Should we trust medicine at all? with Jacob Stegenga
46 minAudio
Send us Fan Mail In this episode, Prof. Jacob Stegenga (Nanyang Technological University), author of the book, Medical Nihilism, chats with us why we should all be more sceptical about medical interventions. Jacob discusses the placebo effect, research malleability, and publication bias. Sinead and Kat ask how we can know when we are being misled by journals and media reports, and encourage Jacob to release his music on Spotify. Remember people, be savvy!
Audio and video play straight from the show's own hosting. PodSubs stores and serves no media.
Transcript
From the show. Tap a time to play from there.
SPEAKER_04The center presents a new podcast, C BME and You, a podcast devoted to promoting awareness and debate around controversial ethical issues in biomedicine.
SPEAKER_02Welcome back to CB Me and You, podcast about biomedical ethics. Joining us today is Professor Jacob Stegener, Professor of Philosophy at Nanyang Technological University here in Singapore and author of the book Medical Nihilism. Welcome, Jacob.
SPEAKER_00Thanks a lot. Great to be here.
SPEAKER_02But in his replacement, we have the honorable Kat Meiskins from CBME. Yeah. I'm here again. Yes. Kat, would you like to do that on us?
SPEAKER_03Yes. We have our traditional icebreaking questions. First of all, coffee or copy?
SPEAKER_00This is an embarrassing question for me because I I think I'm just too shy to even try ordering Kopi. Oh no. Yeah, yeah. I'm intimidated by the entire process. So I I've tried it once. I didn't love it, to be honest. So I guess I'm more of a coffee person.
SPEAKER_02Was it too sweet?
SPEAKER_00A bit too sweet. Yeah. And I don't like the this um instant coffee too, right? Which is also that's just not always instant, is it?
SPEAKER_03Um not always.
SPEAKER_02Okay. Yeah.
SPEAKER_00Okay.
SPEAKER_02Um it's not. We've got wacky on going here, you're probably okay. Uh what's your preferred order for coffee then?
SPEAKER_00Preferred order for what espresso. Um usually I just drink coffee at home with some milk, but if I'm out, I guess I'll order cappuccino.
SPEAKER_02Is that like instant coffee at home?
SPEAKER_00Oh no, brewed coffee. Yeah, no, that's that's just wrong.
SPEAKER_03Stressed. Okay. All right. Well, we know you've been in Singapore for a year, so we'll need to remedy the Kopi order at some point. Um Yeah, okay. You can ask for no sugar. Siudai or Kosong.
SPEAKER_00Okay. Thank you for that.
SPEAKER_03Um okay. So next question. What country do you most want to travel to next?
SPEAKER_00Okay. Um well I'm on my way to Hong Kong in a few days. Um and in fact, I have a busy travel schedule this autumn, so I'll be in Korea, I'll be in Japan.
SPEAKER_01Oh wow.
SPEAKER_00I think I'm most excited about Japan. Okay, that's cool. I have not been. Oh, your first time. Very excited, yeah.
SPEAKER_03Where are you going?
SPEAKER_00Going to Hokkaido and Tokyo.
SPEAKER_03Oh, that sounds nice. Yeah. I've been to Tokyo. I have not been to Haikaido.
SPEAKER_00I've been to Hokkaido.
SPEAKER_03Yeah. Where in where in uh Korea are you going?
SPEAKER_00Going to Seoul.
SPEAKER_02Okay, okay. I recently watched a TV series called Now I can't remember when life gives you 10 dreams, but my husband and I call it Ace on Sobs because the main character just cries a lot. And it's filmed mainly in Jeju Island in South Korea.
SPEAKER_03And so I'm re I want to go so bad.
SPEAKER_00Right.
SPEAKER_03Okay, and third one, my favorite question. Uh, what is your go-to karaoke song?
SPEAKER_00Okay, so this is another embarrassing fact about me, but um for almost my entire life until about a year ago, I didn't think I could sing. And so I just would never sing in karaoke. But when I moved to Singapore, I started a new hobby, which is writing songs and recording them at home. And actually, I think my singing has gotten a little bit better. But these are not karaoke songs, these are my own songs.
SPEAKER_03Oh, that's really cool. Yeah, thanks. That's a really neat neat hobby. Are you on Spotify?
SPEAKER_00I haven't yet built up the courage to uh to release them. Oh do it. Uh but I have put a bit of care into the recording. So one of my rooms in my apartment is all kind of acoustically treated, and I got a decent microphone and stuff like that. Thanks, yeah.
SPEAKER_02Do it. So and the reason I say this is because there is some trash on Spotify.
SPEAKER_00And I don't really like because we can at least reach that.
SPEAKER_02There is this woman who recorded her husband snoring for like 13 minutes, and that has over a million streams. Wow. Um people never listened to anything. Including us. Um thank you very much for going through those questions with us. Uh, we're gonna get into your work, The Book Medical Nihilism, published in 2018. Um, so in that you argue, or your main thesis, which was our takeaway, is that our human confidence, societal confidence in the effectiveness of many medical interventions should be low.
