Tricky numbers - Cochrane says, “Beware”

At some time or another, you’ve heard the latest headlines in the media claiming that the risks of side effects while taking Drug A almost doubles, or that the risks are 2 times more likely or that the risk increases by 50%. Those numbers can sound quite sensational. They’re meant to – they’re the latest breaking news! But when making a decision about your health or when making a decision about whether you’re going to advocate for Drug A to be made publicly available, it’s crucial to understand what those numbers really mean.

Understanding statistics can be difficult. And the way the numbers are presented in the media can make things even more difficult. Take for example Drug A above, the news article reports that the risk of heart attack almost doubles, or that the risk of heart attack increases by 50%. Unfortunately, key information was missing from the article. The article didn’t report how many people were in the study that took Drug A or did not, and how many people had a heart attack.

Consider 2 studies both reporting almost double the risk of heart attack to see what a difference it makes to describe numbers of people.

Study 1:

• 2 out of 100 people had heart attacks with no treatment
• 3 out of 100 people had heart attacks with Drug A.

Study 2:

• 2 out of 1 000 000 people had heart attacks with no treatment
• 3 out of 1 000 000 people had heart attacks with Drug A.

In both cases, the risk is almost double but the impact just doesn’t seem to be the same once you know how many people had a heart attack while taking Drug A or not. Claiming that the risk doubled or increased by 50% is not a lie or exaggeration of the truth, the numbers are just calculated ‘relative’ to each other – but it seems deceiving. Presenting the number of people in the study in ‘absolute’ terms, such as 1 out of 100, seems to provide a clearer picture.

But what can you do when you are presented with ‘doubles’, ‘ % increases’, ‘times more likely’? Look for more information. Most times an exciting new study is broadcasted in a number of newspapers and in other media. More often than not you can do a search on the Internet and find more articles about the study. Usually one of them will list the ‘real’ or ‘absolute’ numbers of people in the study and how many benefited or were harmed from the treatment. Use those numbers to make important decisions.

Fortunately, more efforts are being made to present these ‘absolute’ numbers. The Cochrane Musculosokeletal Group (CMSG), for example, presents both relative and absolute numbers in their systematic reviews so that health care professionals can use whichever they are most comfortable with to make health care decisions. And for the public, the CMSG presents absolute numbers in consumer summaries of their reviews on the Arthritis Society web site.

But when only ‘relative’ numbers are presented, take the time to find more information. The real trick is realising that you need to be aware.

The Cochrane Musculoskeletal Group (www.cochranemsk.org) is 1 of 50 groups in the Cochrane Collaboration. Cochrane is a non-profit international organization ensuring that patients, physicians, researchers, and policy makers have reliable information about health treatments. We gather the evidence and write systematic reviews for about the effects of treatments for musculoskeletal/arthritic diseases.