Reader Question - What Is Number Needed to Treat?

This is the Weight and Healthcare newsletter! If you like what you are reading, please consider subscribing and/or sharing!

This came up in our monthly subscriber Ask Me Anything. When I break down research I often talk about using absolute risk reduction and statistical significance to determine the efficacy of an intervention. Another number that helps determine efficacy is the Number Needed to Treat (NNT). This number helps us understand how many patients need to be treated with an intervention in order to avoid a single negative outcome. 

So, let’s say there is a medication to prevent death from heart attack. If the NNT is 5, then 5 people would have to take the medication to prevent 1 heart attack death. 

NNT is calculated using Absolute Risk Reduction (ARR,) which is the event rate of whatever was trying to be prevented in the control group (those NOT receiving treatment) which is known as the Control Event Rate or CER, minus the event rate of whatever was trying to be prevented in the intervention group (those receiving treatment) which is known as the Experimental Event Rate or EER.

In our example, ARR would calculated as the number of people in the drug trial who died of heart attacks in the control group, minus the number of people who died of heart attacks in the group that was receiving the medication.

Let’s look at a real world example. 

In Novo Nordisk's trial of Wegovy (Semaglutide 2.4mg) to reduce 3 Major Adverse Cardiac Events (death from cardiovascular causes OR nonfatal myocardial infarction OR nonfatal stroke) there was Control Event Rate of .08 (8%) and and Experimental Event Rate of .065 (6.5%), which is to say 8% of the people in the control group experienced one of the three Major Adverse Cardiac Events they studied, and 6.5% of the people in the group that was getting the drug experienced one of the three events over the mean duration of treatment of 34.2±13.7 months intervention, and the mean duration of follow-up of 39.8±9.4.

So let’s do the math:

The equation for NNT is:

NNT = 1/ARR

(Remember ARR = CER - EER)

So in this study:

NNT = 1/(.08 (the Control Event Rate/CER) - .065 (Experimental Event Rate/EER)

NNT = 1/.015

NNT = 67 (NNT is always rounded to the nearest whole number)

So when the math is mathing, 67 people would need to be treated in order to prevent one death from cardiovascular causes OR nonfatal myocardial infarction OR nonfatal stroke.

It gets more complicated, though, since the people who would need to be treated must, at the very least, match the demographics of those in the study. For this study, participants had to be at least 45 years old, have existing cardiovascular disease, not have type 2 diabetes, and have a BMI of at least 27 (In fact, the research showed that even the 1.5% reduction was not statistically significant for women, Black people, Hispanic people, or those in the highest weight categories - I wrote about this study in depth here.)

When we think about NNT, we must also think about harm. Let’s say a drug has a NNT of 100. So 100 people must be treated to prevent 1 negative outcome.

If the 99 people who AREN’T helped by the medication also aren’t HARMED by the medication, that’s a very different scenario then if those 99 people experienced significant negative side effects that impacted their quality of life and/or caused death. Even if they weren’t harmed, if the medication was prohibitively expensive and the positive effect was very small, that’s also something to consider. The formula I offered here is the most commonly used basic formula, but there are also additional, more complex formulas that take harm into account.

While the research teams funded by/comprised of employees from Novo Nordisk, Eli Lilly, and other weight loss research don’t seem to agree with me, I think researchers should make things like the Absolute Risk Reduction, the Number Needed to Treat, and for which demographics the result was statistically significant clear, in the abstract, in front of any paywall. 

Did you find this post helpful? You can subscribe for free to get future posts delivered direct to your inbox, or choose a paid subscription to support the newsletter and get special benefits! Click the Subscribe button below for details:

Liked this piece? Share this piece:

More research and resources:
https://haeshealthsheets.com/resources/

*Note on language: I use “fat” as a neutral descriptor as used by the fat activist community, I use “ob*se” and “overw*ight” to acknowledge that these are terms that were created to medicalize and pathologize fat bodies, with roots in racism and specifically anti-Blackness. Please read Sabrina Strings Fearing the Black Body – the Racial Origins of Fat Phobia and Da’Shaun Harrison Belly of the Beast: The Politics of Anti-Fatness as Anti-Blackness for more on this.