .1 1 .9 2 .2 1 .5 2 .8 2 .0 5 .0 2 .3 1 .7 1 .9 1 .9 2 .3 1 .8 1 .9  0 .8 0 .8 0 .8 0 .3 0 .8 0 .2 0 .3 0 .4 0 .2 0 .3 0.4 0.5 0.8 0 .3 0 .3 0 .4 0 .5 0 .3 0 .3 0 .3  REAR Wheel diam. outside inside 0.04 C .06 0 .04 0 .08 0 .08 0 .10 0 .08 0 .08 0.08 0.06 0.08 0.08 0 .08 0 .05 0 .06 0 .08 0 .08 0 .08 0 .06 0 .04  Caster Wheel diam outside - inside  REAR Wheel outside radius (m)  0` 06 . 0 .04 0 .04 0 .04 0 .04 0 .06 0 .05 0 .06 0 .04 0 .04 0 .04 0 .04 0 .04 0 .10 0 .06 0 .06 0 .05 0 .06 0 .06 0 .04  0.31 0.29 0 .30 0.31 0 .27 0 .27 0 .29 0 .31 0 .31 0 .27 0 .31 0.31 0.31 0.31 0.30 0.30 0 .30 0 .31 0 .30 0 .29  The average acceleration for the 0 kg, 10 kg and own chair conditions for the two groups is shown in Figure 21 . It is apparent that the values for the accelerations for the pediatric group were very small . It is also apparent that the accelerations for the own chair condition was the highest in both groups . It was apparent from these data that the accelerations were all very small (all less 47    than 0 .2 m/sec/sec) . In general, the coast down accelerations were larger for the pediatric group than the adult group which was surprising given the smaller system mass of the pediatric group . However, the values for the pediatric group GROUP PEDIATRIC  ADULT  O kg +10 kg Own  -0 .5  Figure 21 . Summary of average measured accelerations for the coast down tests for the pediatric (n=10) and adult groups (n=10).  were also more variable . The coefficients of correlation for the line fit of the velocity versus time data from which these accelerations were determined to be very low (< 0 .5) in many cases for the pediatric data and in some cases for the adult data . An example of the line fit for the velocity versus time data is shown in Figure 22 .  48    5  4 .75  1►  4 .5  4 .25  -- y = 4 .7445 + -0 .02834x R= 0 .75823 4 0  5  10  15  TIME (sec)    Figure 22 . Example of line fit of velocity time curve for coast down test for a single adult subject in a test chair with 10 kg of added mass. The average accelerations and values for the correlation coefficients for the line fit data are summarized and described in Table 4. Table 4 . Description of the average values for acceleration and correlation coefficients for line fit data. pediatric subject # 1 2 3 4 5 6 7 8 9 10 adult subject # 1 2 3 4 5 6 7 8 9 10  OWN CHAIR -0 .55 -0 .17 -0 .09 -0 .12 -0 .34 -0 .25 -0 .14 -0 .47 -0 .15 -0 .23  R VALUE 0 .76 0 .83 0 .06 0 .56 0 .46 0 .92 0 .80 0 .85 0 .74 0 .79  + 0 kg  -0 .01 -0 .21 -0 .18 -0 .11 -0 .31 -0 .24 -0 .15 -0 .13 -0 .09 -0 .14  0 .30 0 .42 0 .55 0 .18 0 .45 0 .74 0 .80 0 .73 0 .87 0 .47  -0 .10 -0 .01 -0 .03 -0 .01 -0 .04 -0 .01 -0 .03 -0 .03 -0 .01 -0 .02  -0 -0 -0 -0 -0 -0 -0 -0 -0 -0  .05 .06 .05 .01 .15 .02 .02 .01 .02 .03  49  R VALUE 0 79 0 .45 0 .61 0 .41 0 .30 0 .13 0 .33 0 .30 0 .54 0 .61  +10 kg  R VALUE  -0 .01 -0 .06 -0 .03 -0 .02 -0 .02 -0 .09 -0 .01 -0 .45 -0 .02 -0 .02  0 43 0 .91 0 .66 0 .41 0 .14 0 .60 0 .21 0 .76 0 .05 0 .05  0 .85 0 .20 0 .65 0 .16 0 .67 0 .38 0 .62 0 .34 0 .42 0 .34  -0 .06 -0 .02 -0 .03 -0 .03 -0 .03 -0 .02 -0 .01 -0 .03 -0 .02 -0 .02  0 .87 0 .40 0 .74 0 .70 0 .35 0 .04 0 .21 0 .66 0 .23 0 .33    Because of the small absolute value and lack of accuracy in the acceleration data, drag force calculations were determined with the understanding that the accuracy of these calculations would be suspect . The results of these determinations are shown in Figure 23 . In all cases the drag forces were negative and less than 4 Newtons . The drag forces were the highest in both groups for the subjects' own chairs . This is not surprising given the fact that it is reasonable to assume that the subjects' own chairs were not maintained to the extent that the new test chairs were . It is likely that factors such as dirt in the wheel bearing, and toe-in or toe-out of the caster wheels in the subjects' own chairs could have contributed to an increase in drag in these chairs.  D  O kg +10 kg Own   ADULTS  i PEDIATRIC  GROUP  Figure 23 . Summary of average drag force for the pediatric (n=10) and adult groups (n=10).  Summary of results 1 . Subjects did achieve the nominal wheeling velocities of 2 m/sec but the pediatric group were wheeling at a significantly lower velocity than the adult group . Both groups responded to the different test conditions by maintaining the same average wheeling velocity .  50  2. The timing of grab and release and % of cycle spent in propulsion did not significantly change in any of the experimental conditions or between the two groups. 3. The angular kinematics of wheeling in the subjects own chair and the test chairs were not significantly different in either group for three of the four angular parameters . This was despite the fact that the average mass of the subjects' own chairs was 14 kg whereas the mass of the test chairs was 9 kg. This indicates that there was not an effect due simply to placing the subjects in the test chairs. 4. The angular kinematics of wheeling were significantly different in three (elbow, shoulder, and shoulder abduction) of the four angular variables between the adult and pediatric groups indicating absolute angular differences between them . However, despite these differences, the pattern of angular change over time and condition was not significantly different between the two groups . 5. The angular kinematics of wheeling under the mass addition conditions was not significantly different in either group indicating that the two groups responded similarly to the mass additions . Mass did not effect the angular kinematics, wheeling velocities or % propulsion in either group. 6. The within-subject CVs for the angular data (elbow, shoulder and shoulder abduction) were low (between 5 and 10%) . Averaged across the shoulder, elbow and shoulder abduction angles the mean CV was similar for the two groups (adult 7 .2% ; pediatric 11 .8%) . The higher within-subject CV for the pediatric group was due to the greater variability in the shoulder abduction angular data for this group . The within-subject CVs for the trunk angular data was high in both groups (50-75%) reflecting the small mean values for these data. 7. The between-subject CVs were higher (between 10 and 20%) and the differences between the groups were small . The average CV for the shoulder, elbow and shoulder abduction angular data for the two groups was very similar (16 .9% for the pediatric group and 16 .6% for the adult group) . The betweensubject CVs for the trunk data were once again large and greater in the adult group (60-90%) than in the pediatric group (40-60%). 8. The accelerations on coast down tests were very low with low correlation coefficients to the line fit data indicating that the experimental 51  conditions in this study did not permit an accurate determination of drag force. The drag forces determined by this method were also low (less than 4 Newtons) . The drag forces were higher in the subjects own chairs than in the test chairs. DISCUSSION Timing data This study showed that there was no change in the percentage of the cycle spent in propulsion with mass additions in either the adult or the pediatric group . This finding is supported by other studies in the literature . Van Der Woude et al ., (1988) manipulated hand rim diameter in an adult population and found that there was no effect on cycle time or on its subdivisions (push time and recovery time) or the push angle . Rodgers et al ., (1992) studied eleven male paraplegics (T5-T11) wheeling on a laboratory instrumented wheelchair. Resistance was applied for a continuous progressive exercise test and a fatigue test . This study showed that with fatigue there was no change in the temporal factors of wheeling . In the present study, the fact that no change was evident in the timing of critical components in the wheeling cycle supports the conclusion that mass did not affect wheeling style under these experimental conditions. The present study showed a mean % propulsion of 25% for both the adult and pediatric groups wheeling overground . These values are slightly lower than other reported values for adult athletic populations wheeling on an ergometer . Higgs (1986) reports % propulsions of 33 .8% and 37 .7 % for 8 elite track and field athletes . Masse et al ., (1992) reports a mean of 33 .4 0/0 propulsion for 5 elite athletes over a variety of seat positions . Van der Woude (1989) reports % propulsions from 30-45% depending on cycle frequency. Sanderson and Sommer (1985) report % propulsion