My Placebo programme on BBC Radio 4

August 18th, 2008 by Ben Goldacre in bad science, onanism, placebo | 30 Comments »

I’ve been so busy I completely failed to spot that this show went out earlier this evening. It’s a smashing programme I made with Matt Silver from the BBC Radio 4 Science Unit on the placebo effect.

We charge through some of the most fun experiments in the field, and in part two we get all philosophical about what it means for mankind. Here’s the blurb from the BBC:


Our beliefs and expectations about treatments can have a dramatic effect on our health – the so-called ‘placebo effect’. Doctor and writer Ben Goldacre presents a two part series.

Programme 1: The Placebo Effect

When a new drug or treatment is dismissed as being ‘no better than placebo’, we all get the message: any benefits are probably ‘all in the mind’, it’s ineffective, not worth pursuing. Yet studies suggest that the placebo effect can have a significant impact on the course of a wide range of illnesses, including depression, irritable bowel syndrome and angina. It seems that it’s the meaning of a particular treatment to the patient that’s crucial. For example, research shows that the colour of an inert sugar-pill and even the branding on the box, can alter a pill’s effect. In this first programme, Ben Goldacre looks at the growing body of research into the placebo effect, and explores the factors influencing the strength of the placebo response.

Listen again to programme 1

Oh, and, er, here’s a nice mp3 for those of you who don’t like the BBC’s ludicrous inflexible proprietary commercial audio streaming formats:

Audio clip: Adobe Flash Player (version 9 or above) is required to play this audio clip. Download the latest version here. You also need to have JavaScript enabled in your browser.

Placebo BBC Radio 4 Ben Goldacre.mp3

Programme 2: The Implications for Medicine

Studies using placebo or ‘sham’ treatments show that what a doctor says to a patient, along with the ritual of the therapeutic encounter itself, can have a real impact on health outcomes. This raises important ethical issues for those who work in medicine. A doctor’s first commitment is to the wellbeing and health of the patient. Given the undeniable benefits of placebos in the management of many hard-to-treat conditions, can it ever be right to prescribe a placebo without informing the patient? Could complementary therapies, many of whose specific effects are unproven, represent the acceptable face of placebo prescription? Has modern, scientific medicine, with its emphasis on ‘magic bullets’ targeting specific diseases, lost sight of the importance of the ‘art’ of medicine?

Listen again available after Prog 2 Monday 25th August

I’m an enormous fan of BBC Radio 4 science, and I’ve spent a fair amount of time thinking about why they’re so great. As well as resisting the drive to dumb down, there’s also one very important structural factor: around 70-80% of a radio science documentary – by necessity, for auditory colour – is made up of words spoken by the people who have done the science themselves.

This is the kind of unmediated communication which is also so great in blogs by academics, because it is academic scientists explaining things in their own words, and even better, with the help of media people. This is exactly what the print world needs more of: fewer writers, and more editors, helping people who actually know about stuff to express it in a structured fashion for an intelligent audience lacking background knowledge.

Anyway, enough of that nonsense. The programmes’ page is here:

And the listen again link is here:

Part two, same time next week. Y’all come back now, y’hear…

If you like what I do, and you want me to do more, you can: buy my books Bad Science and Bad Pharma, give them to your friends, put them on your reading list, employ me to do a talk, or tweet this article to your friends. Thanks! ++++++++++++++++++++++++++++++++++++++++++

30 Responses

  1. pseudomonas said,

    August 19, 2008 at 1:29 am

    That was very informative – thanks!

  2. used to be jdc said,

    August 19, 2008 at 11:40 am

    Ooh – internal mammary artery ligation and the disprin study look interesting too. I hope your book has more hot placebo action – I’m getting well into this.

  3. JonDurham said,

    August 19, 2008 at 1:36 pm

    Any chance you could tie in a further run of the sold out placebo pills? I sooooooo badly want to have some in the bathroom cupboard.

    And they would make the best Christmas present ever for quite a few people….

  4. Getonyerbike said,

    August 19, 2008 at 8:00 pm

    Instead of lying to patients and tellling them, “Prozac will make you feel better”, doctors should say, more honestly, “If you take this prozac you will feel better”. The empathy inherent in the doctor-patient interaction, the sybolic gift (prescription), the hope (there is something that can be done to treat your depression and what’s more i’ll see you in a week to make sure) and the ritual (taking the Prozac) all contribute to making the patient feel better.

