Number Needed to Treat
How many people do you need to treat with this ___________ (medication, treatment, operation) to save one life? The NNT or number needed to treat is something that I drilled into me since medical school. It is a statistical equation that applies to real life. Unfortunately, we are not given that information many times or it changes or it is distorted by third parties. Well, the video I put up the other day mentioned this site thennt.com and I have to say it is impressive and it is NOT funded by drug companies. Here is what they say:
We only use the highest quality, evidence-based studies (frequently, but not always Cochrane Reviews), and we accept no outside funding or advertisements.
I highly recommend you check it out. For example, how about taking cholesterol medication without any prior history of heart disease? Here is what they come up with:
In Summary, for those who took the statin for 5 years:
Benefits in Percentage
- 98% saw no benefit
- 0% were helped by being saved from death
- 1.6% were helped by preventing a heart attack
- 0.4% were helped by preventing a stroke
Benefits in NNT
- None were helped (life saved)
- 1 in 60 were helped (preventing heart attack)
- 1 in 268 were helped (preventing stroke)
Harms in Percentage
- 1.5% were harmed by developing diabetes**
- 10% were harmed by muscle damage
Harms in NNT
- 1 in 67 were harmed (develop diabetes*)
- 1 in 10 were harmed (muscle damage)
So, the same amount of people who were helped were actually harmed. It makes you rethink the way we use statins for someone especially someone who is young, healthy, eats well, not overweight and exercises but whose cholesterol is a little high. Oh, wait, that’s right, we can’t think. The quality indicators and bonuses by the gov’t/insurers called pay-for-performance demand we get our patients’ numbers in the right boxes. In other words, we are motivated to use drugs. The statin issue above once again proves why this concept is wrong.
Anyway, please check out the site and use it for yourself or if you are a doctor, for you patients.
Guess I should stop my crestor.
This seems to be an apropos time to mention that in residency I was taught the following definition:
Harvard death: All the lab values are normal.