September 7, 2015
All Hail AllTrials
By Michael D. Shaw
Ben Goldacre (from his blog) is a medical doctor, academic, campaigner, and writer whose work focuses on uses and misuses of science and statistics by journalists, politicians, drug companies, and quacks. Among his more high profile targets have been British TV nutritionist Gillian McKeith; anti-immunization ghoul Andrew Wakefield; Brain Gym; and researcher/supplement entrepreneur/conspiracy theorist Matthias Rath.
Perhaps his biggest crusade is directed against what he calls “publication bias,” although he seems to be most concerned with what is more properly termed “positive-results bias,” a particular category of publication bias. As the name indicates, this phenomenon describes the tendency that researchers are more likely to submit—or journal editors accept—positive results, as opposed to those which are negative or inconclusive.
“Publication bias” is a bit more generic, occurring when the publication of research results depends not just on the quality of the research but on the hypothesis tested, and the significance and direction of effects detected. The term is said to have been originated in 1959 by statistician Theodore Sterling, from an article in Journal of the American Statistical Association, entitled “Publication decisions and their possible effects on inferences drawn from tests of significance—or vice versa.”
Here is the pull quote from that old article…
“There is some evidence that in fields where statistical tests of significance are commonly used, research which yields non-significant results is not published. Such research being unknown to other investigators may be repeated independently until eventually by chance a significant result occurs—an ‘error of the first kind’—and is published. Significant results published in these fields are seldom verified by independent replication. The possibility thus arises that the literature of such a field consists in substantial part of false conclusions resulting from errors of the first kind in statistical tests of significance.”
As to Goldacre, he raised a few disturbing issues stemming from positive-results bias, during a TEDMED talk in June, 2012 entitled “What doctors don’t know about the drugs they prescribe.” In this well-constructed presentation, he builds to a climax…
He starts off with the controversial results promulgated in 2011 by social psychologist Daryl Bern of Cornell, demonstrating that precognition really does exist. The work was widely criticized, with respect to methodology, as well as the statistical analysis utilized. Moreover, several researchers attempted to replicate his results, to no avail, but were unable to find a journal willing to publish their negative findings. As it happens, the negative side finally did get published—in the original journal—in 2013, undoubtedly driven to some extent by Goldacre’s TED talk.
Goldacre then moves on to cancer research, citing a 2012 commentary from Nature, entitled “Raise standards for preclinical cancer research.” The key finding is that 47 out of 53 early, laboratory, preclinical studies of potential targets for cancer treatments could not be replicated. The authors suggested, among other things, that:
1. There must be more opportunities to present negative data.
2. There should be transparent opportunities for trainees, technicians, and colleagues to discuss and report troubling or unethical behaviors without fearing adverse consequences.
3. Greater dialog should be encouraged between physicians, scientists, patient advocates, and patients.
Goldacre’s third example is by far the most troubling, involving Lorcainide, an antiarrhythmic agent, developed for recent heart attack victims. After all, such patients are prone to arrhythmia, and it seemed logical to prevent this. So, in 1980, the drug was tested on 95 patients, split into two groups, one group on the drug; the other on a placebo. In the drug group (n=49), nine patients died; in the placebo group one patient died. Based on this, the drug was never brought to market, and the trial results were not published.
Too bad, since some years later, a variety of new antiarrhythmic agents were brought onto the market, in the absence of contrary results, even if the drugs did prevent arrhythmia. Sadly, it is estimated that more than 100,000 people using these drugs died unnecessarily before the connection was recognized. No worries, though. The 1980 researchers eventually published their results in 1993, along with an apology to the scientific community. Thanks, guys!
These and other disasters are why Goldacre is one of the founders of AllTrials, a project launched in the UK, which advocates that all clinical trials should be listed in a clinical trials registry, and their results should always be shared as open data. The motto is “All Trials Registered—All Results Reported.”
On July 29, 2015, it was announced that 50 groups representing 500,000 patients, doctors, and researchers have come together to launch the AllTrials campaign, right here in the USA.
US supporters include the American Medical Students Association, the American College of Obstetricians and Gynecologists, the American Academy of Family Physicians, and patient groups such as amfAR, The Foundation for AIDS Research, the Ovarian Cancer Alliance of San Diego, The Well Project, and many others.
“We provided our bodies, our tissue samples, our data,” says cancer survivor/patient activist AnnaMarie Ciccarella. “I’ve heard the same sentiment expressed many times from patients in clinical trials, ‘This may not help me, but it may help another person.’ It’s time to honor that sentiment.”