of 43 .3%, 34 .7% and 43 .7% for three world class track and field athletes . Cooper (1990) reports in a study of five elite athletes that approximately 33% of the time was spent in propulsion and 67% in recovery . The lower % propulsions reported in this study could be explained on the basis of a non-athletic subject population wheeling overground as opposed to on an ergometer. Angular data The present study showed no changes in the angular data (elbow, shoulder, trunk and shoulder abduction in response to mass additions . The 52  angular excursions measured in this study are consistent with most of those in the literature . These are summarized in Table 5 . This finding is of interest because this study has involved non athletic adults and children with disability. The assumption was made at the outset of this study that anthropometry alone would not affect the angular kinematics . The fact that the kinematic data presented in this study is consistent with other published values taken from adult, athletic populations supports this assumption. Table 5 . Summary of angular ranges of motion in wheelchair propulsion studies RANGE  ELBOW ANGLE  SHOULDER ANGLE  TRUNK ANGLE  SHOULDER ABD ANGLE  present study Wang et at ., (1989) Rodgers et at., (1992) Kobayashi et al ., (1991)  90-140°  20-70°  5-15°  30-70°  ---  ---  ---  65-80°  60° range  67 ° range  7.8° range  45-70°  70-160°  0-70°  0-13°  40-60°  100-160°  ---  4-8°  45-60°  Van der Woude et al ., (1989)  Mass addition data The present study showed no effect on the angular kinematics, timing or wheeling velocities in two groups of subjects with spinal cord injury when mass was added to the low mass test chair . This result was not expected . In a study of children with disability Findley et at ., (1988) stated, "Standard wheelchairs were used in this study and these findings could be different with the new, lighter weight wheelchairs . For example, reduction in wheelchair weight from 20 to 10 kg for a 50 kg child would make a 15% difference in the theoretical effort necessary to go up slopes"(p .860). As discussed earlier, there are few studies which have examined the effects of mass in wheelchair propulsion . The 53  results of this study are, however, supported by Parziale (1991) who also found no difference between 8 subjects with high and low paraplegia in wheeling a standard (Everest and Jennings) and light weight wheelchair (Quickie II) across a short and long distance level course . No differences were found in their study in blood pressure, heart rate, respiratory rate or wheeling speeds between the two chairs . Van der Woude et al ., (1988) also found no major changes in the movement patterns with changing rim diameter and varying velocity despite the concomitant significant changes in physiological responses (oxygen cost, ventilation, heart rate, respiratory exchange ratio, mechanical efficiency) . This group did report an increase in shoulder abduction angle with increasing rim diameter which they speculated might have been the result of "accommodating the arm position with respect to the wheel or it may be a "side effect" of an increasingly active pectoralis major causing internal rotation of the upper arm" (p. 434). There are several explanations of the results of the present study with respect to mass additions . One explanation might be that mass is an important factor in wheelchair propulsion but that its effects can only be measured over longer time periods and over more varied wheeling surfaces . Another explanation may be that mass is not an important factor in determining wheeling style . This explanation would be supported by the large number of trials, subjects, and conditions employed in this study . Another explanation of these data may be that mass did have an effect on wheelchair propulsion but this effect was not apparent on the kinematics of propulsion but on the kinetics . The kinetics of the movement pattern (the application of force and the effectiveness of this force application) could have changed in response to the mass additions without there being any alteration in the kinematics . There is evidence in other types of repetitive cyclical movement patterns (i .e . walking and cycling, Marshall, Wood & Nade, 1990 ; Winter, 1984 ; Hull & Davis, 1981) that it is r w,  possible to have the kinematics of a movement pattern remain unaltered at the same time as the kinetics of the pattern are changing. It is also possible that the subjects did not alter their wheeling style over the short time period that was available to them to adjust to the test chairs. However, this is unlikely given the fact that wheeling style as measured kinematically appears to be remarkably robust over time . Over the past two years at the Biomechanics Laboratory at the University of British Columbia a 54  number of individuals with spinal cord injury have been kinematically assessed repeatedly . Some of these data is presented in Appendix C . Our experience has been that individual subjects have a particular wheeling style which is consistently maintained over time and under a variety of experimental conditions . Also, a short adjustment time to a new wheelchair is certainly similar to what is done clinically for assessment of a new wheelchair . It is unusual for a user and clinician to have more than an hour or two to make the decision regarding wheelchair selection. Pediatric vs adult data The present study shows two different angular styles for the pediatric and adult groups . These are consistent with other reports in the literature . Higgs (1986) described two propulsion styles in a group of sprinters and distance elite athletes . One style was described as resembling the back and forth motion of a shuttle ; the other as a more circular motion . This is similar to the circular versus pump action style of wheeling described by Sanderson and Sommer (1985). Su et at., (1991) describes similar elbow-shoulder angle curves in a study of the differences between the left and right upper extremities in three adult subjects with paraplegia and 3 able bodied subjects . Lamontagne et al ., (1991) studied the angular kinematics of 10 adult bilateral above knee amputees and found differences in pushing style across the subjects . "Five subjects used circular arm action while the remaining five subjects used pump arm action" (p . 423). The differences in wheeling style reported in this study between the adult and pediatric groups could be as a result of differences in anthropometry, motor power or skill . The lengths of body segments and mass distribution (anthropometry) of the upper limbs of pediatric and adult wheelchair users are most certainly different (Jensen, 1987,1988) . The dimensions of the lever arms in movement affect the forces that are generated (Winter, 1990) . However, because it is the kinematics and not the kinetics of wheelchair propulsion that was under study in this case, it is unlikely that anthropometry alone would significantly affect the kinematics of wheelchair propulsion . This hypothesis is supported by previous work with twenty spinally injured adults (Sanderson & Bednarczyk, 1990) in which the kinematics of wheeling were shown to be similar in disabled individuals with varying anthropometry . It is more likely that it is the decreased motor power and skill of the pediatric subjects in this study that resulted in the observed differences in wheeling style . It is unclear why the 55  pediatric group showed significant differences in elbow angle in the test conditions . It is possible that this is a Type I error . It is also possible that the pediatric subjects' own chairs were not appropriately matched for the subjects and this was reflected in the elbow angular changes in the wheeling cycle. The differences between the group ensemble averaged kinematic profiles of wheelchair propulsion are not always consistent with the individual subject responses as can be seen in the individual subject plots shown in Appendix D . There is always a problem in the application of average patterns of movement to individual subjects . Pediatric subjects # 06, #07, and #08 all demonstrate angle-angle plots which are more similar to the adult group wheeling style . These subjects were also the only three to show an effect in response to the mass additions . These three subjects all had ASIA score mid way in the group (54-56) and had ages that were on the lower end of the age scale (ages 8-11) . The mass of two of these three subjects