A Computational Model of Similarity Analysis in Quality of Life Research: An Example of Studies in Poland

Life (Basel). 2022 Jan 1;12(1):56. doi: 10.3390/life12010056.

Abstract

Due to the multidimensional structure of the results of similarity studies, their analysis is often difficult. Therefore, a compact and transparent presentation of these results is essential. The purpose of the present study is to propose a graphical representation of the results of similarity analysis in studies on the quality of life. The results are visualized on specific diagrams (maps), where a large amount of information is presented in a compact form. New similarity maps obtained using a computational method, correspondence analysis, are shown as a convenient tool for comparative studies on the quality of life of different groups of individuals. The usefulness of this approach to the description of changes of the quality of life after the retirement threshold in different domains is demonstrated. The World Health Organization Quality of Life-BREF questionnaire was used to evaluate individuals in Poland. By analyzing clusters on the similarity maps, two groups (employees and retirees) were classified according to their quality of life in different domains. By comparing the structures of the classification maps containing the information about the whole system considered, it is clearly seen which factors are important in the comparative studies. For the considered problems, the uncertainty coefficients describing the effect size and preserving the information on the symmetry of the variables that were used for the creation of the contingency tables were evaluated.

Keywords: World Health Organization Quality of Life-BREF; clustering; correspondence analysis; data analysis; health informatics; quality of life.