The One Number Big Pharma is Keeping Secret From You

Does anyone know how doctors decide whether a pharmaceutical treatment is suitable for a given condition? We’ve all heard of it—it’s the data. Data decides whether we should write one prescription, another, or none at all. Guess who presents this data for the most part? You are going to love this—it’s the very same people who are trying to sell the product: Big Pharma.

I recall being a medical student and diligently learning how these data are presented and what they mean. All doctors are taught to understand terms like absolute risk, relative risk, absolute and relative risk reductions, odds ratios, hazard ratios, and many other statistical terms, as this is the language through which data are presented. Because statistics can be complex, experts in this field working for entities with conflicting interests may choose to obscure the truth, presenting "part-truths" by using out-of-touch statistical lingo.

When I was a young doctor training to be a GP in the UK, we had a powerful presentation on this organised by the GP training provider (West Cambridgeshire). Our group of 30 doctors held differing opinions on the efficacy of a drug depending on how the data were presented. It turned out the conclusions all came from the same data on the same drug; they were just dressed differently. I remember the embarrassment vividly.

There is, however, one statistical term that can separate signal from noise rather effectively, and understandably, its use is shunned by most pharma-sponsored studies. This is called the NNT—Number Needed to Treat. It’s the parameter that makes the most sense and brings reality back down to earth: how many people do we need to treat with this product to see one person benefit? Although often not reported, the NNT for most treatments can be calculated easily because the formula is simple (yes most good things are simple).

Basically, the most honest and transparent way to present data on a treatment must include the NNT whenever possible. When given the NNT, a patient can understand a drug’s efficacy and know what to expect. Some people might think it is worth taking medication even if there is only a 1 in 10 chance of getting better after 12 months; others may feel a 1 in 2 chance is still too low. I don’t think either view is wrong. Instead, I believe it is wonderful to move away from paternalism and let people make informed, individual decisions that empower them and those around them.

Let’s look at one example: statins. If I were to pick the best data on statins (being generous to give the drug its best chance), a five-year course reduces the risk of major coronary events by 24%. That seems like a fair number to some. But if we unpack that statement, it takes five whole years of taking the drug to achieve such a benefit. Moreover, what does a 24% reduction mean if we don’t know our baseline risk? If our starting risk is a 100% probability of a heart attack, a 24% reduction results in a 76% probability after five years. If our starting risk is 10%, our final probability at five years is 7.6%. These distinctions significantly affect a person's decision on whether to take the medication.

The NNT comes to the rescue to pull us out of this confusion. Again, maintaining a generous approach and looking only at the best data while ignoring underperforming studies, the NNT is 25. That means 25 people need to take a statin for five years to prevent one person from having a heart attack. Furthermore, this number is conditional; it is reserved for people who have been unlucky enough to have already had a heart attack. If you have had a heart attack, the best a statin can do—according to the industry's own best data—is prevent 1 out of 25 people from having another event over five years. If I don’t cherry-pick the data, that number rises from 25 to 39. I often ask people plainly: would you take a drug if it only prevented an ailment in 1 out of every 39 people taking it for years?

In reality, many people are prescribed statins without having had a prior heart attack. Your doctors are well-intentioned and have likely seen undesirable results in your blood panels before taking this course of action. However, it would be a dream to see risks and benefits discussed more honestly with patients—and this is what the NNT can do neatly without taking up too much of a doctor’s time.

The NNT to prevent one heart attack in a high-risk patient who has not had a previous heart attack is around 104. That means we need to treat 104 high-risk individuals for five years and cross our fingers that you will be the lucky one at the end of those 60 months. For low-risk individuals, the number is 217. Why would you take a statin if you are low risk? (Mind you, I have heard doctors say everyone over 50 should be on one).

I want to be clear: this post is not an attack against any particular medication. If you are on a statin and are concerned that your prescription needs to be reviewed because of the examples above, please consult your doctor. My intention today is to encourage science and common sense to merge once again, as I believe they used to. Transparency is pivotal in modern medicine, especially as we are surrounded by conflicting interests like never before. If you are prescribed medication to prevent a disease, ensure your doctor talks to you about the NNT. Remember to ask: what do I consider a "good" drug? How many people need to take it for how many years to see one person’s ailment prevented? 100? 50? 10? Or 2? When your decision is informed, whichever way you lean, your heart reclaims its contentment. Long live patient autonomy!