Why is it controversial to tell the truth about health care?
Peter Gøtzsche, physician, medical researcher and leader of the Nordic Cochrane Centre at Rigshospitalet in Copenhagen, Denmark, received the 2016 HealthWatch Award at the charity's 28th Annual General Meeting on Thursday 20th October 2016 at the Medical Society of London.
This article is adapted from a transcript of his talk.
People ask me, why do you look for controversies? And I tell them, I don’t, they come to me. My work is something like that of a medical detective. People come to me if they feel something is wrong in healthcare. When I start looking into these issues, I usually dig very deep. I find skeletons, and when I expose these skeletons, the people who buried them can get very angry.
Even in my most routine work, I dig deep and I still find skeletons.
I started digging many years ago when I looked into drug trials of the NSAIDs – arthritis painkillers. These were head-to-head trials, comparing one drug with another. I collected every trial I could find, and used meta-analytic methods to find out if the results were plausible. I found that they were not, in a statistical sense, and I published my conclusions in my thesis. Doug Altman, professor of statistics in medicine at the University of Oxford, saw my work and in 1990 he sent it to his colleague Iain Chalmers. I’d published three meta-analysis by this time and felt I was quite prolific, but Iain told me his group had published hundreds in Cochrane, a database I’d never heard of. He invited me to Oxford for the 1992 opening of the UK Cochrane Centre, of which he was the director. In 1993 we opened the Nordic Cochrane Centre and four years later I became a full time researcher there, reasoning that as a doctor I could only help one patient at a time, but if I did a good review I could help one hundred thousand.
Shocking results were not unusual
It was felt that my results with NSAIDS had been so shocking that it must be something that only applies to arthritis drugs. But this was the first time anyone had trawled through a whole medical area and done a statistical overview, and we would soon find out that the phenomenon was not unusual at all. In 1997 my wife Helle, a clinical microbiologist, did a study on antifungal agents for people with, for example fever, cancer, or neutropenia, who were at risk of dying from a fungal infection. We did a routine comparison between fluconazole and Amphotericin B, the two most promising drugs at that time. We found that some of the trials were three-armed, with a third arm consisting of nystatin, which is effectively no more than a placebo. The manufacturer, Pfizer, had lumped the data from a drug that didn’t work, with amphotericin B to show that their drug, fluconazole, was better than when you handicapped your comparator. What is more, they had also given amphotericin B by mouth, even though it is poorly absorbed and not prescribed orally with these patients. We tried to talk to the investigators about our results, and they referred us to Pfizer, who wouldn’t answer our questions.1
We couldn’t even see whether one trial publication was re-using the same patients that were included in another trial publication, so we might be counting the patients twice. We asked Pfizer about that, too, and they didn’t reply. We published it in JAMA2 and Helle presented it at a conference on microbiology in Toronto in Canada. There were 800 people and her presentation went way over time, not because she over-ran, but because there were so many questions. Drummond Rennie, one of the JAMA editors, wrote in an editorial that “fluconazole raced against a heavily handicapped opponent”.3 In a response letter to JAMA, Pfizer conceded that the comparison drugs did not work and that it would change how it reported results of trials.4
So, we ended up in the New York Times because we’d reported misleading research being published by the biggest drug company in the world.5 Two years later we were there again.6 My PhD student had studied the placebo effect. We’d thought it could be powerful, as did most doctors. But most doctors make the mistake of asking, what is the placebo effect? And then attempt to answer it by observing what happens to a group of patients treated with placebo. That is wrong. Think about regression to the mean, and the way many conditions heal by themselves.
We couldn't find evidence for placebo effect
Well, that is NOT placebo. To study placebo, you need to take an untreated control group, and compare that with the placebo group, then in most trials you can also have a third group, using an active drug. We collected results from 113 trials, we published in the New England Journal of Medicine and the Cochrane Library. We couldn’t see much effect with placebo. Maybe a little effect on pain, but we didn’t even know if it was a true effect because you cannot blind an untreated control group, and when you measure subjective effects people tend to hope that they’ll get some benefits because they think they are receiving a drug. So, we couldn’t find evidence for placebo effect. It was terrible for the placebo community and made a lot of work for us, dealing with questions and criticisms.
