“Statistics are like bikinis. What they reveal is suggestive, but what they conceal is vital.”
Don’t let the media influence your narrative with misleading statistics. This happens in the health world all the time.
When you read a startling statistic today, pause before sharing it.
Ask:
Compared to what?
Over what time period?
How large is the actual effect?
Is this typical or an outlier?
Those four questions eliminate a surprising amount of misinformation.
Headline
“This drug cuts the risk of heart attack by 50%!”
Sounds incredible, right?
Now let’s ask the four questions.
1. Compared to what?
Compared to people not taking the drug.
Let’s say:
Without the drug: 2 out of every 100 people have a heart attack over five years.
With the drug: 1 out of every 100 people have a heart attack.
That’s where the 50% reduction comes from.
The risk went from 2% → 1%.
2. Over what time period?
This is over five years.
If someone hears “50% reduction,” they might imagine they’re protected immediately.
Instead, the benefit accumulated over half a decade.
3. How large is the actual effect?
Here’s the important distinction.
Relative risk reduction
50%
Absolute risk reduction
1 percentage point (2% → 1%)
Those sound very different.
Another way to think about it:
Treat 100 people for five years to prevent one heart attack.
This is called the Number Needed to Treat (NNT).
NNT = 100.
That doesn’t mean the drug is bad—it may be absolutely worth taking if it’s inexpensive and safe—but it gives the effect its proper scale.
4. Is this typical or an outlier?
Suppose the trial included mostly:
Men
Age 70+
Prior heart attacks
Diabetes
Can we assume the same 50% reduction applies to:
A healthy 35-year-old woman?
Someone with no cardiovascular risk factors?
Probably not.
The study population matters.
Likewise, if five studies show a 10% reduction and one study shows 50%, that dramatic study may be the outlier rather than the rule.
“Wisdom begins when we stop asking whether something is big or small—and start asking, ‘Compared to what?’”