Spatialized composite indices to evaluate environmental health inequalities: Meeting the challenge of selecting relevant variables

Highlights

• Composite indices contain many variables and selection methods for specific needs.

• The absence of a common framework leads to strong subjectivity.

• It is possible to harmonize dimension characterization practices.

• A framework is provided for selection of relevant variables for composite indices.

• An increase in methodological transparency will improve interterritorial comparison.

Abstract

The wide range of factors involved in environmental health and the complexity of interactions between all environmental determinants require the validation of multidimensional approaches. While the development of composite indices is receiving growing attention by scientists and public authorities, the concept continues to lack transposability and robustness partly due to varying conceptualizations and/or methodologies.

This review aims to promote harmonizing practices governing the first step of development of composite index, namely identification and characterization of the dimensions and variables that are included in environmental health indices. A review of available literature (more than 1500 studies) was conducted to identify the composite indices developed to assess territorial determinants from an environmental health perspective. This process made it possible to identify 23 spatialized composite indices and to assess a total of 329 variables. This diversity highlights that the absence of a common framework can lead to a strong subjectivity and limit comparisons between different environmental health indices. The specificity and the availability of certain variables would limit the transposability of indices.

In light of current knowledge, this review proposes a consolidated methodological framework based on a categorization of variables into dimensions and sub-dimensions related to heath, environment, social, economics, services and policy. To characterize the sub-dimensions, several variables are possible and can be chosen according to the availability and/or accessibility of the data. The adaptation of a composite index to a specific territory or to a specific issue would then be effective through the included variables. This also aims to be transposable to any spatial unit (country, region, census tract).

This work is a first step towards a proposal of guidelines designed to provide a consensual framework that could facilitate the exploitation of environmental health indices. This transparency could also increase the understanding and adoption of these tools by public authorities and general public.

Introduction

The WHO first tackled the environmental health issue in 1994 by defining the concept as follows: “Environmental health comprises those aspects of human health, including quality of life, that are determined by physical, chemical, biological, social, psychosocial and aesthetic factors in the environment”. In its definition, WHO considers any health problems that are not caused by genetic factors or due to individual choices. The idea of the environment therefore reflects the different neighborhoods of populations and excludes the effect of individual behaviour and voluntary exposition to factors such as alcohol consumption, smoking or unhealthy dietary habits. Although these behavioural and individual factors are socially distributed, many studies suggest that other factors could also contribute to explaining observed health inequalities (Saib, 2015). The health status of populations is thus strongly influenced by complex interactions between all environmental determinants, which are interconnected and evolve over time and space (Sarkar and Webster, 2017). The environment, considered in this context to be the outdoor living environment, is characterised by a number of determinants such as environmental quality (water, air and ground) and living conditions (e.g. surroundings, access to healthcare) and by socio-economic parameters such as local economic drive, social disadvantage and education. These different environmental determinants can influence the health status of the population in a combined or sequential way. The environment is currently the subject of increasing apprehension and concern (Scarwell et al., 2013), with a particular emphasis on the possible impacts of environmental quality on health. These consequences are difficult to characterize due to the many aspects that can be harmful or beneficial.

The wide range of factors involved in environmental health, the complexity of interactions and their spatio-temporal variabilities all require the validation of multidimensional approaches that would provide more than a simple aggregation of health determinants. One of the recognized methods to explain a complex research problem and follow the evolution of a given phenomenon is based on the creation of composite indices. A composite index is a mathematical combination of variables reflecting one or more selected dimensions that are usually evaluated separatly (Nardo et al. 2005, Nascimento and Carrage 2007). It derived from quantitative or qualitative measurements obtained from field observations (Freudenberg, 2003). In light of the wide range of methodological approaches, the OECD has developed a guidance manual for evaluators, non-specific to environmental health issues (European Commission et al., 2008). This manual confirms the interest of grouping together different variables that measure the multiple facets of the phenomenon and provide a picture of it. A spatial approach, that is the calculation of the composite index for each spatial unit, can be added to provide a further level of information (Beale et al., 2008, Flacke 2015). A wider availability of spatialized data for the environment, health and population has led to an increase in the number of spatialized epidemiological studies. Composite indices are increasingly recognized as a useful measurement tool for not only etiological approaches but also strategy and policy development, and public communication by institutions (Cutter et al. 2010, Saib et al. 2015, Saisana and Cartwright 2007).

Although an increasing number of studies in the literature show interest in composite indices and report on the growing demand for environmental health variables, a number of questions remain to be answered, particularly with respect to methodological issues (WHO, 1999). Informative reviews have been published on specific aspects of methodological constructions and concentrate on mathematical aspects. The findings of these reviews provide guidelines for future environmental composite indices, mainly in terms of the overall uncertainties linked to each step of the methodology. Many focused on weighting and/or aggregation (Becker et al. 2017, Gan et al. 2017, Greco et al. 2018, Wiréhn et al. 2015). Becker et al. (2017) highlighted the complexity of including weighting in composite indices and presented tools to help developers investigate the effects of weights. The main aim of these works was to propose a methodological framework or a means to evaluate the qualitative aspects of the methodology used to develop composite indices. However, no studies to date have considered the potential inclusion of data, as a carrier of information, for use in composite environmental health indices. There is no recognized standard or internationally recognized rule to determine the number and type of variables to be included in the quantification of environmental health (He et al., 2018). The design of composite indices in environmental health has to be optimized and induced variability must be mastered (Burgass et al., 2017).

To this end, this study analyses the literature to identify and characterize the dimensions and variables that are included in environmental health indices. Our analysis will highlight (i) the diversity of the published composite indices, but also (ii) the diversity of variables used in these composite indices. For this purpose, a literature review on the 23 main papers dealing with composite indices in environmental health was conducted. This work is a first step towards a proposal of guidelines designed to provide a consensual framework that facilitates the exploitation of environmental health indices. Their transposability and comparability regardless of the spatial scale were considered as mandatory criteria. The availability or accessibility of data have also been given a special attention.