Did John Stuart Mill Influence the Design of Controlled Clinical Trials?

From the James Lind Library Journal of the Royal Society of Medicine; 2019, Vol. 112(6) 258–260 DOI: 10.1177/0141076819851759 Did John Stuart Mill influence the design of controlled clinical trials? Roger Bailey1 and Jeremy Howick2 1NHS England and Harrold Medical Practice, Peach’s Close, Harrold MK43 7DX, UK 2Faculty of Philosophy, University of Oxford, Oxford OX2 6GG, UK Corresponding author: Roger Bailey. Email: [email protected] John Stuart Mill and causal inference appear to refer implicitly to differences resulting from comparing a treatment (‘instance’) with no The writings of the 19th-century British philosopher treatment and not to differences resulting from com- John Stuart Mill contain logical analyses relevant to paring two treatments. causal inferences about effects but not specifically to assessing the effects of medical treatments. However, Was Mill influenced by earlier writings on because Mill’s analyses are relevant to the evolution fair comparisons of treatments? of thinking about how to assess the effects of treatments, some passages from his writings are included Mill makes one reference to the application of this in the James Lind Library.1 For example: logic to assessing the effects of medical treatments, acknowledging that uncertainties can remain: In the spontaneous operations of nature there is gen- erally such complication and such obscurity, they are Suppose that mercury does tend to cure the disease, so mostly either on so overwhelmingly large or on so many other causes, both natural and artificial, also inaccessibly minute a scale, we are so ignorant of a tend to cure it, that there are sure to be abundant great part of the facts which really take place, and instances of recovery in which mercury has not been even those of which we are not ignorant are so multi- administered: unless, indeed, the practice be to admin- tudinous, and therefore so seldom exactly alike in ister it in all cases; on which supposition it will equally any two cases, that a spontaneous experiment, of be found in the cases of failure. ...Neither, therefore, the kind required by the Method of Difference, is will the instances of recovery agree in the administra- commonly not to be found. When on the contrary, tion of mercury, nor will the instances of failure agree we obtain a phenomenon by an artificial experiment, in its non-administration. It is much if, by multiplied a pair of instances such as the method requires is and accurate records from hospitals and the like, we obtained almost as a matter of course. (p. 431)1 can collect that there are rather more recoveries and rather fewer failures when mercury is administered The ‘Method of Difference’ to which Mill refers in than when it is not; a result of very secondary value that passage is one of the methods he sets out in his even as a guide to practice, and almost worthless as a System of Logic (bk. III, ch. 8, p. 429).1 Mill’s contribution to the theory of the subject. (p. 494)1 ‘Method of Difference’ states: He does not refer to any features of the design of the If an instance in which the phenomenon under inves- ‘artificial experiments’ that he deems necessary and tigation occurs, and an instance in which it does not he appears not to have conceptualised controlled occur, have every circumstance in common save one, trials. Some earlier writers had addressed these that one occurring only in the former, the circum- issues. During early 18th century in debates about stances in which alone the two instances differ, is the safety of inoculation against smallpox, for exam- the effect, or the cause, or an indispensable part of ple, Massey3 had pointed out that ‘to form a just the cause, of the phenomenon. (bk. III, ch. 8, p. 483)2 comparison’ between inoculated and uninoculated people ‘the circumstances of the patients must and One can appreciate the relevance of these ideas to ought to be as near as may be on a Par’. thinking about how treatments should be tested. Proposals to draw lots (random allocation) to gen- However, it should be noted that Mill’s differences erate fair treatment comparison groups were made at ! The Author(s) 2019 Article reuse guidelines: sagepub.com/journals-permissions Bailey and Howick 259 least as early as the 17th century.4,5 Alternation was statistics and the importance of the elimination of used to create unbiased treatment comparison groups chance as an element in his ‘Logic of Medicine’. at least as early as the beginning of the 19th cen- The only evidence of Mill’s influence on medical tury,6,7 and some prominent doctors had started to writers of which we have been made aware has been use alternation during Mill’s time.8 communicated to us by Ulrich Troehler (personal However, we are not aware of any sources indicat- communication, October 2017). In 1852, Friedrich