Applied Ergonomics 1971. 2;4.

221-225

An approach to the design of

wheelchairs for young users

M. G. Kenward

Department of Ergonomics and Cybernetics, Loughborough University of Technology

Sixty six physically handicapped children between the ages of 5 and 16 took part in an
anthropometric study of young wheelchair users. All those involved had at one time
used a wheelchair regularly, and only 9 in the study habitually walked. Using a standard
body measuring instrument, 13 measurements of bodily dimensions were taken with
the children sitting in their usual chairs. An assessment was also made of the degree of
support afforded by the backrest. Five dimensions normally used in prescribing
wheelchairs to patients were used as a basis for working out the sizes of the various
sections of a wheelchair, such as seat, backrest, armrest and frame size. The raw data
indicated that appreciable inconsistencies in body shape existed which were not
allowed for in present wheelchair design. A new approach is proposed in which
wheelchairs would be assembled from interlocking sections of variable sizel This
arrangement would allow wheelchairs to be fitted more completely to the individual.

An article published in a report on "Disabled in Society" School, near Mansfield in Nottinghamshire. It is a

stated that 1 ½ million people, or 3% of the population were
disabled in some way. Another report published in 1966
mentioned that 20% of those substantially and permanently
handicapped were issued with wheelchairs from the
Ministry of Health, and the number being prescribed is
rising annually by 10%. Since the wheelchair is a
mechanical substitute for lost or impaired ambulatory
functions, it is important that it offers maximum postural
support and manoevrability. However, only 25% of
Ministry wheelchairs are built to individual requirements;
the rest of the demand is met by chairs built to a single
pattern with provision for additional features which do not
involve alteration of the basic structure.
This arrangement proves to be unsatisfactory for a
large proportion of younger wheelchair users, who by
virtue of disease are inconsistently proportioned and
therefore not properly supported by the chairs. The
situation is further aggravated for larger or older children
because different sizes of chair vary only in seat width among
the larger models.
Similar sizes of wheelchair made by different companies
have until now been sufficiently different to prevent
interchangeability of spare parts. However, the'Ministry is
at present engaged in standardising the specifications of all
makes of chair, so that the situation should soon change.
This might well provide a good opportunity to review the
sizing of wheelchairs, particularly for the younger age
group. At present, there is very little anthropometric
data of disabled people, and as far as is known this study is
the only source of data on disabled children.
The figures were obtained from Thieves Wood Special
Residential School catering for severely physically
handicapped children between the ages of five and
sixteen. It is one of the earliest schools of its type,
having classrooms, a workshop and physiotherapy section
at one end of a long corridor, and dormitory .andmedical
facilities at the other, with the dining hall and
recreational areas in the middle. The children came from
the East Midlands area mostly, with only a few from
further afield, and they suffered from a wide variety of
diseases.
Methodology and apparatus
Sixty-six of the seventy-five children resident at the
time of the study were measured. Five~ children were either
in hospital or sick bay, and the other four were not
considered suitable for inclusion in the survey, since they
walked everywhere, and had never been issued with a
Ministry wheelchair. A total of 13 measures of body
dimension were taken using a standard anthropometer of
German make (see Fig.l). It could easily be dismantled
into four main sections and the top two were designed for
use separately as calipers. An assessment was also made of
the degree of support afforded by the backrest. The
result sheets included the three manikin sketches shown
in Fig 2, which indicated the measurements and their
order of determination.
Additional information collected during the
measuring process included date of birth, nature of
disability, make and model of wheelchair and how much
they could walk, if at all.

