[Indices based on entropy for measuring social inequalities in health]

As described in the scientific literature, indices used to measure social inequalities in health have positive features, but they also have shortcomings, depending on how they are applied. The objective of this article is to put forward and to demonstrate, in both theoretical and practical terms, the advantages of measurements of inequality based on the notion of entropy, which is well known in physics and in information theory. The article defines and presents the main properties of indices based on the notions of entropy and redundancy. The application of the indices is illustrated with two fictitious data sets and also with real data derived from basic health indicators for the Americas, from the Pan American Health Organization. Indices based on the notion of entropy have properties that include: a) not varying with scale changes, b) being symmetrical, c) incorporating a socioeconomic dimension, and d) being easy to interpret thanks to the condition of equivalence between entropy and a system with two categories.