6. Sensitivity analysis is the calculation of how changes in the underlying assumptions in an economic appraisal or risk analysis would affect the results from the particular analysis. It involves varying the important and uncertain variables in the particular analysis to see what effect this has on the conclusions. If the conclusions are not markedly affected, then the conclusions or results are robust.
7. Sensitivity analysis should be used in circumstances where:
• There are genuine uncertainties in the values ascribed to some variables;
• the outcomes of the particular analysis are sensitive to a number of key variables (i.e. variables with relatively large values); and
• there is no reason to expect significant bias in the estimates of costs, benefits or risks.
8. In a risk analysis, it is good practice to perform sensitivity tests on the risks which have been assessed to have relatively large values and those with uncertain probabilities and/or financial impact. The latter could result from lack of historical information or lack of experience of managing the particular risk.
9. An effective way of presenting the results of sensitivity testing is to calculate the switching value or cross-over point. This is the amount by which the variable(s) under investigation would have to change in order to affect the ranking of the options. For example, it can show by how much the value would have to fall (if it is a benefit) or rise (if it is a cost or risk) to make it not worth undertaking the option. A view can then be taken about the likelihood of the factor turning out worse than the switching value.