 Our first speaker, I'm delighted, I just was able to meet yesterday for the first time. He's a remarkable fellow. Dr. Ben Goldacre is an award-winning writer and broadcaster. He's a medical doctor, and actually, last night we had a small, minor little medical problem at the reception, instantly taken care of, and I noticed we had nine physicians surrounding the person who was feeling poorly, two nurses, a couple EMTs, and I thought, at this conference, the only thing that could get quicker response would be if someone would yell, my laptop won't boot. But Dr. Goldacre is with us today. He writes for The Guardian in the UK since 2003. He's regularly an expert on TV and radio, and we're delighted to have. Please welcome Dr. Ben Goldacre. I had a gig competition with someone last night about who was the bigger nerd, and it was like, you know, PDA battles how well configured your portable devices. And I won by having a small wedgie corner cut out on my right thumbnail so that I can use it as a stylus on the screen. That's one of the most useful things you'll learn in the next 30 minutes. Am I on the screen yet? No. Yes. Ah, so excellent. That's the closest I'm going to get to a magic trick today, I'm afraid. I might be one of the few people on the stage who's never worn a cape, but I can put my tweed jacket around the collar, if you like, and that's the sort of academic equivalent. Right, what I wanted to talk about today is the squabbles that have happened around homeopathy, not just now, but for the past 200 years, really, since its inception. Not because I think homeopathy is the most important thing in the world, because it's not. I teach medical students and doctors on critically appraising clinical trial methodology, or as I like to call it, drug company bullshit. I was very pleased to see that verbatim in the UCL timetable. And what I find interesting about homeopathy is that you can apply the same tools that you use to assess whether a medical drug works to homeopathy, but because the issues are so much simpler in homeopathy, it makes a very good teaching tool. But this won't be my clinical trial methodology lecture. You'll be relieved to hear. And I think there are some very interesting rhetorical issues and historical issues to be found in the ongoing discussions that have happened around homeopathy for the past two centuries. So why would you care about homeopathy? Well, firstly, if I do that, I can see the screen and the microphone. Why care about homeopathy? Well, firstly, it's just interesting and fun to pull things apart and see how they work. That's why we like Linux, that's why we like PDAs, that's why we like thinking about psychics or alternative therapies because when we see how these things really work, then that's often much more interesting than the sort of magic story about quantum distortions in the hysteria field or whatever. And they're interesting even if you can't put them back together again. I think also thinking about alternative therapies is important and interesting because at the moment, particularly in England, alternative therapies have pretty much full spectrum dominance in the newspapers. And by distorting the public's understanding of evidence, by promoting the public misunderstanding of evidence, they undermine all of the good work that's being done by doctors at the moment. And this is a relatively modern phenomenon where doctors try now to work collaboratively with their patients towards an optimum health outcome. That's what it says in the communication skills textbooks that I teach from. But what it means is that doctors are keen now to talk with their patients, discuss the evidence and come to a decision with them. And that requires that people aren't being constantly misled about what a trial means and how evidence works. But honestly, and this is by no means the most important, there are practical consequences obviously. In Britain recently, somebody sent a student to 10 homeopaths and said, look, I'm going to this high-risk malaria area. What shall I do? I'm not very keen on taking malaria tablets. And the homeopaths, 9 out of 10, said, oh, just have some of our homeopathic malaria prophylaxis pills. They'll do you fine. It's not desperately responsible. Well over half of all the homeopaths approached undercover, recommended a mother not to give the MMR vaccine to their child. And the Society of Homeopaths, the biggest EU membership organisation for homeopaths, had a conference in December on treating HIV and AIDS with homeopathy, which is just childish, really. I'm going to try and break this laptop by pulling it up here so I can see it without breaking my airline. If everything goes horribly wrong, I apologise to the owner. Seamless. Okay, so what is homeopathy? Well, it's based on a few basic principles. Firstly, the idea that like cures like. So Samuel Hahnemann, the chap who first concocted the idea of homeopathy, I think concoct or decided, I'd probably the right words rather than discover. He took some quinine and said, oh, well, the symptoms that I'm getting from quinine, which turned out to be an idiosyncratic reaction. But he said, oh, the symptoms I get from quinine are a bit like having malaria. So maybe if you take a microscopically small amount of quinine, then that will help to treat malaria. So like cures like is one principle. But the second and more important principle of homeopathy that everybody would recognise is very extreme dilutions. Now, when we say extreme dilutions, the Society of Homeopaths says 30C, which is the standard homeopathic preparation contains less than one part per million of the original substance. To my mind, that's something of an underestimate. 