Perspective on health inequalities
In their introduction the authors hypothesise that there is an association between poverty and Tuberculosis (TB) in Recife, Brazil, and that this is mediated by accommodation crowding, malnutrition and “other socially determined factors”. However this paper does not include a nutrition variable, and the authors don’t clarify which other social factors they believe are important. They examine numerous possible socioeconomic position (SEP) variables without defining a priori which is of primary interest. They are clear that they believe health inequalities stem from social inequalities at both individual and group levels, and thus we must assume they are considering poverty in relative, rather than absolute terms.
The authors were able to show at an area level that a combined measure of computer ownership and literacy captured the effect of all other SEP variables. This model is based on statistical significance in multivariate analysis, and does not appear to have theoretical underpinnings. However, in the discussion the authors state that this shows “education… is related to all other aspects of poverty”. Given this explicit assumption it is unclear why other education measures were not considered a priori.
It is important to note that the paper does not answer the question posed in the title. It shows that being of low SEP (by the exposure measures discussed), is associated with having TB. To answer the question about effects of relative deprivation, rates of TB in various areas need to be compared, with each census area graded in terms of poverty. This measure of area poverty should then be included in multivariate analysis controlling for individual-level SEP. This analysis could then provide an estimate of TB rates for high SEP individuals in deprived areas and vice versa. The title and introduction, therefore, give a misleading sense of having addressed aspects of inequalities, which are not fully achieved in the paper.
The major constructs explaining health inequalities can be grouped into four types: material, cultural–behavioural, psychosocial and life course
. At an individual level, the author’s use of an asset index, access to water, literacy and short-term work suggests a materialist interpretation of health inequalities - that access to assets and services shapes pathways to health outcomes . In reality these constructs are not mutually exclusive and it’s impossible to disentangle their effects. For example, all material resources relevant to quality of life have some associated psychosocial and cultural-behavioural meaning .
The authors suggest that their approach is ‘eco-epidemiological’, and certainly the use of a multi-level model facilitates recognition that individuals behave in context and that characteristics may cluster in areas. This paradigm then allows analysis across and between levels . In this case the...