SPEAKER_02Uh so if listeners unfamiliar with this, what is the central reason that you think our confidence, we should just take it down a little bit and maybe not trust medicine as much?
SPEAKER_00Yeah, okay. So in the book, there were um three main lines of arguments. Uh one is about the kind of biased way that medical research is carried out, um, quite often, not always, but quite often. So um there's numerous problems with the way medical research um uh is uh uh executed. I use the term malleability in the book, so malleability refers to the kind of ability to bend the methods um at the researchers' will to sort of get the results that the researchers want.
SPEAKER_00Um so that's one big category of arguments in the book, some malleability of medical research. Another big argument is about what I call the a dearth of magic bullets. So magic bullets are medical interventions that are truly effective.
SPEAKER_02I hope they were type of cereal.
SPEAKER_00Is there a cereal called magic bullet?
SPEAKER_02I think magic spoon is I'm sorry. Please keep going. My cultural reference is lost on me.
SPEAKER_00But yeah. So think about like penicillin for bacterial infections, or um uh insulin for type 1 diabetes, or even just something simple like vitamin C for scurvy. So we have this like handful of interventions that are just really they're magic bullets. They're extremely effective. They target diseases very precisely. And it's just a sad fact about medicine that we have very few of those. And the vast majority of them were discovered in the early decades of the 20th century, say like 1910 to 1950, basically. So this kind of golden age of medicine.
SPEAKER_00A golden age is long past us. That's the second category of arguments. And then the third is um when medical research is at its best, so the very biggest and best randomized trials, the very best meta-analyses, uh typically done by academic researchers who are not connected to industry, those trials and meta-analyses give us uh very pessimistic findings on medical interventions. So the effect sizes, it's a technical term in uh in statistics, but those effect sizes tend to be uh extremely tiny in trials that are sort of done done well.
SPEAKER_00So those are three families of arguments.
SPEAKER_02Okay. So um I'm keen to delve into the the first one, and then uh we have many other questions. I read your article on uh the usefulness of antidepressants or depression. Um I found it really interesting uh for for for many reasons, but one of the things I think your main point in that was just that antidepressants are probably as useful as just not using them. Uh-huh. Uh or maybe even taking like placebo is is probably better than taking antidepressants. And you point to several things in here. You talk about how there's uh and you've already mentioned it, conflicts of interest between funding bodies and the researchers who do the research.
SPEAKER_02Um, you talk about how subjects are often carefully selected. So we have a long history of excluding people, uh pregnant women, the elderly, anyone who has a comorbidity that, you know, depression and something else, like maybe diabetes, that person gets excluded. Uh, then there's publication bias, which is where you talked about how more often than not, only the successful studies get published and all the non-successful studies uh don't don't exist, or sorry, don't get published. Then you talked about construct validity and particularly with the uh relationship between the diagnosis of depression.
SPEAKER_02So there was like different factors where you score them and it adds up to depression. So like fit fidgeting, uh sleeplessness, these sorts of things, they get certain points, and if you get enough points or you you have that. And then you also talked about placebo effects and and what was mentioned blind breaking. So when people figure out that they might be on the drug and then that uh does the work for them. And so I wanted to ask you, um, because each of these, you you add them all together and it's like, wow, that sounds really problematic. But then when you look at each of these things, I I just wonder, are these things inherently bad?
SPEAKER_02Like who else has the money to fund studies? And if if we can't do it any other way, then should we just not do these studies at all? And why shouldn't subjects be carefully selected? And how do we know that the that the studies that uh showed that it didn't work actually exist if they didn't get published? And this is this could be very conspiracy theory of and so I I I just wondered about um like how the average lifespan of of of humans has just drastically increased. And a lot of this has been put down to uh medical interventions during childbirth, which reduced uh maternal deaths and things like antibiotics, etc.
SPEAKER_02And so I I just wanted to push you, obviously, um that's the point of this podcast, sure, on what it exactly it is that you're saying about how if antidepressants are unreliable reliable, well, should we trust medicine at all more generally?
SPEAKER_00Okay, yeah. So lots of questions there. So many questions. I'm sorry, sorry. I'll try to take them sort of one at a time. But first, um so th those various features um about the context of medical research that you that you were uh highlighting, like conflicts of interest, publication bias, the careful selection of subjects, and so on, I would say that some of these are inherently bad and some of them aren't. So for example, publication bias looks like it's sort of obviously deceptive. Now, in some areas of basic science, it's sort of routine practice to shelve null results.
SPEAKER_00So you do an experiment in a lab, that experiment might not see the light of day just because you know nothing interesting was found. So if you're studying something sort of um obscure, boutique, that's like not patient-facing or policy-facing in any way, it looks like there's nothing sort of too um bad about that practice. But in the context of studying medical interventions, publication bias looks like it's sort of obviously deceptive. And um, that's been the extent of publication bias in medical research is is really well documented. So, yeah, there I would say that's that's inherently bad. With the careful selection of subjects, this is tricky.
