Eras and paradigms: a new conceptual approach

Susser & Susser (Reference Susser and Susser1996a , Reference Susser and Susser b ) at Columbia University, New York, adapting Thomas Kuhn's theory of changing conceptual frameworks in science, urged the need today for a new such ‘paradigm’ in epidemiology, to follow on from miasma theory and the sanitary movement in the early 19th century, microbial causal theories of disease in the mid-to-late 19th and early 20th centuries and the era of non-infectious disease epidemiology, with its ‘black box’ model of research, up to the present day. The black box model, they argued, is now outdated and due for replacement. Instead the authors foresaw a new era of ‘eco-epidemiology’, which would encompass different levels of organisation, neatly compartmentalised one within another, like a conjuror's nest of Chinese boxes. ‘The practical implication’, they argued, ‘… is that an exclusive focus on risk factors at the individual level within populations – even given the largest numbers – will not serve. We need … to be equally concerned with causal pathways at the societal level and with pathogenesis and causality at the molecular level.’

More recently, March & Susser (Reference March and Susser2006) enumerated the elements central to eco-epidemiology as follows: ‘multiple levels of organization; the individual life course, as reflected by consideration of early antecedents and concepts of embodiment; equal consideration of communicable and non-communicable diseases … the imprint of historical time, and the dynamic relationship between macro causes (e.g. societal change) and micro causes (e.g. genetic mutations).’

This highly ambitious scenario is flawed in a number of respects. Multi-level theorising, as indeed March and Susser remark, is not a wholly new departure, but was present in embryo throughout all eras of epidemiology from Hippocrates onward and explicit, in particular, in the writings of both Snow on cholera (infectious) and Goldberger on pellagra (non-infectious). The black-box model itself is not a scientific paradigm in Kuhn's sense, but simply a metaphor for empirical approaches to the prevention of diseases before their aetiology and pathogenesis have been worked out. ‘Chinese boxes’, another metaphor, is meant to symbolise the various organising systems active in human populations, from the societal to the genetic, but is problematic because actually neither these systems nor the data sets that can be used to represent them are arranged in such neat, orderly hierarchies. Furthermore, a schema that encompasses all such systems, including intra-cellular processes and human gene complements, is so overinclusive as virtually to negate the hypothetical-deductive method of enquiry.

In view of these weaknesses, the notion of an all-embracing multi-factorial causation has failed to gain wide acceptance. Nonetheless, there is growing awareness of the need to apply a more pragmatic concept of eco-epidemiology, in which the study of individual risk-factors is combined with analysis of macro-environmental systems – whether natural or man-made – that mediate population levels of health and sickness. Thus, Koopman (Reference Koopman1996) considered the discipline to be ‘in transition from a science that identifies risk factors for disease to one that analyses the systems that generate patterns of disease in populations’, and the implications have since been spelt out in terms of research design (McMichael, Reference McMichael1999), statistical analysis (Pearce, Reference Pearce2011) and strategies in preventive medicine (Rose, Reference Rose2001).

Implementing epidemiological knowledge in policy and practice

Proposals for change have not been confined to concepts and methods of epidemiology, but have extended to the processes by which its conclusions are implemented. In the history of medicine, translation of new scientific knowledge has been traditionally a hit-and-miss affair, with few clearly established pathways and only a small minority of researchers willing to venture into the intervening ‘valley of death’ (Butler, Reference Butler2008). Things may be changing, however, now that a new construct of translational medicine (TM) has been taken up by industry, national governments and many medical faculties. International symposia have been held, specialist journals set up, degree courses established and pharmaceutical divisions dedicated to advancement of this project (Henderson, Reference Henderson2011).

The purpose of TM as a research programme is to connect up the successive stages of knowledge transfer in product development, usually differentiated as basic scientific discovery, early human studies, early clinical trials, late clinical trials and implementation in practice (Thornicroft et al. Reference Thornicroft, Lempp and Tansella2011). This progression, based on the standard procedure for developing, testing and marketing new pharmaceuticals and vaccines, is in the main ‘product-driven’ and can be applied only in modified form to projects focused on community health and primary preventive measures. Nevertheless, the notion of ‘translational epidemiology’ has emerged as a distinct branch of TM, which contributes to knowledge transfer at each successive stage. Khoury et al. (Reference Khoury, Gwinn and Ioannidis2010) formulated the process as follows:

• • at T0, a discovery is made (e.g. disease agent, biomarker, new vaccine, etc.)

• • at T1, epidemiology explores the use of this advance in developing a candidate test or guide to medical intervention;

• • at T2, epidemiology helps to evaluate the new approach by observation and clinical trials;

• • at T3, epidemiology points to factors that promote or restrict applications in practice;

• • at T4 finally, epidemiology helps to assess the public health impact of such applications.

