Deterministic sensitivity analysis

Deterministic sensitivity analysis (DSA) is a method used to investigate the sensitivity of the results from a model-based analysis to variations in specific input parameters: one or more parameters are changed, and the impact on the model’s output values is observed. The amount that each parameter is varied is typically pre-defined and, where appropriate, corresponds to published uncertainties for that parameter (e.g. based on reported 95% confidence intervals from source studies). In univariate DSA, parameters are varied one at a time, while multivariate DSA involves simultaneously changing multiple parameters. DSA results are often presented as line graphs or bar charts. Tornado charts (a stack of bar graphs) are also commonly used, which are graphs that summarise the impact of several input parameters in univariate DSA, ordered by the magnitude of their effect on the output (with the widest variation on top). DSA is limited in its ability to simultaneously vary numerous parameters, and it is not usually possible to vary more than 4 to 5 parameters simultaneously in combination. Instead, probabilistic sensitivity analysis is required to assess the impaact of simultaneous variation of many input parameters. Univariate DSA should also be interpreted with caution when input parameters are highly correlated (i.e. where parameters correlated to the parameter of interest are not varied together), such as the sensitivity and specificity of diagnostic tests.