were low (24 and 32 kg) . However, pediatric subject # 10 was an eight year old subject with a mass of 22 .9 (ASIA score of 58) who showed no response to the mass additions and had a circular wheeling style . In other words, there is no consistent pattern to those subjects who differed from the group pattern. Despite the differences noted between the group pediatric and adult wheeling style, what is interesting is how much similarity there is between the two groups in all of the variables employed in this study . The two groups showed similar variability in many of the angular parameters and the average CVs were not significantly different in the propulsive versus recovery portions of the wheeling cycle . Winter (1983) has noted similar consistent patterns in gait and states that this is evidence of a motor pattern governing the movement. Variability (CV) data This study showed average CVs for the angular data to be between 7 and 11 % within subjects and 16% between subjects . The pediatric group showed more variability in the shoulder abduction angular data than the adult group . There are few reports in the literature of actual CVs for any data in wheelchair propulsion despite that fact that there are numerous references to large between subject variability in wheelchair propulsion . Su et al ., (1991) describes the variation within each subject across the measurement periods as "small" while the differences between-subjects as "dramatic" (p .198). Sanderson and Sommer (1985) state, "The small variation in the temporal data 56  presented is indicative of how regular the movement patterns of the subjects were, both in terms of the movements at each testing interval as well as across the full 80 min of the testing period ." (p.426) . Masse et al ., (1992) report CVs for the shoulder angular velocities of 50-55%, for the elbow angular velocities of 80-100% . They found large values for the trunk angular CVs which they explained "in terms of the variability among and between the subjects (mean variation of approximately 75 .7 %)" . The between-subject variability noted in some of these studies could have been due to inappropriate grouping of the subjects across different disability categories (Bednarczyk & Sanderson, 1992). The large values for the trunk angular CVs reported in this study were undoubtedly due to the small magnitude of the trunk angles and relatively large measurement errors in this angular parameter. The variability reported in this study in shoulder abduction angles was also noted by Rogers et al., (1992) . They state, "Maximum shoulder abduction angles increased with fatigue for six subjects and decreased for six subjects. This variability may be related to individual differences in propulsion style."(p.204) . Harburn and Spaulding (1986) also noted a high degree of shoulder abduction . The results of their EMG analysis showed that, "the most active muscles monitored during the activity were the middle deltoid, posterior deltoid and in some subjects the triceps brachii . Wheelchair ambulation was found to be an activity-producing idiosyncratic muscle recruitment pattern not only in spinal cord-injured subjects, but in nondisabled subjects as well . These idiosyncratic patterns for each person tested showed low variance from movement cycle to movement cycle, implying that we were observing a learned skill" (p .635). The increased variability noted in this study in the shoulder abduction angular data for the pediatric group may be indicative of more abduction of the upper arm or more internal and external rotation of the upper arm . Because of the way the shoulder abduction angle was defined in this study, it was not possible to separate these two movements. It is possible, that the pediatric group were forced to use more shoulder abduction and/or internal rotation in order to recruit more of the large proximal muscles such as pectoralis major and lattissimus dorsi . It is difficult however to rationalize why the pediatric group showed less variability in the trunk angular data than the adult group . Some members of the adult group appeared to use trunk motion in concert with arm 57  motion in wheelchair propulsion . The pediatric group tended to remain remarkably still in their wheelchairs . One observation made during the data collection phase which is relevant to this point, is that several of the pediatric subjects commented on how anxious they felt in the test chairs about falling. Most of the pediatric subjects had been given little instruction in advanced wheelchair skills such as doing wheelies and going up and down curbs (both of which involve using the trunk as a counterbalance) . In contrast, all of the adult subjects indicated that they were able to perform these tasks . It is possible that the limited trunk movement evident in the pediatric group reflected the more limited skills of this group as compared to the adult group . However, the variability in the trunk angular data was large . This was probably due to combination of measurement error (Appendix B) and between-subject variability. Acceleration and drag data Drag force in the present study was measured by adapting the method of Coutts (1990) . The drag forces reported were all very low (less than 4 Newtons) . Coutts (1992) reports drag force for four wheelchair basketball players from 7 .4 to 9 .5 Newtons. Veeger et al ., (1992) report drag forces for nine subjects in an instrumented test wheelchair (using the towing method of measuring drag) between 9 and 14 Newtons . The correlations coefficients for the line fit data in the present study were reported to be low (often less than 0 .5) in contrast to those reported by Coutts (1991) on average = 0 .995 . There are several explanations for the differences in the data reported in this study . One explanation might be that the present study used video analysis of coast down tests over a limited distance (approximately three meters) within the calibrated volume on the test runway . In Coutts' study, the acceleration data were collected by way of magnets mounted on one wheel over the entire length of the gymnasium floor . It is possible that the distances used in this study to measure coast down tests was not adequate . It is also possible that the frictional coefficients of the new test chairs used in this study resulted in reduced values for drag . Coutts' studies have involved wheelchairs which were extremely well used and were likely to have much higher frictional coefficients due to misalignment of caster wheels and wear of the bearings . The problems encountered with measuring drag in this study were not encountered in pilot studies . This is most likely due to the fact that previous studies done in this 58  laboratory on drag involved subjects wheeling in their own wheelchairs (not new wheelchairs) . Also, the wheeling surface used in the present study (canvas covered wooden gymnasium floor) undoubtedly offered less rolling resistance than previous pilot studies done on concrete covered with a low pile carpet . Whatever the cause, it was apparent that the signal to noise ratio for the coast down data was less than optimal. It is also possible to explain the results of no effect of mass addition on the basis of drag force . Van der Woude et al., (1986) measured drag in four types of wheelchairs . They showed the drag force of a standard hand rim wheelchair (named the R wheelchair ; mass=15 kg) as being more than twice that of a sports wheelchair (named the S wheelchair ; mass=l2 kg) at 0 and 1 degrees of inclination of the treadmill. They state, "Though much lighter the R wheelchair has a high power output, probably due to the small front castors and mass distribution . The S wheelchair shows the beneficial effect of height weight, good mass distribution (with a relatively large weight on the rear wheels) and large castor wheels, resulting in a lower power output at all levels of inclination" (p .1568) . Peizer et al ., (1964) state that, "A reduction in frictional requirements (rather than weight) would be more helpful, however, for vigorous patients since their patterns of operation would probably include longer periods of sustained propulsion" (p .91). It is possible that the new test chairs used in this study had, in addition to a low mass, low frictional coefficients and an effective distribution of mass which resulted in a reduced drag force felt by the user. Thus, even though the mass was increased in the experimental conditions reported in this study, the user was still functioning in an area of drag which permitted him or her to function comfortably thus showing no change in wheeling style. CONCLUSIONS The hypothesis made at the outset of this study was that mass would affect the kinematics of wheelchair propulsion . The data presented in this study showed no change in the kinematics of wheeling in response to mass additions in the 10 kg range in either group under the level wheeling, steady state, test chair conditions of the study . Thus in the short distance, level wheeling condition, mass alone does not appear to be a factor in determining the style of wheelchair propulsion and this hypotnesis was rejected.  