SPEAKER_00So uh, I mean, we know that once a drug is released onto the market, it's going to be used by your sort of average person, and in fact, not just quite your average person, but a person who tends to have comorbidities and tends to be on other pharmaceuticals, tends to be older. Um, on for some classes of drugs like antidepressants, is more like more likely to be female and um so therefore more likely to be pregnant. Um and so that creates a problem of extrapolation. Um that that itself is clear. But of course, when you're doing a trial, you don't want to experiment on subjects who are vulnerable.
SPEAKER_00So subjects who maybe are older or are pregnant or uh are on multiple other drugs and so on. So there are good like ethical reasons to carefully select subjects. There are also good methodological reasons, like it's easier to find a positive signal in noisy data when you have a more homogeneous patient uh population. So there are good methodological and ethical reasons for careful subject selection, but the fact is it creates this problem of extrapolation. So if you find this positive signal in a trial, it becomes a lot harder to infer that that positive signal, this like apparent positive result um due to this drug, will in fact be positive when used in the wild, so to speak.
SPEAKER_00Um so yeah, that's just a kind of small sample of uh addressing that question about whether or not um some of these features are inherently good or bad. You asked about who has the money to fund studies. Um this is a big issue in medical research. Something like 75% of pharmaceutical research is done by um, or yeah, around there is done by private industry. There are huge trials that are done by funded by the NIH, for example, or NIMH in the United States. Um it's not that like central government agencies don't fund trials or meta-analyses, they do.
SPEAKER_00It's just that the vast majority of them happen to be done by um uh pharmaceutical companies today. Now, that's not a principled feature of medical research. So that's something that if we wanted to, we could change, right? So it wouldn't be too hard for regulators to say, all right, if you want us to approve your drug for clinical use, it's gotta be tested by an independent agency. And who's gonna pay for that? You are. Um so you know, you're gonna pay. Yeah, it's so simple, right?
SPEAKER_02So that's how are they gonna get around? This is gonna be a spoiler.
SPEAKER_00I mean, you know, it's it's not so hard to imagine though, right?
SPEAKER_02That you could crazy that I had never thought of that. Yeah.
SPEAKER_00So like an independent agency that's funded by basically user fees, right? Those users, the users are pharmaceutical companies. Anyways, that's one possibility, right? Or you could imagine uh maybe uh because for most diseases today, we have um many possible drugs that are um patented by many different companies. You could also imagine a kind of combative system, a kind of juridical system in which um, you know, this company over here, Glaxosmith Klein, tests out, you know, this company over here's drug and so on. Um so you just take conflicts of interest, sort of at least immediate conflicts of interest out of the hands of the of the researchers.
SPEAKER_00So there are various options that that um that could be entertained that that aren't really what else you want me to talk about. Oh, okay, you mentioned this average lifespan thing. So let me just uh let me just speak to this. So yeah, this is um one form of pushback that I get um with my book. So, you know, you were characterizing this whole line of questioning um with respect to the example of antidepressants, but um this issue about increasing lifespan, of course, is absolutely general, right? So it's like the idea is like how can how can why should we be medical nihilists when we've seen this great increase in um lifespans?
SPEAKER_00Like medicine has surely done some good, right? Um and yeah, I've even gotten that question from some very important thought leaders in evidence-based medicine, which is really ironic because um evidence-based medicine advertises itself as requiring randomized trials to test causality. And this correlation between like medicine developing and increase in lifespan is just one of the grossest kind of correlations, right? So it's true that we've got a lot more medicine now than we do and than we had 100 years ago or 200 years ago.
SPEAKER_00Um, but this uh is just a correlation. So, what some historians say, this is a famous thesis that goes back to this um figure, Thomas McEwen, um he says the vast majority in the increase in lifespans that we've observed since the Industrial Revolution is due basically to more equitable access to nutrition, protein, um, and then some to hygiene sanitation measures, public health measures, and so on. McEwan amassed a very large amount of um population-level data to make that case.
SPEAKER_00Um other epidemiologists, like Michael Marmot, he argues that a major factor for differences in lifespan today is just due to your basically your rank in social hierarchies. Uh it's a lot harder to make the case, in fact, that uh advances in medicine have been responsible for a major proportion of our increase in lifespan. Now it's really clear though, of course, that for for particular conditions, advances in medicine have very clearly helped people with those particular conditions. Um that that that's that's of course true.
SPEAKER_00Yeah.
SPEAKER_03Yeah. Um thank you. Uh I'd like to ask a qu a couple questions about um your book from a little bit different perspective. So we started out with some um, I think this is really like scientifically interesting and um all that. But I want to bring this to more of like a layperson's perspective if they were to read your book. Um, because I think that there could be some bad interpretations, um, especially just with the title itself. It's kind of a provocative, fun title. Um, but there's some people out there who might misinterpret this and into thinking that you should not trust doctors themselves.
SPEAKER_03So I think what your thesis is that we should be less confident in the claims of medical efficacy and um such like that. But uh, how would you separate that out? How would you respond to somebody who thinks, yeah, well, I've always known doctors were just quacks?
