View Full Version : FW: Reply Summary: Controlling walking velocity?

Jody B. Brucker
11-13-2000, 08:08 AM
> We would like to thank all of the people that replied to our
> query on controlling walking velocity during gait trials. Even
> though there is not a definitive answer to the way you should
> control for velocity, it must be accounted for in the design of
> the study, especially, if you are going to compare between
> groups or are repeating measurements. Since Paul and I have two
> different research questions we will control for velocity
> differently. Thanks again to all of those that replied, and
> here is a copy of the messages we received.
> Sincerely,
> Jody B. Brucker
> Paul D. Stoneman
> BYU Physical Education Doctoral Candidates
> ------------------
> ------------------
> Why not have subjects walk at a self selected speed AND at a
> constrained
> velocity. Ultimately, it depends on your focused research
> question but this
> will cover both statistical deign control and allow for
> research protocol
> design control and offers the best option for detecting
> differences in the
> end.
> Michael R. Torry, PhD
> ---------------------
> In our lab we have also found that velocity contaminates our
> interpretation
> of treatment effects on gait. Furthermore, we have found that
> subjects use
> complex combinations of velocity, cadence and stride length to
> achieve a
> variety of objectives that they may have (speed, pain, habit,
> personal
> style, efficiency, etc). Indeed any of those 3 variables can be
> changed or
> controlled by some combination of changes in the other 2
> variables.
> One question that you always have to be concerned about is that
> if you
> control any of these variables, is there gait any longer
> representative of
> anything that you can use to make your judgements about their
> gait. We have
> come to the conclusion that the best way to control these
> variables (if you
> are going to do so) is with a treadmill to control velocity,
> and with a
> metronome to control cadence. This is relatively easy for
> subjects to do and
> seems to least interfere with the quality of their gait. The
> downside is
> that when you control velocity and cadence, you can be sure
> that you are
> also controlling stride length - it may be indirect, but there
> will be very
> little variation in stride length possible if the subjects
> conform to the
> prescribed velocity and cadence. Under these conditions you
> will have to
> rely on kinematic and kinetic variables other than velocity,
> cadence and
> stride length for your analysis, as well as qualitative
> information about
> subject pain, etc under the conditions set.
> I would be interested in hearing what you decide to do and how
> you make your
> decision.
> Peter V. Loubert PhD, PT, ATC
> -----------------------------
> We test in at a fixed velocity and at the subject's average
> velocity. Then
> you can see how much difference there really is. We are
> assessing plantar
> pressure data with an EMED Pedar system.
> Elizabeth J. Higbie, PhD, PT, ATC
> ---------------------------------
> On chatting to a colleague who is a Physio she mentioned that
> in
> clinical settings physio's calculate the cost of gate to allow
> quantitative measures of improvement to be recorded.
> Steve Myers
> -----------
> Jody
> do both.
> if you can.
> at least do a pilot study first with both methods.
> the key is to not end up down the road a year with mush for
> data.
> this problem vexes me when i read running fatigue studies.
> are kinematics changing because person is running slower or
> because
> they are fatigued or both.
> Rodger Kram
> -----------
> Hi Jody,
> I don't have a reference at my fingertips, but I believe time
> normalization
> has been used for gait in both kinematic and kinetic analysis.
> Check authors
> like Winter, Patla,or Perry and some of the other gait
> literature.
> Shaw
> -----
> Dear Jody,
> In 1997 we did some work related to this. We were
> interested in the
> effect of different levels of gait control on planter pressure
> measurement
> (but also force). I attach a PDF file of our paper. I also
> looked at the
> issue of using walking velocity as a covariate in gait analysis
> - this was
> quite successful. It was only produced as an abstract and I
> don't have the
> copy to hand but here is the reference;
> Shaw J.E., Harrison A.J., Entwistle H., Boulton A.J.M., (1996)
> Complex Gait
> Analysis in Diabetic Neuropathy. British Diabetic Association
> medical and
> scientific conference Sept. 1995.
> I think you have a fair degree of choice here - based on our
> experience
> either way will work quite well. It is probably just a matter
> of choosing
> the approach that is most appropriate to your situation. eg. we
> found full
> gat control to be impracticable with subject who had gait
> difficulty eg.
> diabetics older patients etc. but it was ideal with normal
> subjects and
> those with minor gait problems.
> Good Luck
> Dr Drew Harrison
> (the PDF is not included with this Reply Summary, jbb)
> ----------------------------
> Dear Jody,
> You hit at an important point. Speed influences practically all
> gait
> quantities. In order to compare patients before and after
> treatment,
> my opinion is that you should investigate
> 1) whether 'normal' or maximal speed has increased. This alone
> is
> already a very important finding.
