“Positive Results Are Better for Your Career”

Brad Stiritz writes:

I thought you might enjoy reading the following Der Spiegel interview with Peter Wilmshurst. Talk about fighting the good fight! He took the path of greatest resistance, and he beat what I presume are pretty stiff odds.

Then the company representatives asked me to leave some of the patients out of the data analysis. Without these patients, the study result would have been positive.

I guess Serpico-esque stories like this are probably outlier stories, particularly when they have happier endings than what Frank Serpico experienced. Or for that matter, Boris Kolesnikov.

Wow—that’s pretty scary! I had cardiac catheterization once!

The Spiegel interview begins with a bang:

SPIEGEL: In your early years as a researcher, a pharmaceutical company offered you a bribe equivalent to two years of your salary: They wanted to prevent you from publishing negative study results. Were you disappointed that you weren’t worth more?

Peter Wilmshurst: (laughs) I was just a bit surprised to be offered any money, really. I was a very junior researcher and doctor, only 33 years old, so I didn’t know that sort of thing happened. I didn’t know that you could be offered money to conceal data.

SPIEGEL: How exactly did they offer it to you? They probably didn’t say: “Here’s a bribe for you.”

Wilmshurst: No, of course not! Initially we were talking about the results that I’d obtained: That the drug that I had been testing for them did not work and had dangerous side effects. Then the company representatives asked me to leave some of the patients out of the data analysis. Without these patients, the study result would have been positive. When I said I couldn’t do that, they asked me not to publish the data. And to compensate me for the work I had done in vain, they said, they would offer me this amount of money.

I recommend you read the whole thing.

P.S. Full disclosure: Some of my research is funded by Novartis.

7 thoughts on ““Positive Results Are Better for Your Career”

  1. One of my kids was a research assistant at one of the key cancer research centers. Bribes are one thing but the issues are deeper and in some ways more troubling. One, the methods followed at each center in a trial are not the same; though the write-ups sound authoritatively exact, the reality is each trial protocol is a huge document which each center implements differently. One hopes the differences aren’t necessarily material, but since clarifications are requested/demanded on important points and it turns out those issues have not been dealt with, the guess is that all gets hidden in the data. Two, data often sits and passes through more than one set of hands before final review. Take mortality results from late stage trials: there are notes from the time and then there’s a chance the materials/tests/notes/etc. sit for months or even years and then that material is analyzed, perhaps without even access to the original note takers, etc. Third, ever negotiate with company reps about what contributed to death? About what side effects were material? My kid has. It’s the company rep’s job to minimize side effects, to maximize their product’s contribution, etc. These negotiations may be conducted by a recent college grad who has had to read the files. Or they may not even be done, but rather the interpretation of the medical issues may be left to the company reps at some centers.

    • Unfortunately sounds all to familiar to my days in clinical research – but given there are company reps means that this sort of thing is at least at risk of being outed by the FDA. The FDA does audits, makes the many results public and discounts/dismisses studies when things are found to be out of line.

      May not seem like much (and explains the efforts to manipulate the interpretations rather than data) but it is better than here http://statmodeling.stat.columbia.edu/2016/07/05/30596/

      Also, many pharma head offices do seem to try to prevent this but managing others who have been highly incentivized to get positive results ain’t easy.

  2. Also of possible interest:

    1. http://bmjopen.bmj.com/content/4/9/e005253.full

    2. Then there’s what happens when the physician “interprets” the results of a clinical trial. Recently, when I accompanied a friend to a cardiologist appointment, the cardiologist said something like, ” If we could clone you, and gave you this medicine, but didn’t give your clone the medicine, you would live longer than your clone.” (Hmmm — what kind of experiment/analysis could warrant this conclusion?)

    • re: 2 –

      I believe this was made possible with the invention of the Farnsworth Parabox: https://theinfosphere.org/The_Farnsworth_Parabox

      Farnsworth A: Wait a second. If everyone is identical, why did our Leela look into the box and your Leela didn’t?

      Leela 1: Well, to be honest, I tossed a coin. It came up tails so I didn’t look.

      Leela A: That’s weird. Mine came up heads, so I did.

      Farnsworth A: Interesting. Apparently, the key difference between our universes is that coin flips have opposite outcomes.

      Bender A: That explains fruity here! [He points at Bender 1.] I tossed a coin to pick my finish. Fog hat grey!

      Bender 1: Hey! Bite my glorious golden ass!

      Farnsworth A: You people and your slight differences disgust me. I’m going home. Where’s that blue box with our universe in it?

      Farnsworth 1: Oh, you’d like to get back to your evil universe, wouldn’t you? And destroy your box with our universe inside it.

      Farnsworth A: Nonsense! I would never do such a thing unless you were already having been going to do that!

    • With regard to 1 (which brings back memories)

      “Our relative inexperience in dealing with large quantities of information and our lack of familiarity with certain trial documents may limit our ability to assess risk of bias in clinical study reports.”

      Clinical study reports are extremely long and tedious to read – a lot of redundancy and apparent focus on unimportant details – some suggest this might be to wear out the reviewer before they discover any possible defects/problems.

      Apparently overly positive phrases do seem to predict what areas need to be look at very carefully.

      Some Pharma companies hire former/retired regulatory reviewers to read the submission to see what they can get out of the documents.

      “non-industry trials where clinical study reports usually do not exist.”
      So at least there are reports to check in industry trials – they are not written to concisely and clearly communicate important findings but to convince the regulator to approve – but they do exist (and regulators can cross-exam the firms involved to pretty much any level of detail desired).

  3. Here is a memorable quotation from the Der Spiegel interview giving us insight into how age changes us:

    “Wilmshurst: … yes, at the beginning they all want to work for Doctors without Borders. Then, five or six years later when they leave medical school, they want to be a cosmetic surgeon.”

    • I’d be interested to see if there are any economics papers which have estimated how much of this effect is due to economic pressures (student loans), how much to socialization effects in med school, etc.

      This might be similar to Alan Krueger (and co-author)’s study showing that students who are admitted to and go to a prestigious college don’t do any better than students who are admitted to, but do not attend, that prestigious college.

Leave a Reply to Paul Alper Cancel reply

Your email address will not be published. Required fields are marked *