Biomedical Engineering Reference
In-Depth Information
Actually, the deterministic exposure/deterministic effect scenario is not really a risk scenario because
there is no “chance” involved. It would be like saying that releasing a 50 kilogram steel anvil from
1 meter above the earth's surface runs the risk of falling toward the ground! The risk comes into play
only when we must determine external consequences of the anvil falling. For example, if an anvil is
suspended at the height of 1 meter by steel wire and used by workers to perform some task (i.e., a
deterministic exposure), there is some probability that it may fall (e.g., studies have shown that the wires
fail to hold one in ten thousand events, i.e., failure probability of 0.0001); so, this would be an example
of a deterministic exposure followed by a probabilistic effect (wire failure), i.e., permutation number 2.
Estimating risk using a deterministic approach requires the application of various scenarios, e.g., a
very likely scenario, an average scenario, or a worst-case scenario. Very likely scenarios are valuable in
some situations when the outcome is not life threatening or when there are severe effects, like cancer.
For example, retailers may not care so much about the worst case (a customer is not likely under
most scenarios to buy their product) in designing a store, but wants to avoid the risk of losing a likely
customer (e.g., the most potential Armani suit customers are not likely to want to hear loud heavy metal
music when they walk in the store; but conversely loud heavy metal music may be a selling feature
for most customers looking to buy tube amplifiers at a guitar store). The debate in the public health
arena is often between a mean exposure and a worst-case exposure (i.e., maximally exposed and highly
sensitive individuals). The latter is more protective, but almost always more expensive and difficult to
attain. For example, lowering the emissions of a toxic substance from a industrial stack to protect the
mean population of the state from the effects of exposures is much easier to achieve than lowering the
emissions to protect an asthmatic, elderly person living just outside of the power plant property line
(Figure 10.1). While the most protective standards are the best, the feasibility of achieving them can
be a challenge (Figure 10.2). For many biomedical decisions, general rules will vary. For example, a
device or medical protocol may have to be designed for highly unlikely conditions (e.g., two standard
deviations from the mean - very sick people).
Actual or realistic values are input into the deterministic model. For example, to estimate the risk of
a valve malfunction from the materials used, metabolic in a processes, the characteristics of the bodily
fluids, the vulnerability of the valve materials to rupture, and the expected flow through the valve,
from which the risk of failure is calculated. A probabilistic approach would require the identification of
the initiating events and the metabolic states to be considered; analysis of the adverse outcome using
statistical analysis tools, including event trees; application of fault trees for the systems analyzed using
the event trees (i.e., reliability analyses; see Chapter 3), collection of probabilistic data (e.g., probabilities
of failure and the frequencies of initiating events); and the interpretation of results.
Human beings engage in risk management decisions every day. They must decide throughout whether
the risk from particular behaviors is acceptable or whether the potential benefits of a behavior do not
sufficiently outweigh the hazards associated with that behavior. In engineering terms, they are optimizing
their behaviors based upon a complex set of variables that lead to numerous possible outcomes. Engineers
are designing systems to help lower some of these risks and are morally obligated to do so.
In addition, the engineer must weigh and balance any responsibility to represent the client with due
diligence. This diligence must be applied to designs and plans for manufacturing processes that limit,
reduce, or prevent failures. Ultimately, the engineer participates in means to remedy the problem (i.e.,
to ameliorate health, environmental, and welfare damages). The evaluation and selection of the best
alternative is the stuff of ethical decision making.