Applied Ergonomics December 1971 221

 Experimental procedure
All measurements were arranged to be independent of
wheelchair dimensions. The children were asked to sit
comfortably upright, looking straight ahead, and when
possible with arms on the armrests. They did not remove
any clothing, except for blankets, so occasionally
allowances had to be made for calipers, girdles and
braces.
The anthropometer was assembled in front of the
children before being used, because some of the younger
ones seemed to be put off by being suddenly confronted
with the full height of the instrument. It was also found
helpful, in establishing rapport, to use only as many
sections of the instrument as necessary. In most cases
this was three sections, but in some cases two. Although
this was inconvenient when the calipers came to be
assembled, it was less disquietening for the children.
Each measurement will now be briefly described:
1. Measurement of sitting height
The subject was asked to sit well back into the chair,
making as much use as possible of the backrest, and
looking straight ahead. The instrument was placed behind
the subject in the midline of the body, and the caliper
arm brought to rest firmly on the skull, compressing the
hair.

2. Measurement of shoulder height

The subject maintained his posture, and the caliper
was brought down to rest on the acromion of the left
shoulder. It helped standardise the measurement when the
elbow remained unsupported and so whenever possible
Fig 1. Anthropometer similar to the one used in the the left armrest was removed. In this and subsequent
experiment measurements taken to one side of the body the left side

12

Fig 2. Diagrams to show the measurements and their order of determination.

