Page 37 - Biomedical Engineering and Design Handbook Volume 2, Applications

16 MEDICAL DEVICE DESIGN

momentarily to the external fixation device, it requires that the pins be affixed to the fractured bone.
In considering the need, it is recognized that the fracture might represent an athletic injury to a young
healthy person with very strong bones, or it may have resulted from a mild impact to an individual
with relatively weak bones. We would like our device to function over this range, and must choose
test media accordingly. We would start with commercial bone emulation material (e.g., Sawbones®),
but then probably move to some animal models. Careful analysis of this situation is important. Why
is the test carried out? Are we interested in failure mode to better understand the device or are we
proving that this design is superior? Caution is called for.

1.18 TESTING

The testing and evaluation of the final device is a most important activity. This is dealt with here
separately from the prototyping activity because it reaches beyond the design team in many respects
and because the process and attitude required are more distinct from nonmedical products here than
in the prototyping activities that are really internal to the design process.
Once the design is finalized, all of the evaluation and test data become important in the regu-
latory approval process, so each step should be carefully planned, executed, and documented.
Each test should be done with the clear objective of validating some aspect of the design. The level
of performance required should be established in advance. The test conditions should be carefully
spelled out.
Bench or laboratory testing should measure the integrity, precision, and other “engineering”
aspects of the device. They can also verify the production techniques, as at this stage the units being
tested should come from the manufacturing processes that will be used in full production. Besides
testing for normal performance, it is wise to test devices to failure at this stage, to ascertain that there
are no unforeseen failure modes or collateral effects.
Many products move next to cadaver testing, where the biologically driven uncertainties cause
complications. Sometimes it is possible to use animal models first in this phase, which allows for
much more consistent conditions. Human cadaver tests require careful notation of the history and
condition of the cadavaric materials. Results may be correlated with this data if there are anomalies
that require explanation.
Clinical trials with human subjects require extensive documentation. They are most often done in
institutions that have in place extensive procedures to ensure the subjects are informed about any
experimental activities that they participate in. A committee reviews the forms and procedures for
doing this along with the details of the study before institutional approval is granted. These things
require long lead times in many cases, so the planning of the experiments ought to be done well in
advance.
It is common practice to run clinical trials in several settings, often distant from each other. The
point was made earlier that the device should be usable by an average practitioner, not just the ultra-
highly skilled. The trials should include the opportunity to evaluate this aspect, as well as to evaluate
training materials and other ancillary items that contribute to product function. Care and forethought
continue to be the watchwords.

1.19 DOCUMENTATION

Documentation of a medical device design is critical. This applies not only to the usual drawings and
manufacturing documents that accompany commercial products, but also to the needs and require-
ments documents that have been discussed earlier, as well as test data, evaluation reports, results of
clinical trials, and any other pertinent information about the device, its history, or experience.
The U.S. Food and Drug Administration requires that for each device there be a device master
file. The length and breadth of that rule becomes clearer when it is explained that the “file” need not