Our contribution Social epidemiology? No way was meant to be provocative. Unfortunately, the comments made by Macdonald, Krieger, Siegrist, McPherson and Kaufman using qualifications as illogic, narrow, sloppy, and silly indicate that our contribution is taken to be offensive. That was certainly not our intention and we like to stress that we do value the contribution of sociology and psychology to the understanding of health phenomena. Naturally, we realise that not only biological, physical, and chemical factors, but also the social environment and psyche, influence disease occurrence and outcome. We focus in this reaction on the core critical statements of Macdonald, Krieger, Siegrist, McPherson and Kaufman regarding (1) boundaries between disciplines, (2) training and (3) the importance of subject matter knowledge.
The main critique is that it is unnecessary and counterproductive to draw boundaries between disciplines. We agree that cross-fertilization between biological and social knowledge may lead to insights that would otherwise not have been developed. We also realise that these factors may interact and that biomedical, sociological and psychological expertise may be necessary to unravel pathogenesis and prognosis. Multi-disciplinary research teams may be formed whenever necessary or when they are expected to be productive. However, we strongly oppose mono-disciplinary trained professionals using their own tools to make inquiries in neighbouring fields or borrowing tools from the latter without understanding the field for which these tools are made. Shopping in neighbouring scientific fields will only be beneficial when shoppers possess (or combine) expertise in both fields. This is the exceptional case, rather than the rule.
We do not object to use of the term social epidemiology for the exceptional case of a combination of expertise. In other situations where social determinants of disease are studied, it is best to call these activities sociology, because it is clear to everyone that the basis of such activities rests on sociological theory. Using the word epidemiology in such a context would only refer to the method (frequency research) and biostatistics would be a better term. Epidemiology is not defined by its method but by its object (occurrence of disease) and it has never been disputed that the basis of epidemiology rests on (bio)medical theory. It is our impression (and we gave examples) that this important element is not appreciated by the proponents of social epidemiology. Krieger takes the argument one step further, stating that training in solely biomedical sciences is always insufficient. We do not agree. Although social and biomedical determinants may often interact, we do not believe the world to be such a hybrid that research questions cannot be studied without expertise in both fields. If Krieger is right, we would need to redesign almost all epidemiology training programmes in the US and Western Europe.
In contrast to Krieger, Macdonald states that thorough subject knowledge is not necessary to generate and evaluate hypotheses. Indeed, we did not need thorough knowledge about polycyclic aryl hydrocarbons to prove that cigarette smoking causes lung cancer. We needed some biomedical knowledge, however, to know that carrying a cigarette lighter does not cause lung cancer. Logical thinking and common sense without expertise would not have prevented this latter conclusion being reached. Although Macdonald disagrees, we firmly believe that epidemiological research should be based on deductive reasoning and/or biomedical theory, but this is an empirical cyclical process, which may start with little understanding of disease mechanisms. For the same reason, Siegrist's term biomedical epidemiology should also be avoided. After all, epidemiology always deals with improving biomedical theories! Just like sociology always deals with improving social theories. For clinical and community medicine both are important, but should not be intermixed.