Income inequality and elderly self-rated health in São Paulo, Brazil

Abstract

Purpose

To test the association between income inequality and elderly self-rated health and to propose a pathway to explain the relationship.

Methods

We analyzed a sample of 2143 older individuals (60 years of age and over) from 49 distritos of the Municipality of São Paulo, Brazil. Bayesian multilevel logistic models were performed with poor self-rated health as the outcome variable.

Results

Income inequality (measured by the Gini coefficient) was found to be associated with poor self-rated health after controlling for age, sex, income and education (odds ratio, 1.19; 95% credible interval, 1.01–1.38). When the practice of physical exercise and homicide rate were added to the model, the Gini coefficient lost its statistical significance (P > .05). We fitted a structural equation model in which income inequality affects elderly health by a pathway mediated by violence and practice of physical exercise.

Conclusions

The health of older individuals may be highly susceptible to the socioeconomic environment of residence, specifically to the local distribution of income. We propose that this association may be mediated by fear of violence and lack of physical activity.

Introduction

The relationship between income inequality and population health has been the subject of a number of studies. A recent meta-analysis of 23 multilevel studies found that a 0.05-unit increase in the Gini coefficient of income inequality was associated with a 7.8% excess risk of mortality (95% confidence interval, 5.9%–9.8%) [1]. A follow-up analysis from the same authors concluded that the authors of country-level studies found a stronger association between income inequality and health than those in which authors analyzed smaller geographic aggregates (such as neighborhoods) [2]. A similar result was found by Wilkinson and Pickett [3]. In another review, Subramanian and Kawachi [4] concluded that the null results found for smaller areas may stem from the fact that most of these studies were from areas with low income inequality, mostly outside the United States. Lynch et al [5] have stretched this argument further by suggesting that the association between income inequality and health found in the literature could be the result of “American exceptionalism,” more specifically in relation to state-level differences in education and race relations.

The objective of the present study was to test the association between area-level income inequality and individual health by analyzing a sample of Brazilian older adults. Our choice of population advances the literature in two ways. First, our sample comes from a highly unequal area situated outside the United States (for which the results have been mixed). Second, our sample focuses on an age group for which literature has been hitherto mostly unsupportive of the income inequality theory. In an analysis of U.S. and Canadian metropolitan areas, Ross et al [6] found an association between overall mortality and income inequality for all age groups except for the elderly (65 years and older). A similar result was found by Backlund et al [7], leading them to conclude that “income inequality is not a major driver of mortality trends in the United States because most deaths occur at ages 65 and over.”

A focus on older segments of the population is of particular interest because of their greater susceptibility to features of their social environment. Young adults may also be highly influenced by their surroundings, but their capability to adapt quicker to adverse conditions might mitigate the influence of environmental forces. In a review, Yen et al [8] concluded that elderly health is influenced by a range of area-level characteristics such as socioeconomic status, demographics, racial composition, perceived environmental resources, and accessibility to physical activity.

The association between income inequality and violence is among the strongest and most established in the literature [9], [10]. An increase in local violence can also lead to a disproportional increase in fear of crime, particularly for elderly persons because they feel they are not able to protect themselves [11]. Perceived safety has been found to influence the mobility status of the elderly [12], which could have an important effect in their overall health [13].

We propose here that for the relative income theory to hold as an epidemiologic fact there is no satisfying reason for it not work for elderly persons. In fact, its effects should be larger in more vulnerable populations. Our aim in this work was to first test the theory for the sample. Given the results, and taking into consideration previous studies regarding elderly health, we then aimed to propose a causal pathway by which income inequality could affect elderly health.