59  In addition, the pediatric group was found, in this study, to have a significantly different wheeling style than the adult group . However, the responses of both groups to the experimental conditions and over the wheeling cycle were not significantly different, indicating fundamental similarities in the kinematics in both groups . Similar variability was demonstrated in the two groups in the angular kinematics of wheelchair propulsion in both the elbow and shoulder angles . The pediatric group showed more variability in the shoulder abduction angular data and less variability in the trunk angular data than the adult group. The significance of these different patterns of variability are unclear. Also, the kinematic data reported in this study are consistent with other reported data in the literature . This is despite the fact that the present study population was non athletic and considerably younger (pediatric group) than those previously reported. Explanations for these results included the possibility that the effective force application may have changed in response to the mass additions at the same time as the kinematics remained unaltered . It is also possible that the reduced drag of the new test chairs used in this study confounded possible mass effects . The implications of this work for clinical practise are that the principles of user-chair optimization derived from studies of adult wheelchair users with paraplegia may be applied to the pediatric wheelchair user with consideration of the absolute differences between the two groups . It is also clear from the data presented in this study that clinicians and users are not supported in the belief that mass is important to wheeling style in the short distance, level wheeling condition. RECOMMENDATIONS AND SUGGESTIONS FOR FUTURE RESEARCH The present study has examined the effect of mass on steady state, level wheelchair propulsion . It would be of interest to repeat this study with a focus on the long distance wheeling in a larger population of subjects with spinal cord injury . It is also possible that subjects with quadriplegic levels of spinal cord injury would have a different response to mass effects in either steady state wheeling or initiation of movement than subjects with a paraplegic level of spinal cord injury due to their decreased voluntary control and strength of upper extremity musculature .  60  The present study has not directly examined the effect of drag on wheelchair propulsion and the possible interaction effects of drag and mass . It would be of interest to repeat the study presently described with the addition of steady state, level wheeling over two or three other surfaces with greater frictional coefficients (i .e . low pile and high pile carpet) . It would also be possible to use an ergometer, rather than overground wheeling, to directly manipulate drag on the rollers as well as to examine the effects of longer distance wheeling . This type of experimental design has been used in bicycling where the first and last minutes of a long ride are videotaped and analyzed in order to determine if there are changes in the kinematics of bicycling over time. It might also be of interest to examine mass effects in propulsion over other types of surfaces (i.e., incline as well as level wheeling). It would also be of interest to add kinetic measurements to the experimental design described in the present study . There are few laboratories in the world that have reported kinetic measurements in wheelchair propulsion. This is due to the technical difficulties in instrumenting the wheel of a wheelchair. Laboratories in Holland (Van der Woude et al ., 1991), and Ohio (Rodgers et al ., 1991) have recently reported differences in effective force application for adult, athletic populations of subjects wheeling in a laboratory, instrumented wheelchair . Veeger et al., (1991) in a study involving able bodied subjects, states that " propulsive forces on wheelchair hand rims were found to be ineffectively directed" (p . 231) . These findings have been supported by studies in California (Cooper, 1992) involving an instrumented wheel which is transferrable to any wheelchair . However, none of these studies have involved non-athletic or pediatric populations . There is also a problem in kinetic studies with the location of the centre of mass for the body segments for subjects with disability . Some authors (Duval-Beaupere & Robain, 1991 ; Duval-Beaupere, Lougovoy, Trocellier & Lacert, 1983) believe that the location and magnitude of the centres of mass in subjects with disability is significantly different than those reported in able bodied subjects secondary to hypertrophy or atrophy of muscles . In addition, there is further debate as to the magnitude and the location of the centre of mass of the segments in children (Jensen, 1986, 1987; Jensen & Nassas, 1988) . These debates, at the moment, are unresolved and have hampered kinetic descriptions of wheelchair propulsion . There is a need for the resolution of these theoretical and technical problems so that kinetic as 61  well as kinematic studies may be undertaken which will examine the long term effects of both mass and drag in wheelchair propulsion . There is a need for these studies to be conducted on nonathletic populations in order to provide clinicians and users with the information they need to make the most appropriate wheelchair selection. This study has described the kinematics of wheelchair propulsion in a small group of pediatric users with paraplegia as a result of Spina Bifida . There are a large number of pediatric wheelchair users who deserve to be studied in order to determine if the wheeling characteristics outlined in this study are consistent across other pediatric subject groups . These children will spend many years in wheelchairs and it important that the best possible match between user and chair be obtained in order to optimize performance and minimize overuse injuries in the years ahead.  62  BIBLIOGRAPHY Abdel-Aziz, Y .I . & Karara, H .M. (1971) . Direct Linear Transformation from comparator coordinates into object space coordinates in close-range photogrammetry . Proceedings of the ASP Symposium on Close-Range Photogrammetry (pp . 1-18). Anderson, C ., Brattgard, S . O . & Severinsson, K. (1976) . Methods of analyzing the demands of wheelchairs . (1976) . Scandinavian Journal of Rehabilitation Medicine (Suppl) . ., 151-55. Bednarczyk, J . H . & Sanderson, D . J . (1991) . Biomechanics of wheelchair propulsion : Evaluation of user chair interface . American Paraplegia Society. 14(2), 92. Bednarczyk, J . H . & Sanderson, D . J. (1992) . Comparison of functional and medical assessment in the classification of disability in spinal cord injury. Journal of the American Paraplegia Society . 15(2), 128. Bednarcyk, J .H . & Sanderson, D .J . (1992) The effect of subject grouping on the kinematic assessment of wheelchair propulsion . Proceedings of Nacob II . The second North American Congress on Biomechanics, (p .429) Chicago, Illinois, U.S.A. Bennedik, K ., Engel, P ., & Hildebrandt, G. (1978) . Der rollstuhl. Internationale Schriftenreihe fur Rehabilitationsforschung . 15, 237-89. Bostom, A . G . & Bates, R. L. (1987). Ergometer modification for combined armleg use by lower extremity amputees in cardiovascular testing and training. Archives of Physical Medicine, 04), 244-7. Brattgaard, S . O. (1970) . Energy expenditure and heart rate in driving a wheelchair ergometer. Scandinavian Journal of Rehabilitation Medicine, 2,1438. Brown, D. D., Knowlton, R . G ., Hamill, J ., Schneider, T . L., & Hetzler, R . K. (1990) . Physiological and biomechanical differences between wheelchairdependent and able-bodied subjects during wheelchair ergometry . European Journal of Applied Physiolog y and Occupational Physiolo gy . 