Thomas McKeown

Two approaches in medical research




This is an extract from chapter 12 of The Role of Medicine. Dream, Mirage, or Nemesis?



Contrary to what is generally believed, the most fundamental issue confronting medical science is not the solution of one or more of the unsolved biomedical problems: it is evaluation of two approaches to the control of disease, one through an understanding of mechanisms and the other through a knowledge of origins. And as both approaches are needed, what is required is a decision about the distribution of effort between them and, if possible, identification of the circumstances in which each is likely to be rewarding.

Since epidemiologists are concerned particularly with disease origins and clinicians and laboratory investigators with mechanisms, it may be thought that what is in question is the respective roles of epidemiology and other kinds of research. However, this way of looking at the matter is misleading. The discovery of micro-organisms in the laboratory contributed enormously to control of the origins of infectious diseases; and epidemiological investigations of the distribution of blood pressure in the general population were necessary for an understanding of the nature of essential hypertension. The emphasis therefore should be on disease origins and mechanisms, each investigated by the methods - population, clinical, experimental - that seem appropriate to the problem in hand.

There is not much room for disagreement about the contribution of these approaches in the past. Health advanced because of the decline of the infections; and the infections declined mainly because of modification of the conditions which led to them. This conclusion is unaffected by well-founded doubts about the reliability of data (based on certification of cause of death) for individual diseases, and by recognition that understanding of disease mechanisms has made it possible to treat a number of conditions. No knowledgeable person is likely to dispute that we owe the vast increase in expectation of life and reduction of associated morbidity mainly to changes which occurred before effective clinical intervention was possible. We are on less certain ground when we turn from interpretation of the past to speculation about the future. The infections declined largely because of removal of the ill-effects of poverty, and the non-communicable diseases which have replaced them are evidently not responsive to the same changes, otherwise they would not be there. How then are we to decide in what circumstances each of the main lines of enquiry is likely to be successful?

Before trying to answer this question I must refer briefly to the residual disease problems in developed countries. They are often assessed by analysis of causes of death at all ages. Treated in this way they include deaths (the majority) at the end of a normal expectation of life, which have presumably been determined at fertilization, as well as premature deaths not so determined which are in principle preventable. In so far as prevention of death is a goal of medical science, it is only the latter which are relevant. As the genetically determined expectation of life (for those not congenitally handicapped) is about eighty years, it would be better to base the analysis of preventable deaths on deaths at an earlier age, say before seventy. These would inevitably include some deaths which were genetically determined, just as they would exclude some later deaths which were not. They would, however, provide a much better basis for assessment of the tasks of medical science than deaths at all ages.

I suggest that the approaches needed to the residual disease problems can be clarified by consideration of the four-fold classification of diseases outlined in Chapter 7: relatively intractable; preventable, associated with poverty; preventable, associated with affluence; and potentially preventable, not known to be related to poverty or affluence.


[Home] [Top]