An empirical analysis of the dimensions of health status measures.

The objective of this study is to verify empirically the existence of separate dimensions in the overall concept of health status by analyzing 10 variables included in a questionnaire that was applied to all adults in a simple random sample of households in St John's, Newfoundland. The response rate was 85% for a total of 3300 subjects. These data were analyzed by frequencies and by associations with sex, age and education. Nonparametric correlation, factor and cluster analyses on variables were used to verify if health status had identifiable dimensions. All these methods produced similar results showing five distinct factors. The first factor is composed of variables related to disease (disability/chronic conditions/worry about health); the second, to happiness (happiness/emotional); the third, to subjective appraisal of health (physical condition/comparative level of energy/self-rated health status). Finally, the fourth and fifth factors were single variables; restriction of normal activities and social contacts. An interesting finding was that self-rated health status was distributed with almost equal weight in both the first and third factors. A validation of the 10 variables and the 5 factors was undertaken by studying their association with health care utilization. Two measures of utilization were used; number of physicians' visits in a year and number of hospital days in a 4-year period. Number of chronic conditions, disability and self-rated health status were associated with both measures of utilization; factor 1 was the only summary construct showing association with utilization. This paper demonstrates that self-rated health status is valid as a single measure of overall health status in this sample, being associated with both disease and subjective assessment components.