- © 2008 Canadian Medical Association or its licensors
I agree with Finlay McAlister's analysis of the number needed to treat.1 However, I believe there is a missing issue associated with the use of the evidence-based terms we are so accustomed to using. Consider the following 2 scenarios.
In the first scenario, a drug reduces the risk of dying from a myocardial infarction from 3% to 2% over 2–3 years: the relative risk reduction is 33%, the absolute risk reduction is 1% and the number needed to treat is 100. In my experience, many people do not consider this drug to offer sufficient benefit to justify taking it. In the second scenario, a drug reduces the risk of dying from a myocardial infarction from 100% to 99%: the absolute risk reduction is 1% and the number needed to treat is 100. It is likely that many people would take this drug given that the risk of death without treatment is 100%.
These 2 scenarios demonstrate that terms such as relative risk reduction, absolute risk reduction and number needed to treat do not provide patients with any idea of their baseline risk. I have moved away from using these terms when trying to explain the benefits of drug therapy. Instead, I first use a variety of risk estimation tools to provide patients with a rough approximation first of their baseline risk (say 10%) of having an event (e.g., a myocardial infarction or fracture) in the next 5–10 years and then of their risk while taking drug therapy (say 8% assuming a relative risk reduction of 20%); I remind them their risk will not be reduced to 0%. I also outline the side effects and cost of drug therapy and then support them equally whether or not they decide to take the drug.
Most people understand this approach. It removes the need to do mathematical calculations and, most important, it provides people with an idea of what would happen if we do nothing. This is important because people typically believe their baseline risks are higher than they really are. The terms number need to treat, relative risk reduction and absolute risk reduction rarely enter my discussions unless I have to show people why the 20% benefit that they have heard about for a particular therapy translates to only a 2% absolute risk reduction for them.
Footnotes
-
Competing interests: None declared.
REFERENCE
- 1.↵