Medicine's Biggest Secret: Almost Every Medication is More Likely to do Nothing than to Help You

For the video version, click here.

Imagine you're at your doctor's office, and she prescribes you a new medication for an annoying symptom you've been having. Maybe it's insomnia, maybe low back pain, whatever. How likely do you think that prescription is to help you? Few people have that conversation with their doctor, and if they do, they usually get a response like "well, it seems to help most people" or something like that.

The truth is often quite surprising, and the medical establishment has a name for it.

It's called "number needed to treat" – and in my opinion, it's the biggest secret in medicine. 

The number needed to treat is the amount of prescriptions you have to give out to ensure that one person gets better. And if you think that the typical number needed to treat is 1 or 2, you're way off.

In a series appearing in the Milwaukee Journal Sentinel and on, journalists John Fauber, Kristina Fiore, and Matthew Wynn have examined several syndromes that have gotten a lot of press in recent years. The series is entitled "Illness inflation" – the journalists provide evidence that pharmaceutical companies have influenced syndrome definitions to expand the market for certain drugs.

This is a sticky area because the syndromes they are talking about are not "made up". Rather, the strategy appears to be to take a well-established clinical syndrome and expand the criteria to enlarge the population that can be said to be sufferers. But you have to be careful here – the broader the definition of a syndrome is – the less likely a drug is to help.

As a case example, let's take "overactive bladder".  The story notes that overactive bladder is an expansion of the older term "urinary incontinence".  Urinary incontinence had a precise definition. To be diagnosed with urinary incontinence, one had to have leakage of urine. Overactive bladder, as a syndrome, includes urinary incontinence but also includes urinary "urgency" – a feeling of "having to go" even without true incontinence.  Obviously, there is a bigger market for overactive bladder than there is for urinary incontinence so it is in the pharmaceutical industry's best interest to get their drug approved for OAB rather than UI.

Toviaz is a drug used to treat overactive bladder. You need to treat about 4 people to have one experience less urinary incontinence, and 6 people to reduce urinary frequency. Those numbers are actually pretty good as a symptom-relief drug go. But that should still be somewhat shocking. 

Would you feel comfortable if your doctor told you – "here, take this prescription, it has a 1 in 4 chance of helping you"?

To be fair – you might decide to give it a try.  I mean, if it helps, great, if not, you can stop it.  What's the harm?

One of the major undiscussed issues here is the placebo effect.  See, some people are going to feel better taking a drug even if the drug isn't truly helping them – they'd feel just as good with a sugar pill. In this case the sugar pill would be WAY better for them – considering the side-effect profile of your typical prescription medication.

Let's look at Addyi, a drug to treat female hypoactive sexual desire disorder. Its number needed to treat is 12. For every 12 women who take the drug, one will have an improvement in the rate of sexually satisfying events compared to placebo.  But overall, 29% of women given the drug reported an improvement in sexually satisfying events – nearly one in 3.  How do we square those numbers? Because 21% of women who received placebo reported an improvement in sexually satisfying events.

Putting the numbers together, of 12 women who take Addyi, around 4 will see an improvement in sexually satisfying events, but 3 of those would have seen an improvement on a sugar pill.

And the magnitude of the placebo effect is much higher when we're talking about syndromes with prominent symptoms.  Can a placebo reduce your cholesterol level?  Well, a little bit – that's the power of the human mind.  But can a placebo reduce your headaches? Fatigue? Joint pain? Urge to urinate? Absolutely. The brain has much more control over those things than we give it credit for.

So by "just trying" a drug – we run the risk of having a large number of people stay on a medication that they really don't need. This, no doubt, makes the pharmaceutical company happy. It may actually make the patient happy too, considering their symptoms are improved.

But the costs are still there.  Monetary costs are obvious, but so are the risks of taking the drug – from side effects, to interactions with other medications. The average elderly American patient is taking 5 prescription medications. The average nursing home patient is taking 7. The cure for the over-medicating of America just might be good statistical information. 

My solution? Force pharmaceutical companies to report the Number Needed to Treat in their direct-to-consumer advertising. Let patients know what to expect when and if they start a medication. I'm sure we can get at least one out of every two people to agree with that.