Graphical representation of multivariate data, applied to the Chilean healthcare system

Abstract

Objective : To inform public health researches about some proposal of multivariate statistical graphics, such as Chernoff faces, star charts, Andrews curves and radar graphs. Its use is now possible thanks to recent software development. Material and method: Proposed multivariate statistical graphics are used to describe the Systems of Public Health of Chile, year 2010 (n = 29), using quantitative information with dimension p = 7; these graphics are processed using standard statistical software such as R, STATA and SAS.

Results: The description of the 29 systems of public health of Chile based on quantitative information from dimension seven, represented in each of multivariate graphs, allow to determine clusters, standards and trends.

Conclusions: The availability of efficient statistical software indispensable suggests to complement strategies of exploratory analysis of our data with appropriate graphical representations of multidimensional context that can help us to achieve a better understanding of the problem of interest. The graphical summary generated by the application of multivariate statistical graphs to describe Public Health Systems of Chile complements our knowledge of this aspect of national life, may suggest hypotheses, refute others and assist in the interpretation of complex results. Naturally this can be extended to many other situations of interest to the health care professional.