2.5

Taipei Medical University

Skip to 0 minutes and 14 secondsSo, let's now actually calculate an NNT. Again the boxes and this I did make up this was already in the literature but of course it fits with what I like to do that there's boxes that you make. Now remember with measures of Association it's yes no data. So you're gonna have outcome occurred yes outcome occurred no and treatment one and treatment two. Now if you have more than one treatment in a study you're gonna have to have multiple boxes they're comparing to treatment one two treatment two and then to treatment three depending on what the comparisons were in the trial.

Skip to 0 minutes and 48 secondsAnd I put these A, B, C, and D here that you can use to help you with your calculations but there's a couple different ways you can do it you can decide which ones easier for you. So here are the calculations, it's it's not even algebra it's just general math that you're going to do. You're already new relative risk was the event versus the control and you can see the EER is the experimental event rate and the CER is the control event rate wether that people see more or drop. So you can see from the from the table from listed before it's A percent cell a divided by percent of C to get the RR.

Skip to 1 minute and 24 secondsRemember RRR is just one minus RR that are very easy. The absolute risk reduction is thus the percent change minus one minus the other and again which one minus it depends on if it's harm or risk and then for the number needed to treat is just 1 divided by the ARR. So here's an example from the literature that was a study looking at simvastatin versus placebo and whether simvastatin had a positive benefit on patients experience a non-fatal MI or death from CHD. Since the case we see this set up this way the information if they give it to you with the patients you should be able to figure out the percentages If they don't give all of it to you.

Skip to 2 minutes and 4 secondsso you should be able to quickly calculate and the RR, RRR, ARR and NNT. Even if they give you the RR you want to do it to check recheck it. So you can see in here, I did the calculations. I don't need to go over them with you, you can review them later if you'd like. But just those equations that i gave you to come up with a number of 24.4. Now most people say that you need to round down or round up depending on what the number is but we're talking about one patient one way or the other. So the general rule is to round up or down depending on what the standard number is.

Skip to 2 minutes and 38 secondsNow, once you understand how to set up your tables, there's actually a couple apps that I found that that you can plug the numbers in there but I'd only recommend these if you truly understand how to set up the table. Because it's just like a statistical application package and we see this all the time. In the studies that we look at you can plug numbers in there and because something will spit out. It doesn't mean it's the right number, but it'll give you something. So you have to make sure that though you can clearly set up the boxes you understand which one is the benefit what's the harm what's going on to be able to use these websites.

Skip to 3 minutes and 14 secondsSo I took this example and put it in these two websites. Now this one is pretty plain but I actually like this one because it gives you RR, RRR, ARR as well as NNT and NNH. So then that way you can compare all of them to what you bfound to make sure that you did it correctly. So I would do this a couple of times until you feel comfortable because if you put them in the wrong way you're gonna get wrong numbers. And actually when I did this I wasn't paying attention once I did that. I'm like wait a minute that can't be right.

Skip to 3 minutes and 44 secondsWell again if you don't put the numbers in the correct boxes you will get something but it won't be the right something. Now you can see that when their numbers are exactly the same and what I came up with it just depends on how many significant figures you use to calculate each one of them I tend to do two and I think they did four so and more with there. So it just depends on your significant figures but you can see they got the same values. Pretty much the same values that I got from calculating them in the boxes in the previous slides.

Skip to 4 minutes and 12 secondsNow this one looks a little bit more tech-savvy and unique but this one gives you the percentages so you don't have to put them in there. So it's a little bit easier to be able to put those in there and to calculate it. But all it gives you as an NNT and then ARR. It doesn't give you RR or RRR and also you see there NNT explanations missing quite a few key elements there. so it's okay to use this one to get the numbers but again you need to understand what the numbers mean when you get them.

Skip to 4 minutes and 48 secondsSo I'd highly recommend that you do it with the box method a few times and then when you need to do it just save these websites in your bookmarks and then use these when you need to double check in a trial.

How to Calculate RR, RRR, ARR, and NNT/NNH ?

Prof. Mary Ferrill explains how to calculate RR, RRR (Relative RIsk Reduction), ARR (Absolute Risk Reduction), and NNT/NNH.

She starts this video with a table about the exposure and outcome for treatments. We can use this to help us calculate the parameters in the example in this video.

After comprehending Prof. Mary’s explanation, you can try to calculate the example on your own. If you have any questions, please leave them below.

Besides, she provides two websites that can help us to calculate the RR, RRR, ARR and NNT/NNH, and then illustrates the differences. Remember to take the significant figure into account.

We can calculate them with the box method, and use websites to double check.