Alex Broadbent (History and Philosophy of Science, Univ. Cambridge) has a fascinating new article out in Studies in History and Philosophy of Science Part C: Studies in History and Philosophy of Biological and Biomedical Sciences. The title is Causation and Models of Disease in Epidemiology. Here is the Abstract:
Nineteenth-century medical advances were entwined with a conceptual innovation: the idea that many cases of disease which were previously thought to have diverse causes could be explained by the action of a single kind of cause, for example a certain bacterial or parasitic infestation. The focus of modern epidemiology, however, is on chronic non-communicable diseases, which frequently do not seem to be attributable to any single causal factor. This paper is an effort to resolve the resulting tension. The paper criticises the monocausal model of disease, so successful in the nineteenth century. It also argues that a multifactorial model of disease can only be satisfactory if it amounts to more than a mere rejection of the monocausal model. A third alternative, the contrastive model, is proposed and defended on the grounds that it links the notions of disease and of general explanation, while avoiding the philosophical naiveties and practical difficulties of the monocausal model.
I read (and then reread) the paper last night, and will offer some comments. By way of context, I should say that I have been interested in causation for most of my adult life (my undergraduate thesis was on causation in science). Part of the reason for my continued interest is the relevance of illness causality [sic] for the social determinants of health. If the best evidence suggests that the primary causes of health and illness in populations are social and economic conditions, it follows that our public health priorities -- in a normative sense -- ought to be remediation of these conditions.
Turning to the article, as a piece of philosophy, it is a tour de force, one of the best treatments of illness causation I have reviewed since K. Codell Carter's seminal books (Broadbent cites Carter extensively). I have some minor quibbles with the argument structure itself, but it strikes me as generally persuasive, particularly in its documentation of the myriad problems of what Broadbent terms the monocausal model of illness causality.
However, I had two qualms about the remainder of the essay, one fairly technical, the other, more related to the pragmatics of public health policy. Broadbent distinguishes between what he terms the "bare" and the "contentful" models of multifactorial causation. The bare model, he argues, is intended solely as the negation of the monocausal model. Accordingly, he writes,
"saying that diseases are multifactorial -- have many causes -- is trivially true, because everything that is caused is multifactorial. Since it is trivially true of any caused event, the multifactorial approach cannot possibly identify any distinctive feature of disease, and so cannot be methodologically helpful in this way" (306).
I am uncertain that this conclusion follows. Namely, nothing about the fact that diseases have many causes requires that all causes have equal causal significance. I see no inherent reason why an event D could not be caused by causes C1-C12, but where C10, C11, and C12 seem to have much greater causal responsibility in bringing about event D. In fact, this is precisely what seems to be the case in thinking about disease on the population level: lack of access to acute care services may, for example, causally contribute to a latent infection entering an acute phase, but the larger patterns of that disease occurring on the population level -- including in that particular person -- are primarily determined by social and economic conditions.
Perhaps Broadbent might respond by noting that, in its assertions that some causes are more significant than others in thinking about disease, the excessively crude account I have sketched here qualifies as a "contentful" multifactorial model. If that is the case, I have no real objection, but note then that the bare multifactorial model is generally not, at least in the social epidemiologic literature I am familiar with, entertained as a serious candidate for a model of disease causality.
So much for the technical objection. The pragmatic reservation I have concerns Broadbent's positive model of disease, what he terms a contrastive model. The basic idea, as I understand it, is that to have a disease (1) one must have a set of symptoms S which differ in the person P from a contrast class X; and (2) at least some of the causes of S are not causes of the absence of S.
"For example, to have cholera, you must exhibit some symptoms of ill health which a certain contrast class does not have, such as diarrhoea; and these symptoms must be caused by a certain specified cause, namely, the activity of vibrio cholerae in your small intestine" (307).
I have no problems with the argument structure per se, but am concerned that this account gives short shrift to the discursive nature of causal attributions. Specifically, Hume teaches us that causal attributions are inherently subjective in the sense that it is subjects who supply the causal inferences. Since causal attributions are subjective conventions, it follows that a panoply of social and cultural factors affect which causal explanations we approve and reject. This is well-documented in the history of disease explanations, particularly in the work of Carter, Christopher Hamlin, Charles Rosenberg, and Margaret Humphreys. To his credit, Broadbent acknowledges this, but suggests the debate over the value-ladenness of disease causality obscures a "different and perhaps even more important methodological issue . . . ." (310). I see great merit to what Broadbent is trying to accomplish, but respectfully disagree that such an analysis is more important than analysis of the ways in which causal attributions of disease are socialized.
More importantly, however, Broadbent acknowledges that the charge of "biological chauvinism," or the idea that emphasis is directed away from the SDOH to "biological" factors, is equally applicable to his model as it is to the monocausal model. (Sidenote: I eschew the rhetorical tendency to speak of "biological" and "social" causes of illness, since one of the fundamental ontological lessons of the SDOH literature is that there is no such thing as "biological" causes that are truly separable from "social" and "environmental" factors. The biological is the social).
Broadbent's response to this objection is that he is focused on the narrower question of disease causation rather than the larger question of illness causation, and that the concerns of the SDOH proponents may be entirely correct, but do not apply to the study of disease causality as he defines it. This is fine, as far as it goes, but as a public health ethicist, my concern is that it does not go very far. That is, in proferring a philosophically rigorous account of disease causality, Broadbent is performing a very great service indeed. But for those of us focused on identifying our normative priorities in ameliorating human suffering on the population level, the distinction between disease and illness that Broadbent draws (which is quite different from the similar distinction widely utilized in medical humanities discourse) may not have all that much significance. If it is the case that the vast majority of illness burdens in both the developed and the developing world are primarily caused by social and economic conditions, then our normative priority should be to ameliorate those conditions. No?
In any case, these concerns hardly detract from the rigor and quality of Broadbent's article. Highly recommended!
Comments