SPEAKER_00Yeah, right. Yeah. No, this is a really good question. Um so yeah, right. The title itself could be taken um as provocative and the thesis of the book when if you if you're just given the kind of you know two-sentence summary, um, it's easy to uh imagine people would misunderstand that. Um I I think most of my readers have been non-academics, um, and most of the feedback I get from them is you know pretty positive. And um, you know, when people crack the book open, I the book doesn't really lend itself to that kind of misunderstanding.
SPEAKER_00So um early on in the book, I I tried to be careful uh to address this very issue. So I say, all right, uh in the pages that follow, I am gonna be pretty critical of medicine. Um but that doesn't mean that you should say, turn to quackery, um, turn to untested kinds of practices. In fact, if anything, what I'm arguing for in the book is that medicine should be more scientific. Um and yeah, and then another question is okay, what about doctors themselves, right?
SPEAKER_00So the target of the book is um about medical research and the products of medical research, like medical interventions, but I definitely don't, you know, say anything about the trustworthiness of like particular medical professionals. Um I I I was asked to be involved in a debate with this well-known psychiatrist, David Healy, in which we were asked, like, should we trust our doctors? And so ironically, David Healy was defending the view that we should not trust our doctors, we should just trust ourselves. Um and I um disagreed with that.
SPEAKER_00So I think that you know, if you're a sort of an average person walking into your clinic and you're seeking medical care, um, one of the best things you can do is trust your doctor because they're professional. They've been um trained uh in ways that that we haven't. Um now that said, they've been trained in a particular kind of way. Um so many doctors are trained, most doctors are trained to be heavily interventionist, um, to do something rather than nothing is part of the medical ethos.
SPEAKER_00And so uh part of how I conceived of the book was as a kind of intellectual toolkit for people when they're negotiating their care with their doctor.
SPEAKER_03Yeah, so I I think that's one of the things that I found genuinely most interesting about the book, too, is because I've been uh um aware and have experienced that very interventionist kind of attitude. Uh you go into a doctor's clinic here or in the US and doctors are very willing to prescribe you something uh so that you don't walk away empty-handed. I think there's a fear perhaps on their part that they are somehow letting you down if they don't give you something. But uh I think your book is trying to push back on that idea, right? Um so is that one of the applications that you uh see happening?
SPEAKER_03Like if people took your book seriously, if doctors took it seriously, what would they do differently?
SPEAKER_00Aaron Powell That would be my hope, yeah. So so I closed the book with a chapter called Gentle Medicine, and part of what I um part of how I conceive of gentle medicine is uh for medicine to be less interventionist. Now the the issue that you were just pointing to is this um sort of motive on the part of doctors to intervene if they can or when they can. And I think that that is a very widespread sort of ethos in medicine. But I also think that there are institutional structures in place that um promote that, right? So sort of like a legal context in which in which physicians have to do something, otherwise they could be um you know liable.
SPEAKER_00Um so there, yeah, there's a kind of educational context, a professional context, an institutional context that really motivates like strong interventionism. And then I argue maybe medicine should be a little bit less interventionist. Now that raises all sorts of questions, though. So, like hard questions for medicine and medical research, like um, okay, if we take that argument seriously, then should we just come off of our pills if we're already on them? And that's problematic because actually we don't have a lot of very well controlled studies testing what it's like to come off medications.
SPEAKER_00So we have a mountain of evidence about what it's like to go on medications, but coming off medications is a is a different thing. So part of the part of the um the notion of gentle medicine is a call for that kind of research. So drug discontinue discontinuation studies, for example, is something that I think medicine should be doing more of.
SPEAKER_02I had a a question about um uh the criticism of medical interventions uh more broadly. So I'm completely with you on uh the problem of medicalization and pathologization of things that are not medical and are not pathologies in society, and how our society is so happy-go-lucky to just take a drug. Um, and you know, it's it's just the simple stuff being prescribed antibiotics for a um a viral infection. There's just there is nothing that and that contributes to greater problems like antimicrobial resistance.
SPEAKER_02But um now, now, and I'll just like to contextualize this. I've been watching a lot of house lately. I've got lots of words in my head like CDC, stat. My husband and I just walk around the house saying things like that. We just make up things like K, potassium, banana, now. And and um, you know, I think you know, but there are obviously lots of problems about the show and about house's character, and that's for another podcast. Um but one of the things I've sort of realized is how much of uh treating a patient is not necessarily we know what you have, we will try this intervention, and it's almost as if the interven the intervention itself confirms or uh doesn't whether the per the patient has something.
SPEAKER_02So it's not really until you do something that doctors can often know whether or not something. Something was wrong in the first place, or what was in fact wrong. And I think that uh, you know, I've been in this situation myself and I don't like to personalize it because it's very difficult to rebut personal stories. But where it's like, I would I just want to I just I can't sit still, I can't not do something about this. Um, you know, it's just like I'm just gonna try some paracetamol, I'll just try some over ibuprofen. And you know, doctors are the same, you know, let's we'll just try something because I don't want to sit here and deal with it.
SPEAKER_02And this can obviously be even uh the there is obviously greater need for medical intervention when it's life or death. It's like, okay, look, we don't know, we cannot know, there aren't a lot of studies, but you're gonna die anyway. And so I can see there's a scale there as I'm talking you through this.