    Most docs know this anyway. The program didn’t make the point that we (docs) do this all the time. It suggested that white coats and pinstripe suits help. And big scanning machines with lights on help to. But so does an awful lot else in (even an average) consultation

    Clearly the word “Placebo” is woefully inadequate to describe the variety of processes at work.

    Would homeopathy be as effective if the practitioners said, “as a result of the acts of empathy, ritual and symbolism, with a dash of hope and good branding (thank you HRH Prince Charles) you will get better. The tablets are an essential part of the process, but they are completely inert” This reminds me of our recently departed agnostic rector.

    I suspect that it would still be effective, though less effective than if they believe homeopathic pills really were active.

    But we don’t live in a rational post-enlightenment age as the increasing popularity of New-Agism, evangelical religion and advertising demonstrate.

    If people need delusions/ faith/ belief then to deprive them is not to liberate them but to enforce an unnatural rationality upon them, so we need to learn a lot more about healing than how to become evidence-based prescribers. We should be evidence-based practitioners by searching for evidence that takes seriously peoples beliefs. The challenge is to do this without deceit, economic exploitation, or dependency. Tough but interesting.

  5. RS said,

    August 19, 2008 at 10:08 pm

    It doesn’t do the doctor-patient relationship much good to tell people “If you take this prozac you will feel better” and have them come back, not better.

  6. Mrs Magpie said,

    August 19, 2008 at 10:35 pm

    I heard this and really enjoyed it but failed to notice the Ben Goldacre bit! Well, OBVIOUSLY the sign of an excellent presenter. I got the facts, without personality intruding or obscuring.

  7. Pro-reason said,

    August 20, 2008 at 5:39 am

    Great radio programme, Ben. I thoroughly enjoyed it. There were a couple of informative details that I hadn’t heard before.

    I appreciate the italics, too.

  8. denbroeder said,

    August 20, 2008 at 9:02 am

    Unfortunately i Can’t listen to the show at work

    One remark: in the discussion the terms placebo effect and placebo response are usen interchangeble but they aren’t. Placebo responses in RCT’s (% of patients responding in the placebo treated group) can be caused by more than true placebo effect. BTW, also in a lot of epidemiolgy textbooks this distinction is not outlined clearly.

    I understand that placebo responses can be caused by…

    1/ Expectation bias in the patient: “true” placebo effect, patient expect to get better and get better

    …but also through…

    2/ Expectation bias in the observer:
    When a phycisian deliberately prescribes placebo, there should be no expectation as to its mechanism of action, unless the placebo itself is expected to work (There is an interesting selffullfilling prophecy..;). For all kinds of allopaths/ CAM proponents it can be argued that they are as misled as their “patients”, hence expectation bias of the observer could well play a role here.
    Interestingly, in an RCT the message given with placebo treatment is “this might be a real drug or it can be placebo” while placebo is used in clinical practice (antibiotics for sinusitis anyone..?Br J Gen Pract. 1997 Dec;47(425):794-9) with the remarks “this will make you better”. Expectation bias is potentially much larger in the latter case.
    All kinds of design flaws can increase expectation bias, like inadequate blinding for example. In RA we have a drug anakinra that has a side effect of big hematomes at the injection sites in 70% of patients. There goes your blinding…(Ann Rheum Dis. 2006 Jun;65(6):760-2)

    3/ Concomittant treatment:
    I’ve been investigator in several RCT’s in rheumatology in which placebo patients were at greater risk to be treated with steroid in the placebo controlled phase due to the fact that disease activity remained high. These so-called protocol violations have to be extensively documented according to GCP guidelines, but are seldomly reported in the final report (e.g. J Rheumatol. 2002 Nov;29(11):2288-98).

    but the most overlooked is..
    4/ Regression to the mean:
    a combination of inclusion criteria that require a certain amount of disease activity (e.g. joint counts in active RA or systolic tension) and natural fluctuations of this value results in regression of the values after inclusion to the mean lower values).

