An Example to check HR, RRR, and NNT

So let’s go through just using each one of those explaining again what actually it’s telling you, and you can see why we get through to some of these that using ARR and NNT gives you a little bit more needs to be able to explain that data. So RR in that study that we saw about half of the original risk was removed which if you tell somebody that they’d really have to think about it because what was the original risk. RRR is the amount of risk that was removed. You can see in this one since it was about 50% that about 50% of the risk was removed and about a 50% of the original risk is there.

ARR, you take the number the the percent that you get from one minus the other and you take it times the number of patients in the entire study to get the number of patients that are spared an event compared to the other treatment, in this case it was placebo. Now the NNT remember tells us how many patients need to be treated and we put the simvastatin set in 40 milligrams for an average of six point two years to prevent one additional person from experiencing the composite outcome of a non fatal MI or death from CHD compared to placebo.

And you’re going to want to take this number then and compare it to other things to see is 24 a good number or not over a 6.2 period may or may not be. You just need to check and see what other studies have found. So let’s go through an example and see if this can help you in a clinical trial to be able to see how what we do with these values. So you’re asked by your supervisor to review HR, RRR, ARR, NNT, and NNH for the recent DAPT trial to present at the next staff meeting.

In the trial, patients recieved aspirin and a thienopyridine for 12 months followed by an aspirin with a thienopyridine or aspirin plus placebo for another 18 months. One of the primary efficacy endpoints was reccurrence of stent thrombosis Results showed that continued treatment with aspirin and a thienopyridine as compared with aspirin and a placebo reduced the rates of stent thrombosis and they give a 0.4 percent versus 1.4 percent HR is 0.29 ninety-five percent CI zero point one seven two zero point four eight and they gave us a p-value of less than 0.001. So we’re going to want to go and confirm what the HR was to make sure we get the same number. And we do.

And trust me sometimes you don’t and I’ll give you an example where that can’t happen. So it’s always good to check to see what it is. So you want to make sure that you can again state what an HR says and so again it was it takes about 1/4 of the risk was removed compared to the aspirin group. Now again relative risk reduction again just saying what’s there, it’s gonna be one minus a 0.28 and this was is the risk removed from the original risk which was what the RR was I think I said that one wrong RR is the original risk RRR is what was removed.

ARR again you take the total number of patients times what you got from the treatment minus the control and you get the number of patients who were spared an event and then lastly you’re going to calculate the NNT from that and four patients need to be treated to for the with the treatment at the time with the doses in there to see what would happen. I didn’t put the doses in there because they use different drugs. So you’d have to go back and look at the study there was the thienopyridine they use more than one but you can see that for patients for treatment probably and with any disease that four is probably a good number.

But then you’d also want to compare that in context with any of the NNH’s that are significant or severe with this treatment to make sure that benefits are greater than risks. So let’s just go over something I recently found in the news. I always get excited when I find something like this that I can apply in class to to show that drug information and literature evaluation is a clinical practice type of a field and so I wanted to present this one to you.

This was recently in BMJ, and they showed they were looking at the the risks, hospitalization risk for heart failure in patients with NSAIDs was a large case-control trial they looked at many patients although when you look at patients that big of a group you’re looking for the ones that actually come up with that risk would end up being very small which you can see that it was 14.45 percent in NSAIDs groups and those who were considered the control. So they reported their results as 19 percent increase in the risk of hospital missions for a heart failure with use of any NSAIDs without previous 14 within previous 14 days compared with the past use any NSAID and within 183 days.

so nineteen percent that sounds pretty good, but let’s look to see what that really means.