SPEAKER_01Yeah, yeah, yeah.
SPEAKER_02Um, but what what do you have to say about this idea that medicine may not it it's about curing and and helping people live better quality lives, but it's also like let's just let's just try something un until we know for sure. And a lot of medicine is is it's not realized until much down the road. But how could we have known at the time?
SPEAKER_00Yeah, good. No, this is a great question. Um thanks. So yeah, no, no, no, but it's it's it's a hard question. So now I think that the the kind of logic underlying that question is persuasive in certain contexts, especially when we have overwhelming reasons to think that whatever interventions we're trying out in this clinical setting are in fact truly effective. So suppose you know we've got some bacterial infection. Uh we don't have enough it's it's a dangerous infection, so we don't have time to just like wait for the lab results.
SPEAKER_00So let's throw this antibiotic at it or that antibiotic at it and see see what happens quickly. And you know, there are contexts in medicine in which you've got to do something like that. Um so you can't do a full diagnosis. You've you just have to intervene and see what happens. Um and if something, if the patient does get very uh improve a lot um on the intervention, we have really good grounds to think that it was because of the intervention, because we have this background reasons to think that antibiotics are indeed often very effective, right? So I think the logic of that uh underlying your question is persuasive in some contexts.
SPEAKER_00The worry is that that logic, that logic is routinely applied in contexts in which I think it's not compelling. So, in in particular, for diseases that are very placebo-responsive or have a fluctuating and gradually decreasing natural course of improvement, or for which we have interventions that are basically ineffective. Um, so there um that logic of just try something out until it works can can be very misleading. Um, just to give you an example, I think this is a very common practice when treating depression.
SPEAKER_00So a patient comes in, the um doctor diagnoses them with depression, uh, doctor tries out drug one. Uh patient goes away for six weeks, comes back, they're not really any better. So the doctor doesn't say, oh, maybe drug one is ineffective. The doctor infers, oh, this patient is not a responder to drug one. They're not sensitive to drug one based on the particular features of their particular kind of depression. So let's try out drug two. So they try out drug drug two, patient goes away, maybe they gain a lot of weight, or their symptoms don't get any better, they come back, they try out drug three. Maybe drug three seems to work a little bit, maybe they increase the dosage of drug three, patient goes away for a few weeks, comes back, they start to look a little bit better.
SPEAKER_00Now, is is the inference that we should make there that that patient is a responder to drug three at this particular dose? Or is it rather that, well, enough time went on and they tried out enough interventions, and we we have overwhelming empirical reasons in the background to think that depression is very placebo-responsive. So, more likely, what happened here is that the patient sort of gradually got better and responded to placebo after enough intervention. And so we really ought not in conclude in that kind of context that our pattern of just sort of trying out interventions now has given us evidence that um allows us to conclude that this particular intervention was truly effective for this patient.
SPEAKER_03Yeah. Um, so I guess I have a question, a follow-up question about placebos, because you list them as uh a reason that we should be less confident in a lot of the um evidence that's out there, but it's we also know that they're effective for certain conditions. Um so how should we like what should we do with placebos? Should we be giving them to people when we know that they have a placebo um responsive condition?
SPEAKER_00Yeah, it's a it's it's a really nice question. And I think that that that's a that's a normative question. What should we be doing? And uh I don't know if we have very many well-worked out views on that normative question. So I think that's a call for more work in philosophy, right? It would be great to have a have a sophisticated answer to that question. I mean, uh in the book I didn't really develop this uh in much detail, but I I think it is worth floating the idea that because we know that some diseases are placebo-responsive, that's a lever that we can pull.
SPEAKER_00Um so why not try to pull it? Now that raises some questions about like will placebo, will placebo interventions still be as effective if patients know they're placebo? That's an empirical question. There's some empirical evidence on that question. Because you wouldn't probably you wouldn't want to be tricking patients, right? You probably wouldn't want to be saying, this is drug blah, blah, blah, but it's just a placebo. Yeah, exactly. Yeah, yeah, yeah. So probably tricking patients is a bad idea. So if we're gonna harness placebo, we've got to be transparent about it. And then that just raises that empirical question. Are we gonna be able to?
SPEAKER_00Um I can imagine like clever physicians um being able to harness placebo in various aspects of their of the patient-physician interaction that don't necessarily involve tricking them into thinking that you know this this tic-tac is a drug.
SPEAKER_02Maybe we maybe maybe, because I I'm a big fan of limited forms of paternalism. So, you know, patients do not know how their bodies work. They they might know it in vague terms that they have a heart, the heart pumps, blah, blah but do they know how many valves they have? Do they know how many flaps per valve there are between various ventricles and the heart? You know, do they know what you know where the a-order is? You know, all these sorts of things. And so that's why they go to their doctors. You're right, that patients should trust doctors. Doctors are trained in anatomy and physiology and uh and know the body and diseases much better than patients do.
