Relative vs. Absolute Risk

Relative risk measures the percentage difference between two groups. Absolute risk measures the actual probability of an event occurring.

Confusing the two is the most common way to exaggerate drug benefits. It is a marketing trick used by pharma companies and journalists alike to make weak drugs look like miracles.

The Symptom: The "50% Reduction" Miracle

You see a headline: "New Drug Reduces Cancer Risk by 50%!"

This sounds incredible. You imagine that if your risk was high, it is now half. You demand the drug. But then you look at the data, and you realize you've been tricked.

The Mechanism: The Lottery Ticket Analogy

Imagine you buy one lottery ticket. Your chance of winning is 1 in a million. Your friend buys two lottery tickets.

Your friend has doubled their chances of winning.

Relative Increase: 100% (They are twice as likely to win as you).
Absolute Increase: 1 in a million (0.0001%).

The headline "Friend Increases Odds of Winning by 100%" is true. But in reality, they are still not going to win.

The Medical Math

Let's look at a real drug study with 2,000 people.

Placebo Group: 2 out of 1,000 people got the disease (0.2% risk).
Drug Group: 1 out of 1,000 people got the disease (0.1% risk).

The risk dropped from 0.2% to 0.1%.

Relative Risk Reduction: (0.2 - 0.1) / 0.2 = 50%.
Absolute Risk Reduction: 0.2 - 0.1 = 0.1%.

The headline uses the 50% number because it sells papers. The doctor should use the 0.1% number because it reflects reality.

Real World Example: The Contraceptive Pill Scare

In the 1990s, a warning was issued that a new contraceptive pill increased the risk of blood clots by 100%.

Women panicked. Thousands stopped taking the pill. This led to 13,000 additional abortions in the UK alone.

The Reality:

Old Pill Risk: 1 in 7,000 women.
New Pill Risk: 2 in 7,000 women.

The absolute risk increase was 1 in 7,000 (0.014%). It was tiny. But the "100% increase" headline caused a public health crisis because people didn't understand the baseline risk.

The Prescription: Ask "50% of What?"

When you see a percentage in a news story, be skeptical.

Ask for the Baseline: Is the disease rare? If the disease is rare (1 in a million), a 50% reduction is meaningless.
Check the NNT: This stands for Number Needed to Treat. It tells you how many people must take the drug to prevent one bad outcome. (See Number Needed to Treat).
- Low NNT (e.g., 5): Good.
- High NNT (e.g., 1000): The drug has a weak effect.

At PaperScores, we look for Applicability. If a paper only reports relative risk, we flag it. They are hiding the context.