Assessment of disagreement: a new information-based approach.

PURPOSE

Disagreement on the interpretation of diagnostic tests and clinical decisions remains an important problem in medicine. As no strategy to assess agreement seems to be fail-safe to compare the degree of agreement, or disagreement, we propose a new information-based approach to assess disagreement.

METHODS

The sum over all logarithms of possible outcomes of a variable is a valid measure of the amount of information contained in the variable. The proposed measure is based on the amount of information contained in the differences between two observations. This measure is normalized and satisfies the flowing properties: it is a metric, scaled invariant and it has differential weighting.

RESULTS

Two real examples and a simulation were used to illustrate the usefulness of the proposed measure to compare the degree of disagreement.

CONCLUSIONS

Used as complement to the existing methods, we believe our approach can be useful to compare the degree of disagreement among different populations.