Defining Movements
The first consideration when counting number of movements as a measure
of performance efficiency is to understand the nature of the skills that
are being evaluated. In this regard, the field of motor control
kinesiology has typically classified actions as composed of combinations
of movements that can be defined as either discrete or continuous.
Discrete movements are those that have a recognizable beginning and end,
like throwing a ball, turning a doorknob, or flipping a light switch;
they are usually essential to skills that rely of the precise production
of a distinct outcome. Continuous movements, on the other hand, have no
recognizable beginning or end, and will continue until they are stopped
arbitrarily by the performer. Skills composed of continuous movements
can have precision constraints but are often concerned with the
maintenance of an ongoing action. These skills include activities such
as walking, swimming, and cycling.29 In the
surgical-medical domain, the clinical technical performances of interest
can more often than not be characterized as serial actions. This refers
to skills that are made up of a number of discrete movements that must
occur in a very particular order. These actions can appear continuous,
but usually have distinct components with very definitive beginnings and
ends.29 As such, any one clinical technical skill can
be conceptualized in terms of the minimum number of discrete movements
that would be needed for its successful execution. However, as the
number of movements assessment construct suggests, this minimum is not
always achieved.
That performances can contain movements in excess of the minimum
required by the task is fundamental to the use of the number of
movements metric as a measure of efficiency. Simply put, performances
often contain errors or imperfect actions, which require corrections;
and each erred movement and subsequent correction constitutes the
production of additional movements. In this way, the hallmark error
volume associated with novice performances has led to the natural
assumption that new trainees will perform procedures with more
movements, and therefore, more inefficiently. Given this position, the
challenge for assessors of surgical technical skills is to determine
where one movement within a procedure ends and the next one begins. In a
motion capture and analysis protocol, the way that serial movements are
usually disentangled from one another involves plotting the position
function as a displacement profile and then differentiating and
double-differentiating it to generate velocity and acceleration profiles
respectively (Note: if the motion data is captured by accelerometer
technology, then integrations are performed on the resulting profile to
reveal velocity and displacement). From these profiles, assessors look
for determinant characteristics within the action trajectories that
indicate a new movement. Defining these characteristics becomes one of
the most important decisions underpinning the effective use of the
“number of movements ” metric.
Reflecting on the way that errors and corrections emerge in a motor
performance can be helpful in setting the appropriate motion analysis
parameters for determining the onset and offset of a movement. Consider,
for instance, the types of errors that require a correction. For one, an
action can require a correction because the performer selects and
executes the wrong movement. This type of error occurs, for instance,
when the laparoscopic surgery trainee forgets that the display screen is
incongruently rotated with respect to the work space and ends up moving
a grasper to the right instead of the left. To correct these types of
errors, the ongoing movement must be terminated and replaced with an
entirely new movement. Sometimes, this involves reversing direction to
return to where the action started, or stopping to reassess the
situation before initiating a new movement in search of corrective
solution. With this type of error and correction in mind, skill
assessors may set a zero crossing in the velocity-over-time profile as
the end and start points for successive movements. However, this type of
new movement determinant can be insufficient when one considers that a
series of movements can be executed without the limb coming to a
complete stop between each.
It is necessary to understand that noise in the neuromuscular system
means that the production of movements is inherently
variable,30 such that discrete precision actions
usually require a subtle or not-so-subtle correction (or corrections)
towards their conclusion in order to be successful even when the
appropriate movement is selected and executed correctly by the
performer.31 In this regard, new movement determinants
based on zero crossings in the acceleration-over-time profile are also
problematic as they ignore the refined and controlled nature of human
motion.32, 33 Indeed, the findings from over a
century’s worth of experiments on the accuracy of voluntary actions
reveal that any one precision movement includes complimentary impulses
that move the limb toward its goal and then integrate response-produced
sensory feedback to correct the overall movement for
accuracy.34 In this regard, the typical acceleration
profile of a single discrete movement derived via motion capture
techniques has 3 zero crossings, which characterize a large sinusoid
(i.e., the initial impulse) that is followed by a second, smaller
sinusoid (i.e., the corrective impulse).31, 35, 36 The
idea is that because variability in movement execution is so inherent to
motor performance that its management becomes a fundamental challenge
for learners as they move along the continuum of expertise. That is,
novice performers struggle to determine whether their approach to
performance introduces too much variability to correct, while skill
performers understand the inherent variability and develop strategies
that allow them to anticipate the type of corrections required for their
movements. In this way, a single expert movement often includes periods
of deceleration and re-acceleration.31