View Full Version : Re: Bionet controversial topic #6: on the obligation to shareuseful research data

04-11-2002, 11:10 PM
Dear all, the debate about sharing research data has received an
interesting light from the useful suggestion from Gabor to look at the
experience in clinical gait analysis (CGA). Although CGA is certainly a
very widely spread field of biomechanics research, the story there reported
should be taken much carefully because of the specific nature of the
instruments and analyses used. Nevertheless, that experience may be of some
help to all.

Sharing databases of normative (better 'control') human gait data emerged
to be a strong need from the CGA community. In that debate, however,
several researchers have pointed out the high risks in this operation. In
that debate I emphasised how many are the sources of variability: although
marker trajectories are supposed to be almost the same independently from
the motion system utilised, a) marker sets, b) anatomical landmark
identification, c) segment anatomical frame definition, d) mechanical
convention and their mathematical formulation for representing joint
kinematics and kinetics vary significantly from lab to lab. My advice at
that time was to be very careful in comparing own data with corresponding
someone else normative data.

To confirm my concerns on integrity of data going into the repositories (as
also well pointed out by Frank Buczek, particularly true for CGA data), I
have got shocked when Roy Davis, Co-Director of the Motion Analysis
Laboratory of the Shriners Hospitals for Children has reported recently in
the meeting the results of a very revealing test: a single volunteer being
gait analysed in the many Labs of their Organization spread all over the
States. Despite the strict instructions available for long a long time to
all operators, the locally biased practice resulted in differences in gait
analysis final parameters as large as 50 and even 100%, and for some even
in different direction (abducting rather that adducting for example, if I
well remember). Roy was very kind to share this as a public warning
message. All this even when not talking about gender, race, age etc....

But this is CGA. The different experimental practices and methods of
analysis are not the only source of difference. Subtle changes often occur
also related to the experimental original aim and approach, which may
result in pretty different observations and then in drammatically different
deductions and conclusions. The previous debate (Topic 4) on knee joint
DoFs has evidenced this quite clearly. In this respect, the point just made
by Frank Buczek is particularly interesting, pointing out that experimental
data are biased even by the hypothesis in itself, as it must have been, in
my view. Hardly someone else data are useful to me, because this data
simply had been collected and analysed with a different aim, which make a
lot of difference.

The differences both in the methods used and in the original aims affect
the value of these repositories. Although I am not theoretically against, I
am much concerned and my advice would be of a much carefulness. Just my
************************************************** ************************
Alberto Leardini, DPhil
Movement Analysis Laboratory
Centro di Ricerca Codivilla-Putti
Istituti Ortopedici Rizzoli
Via di Barbiano 1/10, 40136 Bologna ITALY
tel: +39 051 6366522
fax: +39 051 6366561
email: leardini@ior.it

"Where is the Life we have lost in living,
Where is the wisdom we have lost in knowledge,
Where is the knowledge we have lost in information."
Thomas Stearns Eliot, Choruses from ''The Rock'' (1934)
************************************************** ************************

To unsubscribe send SIGNOFF BIOMCH-L to LISTSERV@nic.surfnet.nl
For information and archives: http://isb.ri.ccf.org/biomch-l