Empirical Analysis of Health Assessment Objective and Subjective Methods on the Determinants of Health

Abstract

Background: There are several methods for assessing health status. The aims of this study were to investigate the empirical differences between health assessment objective and subjective methods, to identify a possible long-term relationship between methods and health determinants and the influence of these methods on the perceived level of risk according to health determinants.

Methods: Using data from 1970 to 2018 in the United States, health status was assessed by perception of health, absence from work due to self-reported illness, life expectancy at birth and mortality rate. Health determinants were tobacco and alcohol consumptions, number of physicians per 1,000 persons, stay in hospitalization unit, curative care, release of greenhouse gases, per capita gross domestic product (GDP) and urbanization. The differences between health objective and subjective assessment methods were investigated through a Generalized linear model, a structural break date of health methods was investigated by Chow test and the long-term relationship between health assessment methods and health determinants by Engle and Granger cointegration test.

Results: Tobacco consumption was associated with a decrease of life expectancy while no long-term causal relationship was found between them. There was a positive correlation between alcohol consumption and perception of good health with a long-term causal relationship. Although per capita GDP positively influenced life expectancy, there was no cointegration between them. The release of greenhouse gases was positively correlated with both the absence from work due to self-reported illness and the perception of good health. Finally, curative care was associated with a decrease of mortality and absence from work due to self-reported illness and an increase of life expectancy and perception of good health while hospitalization is positively correlated with mortality and negatively correlated with life expectancy with a long-term causal relationship. Finally, the number of physicians per 1,000 persons was not correlated with health assessment methods used.

Conclusion: Our results highlight the influence of health assessment methods on the determinants of health and the fact that the perceived risk of health determinants changes according to the method used. Thus, the impact of health assessment methods must be considered in order to prioritize the determinants of health.

Keywords: cointegration; empirical analysis; health determinant; long-term causal relationship; objective and subjective health assessment methods.