The Evolution of Controlled Trials Before the Middle of the Twentieth Century

Speaker: Sir Iain Chalmers (Editor - James Lind Library)

Contrary to widely believed assertions, the concept of unbiased creation of treatment comparison groups in clinical trials was not ‘a seminal statistical idea’, but was rooted in the much older idea that fair tests of treatments involve comparing like with like, achieved at least 200 years ago using alternate allocation to treatment comparison groups. Strict alternation deals with selection bias as effectively as strict random allocation, but alternation is more likely to result in foreknowledge of allocations among those recruiting research participants. The historical importance of the iconic MRC streptomycin trial published in 1948 is not its use of random sampling numbers to generate the allocation schedule. Rather, it is its clear description of the precautions taken to conceal the schedule, and so secure unbiased allocation. Several hundred reports of controlled trials published before 1948 have already been identified. However, without multilingual, collaborative research the history of the evolution of the controlled trial, a crucially important technology in medical research, will remain seriously incomplete.