| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Editor's Choice |
Entelechy, citation indexes, and the association of ideas
Do you know what ‘entelechy’ means? It was in the title of a manuscript IJE received and it got me wondering what it meant and why the authors had used such an obscure word in their title. Indeed, one of our chosen reviewers (an eminent figure in gerontology) refused to read the manuscript as he did not understand the title. My more curious response was to look it up in the online Oxford Dictionary where all was revealed. Aristotle was the first to use the term as ‘the realization or complete expression of some function’, derived from the Greek meaning ‘to have perfection’. Catalano and Bruckner justify their use of entelechy as the most parsimonious means of communicating the idea of the effect of early-life adversities leading to failure of individuals or populations to achieve their expected lifespan.1 Using cohort life table data from Sweden, Denmark, and England and Wales, they find support for the diminished entelechy hypothesis as the life expectancy of cohorts experiencing higher mortality below the age of 5 years was lower than expected. The effect seemed to be much weaker in women suggesting that they are more resilient than men to adversity, a point taken up by our commentators.2 What does this add to our growing understanding of the early-life determinants of adult health? Will these relatively strong but non-specific signals marked by child mortality on late-life mortality make the detection of specific early-life exposures (e.g. diet, infections) operating at sensitive periods cumulatively more difficult given these pervasive general effects and the likely correlation between them and specific exposures?
The impact factor for IJE has risen yet again and is now 4.045, double what it was when Davey Smith and I took over editorship of the journal. Is this a coincidence? We do not think so, but is it a marker of success? Our historic reprint is Eugene Garfield's 1955 Science paper in which he first describes the rationale for developing a citation index and uses the term ‘impact factor’ for the first time, but, in reference to individual articles and not averaged over the output of a journal.3 As Richard Smith, former editor of the British Medical Journal, points out, Garfield's main perceived need for citation indexing was to avoid the ‘uncritical citation of fraudulent, incomplete or obsolete data’, but this problem remains despite the considerable power of the Web of Science to track citations. Garfield himself commenting here, also notes most scientists have yet to develop ‘citation consciousness’.4 The law of unintended consequences might be invoked to explain the alacrity with which the impact factor has been seized up as a means of providing a number that can be used to ‘value’ the quality of a journal (and by extension, the quality of articles appearing in it). Porta et al.5 provide insight into the numerator denominator bias involved in calculating the impact factor, the covert way in which these are defined journal by journal without any explicit criteria, and highlight the naïvety of using an arithmetic mean (to three decimal places) to compare highly skewed distributions of different journals (Figure 1). They conclude that it just is not scientific! So should we be happy with our rising impact factor? Porta et al.5 suggest that total citations would be more relevant—but quarterly and bi-monthly journals will clearly publish fewer citable papers than weeklies. The sad fact is that a majority of papers published in medical journals are never cited at all, but whether we should read much into International Journal of Epidemiology's rather better zero citation rate (about a quarter) compared with Lancet (just under half) and American Journal of Epidemiology (just over half) is debatable (Table 1). We consider our main editorial function is to make the International Journal of Epidemiology a journal that people want to pick up, dip into, and find at least one or two things that look interesting. Are we succeeding? Could we do better?
|
|
A recent ‘citation classic’ of the International Journal of Epidemiology's—anything with more than 50 cites is a classic for us—was a paper on sample sizes for cluster randomized trials.6 In this issue, Eldridge et al.7 tackle a very common problem for trialists doing cluster randomized trials—unequal cluster sizes that can lead to reduced power. They provide a simple formula for estimating sample size requirements, focusing on primary health care requirements although the methods have general applicability. Hopefully this will prove to be useful for trialists and give us another citation classic.
One of Garfield's purposes for citations was to provide an ‘association of ideas’ index, albeit not a substitute for extensive reading. The current obsession with ‘interdisciplinarity’ in research, particularly among funding agencies who want to shape the research landscape is all about the association of ideas. John Lynch gives his views on interdisciplinarity in an editorial, mapping its rise in bibliographic analysis and highlighting the difficulties in implementing interdisciplinary approaches.8 Epidemiology would be almost unthinkable as a discipline without the involvement and depth of understanding provided by social scientists, biologists, geneticists, ecologists, anthropologists, and psychologists. While we are able to make quite a lot from the associations of ideas between our near neighbours, achieving mutual colonization in the form of sub-disciplines—genetic epidemiology, social epidemiology, behavioural epidemiology, environmental epidemiology—how are we doing in terms of ideas derived from more distant areas of science or even between the sub-species of epidemiology? This is where one might expect really novel and ground-breaking approaches to problems in population health to arise.
