Bipolar blogger Natasha Tracy offers exclusive insight into the world of bipolar disorder.See all posts »
Which Medications are Best – Using the NNT/NNH
When getting treatment for bipolar disorder there are a myriad of options from which to choose and among those are many types of medications. Doctors will all have their favorites, but why does a doctor choose one drug over another?
There are many reasons that a doctor chooses a specific drug for you, including your specific set of symptoms, the drugs you’ve previously tried, your insurance plan coverage (unfortunate, but true), the side effect profile of the drug, and many others.
And one of the things the doctor may be taking into account are variables known as the number needed to treat (NNT) and the number needed to harm (NNH).
Number Needed to Treat
The NNT is the number of patients that need to be treated in order for one to have a successful outcome when compared to a placebo (because some people will get better on placebo—an inert substance).
In an ideal world the NNT would be one—that is every patient given the treatment would have a positive outcome and no patients given the placebo would get better. In a clinical trial, two is the lowest NNT that can be observed. Food and Drug Administration (FDA)-approved drugs for bipolar disorder have an NNT between three and nine. In the case of NNT=3, there is a 33.3-49.9 percent difference in the treatment vs. placebo and in the case of NNT=9, there is an 11.1-12.4 percent difference.
According to Ketter et al. in their paper Treatments for bipolar disorder: can number needed to treat⁄harm help inform clinical decisions?, single-digit NNTs are usually important enough to change routine clinical practice. Although not FDA-approved, drugs with NNTs in the low teens may be considered as alternative treatments if there is good tolerability. Even in the face of extreme tolerability, drugs with an NNT greater than 20 are insufficiently unlikely to help.
In the above graph we can see that the drug with the lowest NNT when treating acute bipolar depression is an olanzapine/fluoxetine combination (Symbax). Following that, lurasidone and quetiapine tie with a NNT of 6. (Note: numbers only available for some drugs.)
Number Needed to Harm
The other number that is critical to consider when looking at drugs is the number needed to harm (NNH). The NNH is the number of people needed to receive the drug to have been exposed to a certain risk factor over those who have taken the placebo.
So, in other words, you want the NNT to be as low as possible and you want the NNH to be as high as possible. The difference between these two numbers often has a lot to do with whether a drug gets FDA-approval. Unfortunately, we can see in the above example that more people are harmed by an olanzapine/fluoxetine combination by weight gain than are helped by it.
In this graph we can see that lithium clearly has the best ratio in terms of sedation as they have a very low NNT and a high NNH. (Note: NNH not available for haloperidol.)
And it’s important to remember that NNT and NNH numbers are only applicable over a certain period of time. For example, if a drug study only lasts six weeks, the NNT only indicates those who got better in those six weeks and not those who may have gotten better after that.
This is clearly seen in the above graph where the NNH regarding weight gain may be low during acute treatment but becomes much higher over time. (Note: Bars in green represent percentages. Data not available for all drugs.) So while a NNH of 8 for weight gain and olanzapine doesn’t sound too bad, knowing that 35 percent of people gain more than 7 percent of their body weight on the drug during the stabilization period is quite shocking.
Knowing the NNT and NNH
Now, it’s pretty hard to get your hands on these numbers in the nice, clean way I’ve presented them to you. It takes quite a bit of research and normally the pooling of data from a number of studies. So if you’re interested in them but don’t know how to find them, just ask your doctor. Even if your doctor doesn’t know the specific number from the study, they likely will have a number (or percentage) in mind that corresponds with their clinical experience.
Reference: Acta psychiatrica Scandinavica, Treatments for bipolar disorder: can number needed to treat⁄harm help inform clinical decisions?. Ketter TA et al. March 2011.