Half of What? How to Turn Any Health Percentage Into Plain Numbers
In the mid-1990s, British women were told a newer birth control pill doubled their risk of blood clots. It did. It went from about 1 woman in 7,000 to about 2 in 7,000. Many women stopped their pill, and an estimated 13,000 more abortions followed in England and Wales the next year.
A percentage can scare us away from something or sell us on it. Either way, it rarely tells us what it means for our own body. Here's how to translate any health percentage into plain numbers you can actually use.
Step 1: Write down the claim, word for word
"Cuts your risk in half." "Reduces heart attacks by 44 percent." "Doubles your risk." Copy it exactly. Watch for the word "risk" with a percentage beside it, and no count in sight.
Step 2: Ask, "Percent of what?"
There are two kinds of risk numbers.
- Relative risk compares two groups to each other: "50 percent lower," "twice as high."
- Absolute risk tells you how many people are actually affected: "2 out of 100," "1 in 7,000."
A drop from 2 in 100 to 1 in 100 can be called "cut in half." It's also one person in a hundred. Both are true. Only the second tells you what it means for you.
Step 3: Find the counts
Look for the actual numbers in each group. A news story or the study summary will often give them. You're looking for two things:
- Without the treatment, how many people out of 100 (or 1,000) had the problem?
- With the treatment, how many?
Here's a real example. In a large 2008 cholesterol drug trial called JUPITER, the headline benefit was 44 percent fewer heart attacks, strokes and related events. In the actual counts, about 2.8 out of every 100 people on placebo had one of those events, compared with about 1.6 out of 100 on the drug. That's a difference of a little more than one person in a hundred.
Step 4: Ask, "Over how long?"
One in a hundred over two years is very different from one in a hundred over a lifetime. If no time frame is given, write that down. It matters.
Step 5: Ask, "How many people take it for one to benefit?"
Researchers call this the number needed to treat. In the JUPITER trial, the researchers' own figures meant about 95 people would take the drug for two years for one of them to avoid one of those events. The closer that number is to 1, the stronger the treatment.
Step 6: Put the cost on the other side of the scale
Every benefit comes with a price: side effects, money, time, and what it does to the body over years. Ask for the same plain counts here. Out of 100 people, how many had a side effect? In the JUPITER trial, for example, slightly more people on the drug were diagnosed with diabetes.
When "average" shows up
The same care applies to the word "average." There are three kinds:
- The mean adds everything up and divides.
- The median is the middle value.
- The mode is the most common value.
A few extreme numbers can pull the mean far away from what most people experience. So when you hear that the "average" person recovers in six weeks, ask which average, and how spread out people were around it. You are not an average. You are one person.
Your own numbers count
The most useful counts in your life are often the ones you keep yourself. When you start something new, write down your symptoms, your sleep and your energy before you begin, and then week by week. That's your own before-and-after, and it belongs to no one else.
A practice for today
Choose one health claim you've seen this week that uses a percentage. Try to fill in these four blanks:
- Without it: ___ out of 100
- With it: ___ out of 100
- Over how long: ___
- Side effects: ___ out of 100
Words to Carry
- I ask, "Percent of what?"
- I ask for plain counts, out of 100.
- I ask, "Over how long?"
- I weigh the cost along with the benefit.
- I am one person, not an average.
With you every step of the way,
Dr. Jocelyn
References
Gigerenzer, G., Gaissmaier, W., Kurz-Milcke, E., Schwartz, L. M., & Woloshin, S. (2007). Helping doctors and patients make sense of health statistics. Psychological Science in the Public Interest, 8(2). https://doi.org/10.1111/j.1539-6053.2008.00033.x
Mitra, A. (2021). Blood clots, COVID-19 vaccines and the contraceptive pill: Are we heading for a repeat of the 1995 pill scare? BMJ Sexual & Reproductive Health, 47(4), 303–304. https://doi.org/10.1136/bmjsrh-2021-201219
Ridker, P. M., Danielson, E., Fonseca, F. A., et al. (2008). Rosuvastatin to prevent vascular events in men and women with elevated C-reactive protein. New England Journal of Medicine, 359, 2195–2207.