30C dilution is one in 100 done 30 times over. So that's 100 to the power of 30 or 10 to the power of 60. Or to put it in the terms of the Society of Homeopaths, less than one in a million, million, million, million, million, million, million, million, million. Just for clarity, you understand. To put that in perspective, in an Olympic swimming pool, if you count up all the molecules, that's your job, not mine, you'll find 100 million, million, million, million, million water molecules. So you'd need about 10,000 million, million, million, million of those to have one molecule. If you imagine a sphere that's diameter is the distance from here to the earth, from here, anywhere on the earth, really, you'll do to the sun. Takes light about eight minutes to travel that distance. In that sphere, and one molecule in there, that's roughly equivalent to a 30C dilution. Well, it's 30.89C, just because I know that I'm in a room full of pedants. And if you imagine a universe filled with water, so that's three times 10 to the 80 cubic meters, I'm aware that the chapel next is from NASA, so I don't want to get these figures too wrong. But three times 10 to the 80, that's a perfect space to start a family. Fill out with water, that would only be a 55C dilution, and commonly you can buy 200C dilutions, perfectly standard from homeopaths, and that's more than a number of atoms in the universe. So is the employability of how it works, the way to go in a discussion with a homeopath? Well, it certainly gets entertaining your childish quite quickly. They say, well, you're barking at the wrong tree because water has a memory. And you go, okay, fair enough. But you don't actually give water, what you do is you drop a drop of your water onto some sugar pills, and then you shake the sugar pills in the little pot, and then you hand those out. So what about the memory of sugar? And also you start thinking through the sort of inconsistencies, like you say, well, how does the water remember the arnica that you're trying to treat me with the homeopathic arnica dilution? Why does it remember that? And not the fact that it was once in the coin's bladder. God bless her. Or Hitler's eyeball, or Nelson's barnum. And they say, ah, wow. You see, I made this point in the garden a little while ago. In a long article, it was all about sort of critical trial, critical approach to the clinical trial methodology, but it made a reference to this ridiculous dilution thing and the inconsistencies. And the homeopath wrote in and said, no. The way that the water knows to remember only the special thing that we're diluting in it is because of succusion. You bang it 10 times firmly against a hard but elastic object, ideally a bespoke wooden striking board covered in water and leather. And because you did not mention that, you made us look stupid. So has this line of argument worked over the years? Well, I would contest no. Firstly, it's very easy to get bogged down in detail around the memory of water issue. I know that there's a subtitle, isn't there? The skepticism in the internet age. And actually there are some very good illustrations of how valuable the internet can be in the memory of water squabbles that have happened over the past few years. What happens is periodically a homeopath finds a junior physicist or just some laboratory equipment and will do a study that shows some small, absolutely minuscule difference between two different sets of water, one of which has been prepared homeopathically and one of which has been prepared, which is just normal water. And they'll do this with exquisitely oversensitive equipment. But in fact, nine times out of 10, it's either a false positive due to a failure to correct for multiple comparisons, which I really promise not to bore you with right now. Although if you like the Bonferroni correction lecture, I could give that later in private. It's not very interesting, Bonferroni correction for multiple comparisons. Or they're just very obvious methodological flaws or obvious if you're the kind of guy who knows how to use a, you know, Raymond's spectrometry machine or whatever, because they prepared the water in obviously different ways that we're inevitably gonna get picked up by very sensitive equipment. And what's great about the internet is that somewhere out there, there'll be somebody who can be bothered to go through these claims, go through these papers, critically appraise them and post on their blog, like Spelman, for example, or Donnie Custina's blog, post on their blog, a critical appraisal of these claims of these papers. Recently, the journal, Homeopathy, published a whole string of papers purporting to show the memory of water. And I got in touch with the publisher, I got their permission to reproduce the contents of them online on badscience.net. And we had a journal club, just like you have, you know, at lunchtime once a week in any academic teaching hospital and most academic units. And people went through, they critically appraised the papers and then we summed up all of the criticisms at the end and this also happened on the JRF forums when everybody came together and they produced a letter which they sent to the journal and which was published in the academic journal, which I think is an extraordinary coup, really, for a bunch of jokers on the internet, you know. There's a very good XKCD joke about people being unable to let things lie on the internet. There's a really hot chick and this guy at his laptop and she's saying, come to bed dying and he says, I can't, somebody is wrong on the internet. We need to harness that power. But the most important reason not to bother with this, you know, the memory of water, the