Why This FDA Change Is Good News for You

So let me go into what went wrong with the Women's Health Initiative and how to interpret the statistics. So this is often the study, the study that was presented as proof that hormone replacement therapy is dangerous for women, increasing risk for breast cancer, heart disease and stroke. However, let's really look at the data. Now, I know I've flogged this course and I have done it a million times, but I want to make sure that we all are clear. So, number one, the Women's Health Initiative was the first time we actually had a women's study that was actually paid for. And it was the most expensive study ever committed to at that time. So. So there was a lot of pressure to show something, right, to show something here. So the Women's health initiative enrolled 16,000 women with an average age of 63. Less than 10% of those women were actually in their late 40s and early 50s. So all of these women were significant outside of the menopausal window. Now, the other thing is the study population wasn't healthy. 70% were overweight or obese and nearly half smoked or had hypertension. All of those are independent, independent risk factors for heart attack and stroke on their own. Doesn't matter what age you are, if you do those things or you have that condition, you're at greater risk. And like I said, only about 10% were in that first 10 year window. Now, now, why did they do this? Right? So when we look at the Women's Health Initiative, the reason why they did this is the goal of the study. So their working hypothesis were these questions. Does hormone replacement therapy in the form of conjugated equine estrogen, which is horse mare urine with 13 different estrogens, or conjugated horse mare urine with a synthetic medroxyprogestin, increase or decrease the risk for heart attack and stroke? That was the primary goal of the study. It was not to look at breast cancer, it was not to look at all these other things. Now, one of the first things you need to understand about studies, the study quality and the method of the study is designed uniquely around answering the hypotheses or questions. Now, one of the things that happens a lot in research is we create a study. I've created a study. I had some very clear questions that I wanted to ask. And my hypothesis was either this is true or not true, right? Or null or not null. Right. Is basically what it shows up in a research study. So everything you do about that study then is designed to answer that very explicit question. So this study was designed to do that, there's a reason why they left out healthy women. We needed to see, in a relatively short period of time, like 10 years, whether stroke and heart attack risk was reduced or increased with the use of those particular medications. Now, there was a group that NEP that weren't on hormones at this time, although some of them had been on hormones previously because that was standard of care. But that was the question they were answering. They weren't answering all these other questions at all. Now, the study was designed around that, so they left out healthy women. Why? Well, if I need to show mortality from heart attack and stroke within a certain period of time or advancing heart disease, if I put healthy women in there, I'm wasting money and time on women who have a delayed risk for that. Right? Now, if you're obese, a smoker, and already have high blood pressure, you already have cardiovascular disease, especially by your 60s, right? We already have risk factors. So they left out healthy women. So the first thing you have to understand is if you are a healthy woman, normal, relatively normal bmi, heart, you know, height, weight ratio, right? You're not obese, you're not significantly overweight, you don't have high blood pressure, you're not a smoker or an ex smoker. You cannot take the data from the study and apply it to your own health because it's not generalizable to a healthy group of women. Because they picked women who were unhealthy and because they wanted to see would they have disease, you know, disease increasing or mortality. So they looked at those women, and in the group that was on conjugated equine estrogen and medroxy progestin, synthetic progestin, they had an increased risk for stroke and heart attack, which stopped the study. Now, in that group, they had a higher risk. And all the studies that we have had since then that looked at medroxyprogesterone keep showing that there is an increased risk with that synthetic form of progestin for heart attack and stroke because it causes increased clotting risk in older women. But, hey, let me tell you something. It also does that in young women, too, that progestin, the synthetic progestin, increases that risk, so they appropriately stop the study. Now, however, what they then did was called data mining. So data mining is when you start digging through data, asking other questions the study was never designed to do. Now, now, they did a lot of that. So they started digging to look for breast cancer risk. They started digging to look for other risks, dementia risk. When the study was Never, ever designed to look for that. Now, the reason why this is really important is when you create a study, part, part of your method is to include and exclude things and also account for what they call confounding variables. Confounding variables are other variables that might influence the outcome of the data you're looking at. Right? So if I'm looking at dementia, do we have family history? Do we have, you know, other, other things? Do we have other risk factors for breast cancer? Because the study needs to be designed around that, and those things need to be identified first before you even start the study. But data mining is done all the time, and it's a way for a researcher or a group of researchers to regurgitate their own data again and again and again. It's also cheaper. So we have to understand that sometimes we do that in research because we can't conduct studies all the time using randomized controlled clinical trials because they're so damn expensive. So data mining is part of that. However, when you data mine, you have to be clear about what you can't control. For. Now, one of the things that they did is when they were data mining, particularly around dementia, they switched endpoints. So the endpoint is the question you're asking. So the endpoint they were originally looking for was dementia, and then they switch when they found no data to support their hypothesis, right? Their beliefs, their bias, they switched the endpoint to mild cognitive decline. Right. And so they kept changing it. The other thing is, is they also changed what groups they were looking at. So when you compare groups, right, so this, this was a comparison between women who took Premarin, Prem Pro, or nothing at all. Now when you do that, you have to keep those groups distinct. But what they were doing with their data mining is they would put groups together to, to get their own outcome. So they would combine a group in one particular, you know, regurgitation of data, and then in another group, they would combine other groups, which then therefore actually skews the data. When you're comparing between groups, you keep them distinct and separate because you can't, you can't extrapolate the data between groups when you're constantly moving people between groups. So what that means is, is they data mined to fit a particular bias to then present what they wanted to present, which was to get away, get, stop the runaway train of hormone replacement therapy for women. And that was their bias. And they kept massaging the data to get to that answer.