60(3), 179-82. Brubaker, C . E . (1981) . Ergonomics . In Wheelchair mobility (pp . 7-8) Virginia : University of Virginia Rehabilitation Engineering Center. Brubaker, C . E . & McClay, I . S . (1984). Effect of seat position on wheelchair propulsion efficiency . Proceedings of the 7th ICRE (p .1214). 63  Brubaker, C . E . & McClay, I . S. (1985) . Determination and analysis of factors affecting wheelchair propulsion on lateral inclines . In : Wheelchair mobility (pp. 10-14) Virginia : University of Virginia Rehabilitation Engineering Center. Brubaker, C . E . (1986). Wheelchair prescription : an analysis of factors that affect mobility and performance . Journal of Rehabilitation Research and, Development ., 22,19-26. Brubaker, C. E . (1990) . Ergonometric consideration . Journal of Rehabilitation Research and Development . Clinical Suppl # 2, 37-48. Cerguilini, S ., Figura, F ., Marchetti, M . & Ricci, B . (1981). Biomechanics of wheelchair propulsion . In : Moreki A, Fideli et al ., (Eds) Biomechanics VII-A, (pp. 410-19) Baltimore, MD : University Press. Cheetham, P . & Sheirman, G . L.(1988) . The video study of motion . Advanced Imaging : Media Horizons Inc ., January. Collins, T . J . & Kauzlarich, J . J . (1988) . Directional instability of rear caster wheelchairs . Journal of Rehabilitation Research and Development, 25(3), 1-18. Cooper, R . A. (1989) . An international track wheelchair with a center of gravity directional controller. Journal of Rehabilitation Research and Development, 26(2), 63. Cooper, R . A. (1990) . Racing wheelchair rear wheel alignment . Journal of Rehabilitation Research and Development . 26(1), 47-50. Cooper, R . A . (1990) . Wheelchair racing sports science : a review. Journal of Rehabilitation Research and Development, 27(3),295-312. Cooper, R . A . (1990) . An exploratory study of racing wheelchair propulsion dynamics . Adapted Physical Activity Quarterly . 2(1), 74-85. Cooper, R . A . (1991) . Comparison of pulmonary functions of wheelchair racers in their racing and standard wheelchairs . RESNA Proceedings, (pp .345-7) Missouri, U .S .A. Cooper, R .A . (1992) Personal communication, September, 1992. Corcoran, P . J ., Goldman, R . F. & Hoerner, E . (1980) . Sports medicine and the physiology of wheelchair marathon racing . Orthopedic Clinics of North America. Z1, 697-716 . 64  Coutts, K. D . & Strogyn, F . L. (1987) . Aerobic and anaerobic power of Canadian wheelchair track athletes . Medicine in Science and Sports Exercise, 19 . 62-5. Coutts, K. D . (1990) . Kinematics of sport wheelchair propulsion . Journal of Rehabilitation Research and Development . 27(1), 21-6. Coutts, K. D . (1991) . Dynamic characteristics of a sport wheelchair . Journal of Rehabilitation Research and Development, _Z (3), in print. Davis, G . D., Plyley, M . J . and Shephard, R . J . (1991). Gains of cardiorespiratory fitness with arm-crank training in spinally disabled men . Canadian Journal of Sports Science . 16(1), 64-72. Dicarlo, S . E . (1988) . Effect of arm ergometry training on wheelchair propulsion endurance of individuals with quadriplegia . Physical Therapy . S$(1), 40-4. Dixon, W . J . & Massey, F .J . (1969) . Introduction to Statistical Analysis . (Third Edition) . California : McGraw-Hill Co . (p.202). Dreisinger, T .E . & Londeree, B . R . (1982) . Wheelchair exercise : a review. Paraplegia, 20 20-34. Duval-Beaupere, G . & Robain, G . (1991) . Upward displacement of the center of gravity in paraplegic patients . Paraplegia . 309-17.  a,  Duval-Beaupere, G ., Lougovoy, J ., Trocellier, L . & Lacert, P. H . (1983) . Trunk and leg growth in children with paraplegia caused by spinal cord injury. Paraplegia,. 21, 339-350. Engel, P ., & Hildebrandt, G. (1974) . Wheelchair design - technological and physiological aspects . Proceedings of the Royal Society of Medicine 67 40913. Fahland, B ., & Grendahl, B . C.(1967) . Wheelchair selection : more than choosing a chair with wheels . American Rehabilitation Foundation . 110-32. Findley, T . W. & Agre, J . C . (1988) . Ambulation in the adolescent with spina bifida II . Oxygen cost of mobility . Archives of Physical Medicine and Rehabilitation . 62(10), 855-6. Franks, C . A., Palisano, R . J . & Darbee, J . C . (1991) . The effect of walking with an assistive device and using a wheelchair on school performance in students with myelomeningocele . Physical Therapy . 71(8), 570-9. Franklin, T . W., Swantek, K . I ., Grais, S. L. . Johnstone, K . S . & Gordon, S. (1990) . Field test estimation of maximal oxygen consumption in wheelchair users . Archives of Physical Medicine and Rehabilitation, 71(8), 574-8. 65  Gaines, R . F ., Zomlefer, M . R. & Zhao, W . (1984) . Armstroke patterns of spinal cord injured users . Archives of Physical Medicine and Rehabilitation . 65 61821 Gangelhoff, J ., Cordain, L ., Tucker, A. & Sockler, J . (1988) . Metabolic and heart rate responses for submaximal arm lever and arm crank ergometry . Archives of Physical Medicine . 0,101-5. Gass, G . C. & Camp, E . M. (1979) . Physiological characteristics of trained Australian paraplegia and tetraplegic subjects . Medicine in Science and Sports. L 256-59. Glaser, R . M ., Ginger, J . F., & Laubach, R . S . (1977). Validity and reliability of wheelchair ergometry . Physiologist . 20, 34. Glaser, R . M ., Sawka, M . N., Lauback, L . L . & Suprasayad, A . G . (1979). Metabolic and cardiopulmonary responses to wheelchair and bicycle ergometry . Journal of Applied Physiology : Respiratory and Environmental Exercise Physiology . 4Qs 1066-70. Glaser, R . M ., Sawka, M . N, Young, R . E . & Suprasayad, G . (1980) . Applied physiology for wheelchair design . Journal of Applied Physioloay : Respiratory and Environmental Exercise Physiology . 4$(1)s 41-4. Gwinn, T. H ., Gass, G. C ., Khouw, W ., Page, N . & Windred, T . (1990) . The effect of wheel location on manual wheelchair propulsive outputs . Supplement to Medicine and Science in Sports and Exercise, 22 (2), 41. Hardison, G . T. & Israel, R . G. (1987) . Physiological responses to different cranking rates during submaximal arm ergometry in paraplegic males . Adapted Physical Activity Quarterly . 4, 94-105. Harburn, K . L. & Spaulding, S . J . (1986) . Muscle activity in the spinal-cord injured during wheelchair ambulation . American Journal of Occupational Therapy . 4Q, 629-36. Hedrick, B., Wang, Y . T ., Moeinzadeh, M . & Adrian, M . (1990) . Aerodynamic positioning and performance in wheelchair racing . Adapted Physical Activity Quarterly, 7 41-51. Higgs, C . (1986) . Propulsion of racing wheelchairs . In : Sport and Disabled Athletes . Sherril, C . (eds) Human Kinetics Publishers, Champaign, Illinois. Hilbers, P . A., & White, T . P. (1987) . Effects of wheelchair design on metabolic and heart rate responses during propulsion by persons with paraplegia. Physical Therapy, 67, 1355-58 .  66  Huberty, C . J . & Morris, J . D . (1989) . Multivariate analysis versus multiple univariate analyses . Psychological Bulletin, 105, (2), 302-8. Hull, M . L., & Davis, R . R . (1981) . Measurement of pedal loading in bicycling, I and II . Journal of Biomechanics, 14, 843-72. Hunter, J . (1987) . Energy costs of wheelchair propulsion by elderly and disabled people . International Journal of Rehabilitation Research . 111 . 10-50. Jarvis, S . & Rolfe, H . (1982) . The use of an inertial dynamometer to explore the design of children's wheelchairs . Scandinavian Journal of Rehabilitation Medicine, 4,, 167-76. Jensen, R . K. (1987) . Growth of estimate segment masses between four and sixteen years . Human Biolo gy . 59(1), 173-189. Jensen, R . K. (1989) . Changes in segment inertia proportions between 4 and 20 years . Journal of Biomechanics, =607), 529-36. Jensen, R . K . & Nassas, G . (1989) . Growth of segment principal moments of inertia between four and twenty years . Medicine and Science in Sports and Exercise . ZQ(6), 594-604. Kauzlarich, J . J . (1987) . A theory of Wheelchair wheelie performance . Journal of Rehabilitation Research and Development , 24(2), 67-80. Kauzlarich, J . J ., Bruning, C . C. & Thacker, J . G . (1984) . Wheelchair caster shimmey and turning resistance . Journal of Rehabilitation Research and Development, 21(2), 15-29. Kauzlarich, J . J . & Thacker, J . G . (1985) . Wheelchair tire rolling resistance and fatigue . Journal of Rehabilitation Research and Development . 22(3), 25-41. Kauzlarich, J . J. (1986) . Wheelchair performance : wheelie curb climbing . In: Bougie T, Davies A . (Eds .) Wheelchairs : Research, Evaluation and Information. (pp . 