Background: The study of the relationship between self-reported morbidity, health status and health care utilization presents methodological problems due to the variety of illnesses and medical conditions that one individual may report. In this article, correspondence analysis was use to analyse these relationships.
Methods: Data from the Spanish National Health Survey pertaining to the region of Catalonia was studied. Statistical analysis included multi-way correspondence analysis (MCA) followed by cluster analysis.
Results: The first factor extracted is defined by self-assessed health perception; the second, by limitation of activities, and the third is related to self-reported morbidity caused by chronic and acute health problems. Fourth and fifth factors, capture residual variability and missing values. Acute problems are more related to perception of poor health while chronic problems are related to perception of fair health. Also, it may be possible to distinguish self-reported morbidity due to relapses of chronic diseases from true acute health problems. Cluster analysis classified individuals into four groups: 1) healthy people; 2) people who assess their health as being poor and those with acute health problems; 3) people with chronic health problems, limited activity and a perception of fair health; and 4) missing values.
Conclusions: Correspondence analysis is a useful tool when analyzing qualitative variables like those in a health survey.