Keep your family healthy | Top medical experts educate and help
Home » Coronavirus » How to understand COVID-19 clinical trials: randomized vs observational. (1)
How to understand COVID-19 clinical trials: randomized vs observational. (1)
See all Expert Centers:
Return to Coronavirus Expert Center:
If we switch to the COVID-19 pandemic, it resulted in an unprecedented focus on clinical trials of medications and vaccines. Clinical trials of new and repurposed medications are moved forward at great speed. Clinical trials use very different methods. Sometimes clinical trials change goals in the middle of a trial based on a rolling analysis of data. And the cost of such rush can be very real. The clinical physicians at the front lines can be confused; conflicts of interest could be present, resulting in lost lives. And let all of us benefit from your enormous expertise in two ways. So first, perhaps we could discuss a few prominent clinical trials that address COVID-19 therapy. And second, let’s get a 10,000 feet view and address the basics of the clinical trial analysis. One of the things, of course, is that we need to try and make sure that we make valid comparisons of treatments. That’s our fundamental target. We would ideally treat a group of people with a treatment, then press a rewind button on time and take them back to before they had that treatment. And then follow them up with an alternative or no treatment so that we could see what happens to people, firstly, under the condition of the treatment, and secondly, under a condition where they had no treatment. Still, we need to press the rewind button on time, because at the end of a period of treatment, they aren’t the same as they were at the beginning. But of course, that’s only a mathematical concept. You can’t do that in reality. So in nearly every instance, we have a group of people who are given a treatment and another group of people who are given no treatment or an alternative, and we try and make sure that the people who are given the time treatment are the same as those who are given the control. Now, if we only observe what happens, and allow physicians to allocate patients to the test treatment, and then simply accept people who’ve got control, this is an observational study. And we don’t know they’re really comparable. So we randomized them to either receive the treatment or the control. And then we can be sure that on average, the group as a whole are similar, who got the treatment and the control, and then we try and follow them up in exactly the same way. Ideally, no one in the study knows whether they’re getting the treatment or the control, but sometimes that’s impossible. And so we have to live with the fact that somebody knows what sort of treatment they’re getting. Their doctor at least knows. In such circumstances, we try and make sure that we have objective measures of what happens to them what we call the outcome. And if we have objective measures of outcome, especially something like mortality, classify classifying someone as dead or alive on the whole is fairly easy. And that means you can’t bring subjective elements into it. If you have subjective elements, then it is much more difficult if you know what the treatment is you have expectations, you hope that the treatment is going to work or you feel that it doesn’t. And so we try and make sure that these comparisons we make are as valid as possible. We also need to make sure that we study enough if I simply have one person on the treatment and one on the control. We know that random variation between people. People might mean that the person on the treatment does badly, but it’s to do with their initial health state and not to do with the treatment. So we need to have enough people that the random variability between people is dealt with. And we have similar groups as a whole. If we have to deal with rare outcomes, we need much larger numbers. If only one in 100 people dying, and we’re studying death, and we only study 90 people, then we obviously won’t see a difference between the treatment and the control. So we need when there is only a 1% mortality rate if we want to have that we’re going to need to study thousands of patients possibly. So we need to have sufficient numbers. We need to have designed the trials properly. Regarding COVID-19, as you hinted at, that is that we need to be aware that sometimes we are wanting to try out something totally new. In other cases, we can use a medication for which we already have quite a lot of experience. And we know that it works in some other condition.
And an example of this is hydroxychloroquine, which has been used for treating malaria and autoimmune diseases. And so we know quite a lot about it. We know the long list of adverse effects that it has. The idea that the medicine has no adverse effects is unrealistic and simply untrue. I sometimes say to students, I have an aphorism that says, Every effective medicine [drug] has unwanted effects that are usually adverse. In our current situation, we are trying to do all of this process as quickly as we can