The Philosophy of Disease

Benjamin Smart

• Publisher: Palgrave Pivot
• Year: 2016
• ISBN:
  9781137552921
• Topic: Philosophy
• Subtopic: Philosophy & Ethics in Science

Disease is everywhere. Everyone experiences disease, everyone knows somebody who is, or has been diseased, and disease-related stories hit the headlines on a regular basis. Many important issues in the philosophy of disease, however, have received remarkably little attention from philosophical thinkers.

This book examines a number of important debates in the philosophy of medicine, including 'what is disease?', and the roles and viability of concepts of causation, in clinical medicine and epidemiology. Where much of the existing literature targets conceptual analyses of health and disease, this book provides the reader with an insight into these debates, and develops plausible alternative accounts. The author explores a range of related subjects, discussing a host of interesting philosophical questions within clinical medicine, pathology and epidemiology. In the second part of the book, the author examines the concepts of causation employed by clinicians and pathologists, how one should classify diseases, and whether the epidemiologist's models for inferring the causes of disease are all they're cracked up to be.

Part I

The Concept of Disease

1

The Concept of Disease in Clinical Medicine

Abstract: This chapter asks the question: ‘what, for the clinician, is disease?’ Of all the tasks philosophers undertake, I consider the conceptual analysis of disease to be of particular importance – and for obvious reasons. For one, one’s disease status can have significant consequences. In the following sections I outline three prominent conceptions of disease, looking for that best suited to clinical medicine: those proposed by Rachel Cooper (2002), Peter Schwartz (2007), and Jerome Wakefield (1992, 1999). I dismiss Cooper’s view based on clear cut counterexamples, but show that significant parts of both Wakefield and Schwartz’s views are compelling. I conclude with a fourth proposal (which draws on both Wakefield and Schwartz), providing an etiological theory of disease in clinical medicine.

Keywords: disease as harmful function; harmful dysfunction; Jerome Wakefield; natural function; Peter Schwartz; the line-drawing problem; What is disease?

Smart, Benjamin. Concepts and Causes in the Philosophy of Disease. Basingstoke: Palgrave Macmillan, 2016. DOI: 10.1057/9781137552921.0007.

Introduction

[H]ow we think of health and disease lies at the very core of medical practice, reflections on bioethics, and the formation of health care policy.

– Engelhardt in Ananth 2008.

This chapter asks the question: ‘what, for the clinician, is disease?’ Of all the tasks philosophers undertake, I consider the conceptual analysis of disease to be of particular importance – and for obvious reasons. To mention just a few: concepts of health, illness, and disease (specifically mental illness) affect whether or not one is legally responsible for one’s actions; a concept of disease is necessary to demarcate the diseased from the healthy (which might distinguish those eligible for medical treatment, from those not¹); not only must medical care be paid for, but one’s state of health affects one’s right to disability allowance, free housing, and so on – so concepts of health and disease have a significant economic effect; and since the decisions made by those in the medical profession affect legal responsibility, eligibility for medical care, eligibility for income support, and the distribution of resources (both locally and internationally), decisions concerning health and disease have enormous ethical implications.

In the following sections I outline three prominent conceptions of disease, looking for that best suited to clinical medicine: those proposed by Rachel Cooper (2002), Peter Schwartz (2007), and Jerome Wakefield (1992, 1999). I dismiss Cooper’s view on the basis of clear cut counterexamples, but show that significant parts of both Wakefield and Schwartz’s views are compelling. I conclude with a fourth proposal (which draws on both Wakefield and Schwartz), providing an etiological theory of disease in clinical medicine, that is not, unlike Wakefield’s etiological account, subject to the line-drawing problem.

The maximally value-laden conception – Rachel Cooper on disease

Cooper provides a tripartite account of disease; that is, she proposes that the concept of disease has three individually necessary, and together sufficient conditions.

1Disease is a bad thing to have;

2We must consider the afflicted person to have been unlucky; and

3The condition can potentially be medically treated. (see Cooper, 2002, 271)

The first criterion states that in order for a patient to have a disease, the condition must be bad for that patient (as opposed to society at large). There are many examples of one and the same condition being good for one individual, and bad for another. Having an unusually fast metabolism, for example, is harmful to the bodybuilder (who wishes to increase muscle mass), but beneficial to the model (who can, as a result, eat a balanced diet, while maintaining a low body mass index (BMI)).² Cooper’s claim thus implies the somewhat counterintuitive conclusion that the same condition can be a disease for one person, and not for another. That said, her example of ‘sterility’ goes some way to justify this – some choose to have vasectomies, yet others, who desperately want children, are sterile for reasons beyond their control. Only the latter group are diseased.

The second criterion states that an individual can be diseased only if they are ‘unlucky as judged by the uninformed layman, that is, roughly, worse off than the majority of humans of the same sex and age’ (2002, 276). She claims that this criterion helps us understand why one can attribute states of health to those with disabilities, and to those with genetic conditions such as Down syndrome. Down syndrome patients, she argues, are healthy ‘in a certain sense,’ just in case they do not have ‘some other’ infection or injury (276).

The third criterion, that the condition must be potentially medically treatable, separates diseases from other harmful and unfortunate states; for example, being robbed at gunpoint, losing one’s job, and so on.