There is too much unreliable research published by people in order to sustain their erroneous beliefs. So it is fundamental that we try to persuade politicians that we cannot leave people to be their own judges. I can’t go to the car inspection with my 18-year-old car and reams of paper and say, I’ve already done the inspections myself, no need for you to do it! Yet that is how we approve drugs. I have worked a lot with psychiatric drugs, and I published a book last year on the subject,7 and there will be documentaries to follow. People are starting to realize that psychiatric drugs are not the solution to psychiatric problems, in fact they make them worse. They disable the brain. People may think it’s useful that they do this because the effect is to sedate the patient, but they also cause many side effects. The clinical trials in psychiatry are the worst I have ever seen. And the more psychiatric drugs we use, the more people end up on disability pensions because they can’t work. There is more and more pressure to establish helplines for people who want to come off psychiatric drugs, so things are moving in the right direction.
On mammography screening: I think it should be stopped. In 1999 a Swedish study compared regions which had taken up mammography screening early, and they compared them with other regions that hadn’t implemented it yet. They couldn’t find any difference in mortality.8 At that time we had screening in 20% of Denmark. The Danish Medical Association asked the Danish Board of Health, can’t we trust the randomized trials which said the screening worked? That hot potato landed in my lap. We were asked to review the randomized trials of mammography screening. Five weeks after I had this assignment, there would be questions in Parliament about whether we should have screening.
A real "Yes, Minister" moment
We worked fast – in four weeks we had produced a report, with meta-analysis, and it concluded that we cannot exclude the possibility that mammographic screening does more harm than good. We took it to the National Board of Health and were immediately called to a meeting, with the result that two days later we were told that our paper was a non-paper. It didn’t exist. This was a real “Yes, Minister” moment. Fortunately, we were able to prove that it really did exist, because we’d sent it to the Danish Board of Health by messenger and its delivery had been recorded! I was told to change my letter, to say there had been misunderstandings, and to say that the report was only preliminary. I didn’t accept that. A promise was made to send it to the Minister of Health. It never arrived.
Two weeks later the scandal broke loose. Journalists got hold of both copies of the letters, my original and one that had been changed and which I had not signed. Both on a full page of the newspaper. As you all know, “Yes Minister” tells you how government discredits an unwelcome report. Step one – refuse to publish in the public interests, saying you are waiting for the results of a more detailed report, still in preparation. If there isn’t one, commission it! And that’s what happened. We were told to prepare a Cochrane review about this, and our original report would not see the light of day. So, we did a little more work, we published our findings in the Lancet,9 then two years later we published the Cochrane review which, naturally, confirmed the original findings. But we were blocked.10 [Editor’s note: the paper was removed completely from the journal website without any formal retraction. The authors later published it in the Danish Medical Bulletin11] You can see why, when life gets tough, I like to watch “Yes Minister”.
I published another paper in the Journal of the Royal Society of Medicine last year, calling for mammography screening to be stopped.12 What does breast cancer screening mean for total mortality? Mortality rates with and without mammography screening are identical. Except, of course, if you include deaths caused by screening overdiagnosis – if you do that, you find that screening increases mortality. Why deaths by overdiagnosis? Well, if you are diagnosed with breast cancer you get chemotherapy and radiotherapy, and that kills some women. Do you think you could have sold mammography screening if this was what you’d shown to the politicians 25 or 30 years ago? And because the NHS leaflet on breast screening was no good, we produced a new leaflet and published it in the BMJ, it’s been translated into many languages and it’s on our website.13
Screening is the best way to earn money, if you are a private doctor
But of course screening – to investigate and find illnesses before people even know they are ill – is the best way to earn money, if you are a private doctor.
So I thought we should also look into health checks, looking at morbidity and mortality from disease. To our big surprise there were already 16 randomised trials with many participants and long follow-ups – 9 years – and almost 12,000 deaths, so much material! And what did we find? No effect whatsoever. Total mortality, around 1, and the same for cancer, cardiovascular mortality. We published in the BMJ.14 Then a huge Danish study was published after our review, that one included 3,000 deaths, and it didn’t find anything either. Extensive screening for risk of ischaemic heart disease followed by repeated lifestyle counselling, a beautiful ten-year study that was also published in the BMJ.15 So there are no lives saved by screening but there is a lot of harm because you make diagnoses and then treat them with drugs that might make your patients impotent, etc., when they might never even have noticed the symptoms of the original condition or just put it down to old age.
