09 BiggestSecretinMedicine

[Music] if you've been with me this far you deserve a little something so i'm going to let you in on a big secret we have in medicine what i call the biggest secret in medicine um to get there though i'm going to need to describe to you the difference between the relative risk and the absolute risk difference and introduce the concept called number needed to treat i'm going to describe some of the ethical implications of number needed to treat and show you how you can apply the number needed to treat to everyday medical decisions but to frame this properly i want to take you to the las vegas strip to a very unusual casino you're walking along the vegas strip one day and you see a sign for a little-known casino the we tell you your chances casino walking in how could you resist you see two slot machines each one costs ten thousand dollars to play but the payouts are a bit different the machine on the left pays out one million one hundred thousand dollars on average every hundred pulls of the lever the machine on the right pays out one million one hundred thousand dollars on average every two hundred poles of the lever which slot machine do you play at first you may think well the machine on the left has better odds of winning in fact on average if i play a hundred times i'll spend a million dollars and win 1.1 million dollars so this is a great deal but then you probably think hold up ten thousand dollars is a lot of money i can barely afford to play once let alone a hundred times i'm going to the all you can eat buffet instead all right so what does that have to do with medications well what it has to do with medications is that when we do medical studies we're assessing the effect of a medication on the population at large like those slot machines which i gave you the odds if you pull 100 times or 200 times we're saying how many lives do we save if we treat a thousand people with this drug or 10 000 people with this drug we aren't really asking what is the chance of saving your life that one pull of the slot machine lever so even if overall if you could pull the slot machine a thousand times you would make money you don't get to you only get to pull the slot machine once and so there's a conflict between the population effects of a drug and the individual effects of the drug and we can resolve that conflict by looking at the difference between the relative risk and the absolute risk difference so to give you an example i'm going to talk about this drug lipitor which you may have heard of also known as a torva statin it's a cholesterol medication and here's an ad that tells you that lipitor reduces the risk of heart attack by 36 percent so what does that mean well if you drill down you see a little asterisk here that says that means in a large clinical study three percent of patients taking a sugar pill or placebo had a heart attack compared to two percent of patients taking lipitor wait three percent compared to two percent that sounds like one percent right not 36 percent where are they getting the 36 percent well the 36 comes from a relative risk and the one percent is an absolute risk difference so right off the bat you can see one of the major differences between these two metrics relative risks are much bigger generally they sound much more impressive than that absolute risk difference but they can be misleading and show you why i want to take you to this study looking at the rate of death among triathletes okay so this was a study that catalog people who died during triathlons and found that the risk of death during a triathlon was 1.74 per hundred thousand competitor days which means that if you had a triathlon with a hundred thousand people in it one point seven four or one or two would die on average during that triathlon now that sounds kind of high actually and indeed it is the background rate of death in the united states is 0.5 per 100 000 people per year triathlon is usually just a single day and so if you do kind of the back of the envelope calculation you can see that your risk of death in a triathlon is a thousand-fold higher than your risk of death on any other day of the year and that should sound very scary and we say oh my gosh why would we ever let anyone do a triathlon but when you look at the raw numbers you see you know 1.74 out of 100 000 is is still a vanishingly small number the vast majority of people are just fine to run a triathlon so the relative risk inflated the sort of magnitude of the problem now how does this work for our lipitor example well the lipitor people told us that among if 100 people took lipitor two of them would have a heart attack and if a hundred people took placebo three of them would have a heart attack and we simply said oh well that's two percent versus three percent so lipitor is one percent better and we are correct that is the absolute risk difference absolute risk difference is subtraction three percent minus two percent is one percent pharmaceutical companies don't like that because one percent sounds pretty paltry right but the relative risk two percent versus three percent two percent divided by three percent two divided by three point six six well that starts to sound pretty good so they can say oh well if the risk in placebo is is one then the risk in lipitor is .66 that's you know 33 less