SPEAKER_02And for that reason, I do think if we understood in an ideal world where patient uh patients uh trusted their doctors and their doctors took genuine care of their patients, not just flicked drugs at them and sent them away. Where I could foresee a situation in which a doctor prof prescribing a uh placebo drug under the guise that it is something else, particularly if there are no side effects. I think that's a key thing there. There aren't actually harmful consequences. I'm on board. Controversial opinion. I'm totally on board with that.
SPEAKER_00It's an idea without exploring earlier.
SPEAKER_02Demedicalization through placebo. I'm so down. So I one of the things I wanted to ask you about though, which is that it seems like a lot of your work is not is not just like, you know, unreasonably attacking medical researchers for doing bad jobs. And, you know, there would there would be bad biomedical researchers out there who are intentionally misleading their data to try to get more funding? Um, and I can see how you might want to manipulate your methods to produce a better result. I I I really do see that. But for the for the most part, let's just be charitable and presume that most researchers are not really bad people.
SPEAKER_02Is your worry then not so much about poor methods and bad by uh bad methodology, but just an over-reliance on drugs and like the exploitation of a medicalized society in which people are just so willing to spend money to have a pill in their hands even if they don't really know what it does?
SPEAKER_00Yeah, good. So, right. I like that way of um summarizing some of my concerns. So I I agree, yeah, yeah, yeah. I agree that like if we're just looking at, for example, the purely method methodological issues in medical research, I agree that um fraud and outright deception are probably so rare that, and you know, this is this is work that occurs in big institutionalized contexts where there's like a thousand eyes on the project, and you've got you know executives and PR people and scientists and epidemiologists, and you know, these are big institutions where uh people are working together.
SPEAKER_00Um and so my charitable assumption is that cases of fraud in medical research are pretty rare. And my bigger concern is about details of methodology that are not nearly at the level of fraud, so to speak, but are really widespread and often quite technical and boutique issues. Like, for example, should these numbers be analyzed using this statistical method or that statistical method? Um and you know, if you if we go this way, then we get such and such number, which looks really impressive.
SPEAKER_00If we go this way, we get this smaller number, which looks less impressive. Um and there's it's just a fact that medical research tends to go with a number that you know looks bigger and better. And that's a very widespread sort of thing. So that's not fraud. That's just a kind of technical detail about medical research that's really widespread. Um but then you you also were asking about as part of the problem, a kind of broader societal norm about seeking particular kinds of medical interventions like pills. Like we want a pill for every problem we have.
SPEAKER_00Um, I've got this section in the book called A Pill for Every Ill.
SPEAKER_02Nice.
SPEAKER_00Yeah, good one. I should write that song.
SPEAKER_02Yeah, it should be a love song, of course.
SPEAKER_00A ballad. Actually, all of my songs are love songs. Yeah.
SPEAKER_02Um that that's interesting. Um, like the the way you and I I just want to um ask about the the way you describe the statistical difference then, because from my understanding, like you do different statistical tests if you have different variables for different purposes. So there are definitely like gold standard recommendations when you are testing for certain things, or if you have three variables and you're looking for the way they interact, and you can you can manipulate that that data and and not not to be like to to to lie or to deceive, but to try to remove um like factors that aren't actually doing anything, or or to try to account for different factors that might be overstating the data or understating that data.
SPEAKER_02And so you there it's it's not and you might have to correct me because I might have made a layperson mistake here. But if you are having three variables, then you should do the three variable test, which would show you this thing. And it would be then problematic to do the other test that makes your gives you the bigger number and then say that you tested for or or have produced the same outcome. And so wouldn't that be picked up in like peer review that you've tested for one particular thing um and then produced a particular number?
SPEAKER_02Should wouldn't someone come along and say, why did you do this test? Like this this this data that you produced is it doesn't actually say anything at all? Or is the peer review system also a bit shot to hell?
SPEAKER_00Okay, yeah. So so let me generalize the question um if I can. So the idea is um, you know, so that this example of the kind of three variables is was just one particular example, of course, but the general idea is um researchers can analyze and reanalyze data in a bunch of different ways. Those ways might not be what they originally planned. And uh if what appears in the paper is a small subset of what they actually did, then that that raises this worry about what's what's sometimes called p-hacking and that's connected with publication bias, then you end up with a worry that the actual results that you see that in a publication are like a spurious and misleading representation of the actual data.
SPEAKER_00And so will would that get picked up in peer review? It would sometimes, so for example, there's this really common practice now of trials being pre-registered. And so when you pre-register a trial, what you're supposed to do is you're supposed to list all of the different outcomes that you're going to be measuring and the statistical tests you'll be applying to those outcomes. And so then if the corresponding publication doesn't match that pre-registration, that looks like there's a problem. Now there's a mountain of second-order empirical work looking at that very thing. So lots of like uh meta scientists uh are comparing these pre-registrations to trial publications and noticing a lot of discrepancies.
SPEAKER_00So, for example, the primary outcome that's measured and reported, for example, often is different. So we said we were going to be measuring, um, I don't know, we we said that our the primary thing we were gonna be measuring was the probability of heart attack in the two different groups, the control group and the drug group. But what we actually reported was a decrease in cholesterol levels. So what happened to the primary outcome? I see, I see, I see. Yeah, this makes much more sense. Yeah, so that's that's one thing that is very commonly happening. Another thing that's very commonly happening is just we see the pre-registrations, but there's no trial.