    In RA RCT it seems to be that placebo responses are increasing. We are currently studying this phenomenom. It makes sense, though. Researchers try to enroll as much patients as possible in a RCT (money, authorships) and patients want to be in the trial, so disease activity is overestimated to qualify fo inclusion. After that disease activity drops drmatically in both placebo and treatment groups (e.g. N Engl J Med. 2000 Nov 30;343(22):1586-93). Patienst are pleased too, because in case of response they get to stay on the drug. Finally, high response percentages are good marketing “our drug has a response rate of 70%” is better than “our dug has 50% response” although there may be no difference if placeboresponse was 25% in the former and 5% in the latter. Mentioning attributive risk (treatment response minus placebo response percentages) corrects this but is not commonly reported. Gee….I wonder why.?

    Keep up the good work.

  9. tom-p said,

    August 20, 2008 at 11:24 am

    The Brand name vs generic aspirin and placebo study that’s described at about 16 minutes is partiicularly interesting.
    Does anyone know what the study is or if the brand-name placebo was better than the generic aspirin?

  10. Jamie Horder said,

    August 20, 2008 at 12:49 pm

    Den : Placebo responses are increased in trials of antidepressants too, probably for similar reasons. There’s quite a literature on it – you should read it if you haven’t.

    To your 4 points I would add :

    5. Experimenter demand effects. These are well known in psychology and are the reason why taking part in psychology experiments is often a rather baffling experience. Researchers go to lengths to avoid letting their volunteers know what the hypothesis is because people generally aim to please and will act as they think they “ought” to act. Which is a real pain.

    So in a clinical trial, if a patient thinks that the doctors expect them to get better, then they may decide to report feeling better, even if they in fact don’t. This is not a “placebo effect” since the patient doesn’t actually feel any improvement – however, if you’re using any kind of subjective outcome measure (e.g. pain, depression, fatigue) there’s no way to tell the difference.

    I’ve often wondered how much of the recorded “placebo effect” is of this nature. I’m not sure how you would go about finding out.

  11. Jamie Horder said,

    August 20, 2008 at 12:54 pm

    Sorry: placebo responses are increasING in trials of antidepressants.

    Here’s a paper

  12. David Mingay said,

    August 20, 2008 at 1:36 pm


    Analgesic effects of branding in treatment of headaches. A Branthwaite and P Cooper. Br Med J (Clin Res Ed). 1981 May 16; 282(6276):

    Both “types” of aspirin were better than placebo.

  13. RS said,

    August 20, 2008 at 6:01 pm

    Jamie – I’ve often thought that patients trying to please the experimenter, or even just feeling a social pressure, are likely to inflate placebo effects.

    I too was wondering how you’d examine it – I presume that getting patients to rate comparator pains (experimentally induced pains – with all the limitations that entails) would be more objective, but not comprehensively so.

  14. Jamie Horder said,

    August 20, 2008 at 11:46 pm

    It’s a tricky one. The problem is inherent in the whole research scenario really – if people know they are being studied they act differently. I suspect that the only way to eliminate such effects would be to make it look like you’re not doing research at all. So rather than relying upon reports of improvement you could see if real aspirin actually sells better than placebo “aspirin” (in a different kind of packet) over a long period of time in a fixed group of people who have the chance to try both and learn which they prefer. You could even do something clever with varying the prices, to see how much people are willing to pay for placebo vs. aspirin, and use that to quantify the effects. Good luck getting that past an ethics committee though…

  15. Craig said,

    August 21, 2008 at 6:07 am

    Just on the subject of “scientists speaking direct to the public”:

    I’d heartily recommend ABC Radio National’s Podcasts at

    The Science Show, The Health Report, All in the Mind, In Conversation, Ockham’s Razor, etc. All worth a listen.

  16. tom-p said,

    August 21, 2008 at 9:17 am

    Thanks very much David Mingay

  17. gazza said,

    August 21, 2008 at 9:04 pm

    Well presented programme covering the fascinating ground on placebos discussed by Ben Goldacre previously on this site.

    What was the name of the branded painkiller used in the painkiller trials? I want to write it on my generic aspirin bottle if it’s that effective!

    I’d be interested to know the effect on trial outcomes of subjects believing that there is a significant chance of them receiving placebos? I presume that the placebo effect is non-existant if the subject knows that they have a 100% certainty of receiving a placebo. But what if they are told there is, say, a 50% chance, or 25% chance, of receiving a placebo. Does this ‘doubt’ significantly reduce the effect?

    A nit-picking continuation of studies in this fascinating area perhaps….

