West of Scotland Coronary Prevention Study – how study statistics are manipulated

19,324 views
| March 18, 2013 | 6 Replies
Print Friendly
Big Pharma uses "relative risk statistics" to inflate miserly study results

Big Pharma uses “relative risk statistics” to inflate miserly study results

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

Pravastatin                                                                               32

Placebo                                                                                     41

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. 

 

 

Tags: , , ,

Category: 'Bad' Science

Comments (6)

Trackback URL | Comments RSS Feed

  1. DoctorSH says:

    Alan

    One more patient interpretation of statin studies I hear is:

    I lower my risk by 22% if I take pravastatin. Patients believe everyone benefits the same amount, not necessarily that only 22% of patients benefit.

    Keep up the fight!!

  2. Nicholas says:

    Is it that only 1% actually benefit? Relative risk is laughable sometimes

  3. Jonathan says:

    Alan, You took this from Calculated Risks by Gerd Gigerenzer. You forgot the footnote. Please add.

  4. Alan Watson says:

    Actually I read about this In a blog from Malcolm Kendrick, UK cardiologist.

  5. Usually I do not find out post on sites, on the other hand would like to express that this specific write-up extremely obligated everyone for you to do that! Your creating tastes have been pleasantly surprised everyone. Thanks, very nice post.

Leave a Reply