|
The outcome of the
expected outcome calculations shows that "Foot Transplantation"
is the preferred option by a narrow margin:
| Decision
Option
| QALYs
|
|---|
| Food
Transplantation | 8.23 |
| Treat
Infection | 8.11 |
|
Amputate
injured foot
|
8
|
|
However, the difference
in QALYs is so small that you can imagine that the preferred option
would depend on the values assigned to the numbers in the decision
tree.
Sensitivity analysis
is the technique by which to discover whether a particular utility
or probability is important in determining the preferred option.
Sensitivity analysis is the key to the power of decision analysis
in situations in which the value of the probabilities is not known
precisely (which is where decision analysis has the most to offer).
Suppose that you don't
know the value of a probability, but you have a pretty good idea
of the lowest possible value and the highest possible value that
it can take. To do sensitivity analysis, simply insert the lowest
value in the range and calculate the expected value of each decision
alternative. Then, repeat the expected value calculation after
substituting the highest value in the range. If the decision alternative
with the highest expected value is the same in both calculations,
the decision is not sensitive to the value of the probability.
Sensitivity analysis
helps you to identify the most important parameters (the probabilities
and outcomes) in the model. You can focus your literature search
on them. Or, you can use sensitivity analysis to set your research
agenda.
Research Opportunities:
Sensitivity Analysis
Keep in mind that there is much opportunity for research that
would establish optimal values for probabilities, and ranges for
sensitivity analyses.
|