SPEAKER_00That means that there's no publication. So you know the trial happened.
SPEAKER_02My mouth has dropped to the floor. Yeah, yeah.
SPEAKER_00That that's very widespread. So um so that's allowing meta scientists to like just observe kind of in real time that publication bias is occurring. Um but then things can be sort of less in your face too, and and more subtle. So let me um let me use this um example that I was referring to just a minute ago about um how to analyze data. So suppose we've got a binary outcome, like dead or alive. Um, there's kind of two families of ways to represent um the numbers that you've got in a trial.
SPEAKER_00There's what's called a relative risk reduction or an absolute risk reduction. Now, of course, I don't, I'm not able to sort of explain what those mean in podcast format. Um, one can just look this up on the internet easily. The main issue is that when outcomes are rare, like dying, for example, um, then it turns out that for the vast majority of interventions, the relative risk reduction just looks a lot bigger than the absolute risk reduction. So I'll give you a real example. In the Bangladesh Mask Study was this huge randomized trial that was done in 2021 looking at whether or not masks slow the spread of COVID.
SPEAKER_00And of course, everybody wanted to know that, right? It's 2021. It's like the question of the year. Do masks slow the spread of COVID? And so there's a massive study, 600 villages in Bangladesh, 350,000 people randomized. And the main outcome was COVID. Um and in the publication, it was published in Science, the world's most impressive science journal. Um, they reported the relative risk reduction. And the relative risk reduction for wearing masks was 9.5%. So it was 9.5% fewer COVID cases in the mask villages compared to the non-mask villages.
SPEAKER_00Now, what the publication didn't do was they didn't actually report the absolute number of cases of COVID. And so statistics blogs asked for this right away. And they got the numbers, and it turns out that in the uh mask villages, there were 1,089 cases of COVID. And in the non-mask villages, there was 20 more. 20 more cases. Over 350,000 people randomized. So when you calculate the absolute risk reduction, it turns out to be something like 0.067%.
SPEAKER_00So here's a case in which you have this really large relative risk reduction and this tiny absolute risk reduction in the publication. The publication reported the relative risk reduction, which is it is that that's not deceptive, right? There's no fraud there, there's no deception. Depends how we define it. Yeah, right, right. I mean, it does kind of look deceptive, right?
SPEAKER_02Um that that's so worrying because in science, like the such a prestigious world-leading journal where scientists themselves read them and then you know go on to do their own research.
SPEAKER_01Yeah.
SPEAKER_02But how can the lay person who die who doesn't even know I don't I don't know the difference between absolute and relative until you explain it just then? How are we supposed to then take like, you know, uh like news articles being like 9.5% reduction, wear your masks?
SPEAKER_00Yeah, right. And so that's that's that's a hard question. And um and you know, this issue about relative versus absolute risk reductions, of course, what I haven't um yet argued during our conversation here is that the absolute risk reduction is the better one, right? The one that's more informative or or more liable to make better decisions and so on. Now in in work that I've done since medical nihilism, um, I have um tried to make that case. In fact, there is a bit of an argument to that effect in medical nihilism, but in in articles since then, I've tried to make the case more robustly.
SPEAKER_00So the short story is absolute measures are informative for decision makers and relative measures aren't. So then the question is, okay, well, what can lay people do? That's just hard, right? I think if you're presented with those numbers, actually we have empirical evidence that shows this. If you give um non-experts um data about a drug presented in relative terms or relative and absolute terms, or absolute terms, they're more likely to ask for a prescription if they're given just relative measures.
SPEAKER_00Then it's a there's a slight decrease in willingness to consume if you give them both relative and absolute. And if you give them absolute, they're much less likely to want to consume.
SPEAKER_03And that's because the absolute the more accurate reflection of what the effect would likely be.
SPEAKER_00Yeah, exactly. Yeah. Yeah. And then when it comes to journals, this is I'm this is something that I've looked into a lot, so I know a lot about this. But so journals have various publishing guidelines. So if you go to the New England Journal of Medicine, for example, like one of the biggest medical journals in the world, they say in their guidelines to authors that um, when reporting the results of a trial, you should report both absolute and relative measures. Then meta scientists have looked at all the major medical journals and at the at the trials that they've published and have just counted the number of publications that report absolute and relative measures.
SPEAKER_00And it turns out that most medical journals publish absolute measures only around 30% of the time. And New England Journal is the worst, they publish it around 22 or 25% of the time, something like that. So they've got their guidelines that they say publish both, but they're not publishing both.
SPEAKER_03There seems to be There's gonna be some angry letters.
SPEAKER_02Angry, angry retweets. Sorry, re-exes. Um I'm I still think that uh so much so much so much of this problem is not gonna be brought back to the individual. Like it's not the individual or layperson's responsibility to amend this. It's not even necessarily the doctor's responsibility. I mean, like there is you you should they should already be couching their prescriptions and drugs, not in absolute and certain terms, um, and doing so in a shared making environment with a patient, you know, where they bring forth the facts of what of what they're trying to do.
