In 1995, the results of the West of Scotland Coronary Prevention Study (WOSCPS) were presented in a press release:
“People with high cholesterol can rapidly reduce … their risk of death 22 percent by taking a widely prescribed drug called pravastatin. This is the conclusion of a landmark study presented today at the annual meeting of the American Heart Association.”
As in most drug trials, pravastatin benefit were reported as relative risk reduction. What does a 22 percent relative risk reduction mean? Many people reading “22 percent” would conclude that 220 people (out of 1,000) benefited by being in the “treatment” group. As we shall learn, in absolute terms, less than 1 percent benefited.
This study randomized 6,595 Scottish men aged 45–64 years who had “high cholesterol” to either placebo or pravastatin 40 mg/day. After five years, 32 people out of every 1,000 who took pravastatin died; of every 1,000 people who took the placebo, 41 died.
Treatment Deaths per 1,000 people with high cholesterol
Absolute risk reduction
To arrive at absolute risk reduction, you take the number of patients who died without treatment (41) and subtract the number who died with treatment (32). The difference after 5 years is 9 per 1,000 – nine people in the treatment group presumably benefited by taking the drug for five years.
- The absolute risk reduction was only 0.9 percent – 9 in 1,000 – less than 1 percent!
So how did the researchers turn less than 1 percent into a 22 percent risk reduction?
They took the number of people in the treatment group who presumably benefited from the drug (9) and divided that number by the number of patients who died without treatment (41):
- 9 divided by 41 = 22 percent – thanks to ingenious wizardry.
Most drug studies are reported in relative terms. In absolute terms, the study may have succeeded only in changing the way some people died.
Category: 'Bad' Science