implausibility argument, I think is that as rhetorical strategies go, it's had 150 years of fail, which is a word I enjoy using as a noun. I'm in America, this is the home of neologisms. So this chap here is John Forbes. I've modeled my facial hair on him, as you can see. He was the Queen Victoria's physician and in 1846, he was talking about the implausibility of homeopathy. So he is, the blah blah is added by me, homeopathic globules side by side, blah blah blah, would reach from the earth to the sun, blah blah blah, 100 million, million, million, million, million, million. This was 162 years ago and it hasn't persuaded anyone since and he wasn't just talking to himself or Queen Victoria. I mean, this chap, you know, he wrote in medical, he wrote, asked us about medicine in mainstream newspapers, you know, like me, except I'm not the Queen's physician. In fact, the Queen's physician is the director of the Royal Women's and Homeopathic Hospital. So that's how things have changed and we lost the empire. Can they be related? I don't know. But more importantly, the argument about the implausibility of homeopathy is completely irrelevant. There are, I think, dozens of doctors here, which is unsurprising. And lots of women, I'm very pleased to see. I mean, that wasn't a, it's not, I'm not, you know. I mean, the doctors will know. You know, when I was a junior doctor, I have cut a man's abdomen open with my own hands. Supervised, I say to her. I've delved in there, I pulled out his appendix, I've chopped out the dodgy bit, sewn everything up again, and he's been fine. We have basically no really good story for how anesthetics work. We just don't know how anesthetics work. But we do know that they work. And that's the difference with homeopathy. We don't know how it works. But actually, when you look at it, we know that it doesn't work. Or at least it doesn't work any better than placebo. Now this is kind of what doctors do for a living. There's an entire industry of people who do evidence-based medicine, medical statistics, you know, tens of thousands, possibly hundreds of thousands around the country, around the world. I've gotta remember I'm outside of England now. I've never been to America before. I'm temporally, culturally, and morally completely disoriented. So this story of how do you know if a pill works, this is bread and butter. You know, it doesn't get any easier than this. So let's imagine that we're having a chat with a homeopathy fan, a patient who likes it, or a practitioner. And you say, look, we should try and reach an accord, you know, and they say, look, all I know is, I feel better with homeopathy. So you go, great. Okay, fine. So perhaps that could be the placebo effect, because the placebo effect is incredibly powerful and also incredibly interesting. Psebo effect isn't just about taking a sugar pill. Psebo effect is about the cultural meaning of a treatment. So for example, we know that four sugar pills are more effective than two sugar pills at clearing gastric ulcers. This is an extraordinary finding. Gastric ulcers are a pretty unambiguous finding. You stick a camera down somebody's throat, and you look there, and they're either there or they're not. Four sugar pills a day clear gastric ulcers faster than two sugar pills a day. We know that saltwater injections are more effective at pain relief than sugar pills, not because saltwater injections have any pharmacological action on the body, but because it's a more dramatic intervention. The placebo effect is possibly the most interesting and extraordinary thing in medicine. We know that the pale color has an effect. And of course, the pharmacy industry know this as well. That's why, you know, anxiety treatment pills are blue and green, not bright red. And we know that packaging, packaging makes a difference. Classic study, Brownthwaite and Cooper. They did a forearm study where they had either sugar pills or pills with proper painkiller in them. And they were packaged either in blank packaging or in the proper big brand name commercial packaging. And they found there was obviously a significant benefit from having painkiller in the pill compared with just a sugar pill. But they also found a significant impact on pain just from the packaging alone. Now this has important ramifications for everyone in this room because you're all clevered skeptics and you like to walk around saying things like, mom, why do you buy brand name Neurofen instead of own brand? I'd be broken when it's exactly the same molecule in both. But actually your mom's right. Neurofen works better than supermarket brand. I'd be broken. She's right, you're wrong. Embrace that. And that's what being a skeptic is all about. So the placebo effect is extraordinary. Works in animals, works in children. There's a biological story about it. You know, we know that you can set up in an animal a conditioned response between taking flavoured water with an immune suppressant in it and then you take the immune suppressant out, just give the flavoured water and you'll still have the immune system changes. We know that you can give human subjects a placebo sugar pill which they're expecting to be a dopaminergic treatment for their Parkinson's and it will change the amount of dopamine being released in their basal ganglia. The placebo effect is probably the most astonishing thing in medicine. And certainly a lot more interesting than any kind of fairy tale about morphic resonance and quantum stuff in sugar pills. So we go back to our long-yielding theme and we say, look, maybe it's the placebo effect and they say, well, look, all of them know is, I feel better with an amyopathy and you go, fine. So maybe it's regression