113-7) Milan, Italy. Kirby, R. L., Atkinson, S . M . & MacKay, E . A. (1989). Static and dynamic forward stability of occupied wheelchairs : influence of elevated footrests and forward stabilizers . Archives of Physical Medicine . 70 681-86. Kirby, R . L., Kumbhare, D . A. & MacLeod, D . A. (1989). "Bedside" test of static rear stability of occupied wheelchairs . Archives of Physical Medicine . L, 241-4. Knowlton, R . G ., Fitzgerald, P . I . & Sedlock, D . A. (1981) . The mechanical efficiency of wheelchair dependent women during wheelchair ergometry. Canadian Journal of Applied Sport Science . 6 187-90.  67  Kobayashi, M ., Rodgers, M. M., Figoni, S. F., Gayle, Q . W., Schrag, D . R ., & Glaser, R . M . (1991) . Multidisciplinary data acquisition and analysis of wheelchair ergometry . Proceedings of the XIII International Congress in Biomechanics, (p . 368) Perth, Australia. Kyle, C . R., Caizzo, V . J ., & Palombo, M . (1978) . Predicting human powered vehicle performance using ergometry and aerodynamic drag measurements. Conference on Human Power and Health . Productivity . Recreation and Transportation, (p . 112) Cologne, Germany. Lakomy, H . K., Campbell, M . A. & Williams, C . (1987) . Treadmill performance and selected physiological characteristics of wheelchair athletes . British Journal of Sports Medicine 21, 130-33. Lamontagne, M ., Giroux, B . & Barclay, H. (1991) . The mechanics of the propulsion for a tilt seat wheelchair . Proceedings of the Canadian Seating and Mobility Conference, Toronto, Ontario, Canada. Lehman, J. F., Warren, C . G ., Halar, E ., Stonegridge, J . B. & Delateur, B . J. (1974) . Wheelchair propulsion in the quadriplegic patient . Archives of Physical Medicine and Rehabilitation . 55 183-86. Lemaire, E . D ., Lamontagne M ., Barclay, H . W. & John, T . (1991). A technique for the determination of center of gravity and rolling resistance for tilt-seat wheelchairs . Journal of Rehabilitation Research and Development . 2a(3), 51-8. Loane, T. D . & Kirby, R . L . (1986) . Low anterior counterweights to improve static rear stability of occupied wheelchairs . Archives of Physical Medicine and Rehabilitation . 67(4), 23-6. Loane, T. D., & Kirby, R . L . (1985) . Static stability of conventional and lightweight variable-axle-position wheelchairs . Archives of Physical Medicine. 66(3), 174-6. Luna-Reyes, O . B. & Reyes, R . I . (1988) . Energy cost of ambulation in healthy and disabled Filipino children . Archives of Physical Medicine and Rehabilitation . M(11), 946-9. Majaess, G . G ., Kirby, R. L. & Ackroyd-Stolarz, S . A . (1991). A new method to assess the dynamic rear stability of occupied wheelchairs . RESNA Proceedings . (pp . 356-7) Missouri, U .S.A. Marshall, R ., Wood, G . & Nade S. (1990) . Effects of ankle arthrodesis on walking : kinematic and kinetic studies . Clinical Biomechanics, a, 3-8. Masse, L . C. & Lamontagne, M . (1989) . Kinematic analysis of wheelchair propulsion for three speeds of propulsion . Proceedings of the XII International Society of Biomechanic (p . 1052) Los Angeles, California, U .S.A. 68  Masse, L .C., Lamontagne, M . & O'Riain, M . D.(1992) . Biomechanical analysis of wheelchair propulsion for various seating positions . Journal of Rehabilitation Research and Development, 29 (3), 12-28. Mattison, P. G., Hunter, J . & Spence, S . (1989) . Development of a realistic method to assess wheelchair propulsion by disabled people . International Journal of Rehabilitation Research . 12(2), 137-45. McLaurin, C . A . (1990) . Current directions in wheelchair research . Journal of Rehabilitation Research and Development . Clinical Suppl # 2 . 88-99. Niesing, R ., Eijskoot, F., Kranse, R ., Veeger, H . E . J & Van Der Woude, L. H . V. (1990) . Computer-controlled wheelchair ergometer . Medicine and Biological Engineering and Computers, 28(4), 329-28. Nitz, J . C. & Bullock, M . B . (1983) . Wheelchair design for people with neuromuscular disability . Australian Journal of Physiotherapy . 2a. 43-47. O'Connell, D . G ., Barnart, R . & Parks, L . (1992) . Muscular endurance and wheelchair propulsion in children with cerebral palsy and myelomeningocele. Archives of Physical Medicine, 73 . 709-11. Parziale, J . R . (1991) . Standard versus lightweight wheelchair propulsion in spinal cord injured patients . American Journal of Physical Medicine and Rehabilitation . Z(2), 76-80. Peizer, E ., Wright, D . & Freiberger, H . (1964) . Bioengineering methods of wheelchair evaluation . Bulletin of Prosthetic Research, 10 77-100. Ragnarsson, K . T. (1990) . Prescription considerations and comparison of conventional and lightweight wheelchairs . Journal of Rehabilitation Research and Development, Clinical Suppl, 8-16. Reger, S. & Hobson, D . (1983). Seat design factors for wheelchairs . In: Wheelchair Mobility, (pp .29-43) University of Virginia Rehabilitation Engineering Center. Reid, M ., Lawrie, A . T., Hunter, J . & Warren, P. M. (1990) . The effect of steering on the physiological energy cost of wheelchair propulsion . Scandinavian Journal of Rehabilitation Medicine 22(93), 139-43. Ridgeway, M ., Pape, C . & Wilderson, R . (1988) . A kinematic analysis of 800 meter wheelchair racing techniques . Adapted Physical Activity Quarterly . (2), 96-107. Robbins, S . (1991) . The 9th annual survey of the lightweights or, how to solve the puzzle and find the right chair for you . Sports 'N Spokes, March/April . 23-50. 69    Rodgers, M . M ., Gayle, G . W ., Figoni, S . F ., Kobayashi, M ., Glaser, R . M . & Gupta, S.C . (1992) Kinematic and kinetic responses to wheelchair propulsion during fatigue in SCI individuals . Proceedings from RESNA International 1992, (p .202) Missouri, U .S.A. Rodgers, M . M ., Tummarakota, S ., Lich, J & Schrag, D . R. (1992) . Threedimensional analysis of joint reaction forces and moments during wheelchair propulsion . Proceedings of NACOB II, (p . 457) Chicago, Illinois, U.S .A. Rozendal, R . H ., Rebroech, S . & Van Der Woude, L . H. V. (1990) . Methodology of user evaluation of hand driven wheelchairs . International Journal of Rehabilitation Research . 1j(3), 265-68. Sanderson, D . J . & Bednarczyk, J . H. (1990) . Biomechanics of wheelchair propulsion in different disabled groups : critical factors in determining effective propulsion . Report #89-55, British Columbia Medical Services Foundation. Sanderson, D . J . & Bednarczyk, J . H . (1991) . Assessment of a clinical tool for the evaluation of a wheelchair user interface : A biomechanical evaluation of effectiveness in wheelchair propulsion . Report # 90-62 . British Columbia Medical Services Foundation. Sanderson, D . J . & Sommer, H. J . (1985) . Kinematic features of wheelchair propulsion . Journal of Biomechanics . 1i 423-29. Shapcott, N ., Heinrich, M ., Brubaker, C . E ., & Ferguson-Pell, M . (1986) . A methodology to assist in the prescription of wheelchairs . Proceedings of the 9th annual Conference on Rehabilitation Technolo gy, June (p 34-45). Smith, A. W . (1991) Within and between-subject variability in normal human locomotion . Proceedings of the XIII International Congress on Biomechanics (p. 462-4) Perth . Australia. Staros, A . (1981) . Testing manually propelled wheelchairs . Prosthetics and Orthotics International, 75-84. Statistics Canada (1986) . Highlights : disabled persons in Canada In The Health and Activity Limitation Survey . (Catalogue # 82-692, pp 2-13 and 51-3 .) Ottawa, Ontario, Canada. Stoboy, H ., Rich, B . W. & Lee,T . R. (1971) . Workload and energy expenditure during wheelchair propelling . Paraplegia . $, 223-30. Stotts, K. M . (1986) . Health maintenance : paraplegic athletes and nonathletes. Archives of Physical Medicine and Rehabilitation, 67 109-114.  70  Su, F .C., Chou, Y. L., Lu, J . W. & Lai, K. A . (1991) Three dimensional kinematics in wheelchair propulsion . Proceedings of the XIII International Congress in Biomechanics, (p . 198) Perth, Australia. Trieschmann, R . B . (1987) . Aging with spinal cord injury . Caliper. Dec :17-20. Tupling, S . T., Davis, G . M . & Pierrynowski, L . (1986) . Arm strength and impulse generation : Initiation of wheelchair movement by the physically disabled. Ergonomics, 29(2), 303-11. Van Der Woude, L. H. V . (1989) . Seat Height in wheelchair propulsion . Journal of Rehabilitation Research Development.a(4), 31-50. Van Der Woude, L. H . V . (1990) . Seat height in hand rim wheelchair propulsion: a follow up study . Journal of Rehabilitation Science, 3(3),79-83. Van Der Woude, L. H . V ., de Groot, G ., Hollander, A . P ., & Van Ingen Schenau, G. J. (1986) . Wheelchair ergonomics and physiological testing of prototypes. Ergonomics . 