How Drug Companies Distort Science: Q&A with Ben Goldacre

What you don't know about how drugs are tested and marketed could hurt you, says author Ben Goldacre in his book Bad Pharma

  • Share
  • Read Later

Think your doctor gets all the scientific evidence on a drug before it gets to market? Not necessarily. Half of the research data on drugs is not readily available to physicians, as the U.K.’s Ben Goldacre reveals in his book, Bad Pharma: How Drug Companies Mislead Doctors and Harm Patients. He discussed the problem— and proposed some potential solutions.

What led you to write this book?

These are all very well-documented problems in medicine. They’ve only really been described in the technical professional literature. I wanted to write about them for a broader audience. If you look at the problem with missing clinical trial data, we’ve known about this for three decades now. And we failed to fix it behind closed doors.

Exactly how much data is missing?

Overall, for the treatments that we currently use today, the chances of a trial being published are around 50 percent. The trials with positive results are about twice [as likely] to be published as trials with negative results. So, we’re missing half of the evidence that we’re supposed to be using to make informed decisions. [And] we’re not just missing any old half, we’re selectively missing the unflattering half.

What can be done about this?

After the book came out [in the U.K.], a Parliamentary Science and Technology Select Committee inquiry was announced to look at the problem of missing clinical trial data. Other Parliamentary inquiries started looking into the problem. We realized that this needed some kind of organizing force, so along with the British Medical Journal and Sense About Science and The Centre for Evidence Based Medicine at Oxford University, I set up a campaign called, asking for three things.

First thing, we want to know about all the clinical trials that have happened on all of the medicines that we use. Including the previous ones. And that’s very important. Not just all the ones from now, because that won’t fix anything for another 25 years.

We need to know [about past hidden trials] because 80% to 85% of all the prescriptions issued this year were for treatments more than 10 years old. So, we’re asking for just knowledge of all the trials that have been done, just to know that they exist.

We want to have the basic summary results and we want, where possible, the full clinical study report. No individual patient data, but the full details of exactly what was done and what was measured. Now, that’s an entirely reasonable thing to ask for. [But] industry in some corners has been up in arms.

We’ve now got support from around 100 patient groups, … academic research [and] also GSK, GlaxoSmithKline, one of the biggest drug companies in the world. [Of course], with a checkered history, a three billion dollar fine for criminal and civil fraud just last year, including withholding data. They’ve made previous comments and promises of increased transparency. [But] what they’ve promised us by signing up with goes further than any of that, to my mind. [Note:  Roche agreed this week to release more data as well].

Has GSK actually started giving you the data yet?

No. This is a reasonably big ask and they’re going to be delivering over the course of the next year of two. People say, how do you know they’re going to deliver? We’ll be watching.

Many people think this problem has already been solved. The medical journals said they weren’t going to publish trials that hadn’t been registered and there were steps by the U.S. government to require it.

This area is absolutely drenched in fake fixes that have actually perpetuated the problem. In 2005, the International Committee of Medical Journal Editors made a promise that they would never again publish a clinical trial unless it had been properly registered on a publicly accessible list of all the trials that were being done.

But unfortunately there was no routine public audit and when one was finally done, many years after this rule came into play, we discovered that academic journal editors hadn’t kept that promise.

[A study published in a major journal] found that half of all the trials published in the top ten journals in the big five fields of medicine weren’t properly registered and [many] weren’t registered at all—and that’s only the ones that we know about.

Why didn’t they keep that promise?

I would say three things. I’d say it’s a combination of chaos [due to medical editors often being academics working part-time for no pay for this work], not taking the problem seriously, financial conflicts of interest, and other conflicts of interest.

Future generations will look back with amazement on this. They will say, ‘You’ve spent tens of millions of dollars, sometimes hundreds of millions of dollars on one trial, to make sure it was as free from bias as possible, to sometimes detect very, very modest differences between two treatments. You went to all of this effort to exclude bias, and then at the final stage, you were perfectly happy to just throw away half the data and not any old half, the biased half of the data. People will look back on this in the same way that we look back on the medieval leeches and bloodletting by doctors.

Some people might take away from this that relying on mainstream medicine may be no better than using unproven alternative treatments.

I don’t think that’s true. My personal judgment is that overall it’s fairly uncommon, though not impossible, for treatments to be widely used that are completely useless or worse than useless. What’s much more common is that we’re misled about how useless something is and also which is the best out of several available treatments in one class.

You write about many ways that drug companies can distort data, including creating trials that make their drug look good by using only a low dose of the competing medication.

It’s certainly not as big a problem as missing clinical trial data but there really are a million different [ways] that I just think [are] brilliant, cruel, and interesting, and funny, how people rig trials by design, in such a way that they get biased and exaggerated benefits for treatments.

How has the drug industry responded to the book?

In the U.K., the book came out a couple of months ago and the response from the [British] industry’s representative body has been that all of these problems are historic and they have all been fixed. Now, that is an extraordinary and completely implausible blanket denial. They’re denying the existence of things [that are well documented in the deceptive design of trials] that are taught in the core curriculum of undergraduate medical education.

To say that that stuff doesn’t exist, is just absurd. It’s like saying that the kidneys don’t exist. It’s completely infantile and frankly, rather obscene. It’s part of a deliberate strategy to prevent public scrutiny. Delay is their product, right? To rewrite the quip from the tobacco industry that doubt is their product.

The intention of the pharmaceutical industry in my view is simply to delay the public becoming aware of the problem and professionals and policymakers addressing the problem and they will defend what they believe is their right to withhold unflattering clinical trial data. They will defend that harder than they defend anything else. You watch.

Marketing, payments to doctors, all of that stuff, they’ll fight on it but to hide the unflattering results in clinical trials, they will fight viciously, tooth and nail, because that is more important to them than anything else. If you can poison the well of medical evidence, then you are made.

But isn’t it in their best interest to actually know if their stuff works or not?

It’s not in their interest for everybody else to know. [That’s why they fight against] head-to-head tests [or comparative effectiveness research]. If you think about it from the perspective of basic economic theory, the more perfect information we have about the relative efficacy of two different treatments, the more accurately we can determine the value of the treatment. As soon as you can perfectly determine the value of the treatment, it has basically become something much like a commodity and the thing that we know about commodities, in marketplaces, like wheat or sugar, is that it’s very difficult to make a profit out of it because they have a price and that is the price.

Industry’s worst nightmare is us having perfect information about which is the best treatment.

How can we improve on the drug development and marketing system?

If we had any sense, we wouldn’t just spend our money on doing individual clinical trials that cost huge amounts of money. The thing that we need to invest in is better information architecture for evidence-based medicine—infrastructure like what we have for sharing x-rays across town. We need infrastructure for drawing together all of the evidence that we already have, summarizing it, getting it to the right doctor, at the right time, to help inform their decisions and make sure that we optimize the use of the evidence that we already have. Now, that sounds really boring, right? But that is the single most important flaw I would say, in the whole of medicine.


There was an error in the section in response to "Why didn't they keep that promise?" The part talking about removing bias and then throwing out data because it's biased is a contradiction. From the context, I assume he meant that the data were thrown out because the results were negative or did not support the alternative hypotheses. The throwing out of results from anything but systematic errors in experiments (which are corrected before running more trials) introduces selection bias. Hence the contradiction.


Capitalism without ethics is one of our biggest problems.