Developing a Small-Area Deprivation Measure for Brazil

Abstract

This report describes the development of the BrazDep small-area deprivation measure for the whole of Brazil. The measure uses the 2010 Brazilian Population Census data and is calculated for the smallest possible geographical area level, the census sectors. It combines three variables – (1) percent of households with per capita income ≤ 1/2 minimum wage; (2) percent of people not literate, aged 7+; and (3) average of percent of people with inadequate access to sewage, water, garbage collection and no toilet and bath/shower – into a single measure. Similar measures have previously been developed at the census sector level for some states or municipalities, but the deprivation measure described in this report is the first one to be provided for census sectors for the whole of Brazil. BrazDep is a measure of relative deprivation, placing the census sectors on a scale of material well-being from the least to the most deprived. It is useful in comparing areas within Brazil in 2010, but cannot be used to make comparisons across countries or time. Categorical versions of the measure are also provided, placing census sectors into groups of similar levels of deprivation. Deprivation measures, such as the one developed here, have been developed for many countries and are popular tools in public health research for describing the social patterning of health outcomes and supporting the targeting and delivery of services to areas of higher need. The deprivation measure is exponentially distributed, with a large proportion of areas having a low deprivation score and a smaller number of areas experiencing very high deprivation. There is significant regional variation in deprivation; areas in the North and Northeast of Brazil have on average much higher deprivation compared to the South and Southeast. Deprivation levels in the Central-West region fall between those for the North and South. Differences are also great between urban and rural areas, with the former having lower levels of deprivation compared to the latter. The measure was validated by comparing it to other similar indices measuring health and social vulnerability at the census sector level in states and municipalities where it was possible, and at the municipal level for across the whole of Brazil. At the municipal level the deprivation measure was also compared to health outcomes. The different validation exercises showed that the developed measure produced expected results and could be considered validated. As the measure is an estimate of the “true” deprivation in Brazil, uncertainty exists about the exact level of deprivation for all of the areas. For the majority of census sectors the uncertainty is small enough that we can reliably place the area into a deprivation category. However, for some areas uncertainty is very high and the provided estimate is unreliable. These considerations should always be kept in mind when using the BrazDep measure in research or policy. The measure should be used as part of a toolkit, rather than a single basis for decision-making. The data together with documentation is available from the University of Glasgow http: //dx.doi.org/10.5525/gla.researchdata.980. The data and this report are distributed under Creative Commons Share-Alike license (CC BY-SA 4.0) and can be freely used by researchers, policy makers or members of public.