 Physicians have a duty to inform their patients about the risks and benefits of whatever they prescribe. However, physicians rarely communicate the absolute risk numbers, such as numbers needed to treat. In other words, how many people are actually helped by the drug? Numbers needed to harm, in other words, how many people are actually hurt by the drug, or prolongation of life. How much longer will it enable you to live, despite patients wanting all this information? If doctors inform patients only about the relative risk reduction, for example, telling patients a pill will cut their risk of heart attacks by 34%, 9 out of 10 agree to take it. Give them the same information framed as absolute risk reduction, though 1.4% fewer patients had heart attacks, and those agreeing to take the drug drops only 4 out of 10, and use the numbers needed to treat, and only 3 in 10 patients would agree to take it. So, if you're a doctor and you really want the patient to take the drug, which statistic are you going to use? The use of relative risk stats to inflate the benefits in absolute risk stats to downplay any side effects has been referred to as statistical deception. To see how one might spin a study to accomplish this, let's look at an example. Here's the incidence of heart attack over five years in those randomized to a placebo. Compared to those getting the drug, a significantly lower risk. If you wanted statins to sound good, you'd use the relative risk reduction, a 24% lower risk. If you wanted statins to sound bad, you'd use the absolute risk reduction, just say 3% fewer heart attacks. Then you could flip it around for the side effects. For example, they found that 0.3%, one out of the 290 women in the placebo group, got breast cancer over those five years compared to 4.1%, 12 out of 286 in the statin group. So a pro-statin spin on this study would be like a 24% drop in heart attack risk and only 3.8% more breast cancers, whereas an anti-statin spin could be like only 3% fewer heart attacks compared to a 1,267% higher risk of breast cancer. Both portrayals are technically true, but you can see how easily you could manipulate people if you picked and chose how you were presenting the risks and benefits. So ideally, you'd use both the relative risk reduction and absolute risk reduction stats. In terms of benefits, when you compile a bunch of statin trials together, it looks like the relative risk reduction is 25%. So if your 10-year risk of a heart attack or stroke is 5%, then taking a statin would take that down from 5% to 3.75% for an absolute risk reduction of 1.25% or a number needed to treat of 80, meaning there's like a 1 in 80 chance that you'd avoid a heart attack or stroke taking the drug over the next 10 years. As you can see, as your baseline risk gets higher and higher, even though you have that same 25% risk reduction, your absolute risk reduction gets bigger and bigger. By 20% baseline risk, you have a 1 in 20 chance of avoiding a heart attack or stroke over the subsequent decade if you take the drug. So those are the benefits. In terms of risk, that breast cancer finding appeared to be a fluke. Put all the studies together and there was no association between use of statins and risk of cancer. In terms of muscle problems, estimates range from just 1 in 1,000 to closer to 1 in 50. If all those numbers are just blurring together, the Mayo Clinic developed a great visualization tool for those at average risk considering a statin. Over the next 10 years of 100 people doing nothing, 10 may have a heart attack. If all 100 took a statin every day for those 10 years, though, 8 would still have a heart attack, but 2 will have been spared. So there's like a 1 in 50 chance taking the drug would help you avoid a heart attack over the next decade. What are the downsides? The cost and inconvenience of taking a pill every day, which can cause some gastrointestinal side effects, muscle aching and stiffness in maybe 5%, reversible liver inflammation in 2%, and more serious damage in perhaps 1 in 20,000 patients. Now note these two happy faces represent heart attacks averted not lives saved. The chance a few years of statins will actually save your life if you have no known heart diseases is about 1 in 250. If you want a more personalized approach, the Mayo Clinic has an interactive tool that also lets you calculate your 10-year risk. You can get there directly by going to bit.ly.com.