Income and health: what is the nature of the relationship?

https://doi.org/10.1016/S0277-9536(98)00385-2Get rights and content

Abstract

The aim of this article is to examine the relationship between income and morbidity, both before and after controlling for other socio-economic variables. We use data from the Health and Lifestyle Survey (first wave), a national sample survey of adults, aged 18 upwards, in England, Wales and Scotland, conducted in 1984–1985. In total, 9003 interviews were achieved. We examine the shape of the relationship between household equivalised income and height, waist–hip ratio, respiratory function (FEV1), malaise, limiting longterm illness.

These indices of morbidity, both self-reported and measured, are approximately linearly related to the logarithm of income, in all except very high and low incomes (this means that increasing income is associated with better health, but that there are diminishing returns at higher levels of income). A doubling of income is associated with a similar effect on health, regardless of the point at which this occurs, providing this is within the central portion (10–90%) of the income distribution. The effect of income on the health measures is comparable to that of the other socio-economic variables in combination. The shape of the relationship found between income and health is compatible with worse health in countries with greater income inequality, without the need to postulate any direct effect of income inequality itself.

Introduction

The association between socio-economic position and mortality has been extensively investigated in the UK and elsewhere (Davey Smith et al., 1990a; Kunst and Mackenbach, 1994). Due to the scarcity of routine data on morbidity, the socio-economic distribution of ill-health has been less extensively investigated, although the same general pattern of less favourable outcomes amongst those in worse socio-economic circumstances is seen (Marmot et al., 1991; Eachus et al., 1996). In the UK, Registrar General's occupational social class has been the most frequently utilised socio-economic measure while, in the US, education and income are more commonly used. This reflects differing traditions of routine data collection and perhaps differences in the ways in which the factors underlying the uneven socio-economic distribution of sickness and death are conceptualised (Krieger and Fee, 1994). There may, however, be reasons why different relationships with health measures will be seen when different socio-economic measures are used. For example, income and occupation can be altered by poor health in adulthood, leading to some degree of reverse causation between socio-economic position and health, while health status in adulthood can only rarely influence the level of achieved education. It is therefore important to investigate the associations between health and different socio-economic measures, with income having been poorly studied in this regard in the UK.

In studies from the US, household income and median income of inhabitants of the residential area have been strongly associated with mortality risk (Sorlie et al., 1995; Davey Smith et al., 1996a, Davey Smith et al., 1996b). In the largest study of household income there is a suggestion of a flattening, or even reversal, of the income and mortality relationship at the high end of the income spectrum (Backlund et al., 1996) while this was less evident in the largest study of income of area of residence (Davey Smith et al., 1996a, Davey Smith et al., 1996b). The latter finding is consistent with a considerable body of data demonstrating that socio-economic differentials in health persist across the whole socio-economic spectrum (Davey Smith et al., 1990b, Davey Smith et al., 1994, Davey Smith et al., 1996a, Davey Smith et al., 1996b; Marmot et al., 1991; Ford et al., 1994; Ecob, 1996; Der et al., submitted for publication). The finally graded association between socio-economic position and health is an important contributor to the formulation of explanations for the existence of health inequalities (Davey Smith et al., 1994; Marmot et al., 1995). The investigation of the association between particular socio-economic indicators and a variety of components of health can contribute to understanding in this area. We therefore undertook a detailed exploration of the exact shape of the relationship between household income and several health outcomes, both before and after adjustment for other socio-economic indicators, in a representative sample of adults in England, Wales and Scotland.

Section snippets

Methods

This analysis is of data from the first wave of the Health and Lifestyle Survey (HALS) (Cox et al., 1987), a national survey of adults in England, Wales and Scotland in 1984–1985. This survey was chosen in preference to other national surveys, such as General Household Survey (GHS, 1997) which has larger sample size and, being annual, is more recent and has better data on income, because of our desire to examine a variety of measures of health, both self-rated and physically measured, which are

Results

The relationship of each health measure to income both before and after social controls is shown in Fig. 1. The graphs are for males in the age group 40–64 except for waist–hip ratio, where females aged 40–64 are plotted. These are the age and sex groups for which, when the relationship to income varies by age or sex, the strongest relationships within working ages (considered as the difference in health measure between the 90th and the 10th percentile of the income distribution) are found.

Income and health

This paper has reported on the relationship of income to morbidity in a national study both before and after controlling for a range of socio-economic variables. Up to now, little has been known about the residual relationship of income to morbidity after controlling for other socio-economic variables. In the UK, this is in large part due to the general omission of income at the individual level from large surveys and the census. The relationship of all measures of health examined, both

Acknowledgements

We would particularly like to acknowledge the advice of Geoff Der in connection with the production of figures in this paper and of Sally Macintyre, Geoff Der and Graeme Ford for comments on earlier drafts; and to thank three referees for useful comments.

References (36)

  • G Davey Smith et al.

    Socio-economic differentials in mortality risk among men screened for the multiple risk factor intervention trial. I. Results for 300,685 white men

    Am. J. Public Health

    (1996)
  • G Davey Smith et al.

    Socioeconomic differentials in mortality risk among men screened for the multiple risk factor intervention trial. II. Results for 20,224 black men

    Am. J. Public Health

    (1996)
  • Department of Social Security, 1993. Households Below Average Income: a Statistical Analysis 1979–1990/1. HMSO,...
  • Der, G., Macintyre, S., Ford, G., Hunt, K., West, P., submitted for publication. The shape of the relationships between...
  • G.J Duncan

    Income dynamics and health

    Int. J. Health Serv.

    (1996)
  • J Eachus et al.

    Deprivation and cause-specific morbidity: evidence from the Somerset and Avon Survey of Health

    BMJ

    (1996)
  • R Ecob

    A multilevel modelling approach to examining the effects of area of residence on health and functioning

    J. R. Stat. Soc. Series A.

    (1996)
  • Ecob, R., Stassinopoulos, G., 1998. Working Paper No. 62: Using Semiparametric Models to Model the Relationship Between...
  • Cited by (0)

    View full text