22, 1561-73. Van Der Woude, L. H . V., Hendrich, M . M ., & Veeger, H . E . (1987) . Manual wheelchair propulsion : effect of power output on physiology and technique. Medicine and Science in Sports and Exercise . 20 t 70-78. Van Der Woude, L. H. V., Veeger, D . J . , & Rozendal, R . H . (1986) . Ergonomics of wheelchair propulsion . In : Bougie T, Davies A . (Eds) Wheelchairs : Research, Evaluation and Information . (pp .101-13) Commission of the European Communities, Milan, Italy. Van Der Woude, L. H. V., Veeger, D . J ., Rozendal, R . H.,Van Ingen Schenau, G. J ., Rooth, F. & Van Nierop, J . (1988) . Wheelchair racing : effects of rim diameter and speed on physiology and technique . Medicine and Science in Sports and Exercise, 20 (5), 492-500. Van Der Woude, L. H . V., Veeger, H. E ., Rozendal, R . H ., & Sargent, A. J. (1989) . Optimum cycle frequencies in hand-rim wheelchair propulsion. Wheelchair propulsion technique . European Journal of Applied Physiology and Occupational Physiology . 31(6), 652-35. Van Der Woude, L . H . V., Veeger, H . E . J . & Rozendal, R . H. (1989) . Propulsion technique in hand rim wheelchair propulsion . Journal of Medical Engineering Technology ., 136-41. Van Ingen Schenau, G . J . (1980) . Some fundamental aspects of the biomechanics of overground versus treadmill locomotion . Medicine and Science in Sports Exercise, 12, 257-61.  71  Veeger, H . E . J ., Lute, E . M . C., Roeleveld, K ., Van der Woude, L. H. V . (1991). Differences in wheelchair propulsion technique between trained and untrained subjects . Proceedings of the XIII International Congress on Biomechanics (p .204-6) Perth., Australia Veeger, D ., Van Der Woude, L . H. V., & Rozendal, R . H . (1989) . The effect of wheel camber in manual wheelchair propulsion . Journal of Rehabilitation Research and Development . 2a(2), 37-46. Veeger, H . E . J ., Van Der Woude, L . H . V ., Drexhage, D ., & Koperdraat, J. (1989) . Peak power output in wheelchair sprinting . Proceedings Biomechanics XII . International Society of Biomechanics, (p .217) Los Angeles . California, U .S.A. Voight, E . D. , & Bahn, C . C . (1969) . Metabolism and pulse rate in physically handicapped when propelling a wheelchair up an incline . Scandinavian Journal of Rehabilitation Medicine . 101-6.  1  Walsh, C . M. , Marchiori, G . E . & Steadward, R . D. (1986). Effect of seat position on maximal linear velocity in wheelchair sprinting . Canadian Journal of Applied Sport Science . 11s 186-90. Wang, Y . T ., Xu, D ., Hedrick, B ., Adrian, M ., & Morse, M. (1989) . Threedimensional kinematic features of wheelchair propulsion across racing speed conditions . Pr. -- .i : • f h- 1 h Ann I M-- no ef h m n n Biomechanics (p .178-9) Vermont, New Hampshire, U .S.A. Warren, P . M. (1990) . Energy costs of one-arm drive wheelchair propulsion : use of one arm versus one arm and ipsilateral leg . Clinical Rehabilitation, 414), 29197. Weber, S . (1990) . Wheelchair ergometer designed to test heart lung capacity of the disabled . The Medical Post . Feb 20 . 21-22. Whiting, R. B ., Dreissinger, T . & Hayden, T . (1984) . Wheelchair exercise testing: a comparison of continuous and discontinuous exercise . Paraplegia . 222.. 92-8. Wicks, J . R. , Olderidge, N . B . , Cameron ,B . J . & Jones, N . L. (1983) . Arm cranking and wheelchair ergometry in elite spinal cord injured athletes. Medicine and Science in Sports Exercise . 224-31.  5  Winter, D . A . (1983) . Biomechanical motor patterns in normal walking . Journal of Motor Behavior . L5(4), 302. Winter, D .A. (1984) . Kinematic and kinetic patterns in human gait : variability and compensating factors . Human Movement Science, 3, 51-76 .  72  Winter, D . A . (1990). The biomechanics and motor control of human gait . (pp .2031) New York : John Wiley and Sons. Wolfe, G . A., Waters, R . & Hislop, H. J . (1977), influence of floor surface on the energy cost of wheelchair propulsion . Physical Therapy, 57(5), 1022-7. Zwiren, L . D .& Bar-or, O . (1975) . Responses to exercise of paraplegics who differ in conditioning level . Medicine in Science and Sports . L 94-98.  73    APPENDIX A . Determination of the validity of the time normalizing and averaging software programs. Two software programs were developed to handle the angular data generated on the Peak System . The first was a program to time normalize the raw angular files (denoted by the extension .3AD) in the Peak system . In every case the raw files were normalized to 23 frames producing 5% intervals of the wheeling cycle . Figure 24 shows an example of a raw file and time normalized file for an adult subject. 31002 .ADN  31004 .3AD  0  2  4  6 8 TIME (MS)  10  0  12  2  4  6 8 TIME (MS)  1 '0  1 12  Fc i .ure 24 . The result of time normalization on the same elbow angle file for adult subject #10 . The .3AD file at the left is the raw file and the .ADN file at the right is the time normalized file. The second program was written to average five time normalized files prior to exporting to the Macintosh platform Table 6 shows the difference between the software program average file (Pgr .AVR) as compared to the hand calculated average file (Calc .AVR) which was generated by hand calculating the average for each of the five time normalized files (denoted by the suffix .ADN) corresponding to five different trials . The data presented here are for the elbow angular data for an adult subject . Similar results were obtained for other angular files and other subjects . In all cases no errors were detected in the software averaging program .  74    Table 6 . Comparison of the 5 trials of time normalized elbow angular data to the averaged file as produced by the averaging software program  1  31002 . ADN 113 .59 101 .15 97 .81 100 .98 115 .87 130 .24 142 .33 138 .39 132 .33 125 .80 125 .90 130 .01 136 .14 140 .38 147 .28 149 .88 147 .91 142 .33 134 .59 127 .84 114 .67  31003 . ADN 107 .63 96 .92 92 .57 94 .54 109 .86 133 .95 144 .20 144 .99 136 .20 128 .79 126 .92 126 .43 130 .33 138 .31 141 .82 143 .03 141 .60 136 .20 130 .18 122 .21 108 .81  31004 . ADN 108 .51 99 .71 97 .72 99 .73 112 .32 136 .57 148 .11 146 .52 132 .36 122 .70 119 .99 119 .58 122 .34 128 .06 133 .05 136 .69 139 .43 139 .60 136 .65 131 .12 117 .65  31005 . ADN 128 .42 117 .70 110 .14 101 .16 98 .32 103 .09 120 .94 141 .23 141 .19 133 .26 121 .49 115 .93 117 .06 124 .37 133,30 140 .01 141 .36 139 .62 134 .13 126 .96 114 .44  75  31006 . ADN 128 .17 116 .27 108 .10 99 .09 96 .56 103 .57 121 .07 133 .03 139 .17 132 .41 116 .40 107 .31 105 .58 109 .31 114 .99 123 .05 123 .03 126 .76 131 .21 131 .02 123 .21  Pgr . AVR 117 .26 106 .35 101 .27 99 .10 106 .59 121 .48 135 .33 140 .83 136 .25 128 .59 122 .14 119 .85 122 .29 128 .09 134 .09 138 .53 139 .07 136 .90 133 .35 127 .83 115 .76  CaIc .AVR 117 .26 106 .35 101 .27 99 .10 106 .59 121 .48 135 .33 140 .83 136 .25 128 .59 122 .14 119 .85 122 .29 128 .09 134 .09 138 .53 139 .07 136 .90 133 .35 127 .83 115 .76  % ERROR 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0    APPENDIX B . Determination of the reliability of the determination of angular and velocity data A reliability study was done by digitizing the same trial for the same subject five times consecutively . Views from both camera angles were digitized in the normal fashion and the angular displacement and the time normalized angular displacement files generated and then exported to the Macintosh computer. Figure 25 . shows the results of this process. #1 ---#2 	#3 ----#4  0  75  80-  `or v 6 wJ • o 40zQ zW W W 20▪ J 0 2 N  0  100  0  25  50 % CYCLE  100  #1 ---#2 #3  z0 T 80 51 1 o  75  ----#4  ----# 5  60-  o  m•z av 240  oJ  C DQ 4200 x co  0  0  25  50 % CYCLE  75  100  0  25  50 % CYCLE  75  100  Figure 25 . Reliability of the angular data (elbow, shoulder, trunk, and shoulder abduction) when the same trial was digitized 5 times consecutively. These data were obtained from an adult subject. The mean and standard deviation was determined for each point in the cycle for the five trials . The average standard deviation for the elbow angle data was 1 .4 degrees, 1 .8 degrees for the shoulder angle data, 1 .2 degrees for the trunk angle data and 1 .1 degrees for the shoulder abduction angle data . The overall mean error for all of the four angular parameters was 1 .4 degrees. 