there's some rounding errors in there you know what they give us in the advertisement but basically a 33 36 reduction in heart attacks which they they're both true but they're expressing the same data in different ways and i think it can sometimes be misleading now if we look at this what we can see is that in both groups we have two people that are having at least two people who are having a heart attack and then there's that that one extra person in the placebo group that has a heart attack so if you look at this carefully and squinting your eyes you can sort of say hey you know really it seems like lipitor is just helping one guy out of a hundred right one person out of a hundred so that brings us to the number needed to treat okay so if we know the absolute risk difference is one percent we can figure out that one out of a hundred people are benefited specifically benefited from from lipitor all right that's the number needs to treat we need to treat a hundred people with lipitor to prevent one heart attack formally the number needed to treat is defined as the inverse of the absolute risk difference so one divided by .01 one percent one divided by .01 is a hundred the number needed to treat is a hundred if i told you that the absolute risk difference um if i told you for instance that the rate of a heart attack was was 20 in the placebo group and 10 in the lipitor group that would be an absolute risk difference of 10 and the number needed to treat would be one divided by ten percent or one divided by point one or ten okay so that's how we can get to the number needed to treat but the real number needed to treat here is one hundred is that a good number needed to treat there's no rule here there's no threshold you're trying to beat or anything like that you have to take into cu into account the risk of the thing that you're doing so to give you some examples if you take you would need to treat 42 people after a major heart attack with aspirin to prevent one death is that worth it well when you ask if it's worth it you say well how expensive is aspirin you know it's pretty cheap death is pretty bad what are the side effects of aspirin not much you know some people get stomach aches and things like that there's some bleeding risk etc but but this seems pretty good and in fact we say yeah sure we'll give 42 people aspirin to save one life that that sounds like a great deal right whereas for other things you might have more pause right so we need to treat 217 smokers with cat scans of the chest to prevent one death from lung cancer now death from lung cancer is terrible but giving 217 smokers cat scans is expensive and it exposes them to radiation that they might not otherwise need to be exposed to so the number you need to treat doesn't give us any hard and fast guidelines but it helps us assess the utility of giving an individual treatment so what's the secret what am i talking about why is this the biggest secret in medicine well if i need to treat a hundred people to save one life i'm essentially treating 99 people unnecessarily and the big secret is i don't know who's who okay my societal obligation as a physician is to save as many lives as i can and that is in opposition to my obligation to try to benefit my patient because yes i want to save as many lives as possible which means i should give drugs like lipitor to as many patients as possible to save those lives but on an individual patient level if my patient says doc is this drug going to save my life the honest answer is well probably not chances are you are not that one out of a hundred people but you might be and is it worth it to you to take this drug just in case you're that one out of a hundred people this does not mean that you should stop taking your medicines if the medicine is cheap and without side effects a 1 in 100 chance it might save my life is totally worth it right talk to your doctor if the drug is you know a vitamin c pill that costs nothing and you have to treat a thousand people to save a life that could totally be worth it it's just totally cheap right but if it's a highly toxic drug and you're vomiting every day you take it and the chance that's going to benefit you specifically is quite low well then maybe you need to talk to your doctor about taking that drug look for the number needed to treat in medical studies some studies will report it and kudos to them when they do if they don't and they report the absolute risk reduction now you know how to calculate the number needed to treat from that to help you make decisions with your doctor about what medications are worth it for you a few take-home points the absolute difference is not as dramatic as the relative risk but it's good information the number needed to treat is the inverse of the absolute risk difference numbers needed to treat have no particular threshold they're all over the map but the fact is that most people will not directly benefit from most interventions the catch is we don't know for which of those medicines you're that one lucky person the number needed street gives you a metric that helps you weigh the risks and benefits of drugs in a fair way thanks a lot i'll catch you next time

Channel: YaleCourses

Views: 7,317

Rating: 4.5581393 out of 5

Keywords:

Id: EFIGM-iYbEQ

Length: 10min 32sec (632 seconds)

Published: Fri Jul 31 2020