We all want to be healthy and live as long as possible. So we think about popping vitamin pills, a daily dose of aspirin, eating foods with probiotics and taking statin drugs. And we puzzle over dietary regimes that claim to result in weight loss, lower blood cholesterol and reduced risk of Alzheimer鈥檚 disease. All of these are supported by some evidence. But the chances of significant benefit are in general much less than what people think.
One of the best way to evaluate such interventions is through a determination of what is referred to as the 鈥渘umber needed to treat,鈥� or NNT. Basically this is a measure of the number of people who need to follow the particular regimen for one to have a successful outcome. Take for example the possibility of reducing the risk of cancer by taking a daily multivitamin. And let鈥檚 not even consider the numerous studies that have failed to show any reduction in risk, lets consider one that did. The Physicians Health Study followed close to 15,000 doctors over the age of fifty who took either a placebo or a multivitamin for eleven years and found an 8% lower risk of cancer in those who took the multivitamin. By comparing this reduction to the normal rate of cancer in such a population, we can calculate that to prevent one case of cancer about thirty people have to be taking the multivitamin. So the chance of prevention is small, but on the other hand, the risk of taking a multivitamin is essentially none.
The situation is more complicated when it comes to medications with potential side effects. The American Heart Association recommends that people who have a greater than 10% chance of suffering a heart attack within the next ten years based on risk factors such as high cholesterol, high blood pressure and family history, take a daily low-dose aspirin pill. Statistics show that this intervention can prevent a heart attack. With what efficacy? Depending on which study one looks at, it is in the range of 1-2000 people who have to be treated to prevent one event. So with the best scenario, about 1000 people have to take the daily aspirin for one death from a heart attack to be prevented. But what about side effects? Aspirin can cause bleeding and over a ten year period in 1000 people there would be about two extra deaths from bleeding. So that doesn鈥檛 look so good. But wait. There is mounting evidence that aspirin can reduce the risk of colorectal cancer. Here the NNT over a ten year period is about 60. So in the case of the 1000 people taking aspirin 16 deaths from cancer would be prevented. Putting it all together, more people benefit than are harmed. But the overall NNT works out to be roughly 50-60, meaning that the chance that a single individual will benefit from a daily aspirin regimen is pretty slim. But if sixty million people followed the regimen, a million premature deaths could be prevented. Most people, though, I think would decide that if their individual odds of benefitting are 60 to 1, the intervention is not worthwhile. Yet they will go out and buy a lottery ticket where the odds of winning are millions to one.
