Super Smart Health

Empowering Health and Wellbeing

Send to a friend

3. What Are the Important Findings?

The first thing to do when looking at a list of findings in research is to consider which findings are most meaningful to what you value. For example, if outcomes of death, hospitalizations, or quality of life are most important to you, you will want to focus on these first.

Next, you need to know how to look at the way the findings are reported. Many researchers and pharmaceutical companies provide impressive looking findings in relative terms. However, to better assess what the finding means to your life it is more helpful to assess the outcomes in absolute terms.

For example if a researcher or a pharmaceutical advertisement claims a 2-fold improvement in the distance walked after taking a particular therapy you would want to know how far that is exactly and then consider whether that is clinically meaningful to your life.

When findings are presented as the percentage of patients experiencing specific events, you will need to know how to crunch some numbers to gain greater clarity.

For example, consider the following figures representing two separate studies reporting the percentage of patients experiencing asthma attacks over a 6-month period and showing what initially may seem like an equally impressive reduction in asthma attacks in each study.

This is the way information is commonly presented in pharmaceutical advertising. Here in both studies you’ve be told there is a 40% reduction in asthma attacks from the “New Therapy”. But looking closely you notice this is not really true. This 40% reduction is a percentage of a percentage – and in Study 1 it’s a 40% reduction from 25%, while in study 2 it’s a 40% reduction from 2.5%. The information has been conveyed in relative terms.

The Relative Risk Reduction

What this 40% represents is something called the Relative Risk Reduction (RRR). It is calculated by comparing the events in the intervention and comparison group as follows:

RRR = (intervention group – comparison group)/comparison group

So in the example in Study 1 above the RRR = (25-15)/25 = 40%

And in Study 2 the RRR= (2.5-1.5)/2.5 = 40%

If we only looked at the RRR, we would have no way of knowing that the impact of the intervention was any different between the two studies.

While commonly used (often because it makes data such as that in Study 2 look even more impressive), the RRR does not convey the real magnitude of the difference between an intervention and comparison groups. That is why it is much better to look at the absolute difference between the groups.

The Absolute Risk Reduction

We calculate the absolute magnitude of difference between groups in the study using the Absolute Risk Reduction (ARR). It is simply calculated as follows:

ARR = (events in intervention group – events in comparison group)

In Study 1 the ARR = 25-15 = 10%

In Study 2 the ARR = 2.5 – 1.5 = 1%

It is clear to see from the ARR that the impact of the intervention in Study 1 is much larger than in Study 2.

The Numbers Needed to Treat

Another way of looking at the absolute magnitude of difference is how many patients would be needed to treat to prevent one event. This is known as the Numbers Needed to Treat (NNT). It is derived from the ARR.

In Study 1, where the ARR = 10%, this means for every 100 patients treated with the new agent instead of placebo you would prevent 10 asthma attacks. Therefore, using simple proportions, for every 10 patients treated you would prevent one event.

Thus, the NNT is easily calculated by the formula, 100/ARR (with the ARR expressed as a percentage).

In Study 1 the NNT = 100/10 = 10

In Study 2, where the ARR =1%, the NNT = 100/ARR = 100/1 = 100. In other words, in Study 2 you would need to treat 100 patients to with the “New Therapy” instead of placebo to prevent 1 asthma attack.

The NNT more clearly helps us understand the impact of the intervention. To simplify, in Study 1 you have a 1 in 10 chance of preventing an asthma attack with the intervention, while in Study 2 you have a 1 in 100 chance of preventing the attack. Hence, the smaller the NNT, the greater your chance of benefiting from the intervention.

One also needs to be mindful that the NNT is dependent on the length of the study. In our Studies 1 and 2 examples it was mentioned that the duration of each were 6 months. Thus, if one expects the benefit of a therapy to continue unchanged over time we could extrapolate that the NNT of 10 in Study 1 over the 6 month study period would be equal to an NNT of 5 over a 12 month period.

The Absolute Risk Increase and Numbers Needed
to Harm

You calculate the Numbers Needed to Harm (NNH) in the same way you calculate the NNT.

Imagine a study where the intervention group has a 6% incidence of nausea and the placebo group has a 2% incidence of nausea.

Instead of an Absolute Risk Reduction (ARR), you have an Absolute Risk Increase (ARI) in the adverse event.

Here the ARI of nausea = 6-2 = 4%. So for every 100 patients treated with the intervention instead of placebo, 4 will experience nausea.

The NNH = 100/ARI = 100/4 = 25. Thus, for every 25 patients treated with the intervention instead of placebo 1 will experience nausea. Here you have a 1 in 25 chance of experiencing nausea.

Extracting the absolute differences from research is one of the keys to making informed choices in healthcare. More specifically, knowing the ARR of benefit and the ARI of harm (or the NNT and NNH) provides the information you need to truly balance the benefits and risks in decision making around therapy. This will be discussed further in the section on Applying the Evidence.

SuperSmart Health Blog - Blogging health and well being

SuperSmartHealth incorporates Evidence-Based Medicine and Integrative Holistic Medicine into a new vision of health.
We provide seminars, wellness programs, coaching, consulting and gateways to knowledge, support and inspiration
to facilitate optimal health and well-being.

SuperSmartHealth does not provide medical advice, diagnosis or treatment. Please see our Terms of Use for more information. Page Updated: Nov 05, 2008