Look at England – there is a universally applied health check for everyone aged 40-74. It was launched in 2009. The Department of Health argued that by spotting people who are at risk from heart attack, diabetes, stroke, kidney disease, these conditions could be prevented. Finally we’d had enough and wrote to the Times, who published on the front page to say that health checks are utterly useless.16 They now claim that, although the programme is not supported by direct evidence from randomized controlled trials, there is an urgent need to handle the growing burden of disease associated with lifestyle behaviours. Six months later, NICE was helping local authorities to encourage people to attend health checks and to support them making changes to improve their health.
Now, this is how the mob operates. You don’t ask the boss questions. You just do what you are told, otherwise you might get a bullet in your head. NICE has prostituted itself to the politicians’ message. This is a side of medicine that students aren’t taught about.
I am kind of fed up with being introduced all over the world as controversial. Why is it controversial to tell what you find when you try to do good science?
Peter Gøtzsche’s books include Deadly Psychiatry and Organised Denial (People's Press, 2015); Deadly Medicines and Organised Crime: How Big Pharma has Corrupted Healthcare (Radcliffe Publishing, 2013); and Mammography Screening: Truth, Lies and Controversy (Radcliffe Publishing, 2012).
The HealthWatch Award is presented annually to an individual who has made significant steps either in medical research or in improving the public’s understanding of health issues by clarifying complicated and often misunderstood medical matters for the general public.
1. Gøtzsche P. Deadly Medicines and Organised Crime. Radcliffe Publishing, London, 2013
2. Krogh Johansen H, Gøtzsche PC. Problems in the Design and Reporting of Trials of Antifungal Agents Encountered During Meta-analysis. JAMA. 1999;282(18):1752-1759 http://jamanetwork.com/journals/jama/fullarticle/192088
3. Rennie D. Fair Conduct and Fair Reporting of Clinical Trials. JAMA. 1999;282(18):1766-1768 http://jamanetwork.com/journals/jama/fullarticle/192075
4. Panzer H. Improving the Conduct and Reporting of Clinical Trials. JAMA. 2000;283(21):2787-2790 http://jamanetwork.com/journals/jama/fullarticle/1030783
5. Grady D. Medical Journal Cites Misleading Drug Research. New York Times, 10 Nov 1999 http://www.nytimes.com/1999/11/10/us/medical-journal-cites-misleading-drug-research.html
6. Kolata G. Placebo Effect Is More Myth Than Science, Study Says. New York Times, 24 May 2001 http://www.nytimes.com/2001/05/24/us/placebo-effect-is-more-myth-than-science-study-says.html
7. Gøtzsche P. Deadly Psychiatry and Organised Denial. People's Press, London, 2015
8. Sjönell G, Ståhle L. Hålsokontroller med mammografi minskar inte dödlighet i bröstcancer. Läkartidningen 1999; 96: 904–13.
9. Gøtzsche PG, Olsen O. (January 2000). "Is screening for breast cancer with mammography justifiable?". Lancet 2000;355:129-134 http://www.thelancet.com/journals/lancet/article/PIIS0140-6736(99)06065-1/abstract
10. Zahl PH et al. WITHDRAWN: Results of the Two-County trial of mammography screening are not compatible with contemporaneous official Swedish breast cancer statistics. European Journal of Cancer 2006 Mar 9 [epub ahead of print] https://www.ncbi.nlm.nih.gov/pubmed/16530407
11. Zahl PH et al. Results of the Two-County trial of mammography screening are not compatible with contemporaneous official Swedish breast cancer statistics. Danish Medical Bulletin 2006;53:438–440 https://web.archive.org/web/20070621235925/http:/www.danmedbul.dk/Dmb_2006/0406/0406-artikler/DMB3890.htm
12. Gotzsche PG. Mammography screening is harmful and should be abandoned. J R Soc Med 2015 Sep;108(9):341-5. http://journals.sagepub.com/doi/abs/10.1177/0141076815602452
13. Nordic Cochrane. Mammography screening leaflet, 2012. http://nordic.cochrane.org/mammography-screening-leaflet
14. Krogsbøll LT et al. General health checks in adults for reducing morbidity and mortality from disease: Cochrane systematic review and meta-analysis. BMJ 2012;345:e7191. http://www.bmj.com/content/345/bmj.e7191.long
15. Jørgensen T. Effect of screening and lifestyle counselling on incidence of ischaemic heart disease in general population: Inter99 randomised trial. BMJ 2014;348:g3617 http://www.bmj.com/content/348/bmj.g3617
16. Smyth C. NHS checks on over-40s condemned as ‘useless’. The Times, 20 August 2013 http://www.thetimes.co.uk/tto/health/news/article3847530.ece