to the mean. So regression to the mean is absolutely fascinating. We know that basically as new ages, like say when things have a cycle, when your back pain gets bad and then it gets better again, your menopause comes on and it goes away because it's self-limiting. You know, menopause doesn't last forever despite what the manufacturers of HRT might like to try and persuade you. You know, if you have a viral cold, it'll get better. And it's natural to assume that anything you do when your symptoms are at their peak are what caused you to get better. So in the case of a viral cold, if you took an amyopathic pill, you might get better. This is an increase in the number of ridiculousness. If you dangled goats and drills around your neck, you might think, oh, that's what made me get better. Brilliant. Result. Or you might get antibiotics off your doctor, which is even more ridiculous than both because they don't treat viral colds. It is a distraction just because I'm obsessed with this area. The correct thing to do when a patient comes to you with a sore throat that's very clearly a viral cold is to empower them and to say, look, it's fine. Just go home, plenty of water, bed rest. You know, look after yourself. That'll all be great. Doctors prescribe antibiotics to whiny patients just to get them out of the door because we think it's going to get us a quieter session in clinic. But that's actually not true. People have studied that. And if you turn away 100 people who are demanding antibiotics for a viral cold and a sore throat, 33 fewer will believe that they're effective. It's a false belief of which they should be disabused. 25 fewer intend to come back and 10 fewer come back the next year. Now, if you're a homeopath, you might turn this on your head, on its head because you want more trade. You just keep giving out the pills. So you go back to your homeopathy fan, you say, look, maybe it's just about regression to the mean and the placebo. And they say, all I know is I feel better with homeopathy and you think, oh, Christ. Is one person's word ever useful outside of a statistical context? I mean, it's an interesting question. There are cases where anecdotal evidence is useful. Maybe if we had a really clear, kind of ambiguous case of someone getting better from cancer, terminal cancer, would that be impressive? Well, the reality is, we're in Vegas. Amazing things happen. People win huge amounts of money. This entire city was built on a mixture of irrationality and that fact. There's a study here from the Australian oncologists where they took 2,337 patients with terminal cancer. They died on average after five months, but 1% were still alive after five years. If they dangled the goats in trails or had the homeopathic sugar pill, they'd be attributing it to that. But amazing things just happen. So to get through this, you do a trial. And I'm sure everybody knows in this room how to do a trial. You get maybe 200 people, split them into two groups of 100, half of them get the proper homeopathy pills that have been treated with all the hocus pocus. And the other half just get normal sugar pills. This is not a new idea. I mean, people have been doing clinical trials for a long time. I don't know much about Christ or any of that stuff, but I do know that Daniel... Is Daniel Old Testament or new? Old? Okay, good. I was gonna say, I don't know much about Christ, but I do know that Daniel is, and then I thought, I don't. So, and Daniel is Old Testament, and I knew that. And here's the first ever clinical trial on record. It's from the Old Testament, Daniel 1.16. The king was saying, you've all got to eat meat to be good soldiers. And he said, we don't want to eat meat. Just give us vegetables. And they said, no, no, you've got to have meat. And he said, it was this eunuch that was sent from the king. And the eunuch said, look, if I don't make you eat meat, the king's going to do whatever it is you do to a eunuch to make his life even worse. And he said, look, we'll do a trial. We'll see what happens, you know, six weeks later. And six weeks later, they were just as strong. So they were let off the meat. Don't know what was wrong with the meat. So here's something with, there have been 200 trials done on homeopathy. I have like 10 minutes to explain the entirety of evidence based mid-sinty. They're going to do a short slot, tough crowd. Here we go. This is my magic trick. 200 trials have been done on homeopathy. People who know about evidence say, well, homeopathy works no better than placebo. Homeopaths say these trials show that homeopathy works better than placebo. How do you explain this mismatch? How is that possible? Well, firstly, there's cherry picking. People just pick out the positive studies because there are bound to be some positive studies just by the play of chance. But that's not enough to explain it. So you put all the trials together. You do a meta-analysis. You stick all of the trials that have ever been done on homeopathy into one big spreadsheet. Now just to hit, this is the emblem for the Cochrane Collaboration, probably the most important new institution in medicine of the last 100 years. This is a program. God's insanity is trying to explain this. But basically, in the 70s and 80s, lots of people were doing studies looking at whether steroids improved pre-term babies' survival. And all of the studies came out being kind of borderline negative and basically negative. You can see the horizontal lines are all touching a big line that goes down the middle. That means they didn't have a significant positive result. It wasn't statistically significant. Somebody's union already. All of these trials were negative and people were