SPEAKER_02It seems like a lot of the problems come into um like the peer review system as well as uh like funding obligations um and conflicts of interest and these sorts of things. Um and it sort of just feels a bit overwhelming.
SPEAKER_00Yeah, I think that's right. I I think that it would be naive to think that um once we have all those problems described in detail, that somehow it's the like the individual patient or the individual doctor who's responsible for managing all that, right? This is just not the remit for your average consumer or your average physician.
SPEAKER_02I think um I've noticed this in Singapore that when I go to the doctor, I get prescribed a lot more drugs than I did back in Australia. So I sprained my ankle and I got like four different kinds of painkillers. And I I remember I I took one of them and I was like, you know, just so nauseous for the rest of the day. I was like, screw this. I'm just sticking to my gold paracetamol, my ibuprofen, and my aspirin. That's it.
SPEAKER_03And this is slightly a different issue, but my sister lives in Korea and she's told me when she gets um prescriptions, they'll be all packaged together. You actually don't know what um pill is which one.
SPEAKER_01Wow.
SPEAKER_03So if your doctor prescribes you like this uh regimen, they're all like in packets, like you take this every day, but not clear which one is which.
SPEAKER_02So you're even if you didn't want to take a specific one, you couldn't.
SPEAKER_03Yeah, she's like Googling things, like, does this pill look like that? The blue pill. Which one am I supposed to take if I don't want this side effect? You know? Um so yeah, there's definitely ways that we can design this better in ways that like empower the patient and the doctor to navigate this better. So like I think it's definitely a macro level um kind of issue. These are not um things that are easy for lay people or especially patients to navigate. It's good to be aware of, though, I think, and to have that um more moderate uh understanding of what can what medicine can do for you, um, not be overly optimistic.
SPEAKER_02I think the words uh medical nihilism uh put made me very worried. I was like, do you just reject medicine? And I think for me, the takeaway was just that there is a system set up to exploit the the pathologization of things that aren't properly diseases and the uncertainty of the guarantees of medical outcomes. Um and we pay the price in cost and in side effects. And so I think the my takeaway from yours was just be more be just be a bit more careful, be a bit more hesitant, um, and and you know, press your doctor for for for numbers if you if you want or if you can.
SPEAKER_02And so that's what I'm gonna do next time. I was like, what does this do? Huh?
SPEAKER_00Or or or a lot more careful. Yeah.
SPEAKER_01Yes.
SPEAKER_03I think I mean maybe the advice for doctors too is to be a little bit more um aware that the tendency to prescribe or to do this interventionist um kind of medicine has negative consequences. And if you're not very confident that this could have a good effect, then maybe nothing is actually better, is preferable.
SPEAKER_00Yeah, I think a lot of physicians, in fact, um do have that attitude. Um and so just go back to the title, Medical Niles is.
SPEAKER_03I would like to meet them. Yeah.
SPEAKER_00I'll introduce you to some of them. Um yeah. So the title is kind of provocative, but it comes from it's got a historical precedent. So the term therapeutic nihilist was a very trendy term in 19th century medicine. And really big name um doctors themselves, like the dean of the Harvard Medical School called himself a therapeutic nihilist. And there was this kind of genre of critical liter literature written by very prominent physicians in the last, say, 20 or 30 years, that looked like uh they were sort of very similar arguments to that kind of 19th-century therapeutic nihilism.
SPEAKER_00So that's where that's where the term comes from. And most of those arguments that that you know we've been speaking about, those just come from medicine themselves, really, like from prominent physicians, um, medical journalists, and so on. And so, yeah, it's a provocative title, but yeah, it is what it is.
SPEAKER_02Yeah, it honestly just sounds more like common sense. Medical common sense. Um, anyway, uh we've we've definitely gone over already, but this has been absolutely fascinating. Thank you so much for coming down from NTU. I know it's a solid hour. It's fine there. Thanks for having me.
SPEAKER_00I'm glad to be here. Thanks.
SPEAKER_03Yeah, thank you so much.
SPEAKER_04Thanks for listening to our new podcast. Subscribe online and tune in every first of the month. In the meantime, listen to CBME and U and be savvy.
Send us Fan Mail We spend much of our lives searching for meaning. Some of us look for it at work. Some in our families and friends. Some look for it at an overpriced spiritual retreat in Bali where some dude named...
Send us Fan Mail In some insurance scheme, you might be required to pay more because you are more prone to disability. In another, you may have to pay more because you're a woman (and are statistically likely to live...
Send us Fan Mail Did you know that genomic sequencing can now screen newborns for hundreds of rare conditions right at birth? That looking too closely at our DNA might actually cause more anxiety than answers? And is...
Send us Fan Mail Did you know that over 80% of people who give birth experience an injury? That anywhere between 30 and 50% of people who give birth find it traumatic? And that up to 15% of people experience PTSD from...
Send us Fan Mail Could a blob of cells in a petri dish one day think? Feel? Have rights? In this episode, Professor Insoo Hyun (Harvard Medical School, one of the world's leading experts on the ethics of stem cell and...