Page 63 - Journal of Special Operations Medicine - Summer 2016

find that completion times by surgery residents with cri- model would translate to another model or to an actual
cothyrotomy experience were significantly improved be- patient. Knowing this, and that the exact anatomy of
tween the first and second attempts without additional each human patient can vary, it may be advantageous
training. 82 to train learners with multiple models. It may also be
advantageous to train learners with varying degrees of
There is insufficient evidence to answer the questions of cognitive load, to optimize performance in a patient-
(1) comparative efficacy, (2) validity of training and as- care environment. 88–90
sessment models, or (3) whether simulator models could
replace LT models. Of the 185 commercially available There are a number of methodology limitations in
synthetic models, studies comparing simulators to LT or many of the studies reviewed. The majority of the stud-
cadaveric models were sparse. A number of LT and syn- ies assessing performance scores used assessment tools
thetic models showed acceptability. without reliability or validity evidence, making the re-
sults of these studies difficult to interpret and compare.
The cadaver model, with its inherent anatomic fidelity, Many studies did not include an a priori power analy-
is assumed by many to have implicit content validity sis or sample-size calculation, making the any results
and predictive validity, as evidenced by its consistent use without a significant difference less clear. A number of
as a testing model. 23,83–87 However, we did not find any studies were based solely on reports of the trainee’s self-
studies investigating the transfer of skills from cadav- confidence or self-efficacy, which we know is often not
ers to live patients. Furthermore, the time of death and an accurate reflection of a learner’s actual competence,
the fresh versus frozen state of the model often is not and can be especially unreliable in less experienced, nov-
controlled and likely influences tissue fidelity; this vari- ice, or underperforming learners. 91–101
ability does not allow for a true “standardized” model.
The high cost and scarcity of donated bodies may also The remaining validity evidence is also sparse, with some
be problematic, especially when attempting to do large- evidence of acceptability and face validity for certain LT,
scale training or assessment. synthetic, and cadaver models, knowing that cadavers
may have a number of logistic barriers and variability
The only study that compared LT and synthetic mod- between models. However, this variability mirrors actual
els (caprine versus TraumaMan) for cricothyrotomy patient care and may be desired. There is a lack of inves-
and TT placement (with objective performance test- tigation into content validity evidence for the LT model;
ing on a cadaver model) found no difference between content validity is often implicit in the cadaver model
training models. Unfortunately, the fact that it was and has been questioned based on various measure-
23
underpowered, due to difficulties obtaining the requi- ments of or between anatomic structures in eight of the
site number of cadavers for testing, only allows us to simulator models. 23,54,67,68 Despite these measurements,
conclude that the investigators were unable to detect a that ETI success was similar for two of these simulators
very large difference between these training modalities; and with actual patients calls into question the clinical
any smaller differences may not have been elucidated significance of these measurement differences. Further
61
in such an underpowered study. While the primary foci investigation is needed to assess the relevance of, and
of this study were appropriate (measuring the subjects’ subsequently define and compare anatomy, physiology,
ability to complete the tasks and time to completion), and tissue property characteristics of, the various train-
performance scores generated based on critical elements ing models to assure proper translation to the human
of the procedure were not examined. Therefore, the trauma patient. Convergent and discriminant validities
23
investigators may have failed to notice smaller improve- were only described in three studies 22,23,55 , suggesting
ments in performance if the procedure was not success- more investigation is required.
fully completed.
The data are similarly scarce regarding learning curves
True translational outcomes (i.e., performance on ac- and skill retention or decay. There is a body of evidence
tual patients) following educational interventions were regarding ETI learning curves of trainees. The medi-
only discussed in six studies (five for ETI, one for TT). cal/paramedic student studies seem to indicate greater
Other studies used performance outcomes on simula- success rates after fewer attempts than does some an-
tors and some investigated performance outcomes on esthesia-trainee literature; however, all of these student
LT models. 22,37,56,57,81,82 We are unsure of the utility of studies used simulation training prior to OR attempts,
measuring performance outcomes on synthetic or ani- and two of the three studies included ongoing simulation
mal models, especially when testing on the same model training throughout the course of the study. It is possible
used for training, since there is a learning effect that that in addition to the less stringent requirements for de-
occurs when using any one model. 64,66,82 It is difficult fining success in the student studies, bolused or ongoing
to ascertain whether performance on any one training simulation training could accelerate the learning curve

Simulation Versus Live Tissue for Training Trauma Procedures 49