Understanding Benefits and Risks of Treatments

We all know people for whom medications have improved their quality of life, and others where medications have actually kept them alive. Yet anything that is designed to have a significant effect in changing your biochemistry, has the potential of causing other consequences that aren’t wanted: in other words, side effects. And many of us know people who have reduced or quit their medications because they have experienced some of these!

Before drugs can be put on the market, trials must be done to check whether they work to address a medical condition, or a marker of that condition, such as a blood test result. Trials also measure whether the drugs trigger side effects, and how common and intense those side effects may be. These can range from a dry mouth to something life-threatening. All that information is statistically analyzed and expressed in ways that many people find confusing.

Many medications are aimed at reducing the risk of something bad happening, such as a heart attack or stroke. We need statistical analyses of how many they reduce, as we don’t notice when things don’t happen, only when they do! That risk reduction can be expressed in different ways, that can make a treatment seem much more or less appealing.

Informed consent is the aim in all medical treatment, and it means that the person who will use the treatment understands the risks and benefits of doing so. But if you don’t know how to make sense of the numbers, you can’t be fully informed. We hope this information will make it easier for you to make a better informed choice.

Absolute Risk Reduction (ARR) vs Relative Risk Reduction (RRR)

Here’s a theoretical example: If we could expect 5 out of 100 people with a certain illness to become blind within 5 years without treatment, and trials find that a medication reduces that number from 5 people to 3, the absolute risk is reduced from 5 to 3, which is 5% - 3% = 2%.

So the Absolute Risk Reduction for this treatment is 2%.

If you look at it relatively, a risk that goes down from 5 people going blind to 3, has gone down by 40% of the original figure.

So the Relative Risk Reduction is 40%.

If you were marketing a medication and encouraging people to use a treatment, it’s obvious which figure would be more persuasive! Both are true in their own ways, but most people don’t know which kind of number they are being quoted, let alone what the other number is, or that it even exists! More information gives a better overall picture from which to make an informed decision about taking a medication.

Here's a real-life example: a meta-analysis on using statins for heart attack and stroke prevention, published in the Journal of the American Medical Association in 2022, calculated that the Relative Risk Reduction of using statins is a 9% reduction in deaths overall, a 29% reduction in heart attacks, and a 14% reduction in strokes.

In comparison, the Absolute Risk Reduction for statins is a 0.8% reduction in deaths, a 1.3% reduction in heart attacks, and a 0.4% reduction in strokes.

To prevent a heart attack, people may be willing to tolerate more side effects for a risk reduction of 29% (RRR) than 1.3% (ARR). Two very different numbers reporting the same result in different ways. Somepeople experience side effects that significantly reduce their quality of life, and understanding these numbers can help them decide how they choose to balance quality of daily life, with the probability of avoiding a heart attack or stroke in future.

Relative risk reduction is always larger than absolute risk reduction, so if you have both numbers you can tell which is which, but it is rare for both to be quoted. If you only have one number it is harder to work it out.

The oral contraceptive pill is spectacular in this regard: if taken correctly it results in a 99% reduction in pregnancies. That’s very impressive for both absolute and relative risk reduction!!

Number Need to Treat (NNT) and the Number Needed to Harm (NNH)

This can help you assess the potential risk /benefit of using a treatment.

The Number Needed to Treat means the number of people who would, statistically speaking, need to use that treatment, for at least one personto get the expected benefit.

In the example above, if 2% of people avoid blindness by using the medication for 5 years, you would expect to have to give the treatment to 50 people to save the sight of one person. 2% is 2 in 100, which is the same as 1 in 50, so the Number Needed to Treat is 50.

The Number Needed to Harm is the number of people treated before one person is harmed by the treatment. If 1% of people are harmed, you would expect that if you gave the treatment to 100 people, only one of them would be harmed. So the NNH would be 100.

Clearly we want medications where the Number Needed to Treat is small, because then you only have to give it to a few people for one to benefit. That also means that if you give it to lots of people, a large proportion of them can be expected to benefit.

You also want the Number Needed to Harm to be a large number, so you know that most people who take it won’t be harmed. It is also helpful to find out what that harm may be: you would want it to be a minor one, not a major one. You may be willing to tolerate say, a dry mouth, but not something that seriously affects the quality of your life.

This analysis can be applied to natural treatments too. We tend to have fewer figures on natural treatments, as natural supplements aren’t required to be trialed before they go on the market. That makes sense because in general, they more supportive to the body and less potentially dangerous, than drugs that are designed to impact the progress of a disease condition. But it means we often have less information to work with.

If you recall the post on calcium and bone density, taking just calcium, or just calcium and vitamin D, wasn’t a good idea. The British Medical Journal published a meta-analysis of trials in 2011, which found that the Number Needed to Treat (NNT) for preventing fractures was 302, and the Number Needed to Harm (NNH) by causing additional heart attacks or strokes was 178. Where the NNH is greater than the NNT, that treatment should clearly be avoided, or you are just swapping one problem for a greater chance of a different one. Better approaches are needed!

The NNT and NNH of any drug are not always easy to find, but they are worth asking about. If you are considering a new treatment you may want to know more about how likely it is that you might experience any side effects that would harm you, and how likely it is to help you.

We notice that people tend to expect any drug on the market to work for everyone, whenever they use it. As you can see, there are no guarantees. It’s a good idea to understand the numbers and know the probability of it working for you.