> 2) whether the walking pattern, i.e. the time course of the
> relevant
> quantities, moments, forces, angles etc. etc. has significantly
> changed. To separate treatment effects from effects due to the
> speed
> increase, patients should be compared at equal speeds. A
> proposed
> protocol would thus include a) pre-op: at 'comfortable' or
> 'maximal'
> speed, b) post-op: again at the new 'comfortable' or 'maximal'
> speed PLUS recordings at the speeds of a)
> 3) When patients are to be compared with a normal population,
> care
> should be taken that the normal values represent the whole
> range of
> speeds to be encountered in the patients.
> 4) To compare speeds in subjects of various stature, speeds
> should be
> scaled. This gives me one more opportunity to advertize my
> paper
> in Gait and Posture 4:222-223 (1996).
> It is a pity of much otherwise good research in this field,
> that
> these effects are insufficiently taken care of.
> Good luck with your work,
> At Hof
> --------
> I, too, am running test subjects through a gait lab and
> understand there may
> be effects of walking velocity on their GRFs. I would be
> interested in the
> responses you collect from your inquiry.
> Thanks.
> Jacqueline Schmit
> -----------------
> Jody,
> An interesting article by Grabiner in J Biomech (1995 Sep; 28
> (9): 1115-7)
> looked at targeting and ground reaction force variability and
> concluded
> that "targeting would not be expected to influence the
> variability of
> calculated kinetic variables that are subservient to GRF".
> In my own work subjects walk across a force plate but I control
> their
> walking velocity via a metronome. Subjects choose a
> self-selected pace and
> begin a self-selected distance away from the force plate so
> that they are
> walking at a constant velocity upon heel-strike and the foot
> strikes in
> relatively the same place on the force plate. My particular
> group of
> subjects then under go arthroscopic ACL surgery and return 3
> and 6 months
> later for post-testing. I then use the same self-selected
> velocity and
> distance for each of the 2 post-tests as was used in the first
> test sesion.
> There have been few alterations in velocity and no deviations
> in distance
> thus far which may or may not play an important role in
> subsequent data
> analysis according to Grabiner.
> This may not answer your question regarding inter-subject
> variability and
> what to control. I chose to be more concerned with
> intra-subject
> variability and collect from as many subjects as possible then
> look to see
> how to generalize the results. Essentially, I have a
> single-subject design
> and will look for possible group similarities afterwards. In
> this fashion,
> intra-subject variability was more my concern and collecting
> numerous
> trials more important than inter-subject variability and
> controling for
> cadence and stride length.
> I hope this is useful.
> Reed
> --------
> there is some work by Messier in the area of OA gait where they
> controlled
> the gait & looked at kinematics & ground reaction forces. This
> is always a
> debate. The other possibility if you have the time is to do a
> few trials
> each way & see what the changes are.
> Karen Perell
> -------------
> If you use their preferred walking speed and possibly match
> your control group
> for height and leg length, then changes in the preferred
> velocity (step length
> and cadence) could be attributed to the treatment or recovery
> in general. If
> you are measuring your pre and post treatment on the same day,
> this conclusion
> is tighter. Try some searches on preferred rate with respect
> to gait.
> Beth Boyd EdD, PT
> -------------------
> Hi Jody,
> Suggest you use 2 velocities: self paced and one standardized
> velocity. You
> can always normalize afterwards.
> Shaw Bronner
> ---------------
> I have just about made up a simulation software for a biped
> walking model
> where I may compute the impulses in order to maintain symetry
> and a desired
> walking velocity in two dimensions.My model, while still
> elementary,
> produce an illimited number of steps, reaching a sort of stable
> limit
> cycles. I have made a few conferences so far with this model
> but I am in
> the process of writing one or two journal papers. If you can
> supply me with
> some recorded data on foot impulses and velocity, I might be
> able to see if
> there is some validation coming out for my model. Of course, I
> would not
> publish anything with your data unless authorised and I will
> keep them
> confidential.
> Yours very truly,
> Paul Bourassa
> ------------------
> I've been a subject in projects where the researchers have
> tried to vary
> velocity as a continuous variable- *extremely* difficult,
> especially at
> higher speeds. Has to do with the leg as a pendulum and all
> that. My
> inclination is to treat velocity as a categorical variable
> (slow walk, fast
> walk, trot, run, that sort of thing) and figure that the actual
> velocity of
> each individual is going to be fairly consistent from session
> to session,
> but that there's going to be outrageous inter-individual
> variation. Might
> be a good correlation with their inseam length. A literature
> search on
> Gregory Breit might help here.
> Clifford M. Les, DVM, PhD
> ---------------------------

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