The road to randomisation: patulin and the common cold
Randomised clinical trials (RCTs) are regarded as a ‘gold standard’ method for assessing the effects of treatments. But how did they come to be accepted? Here Sir Iain Chalmers describes an often overlooked MRC trial of a potential treatment for the common cold, and explains the role it played on the road to randomisation.
When people look back on the history of clinical trials, the MRC trial of streptomycin for pulmonary tuberculosis reported in 1948 tends to receive a lot of attention. But there’s an earlier MRC trial that helped to develop methods to ensure that there is no bias in how patients are allocated to particular treatment groups, to ensure that ‘like is compared with like’. And, unlike the streptomycin trial, this earlier trial used placebos to reduce bias in measuring treatment outcomes.
The trial had been prompted by conflicting reports about possible beneficial effects on the common cold of patulin, a fungal toxin often found in rotting apples. Between 18 January and 11 April 1944, volunteers at three London units of the General Post Office, along with four Royal Ordnance factories, and seven other factories participated in the MRC’s first well-controlled multicentre clinical trial.
The results of an earlier, small, controlled trial involving sailors had been encouraging; but another among soldiers did not detect any benefit. Despite these uncertainties about the value of the drug this did nothing to moderate enthusiastic reports in the Sunday Express of the drug’s supposed benefits.
In October 1943, after the navy and army trials had finished but before either had been published, the manufacturers of patulin, the Therapeutic Research Corporation, asked the MRC whether it would be willing to conduct a trial on a large scale. The MRC agreed, on condition that no other large trial was done, and set up a committee to plan the study. Professor Harold Himsworth (Head of the Medical Unit at University College Hospital) was appointed chair; Dr Philip D’Arcy Hart, Director of the MRC Tuberculosis Research Unit at Mill Hill*, secretary; and Dr Joan Faulkner (later Lady Doll), assistant secretary.
The committee met in November 1943 and approved a design similar to the smaller navy and army trials. Alternation (or more precisely, rotation) was used to allocate 1,348 Post Office and factory workers to the patulin or placebo groups, but with the twist of using two treatment groups and two placebo groups. This was done to help prevent those involved in allocating patients from knowing or predicting correctly which allocation was next in line; knowing to which group a person is scheduled to be assigned can tempt researchers to depart from the trial schedule. The rotation scheme among the four groups proved very successful, as Philip D’Arcy Hart told me in an interview in 2003:
“We thought we were very clever in setting up a scheme with two patulin groups and two placebo groups using letters to designate each of the four groups, then using rotation to allocate people to the different groups … We thought we were doing something completely new. We wanted to muddle people up. In fact we succeeded in muddling ourselves up. We didn’t always remember what the letters stood for. None of us was a statistician, but we felt that the patulin trial was the first decently controlled trial the MRC had done.”
The trial showed that any beneficial effect of patulin was likely to be small, negligible or non-existent. Indeed, the results suggested that, if anything, the placebo might have been better than the drug.
The report of the patulin trial is a model of clarity. Its introductory paragraphs set out the methodological challenges faced by the investigators and the way these had been addressed — by studying sufficiently large numbers of patients, ensuring unbiased allocation to drug or placebo controls, and by developing a simple, standard way of recording symptoms.
The exceptional methodological quality of the patulin trial was one of the reasons that, a few years later, the MRC asked Philip D’Arcy Hart and Marc Daniels in its Tuberculosis Research Unit and Austin Bradford Hill from its the Statistical Research Unit to organise the multicentre trial of the then limited supplies of streptomycin for treating tuberculosis. The tuberculosis trial used an allocation schedule based on random numbers, to further reduce the likelihood of foreknowledge of assignments among those recruiting patients, an essential feature in controlled trials today. These two trials ushered in an insufficiently celebrated golden decade of pioneering clinical trials done under the aegis of the MRC.
In his article published in the James Lind Library, Professor Mike Clarke sums up the lessons of the 1944 patulin trial for today:
“The MRC trial of patulin, as a possible treatment for the common cold, is an exemplar of researchers, research funders, manufacturers, patients and government working together with a common purpose to pose and answer an important healthcare question. To do this within less than two years seems remarkable today, and is something that everyone currently involved in healthcare research, policy and decision-making would do well to learn from.”
Sir Iain Chalmers
Sir Iain Chalmers is a co-founder of the Cochrane Collaboration and coordinator of the James Lind Initiative. We invited him write about the MRC’s role in introducing randomised clinical trials because, in a paper published last year, he noted that the MRC was ‘curiously silent’ about its important role in developing clinical trial methods in the 1940s and 1950s.
*The MRC Tuberculosis Research Unit was actually in Hampstead, London.