Editor's Choice |
The future of modern epidemiology: genetics, methods, and history
E-mail: Shah.ebrahim{at}lshtm.ac.uk
‘It's all genetic’. A confusing aspect of genetics is the concept of heritability. Obesity is highly heritable, as shown by its higher concordance among identical compared with non-identical twins. But obesity appears to be environmentally determined as its rising prevalence indicates. In this issue we reprint Lewontin's1 article, which provides an important discussion on causation in human genetics. He distinguishes two categories of causal question: first, that of discrimination between two alternative and mutually exclusive causes that underlie an observed phenotype—that is distinguishing major gene effects from polygenic effects. The second type of question is concerned with analysing the interaction between environmental and genetic components in determining a phenotype. His sketch of the purpose of genetic analysis should be helpful to the current generation of genetic epidemiologists in defining their public health role—’Together with a knowledge of the relative frequencies of different human genotypes, a knowledge of norms of reaction (i.e. the relationships between genotype, environment and phenotype) can predict the demographic and public health consequences of certain massive environmental changes'.
In previous issues we have explored the failures of traditional epidemiology to deliver the correct answer.2,3 Here Buchanan et al.4 in a point–counterpoint debate draw an analogy between the search for causes of chronic diseases with the alchemist's quest to discover the Philosopher's stone that would turn dross into gold. They chronicle the reasons for our difficulties in estimating weak effects, highlighting assumptions about design and interpretation that result in mistakes, and question the scientific orthodoxy that simply throws more dollars at the same research groups doing the same thing on a bigger scale. They offer by way of a partial solution, the possibility of intervention without full knowledge of the mechanisms involved. Not everyone would agree with epidemiology's contemporary record of failure. Schwartz and Susser5 argue that discoveries of public health relevance in child health have been made using standard epidemiological methods and that, provided we are realistic about what risk factor studies can achieve, they still have an important place and we can develop methods that consider context, contingency, and non-linearity. Merikangas et al.6 focus on the need to integrate better the component disciplines of epidemiology, genetics, and clinical medicine, and suggest that the audience who need to take notice of Buchanan and colleagues views are a ‘a small number of highly vocal gene hunters’. Ioannides,7 calm at the prospect of studies examining 500 000 candidate SNPs with different phenotypes, provides a proposed grading of credibility of evidence from genetic epidemiology studies using axes of effect size, replication, bias, biological credibility, and relevance. The divergent views of the success or failure of contemporary epidemiology highlighted in this point–counterpoint are remarkable, suggesting that we need to take stock of where we actually are.
Developments in epidemiological methods are needed in examining the causes of complex diseases. Pearce and Merletti8 discuss the application of complexity theory to epidemiology, reflecting on our growing understanding that historical, cultural, and social contexts have a major bearing on individual behaviours, and that we will need methods capable of examining non-linear effects, feedback loops, and adaptive systems. Bayesian methods are the subject of Greenland's9 paper in which he notes that practical issues such as complex software and concerns over subjectivity in defining priors are obstacles to their understanding and use. His papers (this is the first in a series) should provide useful ideas and practical examples for basic epidemiology courses.
Oppenheimer10 explores the historical differences in approach between cancer and coronary heart disease epidemiology in the USA. The former was dominated by the strong effect of smoking on lung cancer, whereas the latter examining weaker and multiple effects established cohort studies, studying individual-level risk factors potentially manipulable by clinicians. The invention of the ‘coronary-prone’ individual provided clinicians with a new source of patients and partly explains the uptake of primary prevention within clinical practice. However, the bigger picture of the upstream determinants of health behaviours and individual risk factors remained largely unstudied resulting in population-level control of coronary heart disease lagging behind lung cancer.
References
1 Lewontin RC. The analysis of variance and the analysis of causes. Int J Epidemiol 2006;35:536–37.
2 Davey Smith G, Ebrahim S. Epidemiology—is it time to call it a day? Int J Epidemiol 2001;30:1–11.
3 Davey Smith G, Ebrahim S. Mendelian randomization: can genetic epidemiology contribute to understanding environmental determinants of disease? Int J Epidemiol 2003;32:1–22.
4 Buchanan AV, Weiss KM, Fullerton SM. Dissecting complex disease: the quest for the Philosopher's stone? Int J Epidemiol 2006;35:562–71.
5 Schwartz C, Susser E. What can epidemiology accomplish? Int J Epidemiol 2006;35:587–90.
6 Merikangas KR, Low NCP, Hardy J. Understanding sources of complexity in chronic diseases—the importance of intergration of genetics and epidemiology. Int J Epidemiol 2006;35:590–92.
7 Ioannidis JPA. Grading the credibility of molecular evidence for complex diseases. Int J Epidemiol 2006;35:572–78.
8 Pearce N, Merletti F. Complexity, simplicity and epidemiology. Int J Epidemiol 2006;35:515–19.
9 Greenland S. Bayesian perspectives for epidemiological research. I. Foundations and basic methods. Int J Epidemiol 2006;35:765–75.
10 Oppenheimer G. Profiling risk: The emergence of coronary heart disease epidemiology in the US (1947–1970). Int J Epidemiol 2006;35:720–30.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||