The geography of the highest mortality areas in Spain: a striking cluster in the southwestern region of the country

Abstract

Aims: To determine the areas of Spain with the most increased mortality risk. Methods: Age adjusted relative risk of death by gender was estimated in each of 2218 small areas of the country using a non-parametric empirical Bayes method. To determine areas with ''significantly increased risk'' a Poisson based score test was used. Results: Mapping of the highest risk areas showed a striking geographical clustering in the southwestern region of the country. This region, comprising 8% of the Spanish population, accounts for about one third (2884 deaths) of the total excess mortality. M ortality inequalities across geographical localities have been well documented in various developed countries.1 Our research in Spain has shown that mortality inequalities at the small area level are a widespread phenomenon.2 Here we report a striking observation in Spain that the areas with the highest increased risk of death are clustered in one specific region of the country. This presents a pressing public health concern with respect to the country's health policy and calls for attention from researchers and policy makers. METHODSTo assess the mortality risk by locality, we estimated age adjusted relative risk of death by gender in each of the 2218 small areas of Spain established in our previous research.3 To estimate the relative risk in each small area we obtained, by age and gender, all-cause mortality counts aggregated for a nine year period and the 1991 census population counts. Based on these data, age-gender specific mortality rates of Spain were estimated by a Poisson regression model with the generalised estimating equation (GEE) modification to account for within-area correlation of mortality across ages and genders.4 These age-gender specific mortality rates of Spain were used as ''reference rates'' for calculating expected counts of death by gender in each small area. The relative risk in each small area was estimated by a non-parametric empirical Bayes method for a standardised mortality ratio (SMR), the ratio of observed versus expected counts of death.5 6 The comparison of SMRs across small areas assumes a uniform rate ratio across age-sex strata within each small area. We used the SMR approach under this required assumption, instead of the direct standardisation approach, because stratum specific rates were unstable in our small areas and the application of the direct standardisation was untenable.To determine small areas with ''significantly increased risk'' we tested the null hypothesis of the age adjusted relative risk equalling 1.0 by a Poisson based score test that takes into account the estimated amount of overdispersion from the GEE model. Let O and E be the observed and expected deaths in each small area and w be the estimate of the overdispersion parameter based on the Poisson assumption. The significance test statistic is defined as: T = (O 2 E) 2 /wE, which is compared to a x 2 distribution with one degree of freedom. The small area was considered as a significantly increased risk area if the departure...

Bringing the individual back to small-area variation studies: A multilevel analysis of all-cause mortality in Andalusia, Spain

“…Our conclusion critically questions previous ecological/spatial studies of small-area variation published hitherto in Spain (Benach et al, 2003;Benach et al, 2004;Borrell et al, 2010;Collaboration, 2010;Ocana-Riola & Mayoral-Cortes, 2010;Ocana-Riola et al, 2008a). These ecological/spatial studies of small-area variation are only justified if we implicitly assume that the geographical variation between municipalities and between census tracts represents an important factor for understanding individual inequalities in mortality risk.…”

“…However, many multilevel analyses performed so far have been mainly focused on the study of associations between contextual variables and individual health, considering the analysis of variance as secondary information (Blakely & Woodward, 2000;Diez Roux, 2008). In contrast, others scholars have explicitly concluded that the analysis of variance provides indispensable information for understanding place effects on health (Boyle MH & Willms JD, 1999;Clarke P & Wheaton B, 2007;Duncan et al, 1993;Merlo, 2003;Merlo et al, 2004;Merlo et al, 2009;Riva et al, 2007) Moreover, all around the world, a persistent amount of observational information on place effects is still being obtained from ecological/spatial studies of "small-area variations" , frequently in the form of coloured atlases and disease maps (Benach et al, 2003;Benach et al, 2004;Borrell et al, 2010;Collaboration, 2010;MacNab & Dean, 2002;Middleton et al, 2008;Ocana-Riola & Mayoral-Cortes, 2010;Ocana-Riola et al, 2008a;Pickle et al, 1999;Shaw, 2008;Turrell & Mengersen, 2000). From an empirical perspective, the advantages of multilevel versus ecological regression analyses were clearly identified by the seminal work performed by Aitkins and Longford (Aitkin M & Longford N, 1986) as well as by Jones,4 Duncan, Moon, Subramanian and colleagues (Bullen et al, 1996;Duncan et al, 1993Duncan et al, , 1995Duncan et al, , 1996Duncan et al, , 1998Duncan et al, , 1999Jones et al, 1991;Subramanian et al, 2009;Twigg et al, 2000).…”

Geographical mortality patterns in Italy

“…This can be useful in enabling us to better plan and manage the resources of a public health system (for an in-depth review on disease mapping methods and applications see Marshall 1991aand Lawson et al 1999. In recent years, there has also been considerable interest in these kinds of studies from a demographic perspective, and a great deal of effort has been put into developing and applying statistical methods which take into account mortality information about small geographical areas (see for example Benach et al 2003Benach et al , 2004Caselli and Lipsi 2006).…”

“…This can be useful in enabling us to better plan and manage the resources of a public health system (for an in-depth review on disease mapping methods and applications see Marshall 1991aand Lawson et al 1999. In recent years, there has also been considerable interest in these kinds of studies from a demographic perspective, and a great deal of effort has been put into developing and applying statistical methods which take into account mortality information about small geographical areas (see for example Benach et al 2003Benach et al , 2004Caselli and Lipsi 2006).Within this framework, statistical results become more or less pertinent depending on the degree to which the areas display homogenous environmental, demographic, social, and economic characteristics, as these characteristics may influence the phenomenon of interest. In this way, the average value of the mortality incidence observed with respect to any area effectively represents the risk that the majority of the corresponding population is exposed to.…”

Distribución y tendencia de los fallecimientos en el medio hospitalario en España durante el periodo 1997-2003

“…Por otra parte, estudios previos realizados en España han descrito importantes desigualdades socioeconómicas y en la oferta de servicios sanitarios entre las distintas comunidades autónomas (CCAA) [18][19] , así como la influencia de diversas característi-cas sociodemográficas sobre el riesgo de muerte, como el desempleo, el nivel educativo, el estado civil, el número de miembros del hogar, la actividad laboral o el apoyo social 18,[20][21][22][23][24] . Nuestra hipótesis era que las muertes en hospitales serían más frecuentes en las CCAA con menos población anciana y rural, y que la proporción de muertes en urgencias se relacionaría directamente con la frecuentación hospitalaria e inversamente con las dotaciones de camas.…”