Sir Austin Bradford Hill’s own bouts with illness inspired him to set the standard for safety testing.
by Dr. Robert E. Marx
The standard test protocol for safety and efficacy of new drugs and treatments, the randomized, controlled trial (RCT) was introduced by a professor of medical statistics at the archaically named London School of Hygiene and Tropical Medicine.
Austin Bradford Hill had a special interest in tuberculosis, which was at the time the most common lethal infectious disease in developed countries. As a young man, he had to abandon his aspiration for a medical education when he developed tuberculosis in 1917. Although most tuberculosis patients of that era succumbed to the disease, he miraculously responded to the only available treatments at that time–bed rest and induced collapse of the lung–and lived to the age of 93.
His personal tuberculosis history and his appointment as Professor of Medical Statistics came along with the introduction of streptomycin, which was shown by US researchers in 1944 to be capable of killing the tubercle bacillus. To set up a properly conducted trial, Bradford Hill then joined the British Tuberculosis Trial Committee in 1946.
As in many medical studies, serendipity assisted the proceedings when the supply of streptomycin was limited by financial constraints. There was literally only enough streptomycin available for half of the patients in the study. Bradford Hill convinced the trial committee that the choice of who would receive streptomycin should be made by a fair random chance, and those who did not receive it would act as a control group. Although this was not the first RCT comparing the outcomes of two groups (there had been at least two previous studies), it was the one that established the value of a planned RCT. In it, 55 patients were given streptomycin for four months, and 52 were treated with the standard bed rest and , if needed, an induced pneumothorax.
The results of this landmark study are intriguing, because they show one of the great values of a properly conducted RCT–and one of its fallacies which is too often duplicated in today’s RCTs. After six months, 28 of the 55 patients (51%) who received streptomycin had improved significantly, and only four had died (7%). None of the 52 patients randomized to the control group had improved, and 14 (25%) had died. Bradford Hill reached the legitimate conclusion that streptomycin was effective in treating tuberculosis, and that the improvements in the patients could be attributed only to the drug.
However, the validity of this conclusion was soon tempered when the condition of all the streptomycin recipients began to deteriorate. An unforeseen drug resistance had developed. The enthusiasm generated by the results at six months required an updated publication three years later. The track record identified that 32 of the 55 streptomycin recipients (58%) had died, and 35 of the 52 controls (64%) had also died–a nearly equal outcome.
That outcome is eerily similar to the Fen-Phen, Vioxx, and bisphosphonate track records, where the early therapeutic advantages were later reconsidered due to serious side effects and complications. Bradford Hill’s approach made sense in theory and was adopted as a reasonable way to evaluate the flood of new medicines and therapies that arose in the 1950s. Indeed, its ascendancy to gold standard status occurred in reaction to thalidomide-related birth defects that emerged in 1960. Governments began demanding that all new drugs be tested for safety and efficacy through RCTs before a product license would be granted. Nevertheless, this failed to prevent the thalidomide-like tragedies of recent times, involving Vioxx, Fen-Phen, intravenous bisphosphonates, and now oral bisphosphonates.
In 1950, Sir Bradford Hill, together with British physiologist Sir Richard Doll, published the first paper demonstrating a link between cigarette smoking and lung cancer. He also pioneered, in 1965, several aspects of causality in medicine which are still used by epidemiologists today.