  18. used to be jdc said,

    August 21, 2008 at 9:51 pm

    gazza – as someone who is not fond of advertising and marketing, I find it hard to type the name of the product. It is mentioned in one or two previous blog posts though, like this one:

    If you want to know about trials where subjects know they are receiving a sugar pill, Lee Crandall Park might not be a bad palce to start:


  19. gazza said,

    August 21, 2008 at 10:39 pm

    used to be jdc – thanks for the feedback.

    I’ve now got a newly branded pill bottle bought originally at discount aspirin prices – I’m looking forward to the improved potency!

    That paper you flagged up for me describing a small sample of subjects who knowingly received placebos just shows what a complex mind game can be set up between the receivers and the administrators of a pill popping environment!

    So a few of them had improved apparently because they persuaded themselves there really was an active ingredient present; others improved for a range of reasons despite believing the pills were placebo but I would hazard a guess it was to do with the ‘ceremony’ of the interaction between testers and subjects referred to in Ben’s radio programme. Fascinating.

  20. trickcyclist said,

    August 21, 2008 at 11:31 pm

    I have vague memories of a paper that tried to get round the ethical problems with deceiving people associated with giving a placebo by first giving the active drug, and telling them that at some stage they would be switched to placebo, but not telling them when. The upshot was that the placebo worked despite people knowing that they were taking one. Does that one ring any bells with anyone?

  21. Ben Goldacre said,

    August 21, 2008 at 11:36 pm

    it’s an adhd trial, and we discuss it in the second show! i should post a refs list for this show but snowed under, and it’s all in the book, ooh which is out in about a week now.

  22. trickcyclist said,

    August 21, 2008 at 11:56 pm

    Excellent! I shall wait impatiently for both. Will you be doing a tour and signings? Richard and Judy book club, I say…

  23. gazza said,

    August 22, 2008 at 9:07 am

    Presumably it’s as evidence based an analysis as he is used to doing – know the results you want (ie crap book) and marshall the facts accordingly (my opinion of Ben Goldacre is sufficient to tell me it’ll be a crap book).

    If he has friends that know how sulky and bitter that review appears to people I hope they tell him – I’d hate to think he really was like that….

    Why does Amazon already show such a heavy discount on the RRP of the book? It’s half price and not even printed yet! I hope that doesn’t affect the author’s share of the takings.

  24. Maarten Van Hemelen said,

    August 22, 2008 at 1:48 pm

    “Funny he never attacks the many disasters perpetrated by mainstream big pharma.”
    Bad English writing skills aside, this goes to show he has never even looked at Ben’s blog.

  25. used to be jdc said,

    August 22, 2008 at 7:51 pm

    @Ben – “…it’s all in the book, ooh which is out in about a week now.”
    Like trickcyclist, I shall be waiting impatiently.

    @gazza – “Presumably it’s as evidence based an analysis as he is used to doing – know the results you want and marshall the facts accordingly”
    Haha – spot on.

    @Marten – “Bad English writing skills aside, this goes to show he has never even looked at Ben’s blog.”
    I’m glad someone made that comment. Another one that’s spot on. I wouldn’t have dared criticise his Bad English though – particularly given my claim that “Lee Crandall Park might not be a bad palce to start” looking into unblinded placebo provision.

  26. Lemonade Lily said,

    August 24, 2008 at 4:32 pm

    Nick 127 – I was about to try holding my stereo up to the computer when I noticed that Ben has included a link to an mp3 version, which I dutifully downloaded this afternoon. Much easier! It is right in the middle of his article above….

  27. Ian said,

    August 28, 2008 at 11:05 am

    there’s a nice little freeware program that enables you to create mp3 (and other files) from the computer audio output channel – audacity

    Pleeeease can someone put up the second placebo mp3 file as I cannot get bbc streaming audio!!

  28. Yam69 said,

    August 31, 2008 at 6:29 pm

    Look here, Ian:

  29. longyan said,

    October 28, 2009 at 12:42 pm

    The ugg knightsbridge world is a comedy to those who thinks, a tragedy to those ugg classic cardy who feels.Look not mournfully into the ugg boots past, it comes not back again. Wisely improve the present, it is ugg lo pro button thine. Go forth to meet the shadowy future, without fear, and with a manly ugg bailey button heart.

  30. tarling said,

    December 1, 2009 at 10:01 am

    Programme 2 audio here: