Common errors in disease mapping

Abstract

Abstract. Many morbid-mortality atlases and small-area studies have been carried out over the last decade. However, the methods used to draw up such research, the interpretation of results and the conclusions published are often inaccurate. Often, the proliferation of this practice has led to inefficient decision-making, implementation of inappropriate health policies and negative impact on the advancement of scientific knowledge. This paper reviews the most frequent errors in the design, analysis and interpretation of small-area epidemiological studies and proposes a diagnostic evaluation test that should enable the scientific quality of published papers to be ascertained. Nine common mistakes in disease mapping methods are discussed. From this framework, and following the theory of diagnostic evaluation, a standardised test to evaluate the scientific quality of a small-area epidemiology study has been developed. Optimal quality is achieved with the maximum score (16 points), average with a score between 8 and 15 points, and low with a score of 7 or below. A systematic evaluation of scientific papers, together with an enhanced quality in future research, will contribute towards increased efficacy in epidemiological surveillance and in health planning based on the spatio-temporal analysis of ecological information.

Disease mapping and spatio-temporal analysis: importance of expected-case computation criteria

“…Historically speaking, the use of the SMR has a long tradition and is the indicator most frequently used for producing disease atlases and undertaking small-area studies. It is also an indicator that is highly criticised due to the intrinsic difficulty of comparison among SMRs, since the weighting of each SMR depends on the age structure of the study population (Julious et al, 2001;Rothman et al, 2008;Ocaña-Riola, 2010). Nevertheless it continues to be used in practice.…”

Geospatial Health: the first five years

“…4 reveals that the publication record in the first five volumes of Geospatial Health covers all parts of the world. We identified seven contributions (6.0%) that had a global focus, five of which were reviews (Rinaldi et al, 2006;Martin et al, 2007;Bernardi, 2008;Ocaña-Riola, 2010;Machault et al, 2011) and the remaining two original research articles (Si et al, 2009;Catelan and Biggeri, 2010). Approximately two thirds of the research published so far had a geographic focus on Africa, Asia and Europe, with similar shares for each of these three continents.…”

“…1). The second most important publication type (n = 10, 7.5%) were the reviews (Fuentes, 2006;Rinaldi et al, 2006;Gazzinelli and Kloss, 2007;Leonardo et al, 2007;Martin et al, 2007;Bernardi, 2008;Marechal et al, 2008;Stensgaard et al, 2009;Ocaña-Riola, 2010;Machault et al, 2011). So far, four editorials have been published, two invited guest editorials in the inaugural issue (Cline, 2006;Crump, 2006), and two unsolicited editorials from journal board members (Brooker and Utzinger, 2007) and other contributors (Seifter et al, 2010).…”

“…The distribution is slightly skewed with a peak of five authors per paper and a median of six. There are 10 single-authored pieces, half of them editorials and reviews (Cline, 2006;Crump, 2006;Fuentes, 2006;Bernardi, 2008;Ocaña-Riola, 2010), and the other half presenting original research (Aron, 2006;Lwasa, 2007;Murad, 2008;Estrada-Peña, 2009;Livingston, 2010). Eight papers, all of them original research, where co-authored by at least 10 and up to 15 individuals, interestingly often stemming from P.R.…”

Spatio-temporal dependencies between hospital beds, physicians and health expenditure using visual variables and data classification in statistical table

“…The signifi cance of cartographic visualisation in medicine relates to the presentation as a sui generis document concerning the spatial scope of diseases (Schweikart and Kisteman, 2013;Ocaña-Riola, 2010), or parts of reports elaborated for specialists, e.g. incidences of malignant neoplasms.…”