At a more practical level, bibliographic databases provide a means of assimilating common threads of ideas and data. Systematic reviews of aetiological and intervention effects would be impossible to conduct without them. Most epidemiologists, if challenged on the effect of infant size on adult disease would confidently cite David Barker's early life origins of adult disease hypothesis, and conclude that larger size is associated with better outcomes. Fisher et al.9 have pulled together the evidence on the effects of infant size and growth on adult disease burden using rigorous methodology and demonstrate that ‘no single optimum pattern of infant growth was associated with beneficial health outcomes’. Surprised? I confess I was. The reduced risk of ischaemic heart disease in larger infants was confirmed in this review but larger size was associated with increased risk of insulin-dependent diabetes. They also demonstrated that the associations between infant size and growth with other major causes of global burden of disease remain under-researched.
Paulo Vineis and Marriane Berwick provide us with ideas on cancer aetiology drawn from evolutionary theory and cell biology and come up with a ‘Darwinian’ model of carcinogenesis considering a population of cells rather than organisms as the object of natural selection.10 This idea predicts that avoiding exposure to mutagens would be insufficient to prevent cancers as some mutations may be present at birth and certain environmental conditions may result in selective advantage for these cells. Moreover, they argue that simple linear analysis is inadequate and more complex non-linear models that better reflect the underlying biological processes are needed. Breivik expands on these ideas suggesting that carcinogenesis is an evolutionary process and helpfully distinguishes selection of genes through the germline for their control, coordination, and stability effects, whereas for somatic cells, genetic information is selected for its ability to promote cell proliferation—cancer arises as a breakdown of germline encoded control mechanisms.11 Weiss, in a commentary on this work, notes that our educational system is failing us in producing scientists ‘...with a short memory and too much technical specialization, that drive out a more nuanced biology of which evolution is a vital part.’12
Perhaps our continuing professional development as epidemiologists would be enhanced if we took Garfield's advice on being aware of citations to our own work: ‘...an author could readily determine which other scientists were making reference to his work, thus increasing communication possibilities...the individual scientist might become aware of the implications in his studies that he was not aware of before.’3 At a recent meeting in India, the importance of communication between scientists was emphasized thus: I have an idea and you have an idea. We talk and discuss our ideas. We both go away with two ideas where before we each had only one idea. Dip into this issue and I can guarantee that you will be rewarded with plenty of new ideas!
References
1 Catalano R, Bruckner T. Child mortality and cohort lifespan: a test of diminished entelechy. Int J Epidemiol 2006;35:1264–69.
2 Crimmins EM, Finch CE. Commentary: Do older men and women gain equally from improving childhood conditions? Int J Epidemiol 2006;35:1270–71.
3 Garfield E. Citation indexes for science. a new dimension in documentation through association of ideas. Science 1955;122:108–111. (Reprinted Int J Epidemiol 2006;35:1127–28).
4 Smith R. Commentary: The power of the unrelenting impact factor Int J Epidemiol 2006;35:1129–30.
5 Porta M, Fernandez E, Bolúmar F. Commentary: The "bibliographic impact factor" and the still uncharted sociology of epidemiology. Int J Epidemiol 2006;35:1130–35.
6 Hayes RJ, Bennett S. Simple sample size calculation for cluster-randomized trials. Int J Epidemiol 1999;28:319–26.
7 Eldridge SM, Ashby D, Kerry S. Sample size for cluster rqandomized trials: effect of coefficient of variation of cluster size and analysis method. Int J Epidemiol 2006;35:1292–1300.
8 Lynch J. Editorial: Its not easy being interdisciplinary. Int J Epidemiol 2006;35:1119–22.
9 Fisher D, Baird J, Payne L et al. Are infant size and growth related to burden of disease in adulthood? A systematic review of literature. Int J Epidemiol 2006;35:1196–1210.
10 Vineis P, Berwick M. The population dynamics of cancer: a Darwinian perspective. Int J Epidemiol 2006;35:1151–59.
11 Breivik J. Commentary: Cancer – evolution within. Int J Epidemiol 2006;35:1161–62.
12 Weiss KM. Commentary: Evolution in action in cells and organisms. Int J Epidemiol 2006;35:1159–60.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||