Citation

Coles OB. Public Health 2005; 119(1): 71-72.

Affiliation

12 Mayorswell Close, Durham, DH1 1JU, UK.

Copyright

(Copyright © 2005, Elsevier Publishing)

DOI

10.1016/j.puhe.2004.10.002

PMID

15560905

Abstract

Original article published in Public Health 118 (5), 346-348, doi of original article:10.1016/j.puhe.2003.10.00)

This short communication prompts comment because its manner of presentation arguably makes it unduly difficult for the reader to obtain a clear overview of the scope of the research activity it draws upon. Even after allowing for the need for brevity, findings are presented in a way that leaves uncertainty as to the true scope and approach of the research, and suggests that findings of central relevance to the topic may have been excluded. This is despite the work having a

very clear stated aim, namely 'to analyse the relationship between suicides rates and economic development within the theoretical framework of epidemiological transition.' (p. 347) and its methods being characterised as exploratory. The few reported results are, arguably, only a fragment of those implied by this research brief. The core difficulty appears to be that of identifying the implications for quantitative analysis of the authors' theoretical assumptions.

It is unclear from the textwhat the precise theoretical implications of the epidemiological transition perspective are. The impression is given of broad generalisations about patterns and overall prevalence of physical illness and injury in three, or possibly, four historical eras. However there is no discussion of how the transitions between these affect patterns of mental illness, and still less of the broad range of factors that predict reported successful suicides. Not discussed either is how the epidemiological transition perspective relates to other relevant theoretical frameworks for the explanation of how economic development affects the incidence of suicide, and how the perspective complements or rejects elements of these. Examples of questions not explicitly addressed but arguably meriting attention are:

How does this model accommodate within any single phase of epidemiological transition divergences in national economic policies in such matters as economic and other inequalities, employment conditions, and unemployment rates?

How does it accommodate national variations in other policies relevant to the production or prevention of suicides? Examples are social policies. bearing on the determinants of mental health such as housing and education and health service policies.

Underlying this is the more fundamental question of whether the theoretical perspective is intended to generate testable, and thus falsifiable, hypotheses about associations between economic development and morbidity patterns. There is, further, the question of the operational criteria used to place a state in one of the three/four stages of economic development. It is unclear from the use ofmean per capital gross national product in the only results reported whether this is the authors' preferred indicator, or one chosen for want of access to data for unspecified superiormeasures. In any event, no series of 'income-brackets' for per capita GNP is given as a means of placing individual economies in one of the stages of economic development.

The authors write of their work that 'this type of study is the simplest and least expensive, and can serve an explorative purpose. It can also generate hypotheses that can later be tested in more rigourous efforts.' (p. 347). In the light of this a case can be made for their reporting far more than a pair of linear regression equations (for suicide rate by gender) with a single independent variable (mean per capita GNP). An exploratory quantitative approach suggests an examination of associations between a large numbers of predictors of suicide rates suggested by the literature. Examples include unemployment rates, income inequality, patterns of causes of death and morbidity, and measures of health care provision and access. A correlation matrix for such variables would have been a highly effective use of a very small investment in page space.

It arguably would also be valuable to report direct comparisons of suicide rates in aggregations of countries grouped, by some explicit criterion, as being at the same stage in the epidemiological transition. Reporting both the magnitude of contrasts in mean suicide incidence between the three or four epidemiological �¢??ages�¢??, and the extent of the range of mean values for individual countries within each group would provide the reader with an useful overview of the global transformation under discussion.