Objections to Cooper’s model

Although the three conditions Cooper proposes do seem to apply to most diseases, the theory is subject to a variety of counterexamples. First, it is questionable whether all three necessary conditions are, in fact, necessary. I refer in particular to the second criterion, in which she states that one must be ‘worse off than the majority of humans of the same sex and age’ in order to be diseased. Cooper suggests that ‘90-year-olds who can’t walk as far as when they were younger are not diseased because we expect old people to become increasingly frail’ (276). This seems fair. However, there is a clear correlation between age and dementia; in an ageing population, it is worryingly possible that there will soon be an age group in which dementia is the norm. Anyone who knows somebody with dementia will confirm that the condition should always be deemed pathological, yet Cooper’s conception implies that, for the very elderly, it might not be. This second criterion is at best dubious, but there are more serious problems with the thesis. Cooper presents her view as a set of necessary and sufficient conditions, yet there are numerous non-pathological states that satisfy all three.

Cooper presents and refutes a counterexample of her own – that of unwanted pregnancy. She argues that we consider this a counterexample, only because our intuitions tend to lag behind the facts. Becoming pregnant can now (assuming one is taking contraceptive measures) be ‘unlucky’, so unwanted pregnancy is a disease. The implication is that, in time, we shall appreciate this.

Whether her response is satisfactory is questionable, but I shan’t dwell on this example, since one can imagine far more robust counterexamples without difficulty. I outline two possibilities below, but there are many more.

First, consider the pre-operative transsexual. A pre-operative transsexual feels he or she was either ‘a man born in a woman’s body’, or vice versa (unlucky and bad), and one can now have a sex change (medically treatable). A pre-operative transsexual is thus, according to Cooper’s model, diseased. But of course, one does not consider the transsexual to be diseased any more than the homosexual. He or she is perfectly physically able in every respect, and the transsexual is not mentally ill.³ Perhaps Cooper might respond that our intuitions are lagging behind relatively recent developments in medicine, and that soon, now that the condition is medically treatable, we will consider the ‘condition’ of ‘pre-operative transsexual’ a disease – but I suspect not.

Second, suppose Sam decides to get a tattoo reading ‘peace and love’ on her arm, and, as is the fashion, decides to have it written in Chinese characters. This is explained to the tattoo artist, whom Sam knows to be very skilled, honest, and reliable. On this occasion, however, the tattoo artist gets confused, and instead of ‘peace and love’, Sam finds herself wandering around with ‘I hate China’ decorating her upper arm (albeit in aesthetically pleasing symbols). Given that Sam is about to go to China, this is definitely bad and unlucky, but fortunately for her, this is also medically treatable. Nonetheless, it is not a disease. Cooper might respond that the tattoo is, in fact, pathological, but she would be clutching at straws.

Let us assume that neither Sam nor the pre-operative transsexual is diseased. Why do we take this to be the case? What strikes me is that neither Sam nor the transsexual has damaged or poorly functioning physiological traits (furthermore, neither is mentally ill in any respect). On the face of it, a plausible conception of disease in clinical medicine must accommodate this intuition.

The pure statistical conception

In this section I consider whether a value-free conception of disease will resolve the problems with Cooper’s tripartite account. Although the statistical thesis does rule out cases like unfortunate tattoos and the pre-operative transsexual, the view ultimately fails (and for countless reasons). However, the basic model plays a big role in both Schwartz’s account of disease (to be discussed shortly), and in Boorse’s (1977) biostatistical theory (BST),⁴ so a detailed exposition is warranted nonetheless.

Given the qualities expected of naturalist (value-free) views, and in particular, the thought that scientific theories often involve statistical analysis,⁵ it is unsurprising that some have tried to differentiate between health and disease states using mathematical models. The pure statistical model focuses on the quantifiable properties of an organism’s physiological subsystems. The precise qualities to be measured differ depending on the subsystem in question, of course, but the underlying method is (roughly) the same.

The statistical conception involves gathering data over large populations and applying formal methods to determine disease-status. Often (but not always), the measured values of a characteristic (e.g. quantities of a hormone or cell count; sizes and shapes of organs/tissues/cells; blood pressure, etc.) produce a normal distribution curve (see Figure 1.1). According to the basic statistical model, some physiological subsystem is dysfunctioning, and disease-status is met, when the measured value of the characteristic under investigation is beyond two standard deviations from the mean (Ananth 2008).

One might think this gives the thesis a normative quality, since the precise value at which ‘the line is drawn’ (two standard deviations from the mean) seems somewhat arbitrary; that is, nothing in nature ‘fixes’ this value. Perhaps surprisingly, however, this is not particularly unscientific – at least insofar as it is not inconsistent with standard practice in other disciplines widely regarded as sciences. Central to testing significance in psychological and epidemiological studies, for example, is the Null Hypothesis Significance Test (NHST).⁶ When one attempts to establish a non-chancy relationship between two variables, X and Y, the null hypothesis is that any correlation the data shows is accidental. The NHST states that a result is significant only if P, the probability of the null hypothesis being true (given the results of the study), is less than 0.05. This value, although not entirely arbitrary (insofar as it would not be sensible to fix the value at 0.4 or 0.8 etc.), was ultimately chosen by R.A. Fisher early in the 20th century (1925), not by nature.

The NHST is deemed a scientific method, yet the ‘line-drawing’ it involves is no less arbitrary than that used in the ‘pure statistical method’ of picking out disease states. If it is normal within the sciences to choose such arbitrary values, then there is no reason to think doing so should fall outside the remit of naturalistic conceptions of disease; that is, conceptions committed to employing on...

Table of contents

• Cover
• Title
• Introduction
• Part I  The Concept of Disease
• Part II  Disease and Causation
• Conclusion: Context Dependency is Rife
• Bibliography
• Index