Historically, people with high cholesterol levels were presumed to need drug treatment. But do they? The way drug companies pseudo-scientifically describe the benefits of medications – framed in terms of "relative risks" – powerfully and systematically exaggerates the benefits of those drugs, and inflates their market value. Thus, patients frequently buy and consume medicines that do very little actual good. An alternate way of describing the benefits of drug (or other) therapy could change that.
Take cholesterol-lowering drugs as an example. In 1995, the prestigious New England Journal of Medicine published a study regarding cholesterol-lowering drugs called statins. Researchers reported a 31% reduction in the risk of heart attack among men taking the statin “pravastatin” (sold under the brand name Pravachol). Due in large part to this study, Pravachol now grosses more than $2 billion in annual sales for manufacturer Bristol-Myers.
A 31% reduction in heart attacks does seem impressive, but what does it really mean? In that 1995 study, it meant that taking Pravachol every day for five years reduced the incidence of heart attacks in the study group from 7.5% to 5.3%. By that measure, taking the drug did result in 31% fewer heart attacks in the patient population. However, for any given person, Pravachol reduces the "absolute risk" of heart attack by only 2.2 percentage points (from 7.5% to 5.3%). Consider the numbers ... suppose that 100 people with high cholesterol took statins. Of them, 93 will not have heart attacks anyway. 5 people will have heart attacks despite taking Pravachol. Only the remaining 2 out of the original 100 can avoid a heart attack by taking the daily pills. Since the study determined that 100 people needed to be treated to avoid two heart attacks, the number of people who must receive treatment for a single person to benefit is 50. This is known as the "number needed to treat" (NNT).
Developed by epidemiologists in 1988, the NNT is an objective tool to help patients make informed decisions. The NNT is intuitive, and avoids the confusing distinction between "relative" and "absolute" reduction of risk. To an intelligent, otherwise-healthy person with high cholesterol (that didn't decrease in response to diet and exercise), a doctor could say, "A statin might help you, or it might not. Out of every 50 people who take them, one avoids getting a heart attack. On the other hand, that means 49 out of 50 people don't get much benefit."
Drug companies do not want people thinking that way, and they frame discussions of drugs only in terms of “relative” risk reduction. (For instance, Pravachol’s package insert does not mention NNT. Similarly, Pfizer's literature about Lipitor and Aventis’ press release regarding Actonel both heavily promote “relative” risk reduction while making no reference to NNT). The reason is simple. Big numbers encourage people to purchase drugs. Also, those numbers even encourage medical professionals (who should know better), to aggressively prescribe drugs. In 1991, researchers surveyed faculty and students at Harvard Medical School – a group that should understand health statistics. When presented with identical information (but in two different formats) about a drug, the doctors who received information about the relative risk reduction had a "stronger inclination to treat patients” as opposed to those doctors who received absolute risk reduction (NNT) numbers.
When a therapy is extremely effective – such as surgery for acute appendicitis or insulin for juvenile diabetes, for instance – NNTs are not an important factor. But medical interventions vary in effectiveness (and most are not “home runs”), so NNTs are a highly useful tool in determining if specific therapies may be worthwhile, medically and economically. Some examples: 1) For shoulder pain or stiffness, the NNT for a cortisone shot is 3, which is pretty good. But that still means two out of three patients won't feel any better after the needles. 2) Pediatricians routinely treat childrens’ ear infections with amoxicillin, but the NNT for antibiotics used to shorten the duration of fever is more than 20 … so at least 19 out of 20 parents pour the stuff down their toddlers' throats for no reason. 3) Taking Proscar for four years to treat an enlarged prostate carries an NNT of 18. The drug costs $100 per month per person, so a health insurer will spend $86,400 on drugs to prevent each $28,000 prostate surgery. 4) Taking aspirin to help avoid a heart attack carries an abysmal NNT of 208.
Bear in mind that none of the above figures include the risks of side effects.
In some cases, drug companies’ not-entirely-honest messages about public health aren't necessarily a problem. Consider statins again. Although any one individual with high cholesterol has little reason to take them (since 49 out of 50 get no benefit), when millions of at-risk people consume the drugs, even low-effectiveness rates (multiplied by those huge numbers) can still result in quite a few averted heart attacks. Therefore, well-meaning public-health authorities routinely tolerate such exaggerations of relative risks … that is, as long as the touted intervention is fairly painless and readily accessible. But the NNT calculations become important in the case of an expensive drug.
NNT calculations are also revealing about things people do, like breast-feeding. In June 2006, the New York Times ran an article headlined, "Breast-Feed or Else." It declared “breast-fed babies are at lower risk for sudden infant death syndrome and serious chronic diseases later in life." Yet, the article never mentions the NNT to prevent these scary diseases (such as asthma, diabetes, leukemia and lymphoma), and there’s a significant reason for that omission ... the NNTs are astronomically high. Reasonable women might well (and quite rationally) decide that breast-feeding isn't worth the trouble – a conclusion that those who heatedly promote “breast-feeding at any cost” do not want the general public to draw.
In the end, however, the argument that it is acceptable to promote expensive and largely ineffective drugs/interventions for the sake of the greater good just doesn't wash. Nor does the excuse that NNTs are difficult to understand, or that the math is too hard. Patients look to doctors to translate and/or interpret complex (and often-conflicting) information from drug companies, medical journals, and the media ... and NNTs are an excellent tool for doing precisely that. Doctors need to make the use of NNTs a standard part of helping patients adequately understand their choices.