going, well, we don't know if this works or not. So people generally weren't doing it. Then, after 10 years of people ignoring this treatment, somebody came along and did a meta-analysis. That little diamond shape at the bottom doesn't touch the middle line. They showed that in the aggregate, when you pool all of that data together, actually steroids do save lives in pre-term babies. The evidence was always there, but nobody had bothered to add it up in that special way. We had to wait for Excel to be invented or something. I don't know. So these ideas are important. They save lives. But there's even dispute over the meta-analysis because the homeopaths say, oh, the meta-analysis show that homeopathy works better than placebo and people like us say, well, no, the meta-analysis show that homeopathy doesn't work better than placebo. And that's because it's not just about doing a trial. There are good trials and bad trials. I'll skip that slide. This is a lot of trials with quite serious methodological flaws on another day. So what are the kind of flaws you can have? Well, actually, on this one, so third one down, you can see on the right-hand side, there are 41 dropouts. You can see that very clearly. 41 dropouts out of 100 people in a trial. That's a phenomenal dropout rate. We know that when people drop out, they're more likely to have done badly. They might not come back to clinic, for example, because they didn't do one with tablets. They had horrible side effects. They couldn't be bothered to take them. Or they're dead. So lots of dropouts is a bad sign in a trial. Lots of these trials, there were no statistics done. The numbers were all too small. Sometimes you can get things like inadequate blinding. If there's inadequate blinding, then people don't know. People can't know if they're on the tablets or if they're on the real treatment or the fake treatment, right? Because if they know, then that spoils everything. So here is a study on acupuncture. You can see on the left, when people didn't know whether they were getting real acupuncture or sham acupuncture, there was no difference between the two. But on the right, you can see the two bars, if they knew that they were getting real acupuncture or if they were told they were getting fake acupuncture, who's going to get better when they're being told they're being given fake acupuncture? Where's the dignity in that? And that's the problem in all trials. We know that in all trials blinding, failed blinding overestimates treatment benefits by 17%. Or even a bad randomization. It's, you've got to randomize people into the treatment group or the placebo group because if you're allowed to put people in whichever group you want, then that's ridiculous. Obviously you're going to put all the punters and I use the word punter advisedly in medicine because it reflects the element of chance in all clinical contact. If you put, if you get to choose which group you put people in, then you're going to put the patient so you know we're going to do well in the treatment. I'm the patient you know we're going to do badly in the placebo home and then suddenly your trial looks really great. Or you might do it unconsciously. Randomization isn't a normal idea. 1662, that's not bad is it? Why would you ignore an idea from 1662 when it works so well? Or you get dodgy randomization in trials. And there's no point in doing a trial with bad randomization. Alternative therapists of all flavors always say, well you know we don't have the money to do proper trials as if it's really expensive as if it's really difficult. Common sense doesn't cost anything. Randomizing properly doesn't cost anything. Doing trials with inadequate randomization is a waste of money and a waste of the patients who have given their bodies over to you for you to do an experiment on them on understanding that you're going to do an experiment that produces information. That's the ethical transaction. Bad research is unethical. So overall, the dodgy trials, the trials that are methodologically inept, the trials which are not fair tests, the ones on the left-hand side of this graph, show that homeopathy is better than placebo. As the methodological quality of the trials improves, so the result tends towards homeopathy is no better than placebo. And if you look at the meta-analyses that homeopaths quote, they say, exactly the same meta-analyses which I would quote as negative because I only quote the figure for the most rigorous trials, they quote the figure for all of the trials, including the unfair tests. And that is a spectacularly ill and detailed explanation of the homeopaths in the room. Are they lying? Well, I think that's a very interesting question. There's a man called Henry Frankfurt who's a philosopher at Princeton. He wrote a very interesting monograph called on bullshit. He says, it's impossible for someone to lie unless they know the truth. The honest man says what he believes to be true. The liar considers his statements to be false. For the bullshit, however, all bets are off. He's living on the side of truth, not on the side of the false. The bullshit seeks only to impress us. These trials for homeopaths are window dressing. But more importantly than anything to do with homeopathy, the background issues, the placebo effect is amazing. It's amazing. It's so much more interesting than homeopathy. People with terminal cancer sometimes live. And that's amazing too. Evidence-based medicine is extremely interesting as you can see there. You enjoyed a bit about the blubbergram and the big diagrams of meta-analysis. And lastly, skeptics and homeopaths have been having the exact same score goals for 150 years. Thank you very much.