This schema is logical as far as it goes, but actually is only the beginning. How and when new knowledge is put into effect in the real world will depend to some extent on policy-makers and service managers, but probably more on gradual changes in professional and public attitudes, mediated largely through unofficial channels (Ogilvie et al. Reference Ogilvie, Craig, Griffin, Macintyre and Wareham2009). Even so, the TM band-wagon is now rolling and has powerful backing. The US National Institutes of Health, for example, earmarked an annual budget of US$500 million for TM research, while in Britain the Medical Council established six academic centres for the same purpose, including one for translational epidemiology (Woolf, Reference Woolf2008).

Adaptation of epidemiology to a changing world economy

This third trend is less clearly differentiated than the first two, but potentially the most significant, because it is directly related to the moral and ethical basis of public-health medicine.

We are living in an era of growing global inequality, fraught with heavy consequences for human society. Loewensen (Reference Loewensen2004) summed up the position a few years ago as follows:

An imbalance so gross must pose serious problems for public-health medicine. The fundamental problem for low-income countries in particular is shortage of resources in money and trained manpower for health services and research, including epidemiological studies that could help to narrow the so-called ‘10/90 gap’: that is, the chasm between their contribution to health research and their share of the global disease burden. The resulting disparities are especially flagrant for psychiatric disorders, because the poorest countries spend the lowest percentages of their overall health budgets on mental health (Saxena et al. Reference Saxena, Thornicroft, Knapp and Whiteford2007).

A partial solution to the dilemma lies in promotion of epidemiological research within global health programs, in which high-income countries bear the brunt of costs, while low- and middle-income countries benefit in terms of training and development (Razzouk et al. Reference Razzouk, Sharan, Gallo, Gureje, Lamberte, de Jesus Mari, Mazzotti, Patel, Swartz, Olifson, Levav, de Francisco and Saxena2010). Although such projects are now being realised, in many instances they rely upon public–private partnerships in which government-supported agencies work directly with both charitable foundations and commercial groups to promote health care projects.

This trend accelerated following the 1992 UN Conference on Environment and Development (Richter, Reference Richter2004), at which leading trend-setters from the World Health Organization (WHO) and United Nations argued that it was time for all ‘stakeholders’ in health care to move away from confrontation with one another, towards co-operation and dialogue. Since then, both WHO and the World Bank have collaborated extensively with non-governmental organisations, private institutions and the pharmaceutical industry (Nishtar, Reference Nishtar2004).

The fact that such collaborations, capable of raising billions of dollars, are being targeted on epidemic diseases – including tuberculosis, AIDS, malaria and cholera – underlines their importance for the developing world, but at the same time also their ultimate dependence on population-based research. Such projects, however, are focused primarily on product development; that is to say, on making available effective medication, vaccines, diagnostic biomarkers and other technologies (Mahoney, Reference Mahoney2011). Commendable as are these objectives, the context is one in which epidemiology – and in particular current moves towards eco-epidemiology – can easily come to be regarded as of only marginal interest.

For this reason, calls are growing louder for international programs in health research and development that will be truly independent of marketing influences. Richter (Reference Richter2004), for example, has advocated safeguards to ensure that public-health interests remain at the centre of all financial relationships and joint ventures between publicly funded agencies and the commercial sector or wealthy corporate donors. Concrete steps, in her view, should consist in:

1. 1. replacing the ‘partnership’ appellation by ‘agreement’, ‘interaction’, or other more neutral terms;

2. 2. identifying nodal points at which conflicts of interest are likely to arise, and

3. 3. establishing firm regulations to ensure that the public interest remains paramount in all such interactions, including management and oversight.

Psychiatric epidemiology: still a backwater?

On turning to the mental health field, one is confronted by quite other research concerns and priorities. Here, attention is concentrated on issues of case-definition, classification and diagnostic validity. As Kessler (Reference Kessler2000) observed over a decade ago, this branch of epidemiology traditionally lags behind mainstream developments because of its difficulties in conceptualising and measuring mental disorders, and in consequence remains largely descriptive. The current debates in general epidemiology are hardly reflected, or else take on different tones. The one major paradigm-shift in modern psychiatry – basically from developmental psychology to neuroscience – was satirised by Eisenberg (Reference Eisenberg1986) as a switch from ‘brainlessness’ to ‘mindlessness’. Translational psychiatry has emerged in its own new journal as a branch of biomedical informatics, or as the link between neuroscience and clinical treatment, but with very little bearing on public health.

In the developing world, mental disorders constitute a neglected and under-researched area for public-health research. A survey of studies in this field, covering 114 low- and middle-income countries, found that over half had contributed fewer than five articles each to a total of over 10 000 relevant publications in the period 1993–2003, even though participating research groups ranked epidemiological studies of illness burden and risk factors among their highest priorities (Sharan et al. Reference Sharan, Levav, Olifson, de Francisco and Saxena2007). Impressive progress, it is true, has been made in international collaborative research, exemplified by the WHO World Mental Health Survey (Kessler & Ustun, Reference Kessler and Ustun2008) has yielded a mass of statistical information. The wide variation among countries, however, both in survey response rates and in reported prevalence and severity distributions, must raise questions about the validity of these data.