222 AppliedErgonomicsDecember 1971

was preferred because the scales became easier to read.
Except in the case of electric chairs it was more
convenient to manoevre the chair rather than the
anthropometer, when different regions of the body had to be
measured.
3. Measurement of elbow height
This was determined by placing the anthropometer arm
hard up against the base of the humerus, in the upper arm.
The forearm was parallel to the floor and at right angles to
the upper arm. It was essential that the elbow received no
support from those armrests that could not be removed.
The scale on the anthropometer relates to the inside edge of
the caliper arm, and since elbow height and seat height
involved the outside edge, one centimetre had to be added to
these values to allow for the thickness of the arm.
4. Measurement of seat height
This was the last of vertical measurements, and was
taken in a manner similar to sitting height. The arm
was pressed upwards against the canvas of the
collapsible chairs, and the correction explained above applied.
The thickness of the canvas was not taken into account,
since measurements were taken to within ½ cm and the
canvas was much thinner than this. In the case of electric
chairs the inside edge of the anthropometer arm was pressed
down against the seat squab, and this needed no
correction.
5. Measurement of buttocks to knee
Having taken the last of the vertical measurements, the
top two sections of the instrument were assembled into
calipers. For the first dimension with this arrangement the
fixed caliper arm was located against the lowest point of
the bulge on the backrest. The other caliper was moved
along the tube and positioned so as to press against the
lower edge of the patella. In order to read this scale the
measurement was taken as usual on the left side of the body.
In this and all other measurements taken with calipers the
reading was noted before the instrument was withdrawn.
6. Measurement of foot to popliteal
This measurement involved locating one caliper in the
welt of the shoe and sliding the other arm until it pressed
securely into the angle of the knee, or popliteal. The
calipers were held parallel to the leg during the measurement.
This was obviously not a true anthropometric measurement,
since it was not taken between bony landmarks. However,
it was considered a valid and accurate dimension from the
point of view of fitting subjects to wheelchairs. There was
in fact little alternative since it was not practical to remove
shoes or bandages. An additional centimetre had to be
added to this reading for reasons that applied to elbow
height.
7. Measurement of foot to knee
In this measurement was located in the same position
as above, and the upper caliper was brought to bear on the
upper edge of the patella. The body of the instrument was
once again parallel to the lower leg. After this
measurement the calipers were reduced in size since only
one length of the tubing was needed hereafter.
8. Measurement of ankle to toe
The length of the foot was then measured, and because
footwear was not removed, this again could not be
regarded as a true anthropometric measurement. It was
more representative, however, of the normal situation as
in the example above. If footwear other than shoes were
worn such as, for example, socks or bandages, a note was
made to this effect on the result sheet.
9. Measurement of popliteal to knee
To facilitate accurate measurement, subjects were asked
to move forward in their chairs when possible, so that the
knee was well clear of the forward edge of the seat canvas.
The caliper arms were then located in place, one into the
angle of the knee and the other on the lower edge of the
patella.
10. Measurement o f forearm length
An armrest was needed for support for this measurement,
and so the right arm was measured for this reason. It was
also easier to read the scale from the right hand side. The
fist was clenched in order to lock the wrist, and one
caliper arm was aligned with the forward edge of the
styloid process on the ulna bone. The other caliper arm
was pressed against the olecranon process of the elbow.
11. Measurement of buttocks width
It was originally intended to measure hip width, but
from the outset this was found completely impracticable,
especially with some examples of muscular dystrophy.
A small number of lean subjects, however, could be
measured across the hips, these measurements had to be
converted to buttock width by a correction factor. The
figure decided upon was 4 cm, since this was suggested as a
good approximation by Roberts (1878).
12. Measurement of scye width
This refers to the width of the trunk at the level of the
armpits. Both buttock width and scye width were taken
from the front of the subject, and afterwards the left
armrest was replaced.
13. Measurement of shoulder width
This was found easier to measure from behind the subject.
The elbows were positioned well back along the armrests in
order to obtain the maximum representative value. The
calipers were pressed against the acromion processes of the
shoulder.
14. Assessment of backrest support
At this stage, any cushions previously removed were
replaced and the subjects asked to assume their usual posture
It was noted where the lowest point of contact with the
backrest occurred, and a comment was made on the result
sheet if the subject was likely to benefit from further
cushions or special padding.
Results
No correlation could be found between the data and any
criterion such as age, sex, disability or size of wheelchair.
In addition, a large percentage of children received
Applied Ergonomics December 1971 223
inadequate support from the chairs in one or more areas of Table I Results and sizes proposed
the body. The only feasible solution appeared, therefore,
to favour construction of wheelchairs on a modular basis, Body Range Sub-range Chair Chair
involving various sizes of the basic elements of frame, seat, parameter (cm) code parameter dimension
backrest and armrest assembled easily to produce a well-
fitting chair. 1Buttock 20-24 A Seat 28
The five body measurements normally used in prescribing width 25-29 B width 33
30-34 C 38
wheelchairs were derived for each subject, and the total
35--39 D 43
range found for each. These were then divided equally into 40-44 E 48
five sub-ranges and labelled A, B, C, D, and E. A Wheelchair
Prescription Indicator (Everest and Jdnnings, 1968) was Buttock 29-33 A Seat 27
then used to relate the sub-ranges of each body parameter to 34-38 B depth 32
popliteal 39-43 C 37
to an appropriate wheelchair dimension. The results are
44-48 D 42
shown in Table 1. 49-53 E 47
An attempt was made to correlate seat width and seat
depth, with a view to producing a number of overall seat Foot to 18-22 A Seat 24
sizes. It became apparent that, when different code letters popliteal 23-27 B height 29
28-32 C 34
for width and depth occurred, the width could not be
33-37 D 39
reduced, and neither could the depth be increased. In 38 - 4 2 E 44
addition, very few examples occurred where the code
letters were more than two sub-ranges apart. As a result the Shoulder 15-20 A Backrest 18
number of combinations o]F seat width and depth have to elbow 21-26 B height 24
been rationalised to ten, as shown in Table 2. 27-32 C 30
33-38 D 36
Armrest height was further investigated and a high 39-44 E 42
correlation found to exist b~tween arm length and the
dimension elbow to seat. In consequence a single letter Elbow to 9--12 A Armrest 13
could be used to denote an armrest fitment of given seat 13--16 B height 17
length and height. The sizes are displayed in Table 3. 17--20 C 21
21-24 D 25
25-28 E 29
Recommendations
By defining the size of a wheelchair by five letters as Table 2 Seat sizes
explained above a patient is clearly more likely to be
issued with a well fitting appliance. This could quite Code letters Seat dimension (cm)
easily be accomplished using existing methods of producing (Width) (Depth) (Width) (Depth)
wheelchairs out of tubular steel. The basic frame would be in A A 28 x 27
five sizes, differing mainly in seat height rather than length. B A 33 x 27
B B 33 x 32
This would enable the tubular part of the armrest to be of
C B 38 x 32
constant length, varying only in height, and with the C C 38 x 37
appropriate length of padding bolted on to it. Although D B 43 x 32
this arrangement would tend to be wasteful of tubing for D C 43 x 37
the smallest sizes of chair, the advantages of standardization D D 43 x 42
would, hopefully, more than compensate for the small E D 48 x 42
numbers of wheelchairs with Group A seat frames. Only E E 48 x 47
two subjects in the study needed this size of frame. The
correct size of seat canvas once rivetted in place, would
naturally dictate the width of the chair, and the backrest Table 3 Armrest sizes
could be similarly attached at the correct height for the
Code letter Armrest size (cm)
individual.
(Height) (Length)
An example will illustrate the procedure for
prescribing a wheelchair for a child. Firstly, frame size is A 13 x 26
determined from measurement 6 - foot to popliteal; 5 cm B 17 x 29
are added to allow for minimum floor clearance and 2 cm C 21 x 32
D 25 x 35
E 29 x 38

