The need for,and value of,a spatial scan statistical tool for tackling social health inequalities

Background

Despite reports documenting the links between social determinants and health, disparities remain a major public health issue (^1–3). Today, the European Union, supported by the World Health Organization (WHO), recognizes that it is time to move from research about risk factors of health disparities to actions which aim to reduce them. Several countries, mainly the UK government, have put ‘tackling health inequalities’ at the heart of their health agendas and issued a number of policy documents and related targets ( ⁴ ). For instance, reducing the gap in life expectancy between the most and least advantaged remains a key priority for the Northern Ireland Executive and is reflected in its Program for Government ( ⁴ ).

In the majority of epidemiological research investigating health inequalities, regression analysis is used to measure the strength of the association between variables and outcomes. These studies control for covariates such as socioeconomic status; however, they do not investigate how the spatial distribution of risk changes due to the adjustment. Today, a growing number of health inequalities studies base their analyses on spatial or geographical approaches (^5–8). Yet none of them have discussed or suggested the added value of spatial approaches to tackle social inequalities in health. Among these approaches, we believe that the Kulldorff scan method, which is based on spatial scan statistics (^9,10), could be a useful tool to tackle social inequalities in health by guiding policy makers. Several steps are required to promote the interventions that would maximize health benefits among the general population, reduce health inequalities among disadvantaged or vulnerable groups who are not equally distributed over a spatial area, and respond to life-threatening situations.

More precisely, the geospatial approach can:

Geospatial approach and priority setting to tackle social inequalities in health

Using the spatial scan statistic implemented in the SaTScan free software ( ¹⁴ ), the application of spatial statistics to health outcomes is now being widely used to provide novel ways of investigating disease patterns. More precisely, it aims to determine and visually represent the spatial distribution of the relative risks using cluster analysis (^9,10,15).

When applying the scan statistical model in 1995, Kulldorff and Nagarwalla found that the most likely breast cancer cluster was located in a region encompassing the New York City–Philadelphia metropolitan area ( ¹⁰ ), with a mortality rate 7.4 percent higher than in the rest of the Northeast. In Massachusetts, Sheehan et al. in 2004 identified several geographic areas with invasive breast cancer incidence higher than expected ( ¹⁶ ). The most statistically significant area of excess was found in the west of Boston counting 15% more cases than would have been expected plus two other high areas in the east of Boston ( ¹⁶ ). In 2011, Dahly and Gilthorpe detected clusters of overweight and obesity located in urban areas ( ¹⁷ ), but extended into peri-urban and rural neighborhoods. More recently, in France, Kihal-Talantikite et al. identified the locations of clusters of high risk for several health outcomes including infant mortality and end-stage renal disease (ESRD) at different regional levels (^18–20). These examples illustrate the usefulness of the geospatial tool to support policy makers in planning more focused community interventions in appropriate areas and to choose if public health interventions should be implemented either at a national level, at a local level, or both (^21,22).

Geospatial approach and appropriate local prevention/intervention programs

Spatial approaches appear to be adequate to examine the spatial implications of neighborhood characteristics which may affect health inequalities ( ¹⁵ ). For instance, Dahly and Gilthorpe in 2011 concluded that clusters of high risk of overweight and obesity in males, but not females, were explained by the spatial distribution of socioeconomic status ( ¹⁷ ). More recently, studies revealed that spatial distribution of infant mortality risk in the Lyon metropolitan area may be explained by deprivation level, noise exposure level or green space level (^19,20). These same authors suggested that spatial analysis of ESRD incidence indicated that beyond age and sex, neighborhood characteristics, including socioeconomic deprivation index, healthcare supply, and rural/urban typology, partially explained the spatial distribution of ESRD incidence ( ¹⁸ ).

Demonstrating differences in the socio-demographic factors underlying the identified clusters can assist policy makers in understanding the spatial epidemiology of diseases, risk factors or health issues. It can also aid in the development and implementation of interventions as well as the reorientation of health services.

Therefore, initially targeting only the neighborhood determinants which explain the spatial distribution of air pollution or spatial clusters with risk factors may be a cost-effective approach to improve health. For example, the SaTScan methodology has been used in Baltimore to identify clusters of hypoplastic left heart and investigate genetic and environmental factors contributing to hypoplastic left heart ( ²³ ), in China to plan regional tuberculosis programs, after identifying clusters of high incidence of tuberculosis ( ²⁴ ), and to design antimalarial interventions at the household level ( ²⁵ ).

This geospatial tool, combined with GIS (Geographic Information System), allows us to extract local knowledge and understanding of the neighborhood characteristics of the identified geographical zone, including:

Local diagnoses can assist policy makers to focus the scope of prevention/intervention programs and changes to the health care system, thus providing more effective interventions in order to meet to individual needs, and public resources can be distributed more efficiently. Therefore, this spatial tool may assist policy makers in their efforts to tackle the social gradient in health if they choose to apply the strategy of ‘proportionate universalism’ described by Marmot in 2010 ( ¹³ ).

Geospatial approach and health equity of public policies

To date, one of the biggest challenges is to place health equity at the heart of all policies and to assess the health equity effects of strategies and policies. Among studies which investigated health benefits of policy interventions such as air quality management (^26–30), sophisticated statistical analyses were used to quantify the days of lifetime gained for each individual over 2 years ( ²⁶ ), and the decrease in the cardiovascular death rate ( ²⁷ ). However, none of these studies was interested in quantifying the health gains across different socioeconomic groups. More recently, in London, Tonne et al. in 2008 suggested that the Congestion Charging Scheme ( ²⁸ ), a localized scheme targeting traffic congestion, could lead to the greatest reductions in air pollution in the more deprived areas and was likely to reduce socioeconomic inequalities in air pollution impacts. However, no clear analytical approach to explore the spatial distribution of gain was applied. In a systematic review addressing equity in air quality interventions, the authors emphasized the methodological and theoretical challenges when assessing equity in interventions to reduce air pollution and suggested the need to develop this area of research ( ³¹ ).

To our knowledge, no study proposes a spatial scan tool to explore the spatial distribution of health benefits resulting from public health policies. In this context, we propose to repurpose this spatial tool from its original use to analyze the spatial distribution of the days of lifetime gained for each individual or years of life gained (YLG) modeled by a quantitative risk assessment approach ( ³² ) and thus, to evaluate the benefits to health and wellbeing of health action. This spatial tool may guide and support policy makers to assess, before developing and implementing interventions to tackle health inequalities, whether the benefits of these local policies are equally distributed according to social conditions ( ³³ ).

Conclusion

This paper argues that a spatial tool could be an appropriate tool to investigate social inequalities in health by visually representing the spatial trends of these social inequities in health and the distribution of their determinants at a fine scale. In addition, this paper introduced various pathways by which Kulldorff’s scan method may help and guide policy makers to tackle social inequalities in health. Two main aspects have been highlighted: the need for information to support equity-oriented policies and the monitoring of equity aspects of various policies in health and other related sectors.