Archie Cochrane will be remembered for many things, but perhaps more than anything for his enthusiastic promotion of the randomised controlled trial in clinical research, and for his call for periodic overviews of such trials. Yet Archie himself never conducted a randomised controlled trial!

• Download figure
• Open in new tab
• Download powerpoint

In his paper “Sickness in Salonica: my first, worst, and most successful clinical trial”1 Archie described a trial of yeast in the reduction of starvation oedema in 1941. His experimental subjects were 40 prisoners of war, all of whom were “emaciated above the waist and had pitting oedema to above the knees”. These men were allocated alternately to two huts. Each of the men in one hut received two spoonfuls of yeast daily. Men in the other hut received a tablet of vitamin C daily. The outcome measure was the number of buckets of urine carried out of each hut each day. By the fourth day there was a clear difference in favour of the hut in which the men had received yeast. In his paper Archie admits that this was not a good trial. The wrong hypothesis had been tested, the numbers were too small, the time too short and the outcome measurement was crude. Although Archie states “it could be argued that the trial was randomised”, it was not. Nevertheless, as the title of the paper indicates, the trial was successful!

I persuaded Archie to give a talk about this trial at a dinner I had arranged for speakers and others involved in a conference on byssinosis. He was most reluctant, saying that he had never been able to come to terms with his experiences in the prisoner of war camp. However, he agreed. As he talked about the men and the conditions under which they had had to live in the camp, Archie faltered, and nearly broke down emotionally. Afterwards, he thanked me, saying that talking about the trial had been very helpful.

Very shortly after this, Archie wrote the paper and sent it to theBMJ.