Table 4 Prescription of a wheelchair

Body parameter Dimension no Dimension (cm) Chair parameter Sub-range code

Buttock width 11 28 Seat width B
Buttock-popliteal 5--9 40 Seat depth C
Foot-popliteal 6 30% Seat height C
Elbow-shoulder 2--3 17 Backrest height A
Elbow height 3-4 18 Armrest height C

224 Applied Ergonomics December 1971


32cm
/-J
21cm
37c
Fig 3 Diagrams to show dimensions of the wheelchair prescribed in Table 4
subtracted to elevate the thigh above the front edge of the
seat. FromTables 4 and 1,the correct seat height can be
seen to be "C".
Seat width is determined by adding 5cm to buttock
width, and 4 cm are subtracted from the dimension
buttock to popliteal to determine seat depth. Reference
to Tables 1 and 2 show the correct seat size as "BB".
Armrest height is found by adding 2½ cm to elbow height
and again Table 1 indicates Group "C" as the best fit.
Backrest height is calculated as being half way between
shoulder and elbow at the apex, and this is found to be
size "A". Fig 3 shows the scaled diagrams of the completed
chair.
So by the above process, an accurately fitting
wheelchair can be quickly prescribed for a patient and easily
put together by existing methods of manufacture.
Discussion
One important factor to take into account is the
relative frequencies of the various diseases. In the sample
studies the incidence of polio was becoming increasingly rare,
but the relative proportion of spina bifida cases was
greatly increasing. Young children with this disease
presented fewer problems with back support, since they
had been operated on at birth to remove the meningocele
that is characteristic of all such children over fourteen.
There were indications that muscular dystrophy was
becoming less common, although this may not be true in
practice.
Consideration would also have to be taken of rate of
growth, or rather increase in size. There is a very large
[
33 cm --[
II
variation, depending upon the nature of the disease. It must
be stressed that the sample was extremely small and
specialised, and the results may not be representative of
the whole population. The study should not be regarded
as final, and more work in the anthropometric field is
necessary for a completely definitive solution.
Nevertheless, it is suggested that some real improvement
can be made in designing wheelchairs for the younger
generation by applying the approach proposed here.
Acknowledgements
The author is very grateful to the National Fund for
Research into Crippling Diseases, and to the Industrial
Design Section of the Royal College of Art for their kind
assistance and cooperation in producing this report.
References
Everest and Jennings
1968 'Measuring the Patient', Everest and Jennings, Inc,
California.
Parkes, K.
1966 Paraplegia, 4, 2 4 - 2 6 , A pilot study of the space
requirements of wheelchair users.
Roberts, C.
1878 'A Manual of Anthropometry'. London.
Siegal, S
1956 'Non-Parametric Statistics for the Behavioural
Sciences'. McGraw Hill.
M.G. Kenward 1971
Applied Ergonomics December 1971 225