ing that Mill was aware of any of these developments Oesterlen, a German physician, referred to Mill and in thinking among doctors or that he saw its rele- aimed to transfer his teachings to medicine in his vance to his analyses. book Medicinische Logic.15 Did Mill’s writing influence the development I am indebted, however, to J. Stuart Mill’s excellent of fair comparisons of treatments? work for the greater part, the true pith and thread of my system, and the following is, in a great meas- Some philosophers have claimed that Mill’s writings ure, but an application of his doctrines to the are relevant to the evolution of controlled clinical particular questions and objects of our department trials. The most specific claim has been made by of science. (p. X)15 Cartwright and Hardie9: Oesterlen referred to the significant contribution of An RCT [Randomized Control Trial] is a study statistics in general terms as ‘the essential link in design based on John Stuart Mill’s method of differ- our research on truth based on experience’. He fur- ence for making causal inferences. ther emphasised that this held, provided one kept to the rules of extremely precise observation, compared Other philosophers, while drawing attention to Mill’s like with like, considered the natural course of dis- methods, have been less specific, perhaps because eases, collected large numbers, and was appropriately they are aware that random allocation for controlled cautious in making generalisations (pp. 129–140). trials was not widely adopted until a century after Oesterlen’s book dealt with medical observation and Mill and that alternation had been used to create thought: the concepts of induction, deduction, gener- similar treatment comparison groups 50 years alisation, experiment and statistics. However, before that.10 Examples of these less specific state- although Oesterlen’s book made a major contribu- ments include the following: tion to methodological discussions in Germany,16 Ulrich Troehler has not found any evidence that it Mill’s ‘method of difference’...seems to lie at the penetrated the clinical world (personal communica- heart of EBM [Evidence-Based Medicine] because it tion, October 2017). lies at the heart of clinical trials. (ch. 6, part 3)11 Comparative clinical studies have been referred to as Conclusion Mill’s methods ..., the ‘numerical’ method ..., the ‘statistical’ method ..., and ‘difference-making’ evi- Mill’s writings are important in their own right, dence. (p. 124)12 regardless of whether they were influenced by prior Mill’s method of difference can be seen as the basic writings on controlled trials or whether they have principle behind the controlled clinical trial. (p. 212)13 influenced the subsequent development of controlled trials. However, commentaries on and examples of Neither Cartwright’s claim or any of the above state- prototype controlled trials antedated Mill; there ments is accompanied by references to supporting does not appear to be evidence that he was aware evidence that Mill’s ‘Method of Difference’ had any of them. Nor does there appear to be evidence that subsequent influence on those promoting or doing Mill influenced the development of controlled trials, controlled trials. Had Mill been influential in the evo- even if the way they are conducted is in keeping with lution of controlled trials one would have expected to his methods. Of course, it may be the case that evi- find references to him in the writings of pioneers of dence of these possible influences on and of Mill controlled trials using alternate allocation during the remains to be discovered. Yet until such evidence is first half of the 20th century – Haffkine, Choksy, discovered, philosophers should acknowledge the Cecil, Bullowa and Hill, for example – and of lack of it. random allocation during the second half of the century. As it is, not even Alexander Bain14 (p. 362), a Scottish philosopher and close friend of Mill, referred Declarations to Mill’s writings in his emphasis on the use of Competing Interests: None declared. 260 Journal of the Royal Society of Medicine 112(6) Funding: None declared. Alsop at the two Sugar-loaves over against Ethics approval: Not applicable. St. Antholins Church at the lower end of Watling Guarantor: RB. Street, 1657. 6. Lesassier Hamilton A. Dissertatio Medica Inauguralis Contributorship: Main contributors: RB and JH, with edits De Synocho Castrensi [Inaugural Medical Dissertation from Iain Chalmers. on Camp Fever]. Edinburgh: J Ballantyne, 1816. Acknowledgements: The authors are grateful to Ulrich 7. Lo¨hner G, on behalf of a Society of truth-loving men. Troehler and Alexander Bird for commenting on earlier drafts of Die Homooöpathischen Kochsalzversuche zu Nu¨rnberg this paper. [The Homeopathic Salt Trials in Nuremberg]. Provenance: Invited article from the James Lind Library. Nuremberg, 1835.

Load more