76    APPENDIX C . Determination of the robustness of the angular kinematics over time. Two of the subjects involved in this study were involved in other studies over the past two years . The study in 1991 involved wheeling moderate distances overground in a different laboratory space in which the wheeling surface was low pile carpet . Three dimensional video analysis of wheeling was used in this study . The results of a comparison of the elbow and shoulder an g ular data for two sub'ects over a two ear time s•an are shown in Figure 26. SUBJECT # 25 SUBJECT # 25  0  SUBJECT # 26  25  50 CYCLE  75  100  SUBJECT # 26  60 0  25  50 % CYCLE  75  100  Figure 26 . Comparison of the angular data of two adult subjects collected over a two year time period. What is interesting however is that the angular profiles for these two subjects are quite different . Notice the rapid decrease in elbow extension in subject #25 during the recovery phase . In contrast the elbow remains extended throughout much of the recovery phase for subject # 26 . This same pattern 77  remains remarkably consistent across a twelve month time span (1991 and 1992 data) and across different laboratory conditions . These data would support the hypothesis that wheeling style as measured kinematically is quite robust.  78    APPENDIX D . Individual subject angular data.  PEDIATRIC SUBJECT #01  80  -+ 0 k  + 5 kg ---+10 kg  0  25  50 % CYCLE  25  50 % CYCLE  75  75  0  100  Oj 0  100  60 , 20 35 SHOULDER ANGLE  25  I  50  65  50 % CYCLE  25  50 % CYCLE  75  75  100  i 100  r 80  EXTENSION -1n  Fiaurg 27 . Individual angular data (elbow, shoulder, trunk, shoulder abduction angle, and shoulder angle-elbow angle, in degrees) for pediatric subject #01.  79    160-  PEDIATRIC SUBJECT #02  . ..~•  120-  z  z  2100-n , z  1 _, .0  0  w  = u, 20-  --+10 kg 25  40-  o m  '~  0 kg  w 80-  o  N`  50 % CYCLE  75  --50 75 % CYCLE  I  100  0  25  100  0  25  50 % CYCLE  I  25  75  100  I  50 % CYCLE  75  100  1601401 20 -  Propulsion 60 20 40 SHOULDER ANGLE  60 80 EXTENSION-.  Figure 28 . Individual angular data (elbow, shoulder, trunk, shoulder abduction angle, and shoulder angle-elbow angle, in degrees) for pediatric subject #02.  80    PEDIATRIC SUBJECT #03  60	 I 0  25  50 % CYCLE  75  100  0  15  25  50 % CYCLE  75  80-  60-  10-  40-  0  1  25  50 % CYCLE  75  100  0  25  50 % CYCLE  75  100  1  60  20 35 50 SHOULDER ANGLE  65 EXTENSION  80  Figure 29 . Individual angular data (elbow, shoulder, trunk, shoulder abduction angle, and shoulder angle-elbow angle, in degrees) for pediatric subject #03.  81     PEDIATRIC SUBJECT # 04  ll: 600  25  50 % CYCLE  75  25  0 CYCLE  -15  --i 100  25  ----1 100  25  50 °le CYCLE  ---T 50 % CYCLE  75  100  T---  75  100  160 ; eV,:  1  14 ; -;  1201  u'1 00 (I) z  80-1 uI  Propulsion  6o 20 35 SHOULDER ANGLE  50  65 80 EXTENSION --No  . Figure 30 . Individual angular data (elbow, shoulder, trunk, shoulder abduction angle, and shouider angle-elbow angle, in degrees) for pediatric subject #04.  82    PEDIATRIC SUBJECT #05  80  60  40 0  w z r. 20 w x w 0 0  25  50 % CYCLE  75  100  15  z  0 I-  U O m Q  0  % CYCLE  Propulsion  I  60  20 35 50 SHOULDER ANGLE  i  I  65 EXTENSION  80 --~  Figure 31 . Individual angular data (elbow, shoulder, trunk, shoulder abduction angle, and shoulder angle-elbow angle, in degrees) for pediatric subject #05.  83    PEDIATRIC SUBJECT #06  160  1  140  120 z 0 z 100 w ix w 80 60 20 40 60 80 SHOULDER ANGLE EXTENSION  100 ►  Figure 32 . Individual angular data (elbow, shoulder, trunk, shoulder abduction angle, and shoulder angle-elbow angle, in degrees) for pediatric subject #06.  84    PEDIATRIC SUBJECT #07  15 80  A 60  z O_ X W  J LL  5 a  I   0 0  25  50 % CYCLE  I  I  75  100  0 0  25  50 % CYCLE  75  100  160  A 140  120  0100  wz 1-x  Propulsion  80  w  60 20 30 40 SHOULDER ANGLE  50  60 70 EXTENSION  80  Figure 33 . Individual angular data (elbow, shoulder, trunk, shoulder abduction angle, and shoulder angle-elbow angle, in degrees) for pediatric subject #07.  85    PEDIATRIC SUBJECT #08  25  50 % CYCLE  75  100  75  100  10  z 05 x w J  -  LL  0—  0  	 I 25  I  0  50 CYCLE  0  25  50 % CYCLE  75  100  160 140 120 z 2100 to z w 80 w  x  Propulsion  60 20 40 SHOULDER ANGLE  60 EXTENSION  80  Figure 34 . Individual angular data (elbow, shoulder, trunk, shoulder abduction angle, and shoulder angle-elbow angle, in degrees) for pediatric subject #08.  86    PEDIATRIC SUBJECT # 09  15  z 0 W  5  J LL  0 0  25  50 % CYCLE  75  100 % CYCLE  160  Recovery  140 120 z  0  1 00  z w Propulsion  60 — 20 35 50 SHOULDER ANGLE  65  i 80  EXTENSION  Figure 35 . Individual angular data (elbow, shoulder, trunk, shoulder abduction angle, and shoulder angle-elbow angle, in degrees) for pediatric subject #09.  87    PEDIATRIC SUBJECT #10  80 60  40  20  0 0  25  50  75  100  % CYCLE  Recovery  z co z  100  w  x  W  80  Propulsion  60 20 35 50 SHOULDER ANGLE  65  80  EXTENSION  Figure 36 . Individual angular data (elbow, shoulder, trunk, shoulder abduction angle, and shoulder angle-elbow angle, in degrees) for pediatric subject #10.  88    ADULT SUBJECT #01  80 A 60 z 40 O_ zz w xx - 20 w  0 25  50 % CYCLE  75  z  o  - 20 0 O  m 0  L----L-  0  25  50 % CYCLE  75  25  100  50 % CYCLE  75  100  160  w  Propulsion  80 60  20 35 50 65 SHOULDER ANGLE EXTENSION  80  Figure 37 . Individual angular data (elbow, shoulder, trunk, shoulder abduction angle, and shoulder angle-elbow angle, in degrees) for adult subject #01.  89    ADULT SUBJECT #02  80 60 40  r-  0 25  50 % CYCLE  75  100  I  0  25  50 % CYCLE  75  100  80  15  60 40  A  160 140  Propulsion  w 80 60 20  35 50 65 SHOULDER ANGLE EXTENSION  80 ►  Figure 38 . Individual angular data (elbow, shoulder, trunk, shoulder abduction angle, and shoulder angle-elbow angle, in degrees) for adult subject #02.  90    ADULT SUBJECT #03  15  80  60  z40 0  I  ~_  0 0  25  020 Q  v  I	I  50  75  100  % CYCLE  120 z  4 100 N z  w  xW  Propulsion 80  60  I 0  20  40  SHOULDER ANGLE  60 EXTENSION  80  Figure 39 . Individual angular data (elbow, shoulder, trunk, shoulder abduction angle, and shoulder angle-elbow angle, in degrees) for adult subject #03.  91    ADULT SUBJECT #04 80  60  0  A 160  Propulsion  60  I  I  0 20 40 60 SHOULDER ANGLE (degrees) EXTENSION  I  80 ►  Figure 40 . Individual angular data (elbow, shoulder, trunk, shoulder abduction angle, and shoulder angle-elbow angle, in degrees) for adult subject #04.  92    ADULT SUBJECT # 05 80 w J C, z  60  40 wz co J V,  0 w 20 Nw  0  80  60  z 40 0 o0  0 20 Q  0 0  25 %  50 CYCLE  75  100  0  25  50 % CYCLE  75  100  160  A 1 40 120 z 0  0 1 00 w  Fx  w  Propulsion  80  60 — 0 20 40 60 80 SHOULDER ANGLE EXTENSION --~  Figure 41 . Individual angular data (elbow, shoulder, trunk, shoulder abduction angle, and shoulder angle-elbow angle, in degrees) for adult subject #05.  93    ADULT SUBJECT #06 160  % CYCLE  z  120  O  N  w 100 w w  Propulsion 80 60 0  20 40 60 SHOULDER ANGLE EXTENSION  80  Figure 42 . Individual angular data (elbow, shoulder, trunk, shoulder abduction angle, and shoulder angle-elbow angle, in degrees) for adult subject #06.  94    ADULT SUBJECT# 07 80  60  40  20  0 25  50 % CYCLE  75  100  0  25  25  50 % CYCLE  75  100  0  25  50 % CYCLE  75  15  0 0  50 % CYCLE  75  100  160  140  120  Propulsion  60 0  20 40 60 SHOULDER ANGLE EXTENSION  80 —>  Figure 43 . Individual angular data (elbow, shoulder, trunk, shoulder abduction angle, and shoulder angle-elbow angle, in degrees) for adult subject #07.  95    ADULT SUBJECT # 08  Propulsion  60 0  20 40 SHOULDER ANGLE  60  80  EXTENSION  Figure 44 . Individual angular data (elbow, shoulder, trunk, shoulder abduction angle, and shoulder angle-elbow angle, in degrees) for adult subject #08.  96    ADULT SUBJECT # 09  80  60  z 0 z 1-  40  20  x w 0  15  10 z 0 x— w  5  J U.  0  0  Recovery  Propulsion  60 0  20 40 SHOULDER ANGLE  60  80  EXTENSION -->  Figure 45 . Individual angular data (elbow, shoulder, trunk, shoulder abduction angle, and shoulder angle-elbow angle, in degrees) for adult subject #09.  97    ADULT SUBJECT #10  25  80  60  z 40 0 0  m 20 m 0  0 0  25  50 % CYCLE  0  75  100  0  25  50 CYCLE  75  100  120  N z X  100  w  Propulsion  80 60 0  20 40 SHOULDER ANGLE  60  80  --► Figure 46 . Individual angular data (elbow, shoulder, trunk, shoulder abduction angle, and shoulder angle-elbow angle, in degrees) for adult subject #10.  98  EXTENSION  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0077075/manifest

Comment

Related Items