Randomized, Controlled Trials—Not What They Were in 1946—Need Restructuring
Sir Austin Bradford Hill, the father of the randomized, controlled trial (RCT), was neither a physician nor a trained statistician. He was, however, mentored by Dr. Major Greenwood at the London School of Hygiene and Tropical Medicine, where he became Professor of Medical Statistics after his mentor’s retirement.
Since their development in Sir Bradford Hill’s time, RCTs have become accepted as the highest level of scientific evidence, and they even maintain an air of infallibility in some spheres. However, professionals and the public at-large are well-reminded that Fen-Phen, Vioxx, both intravenous and oral bisphosphonates, and others were cleared by the FDA, relying on such RCTs, and the drugs nevertheless went on to cause serious adverse effects. Patient suffering, even deaths; the results of missed signals in their respective market clinical trials.
What went wrong? In large part, the problem has been the trials’ sponsorships by the very pharmaceutical companies seeking positive outcomes for their drugs. The respect–and therefore the level of evidence–attributed to RCTs has been corrupted by a reliance on drug company sponsorship.
Sir Austin Bradford Hill’s original study on the tuberculosis medication streptomycin was not sponsored by the producers of streptomycin, and Sir Bradford Hill duly reported the failures of the drug when the tuberculosis recurred. He also wrote and conducted the study. He published his initial results and later those of the drug failures, with no help or financial support from the drug manufacturer or distributor. (Compare that to today, when the FDA enlists assistance from the pharmaceutical companies, even when the task is the mere wording on a warning label.) Sir Bradford Hill did not attempt to hide drug failures with repetitive denials, attributing recurrences to the comorbidities and confounding variables of that time, which included inadequate ventilation, inadequate sanitation, and poor working conditions.
As a matter of practical necessity (born, in part, from a lack of political will to properly fund the agency), the FDA cautiously accepts drug company-sponsored studies. But the medical and dental communities of clinicians who must treat patients who entrust them to make informed decisions certainly must not. Trials sponsored by the pharmaceutical companies, then, ought no longer be valued as the highest level of evidence, though to date they have been. The more appropriate levels of evidence are as follows:
Level I: Unsponsored, randomized, placebo-controlled trials.
Level II-1: Unsponsored, randomized, open-label trials.
Level II-2: Unsponsored, well-designed, non-randomized or cohort studies.
Level III-1: Unsponsored, multiple-series studies.
Level III-2: Retrospective studies.
Level IV: Sponsored, randomized, placebo-controlled studies and sponsored randomized open-label studies.
Level V: Opinions of respected authorities based on their clinical experience, descriptive reports, or reports of expert committees.
The scientific medical community ought to be perfectly outraged over the slow and insidious corrosion of reliability in results and conclusions published by the drug companies today. In so writing, I don’t intend to indict randomized clinical trials, per se. Rather, I plead for a return to what they were meant to be originally, as set out by Sir Bradford Hill; written by the principal investigator, unsponsored, and unbiased.
Copyright 2015 Dr. Robert Marx. All Rights Reserved.