Understanding Clinical Trials: What You Need To Know

Subscribe: Apple Podcasts | Google Podcasts | Stitcher | Spotify

Keeping up on the latest medical research can be a benefit but the sheer amount of information – and the sometimes contradictory results – can be wearying. How do you know which data to trust? Afshin Dowlati, MD, Director of UH's Phase I clinical trials program, describes how to separate good-quality, trustworthy information from data that is less reliable.

---

Podcast Transcript

Pete Kenworthy

The latest studies show drinking coffee is good for your health.

Macie Jepson

What study is that?

Pete Kenworthy

Oh, I don't know. I just heard latest study and I thought, is it good for you? What about your heart?

Macie Jepson

I don’t know either. I can't keep up anymore, but I do know this: I can eat dark chocolate.

Pete Kenworthy

Right.

Macie Jepson

And at the end of the day that is all that counts. Right?

Pete Kenworthy

Right. You’re good.

Macie Jepson

But I don't know. I mean, if it was a study of 20 people, that's not a good thing either. So, Pete, a lot of people know this who listen to this podcast. We were in the news media for a long time. Tell me the truth. All the latest research, is some of them that you shared on the news sexy headline or did you look into it and figure out if it was a legit study?

Pete Kenworthy

Yeah, honestly I think I probably left it to the producers, which maybe wasn't the right thing to do either. But honestly, it can be very confusing. How do you know which studies to trust? Hi, I'm Pete Kenworthy.

Macie Jepson

And I’m Macie Jepson. This is Healthy@UH. You know, having the latest medical information at our fingertips every waking moment can be a blessing most of the time. But it can also be a curse when you're getting conflicting information. Nobody wants to knowingly do harm to our health. So, we get jerked around. You know, screen time -- is it good or bad? I don't know anymore. Eggs. Can I eat them or not? I really don't know. It's confusing.

Pete Kenworthy

Yeah. The good news is we have someone here to help us sort it all out so we can make heads or tails of all this research. University Hospitals’ Director of Phase One Clinical trials program, Dr. Afshin Dowlati. Thanks for being with us.

Dr. Afshin Dowlati

Thank you very much for inviting me.

Macie Jepson

Dr. Dowlati, many studies we see pair two things: one recent one that linked to developing cancer and hair dye use. It could be a bit of a leap, but how do we separate correlation and causation?

Dr. Afshin Dowlati

That can be very difficult. So, when we try and interpret a clinical trial or a study, we basically have a pyramid. And the top of the pyramid indicates that the data that we're receiving or interpreting is very circumstantial and difficult to interpret. And as we go to the bottom of the pyramid, the data is much more trustworthy and more understandable and applicable to every person. So, at the top of the pyramid is an anecdote. I took this drug, and I felt better, right? One person, someone in your family did that, someone in your friends did that. That is not science. That is one experience. And utilizing that data to do something for yourself is not necessarily the right thing to do, right? Because it may have been circumstantial, may have been by chance it worked or it just was by hazard that person felt better because they took it at that particular time.

At the bottom of the pyramid is what we call randomized clinical trials, blinded randomized clinical trials where a group of patients generally in the hundreds or thousands received study drug A, and the other group did not receive it. And they look at outcomes. And there's many examples of those. For example, in breast cancer, if you receive a certain hormonal therapy as opposed to not receiving it, your outcomes are much better. Those are what we call blinded randomized trials. Now, in between those, the top of the pyramid and the bottom of the pyramid, you'll find everything on the internet. You'll find various interpretations of those. And we can go through those one by one, but basically as you go from the top to the bottom of the pyramid, scientists have weighed on that particular trial. That trial at the bottom of the pyramid has been designed by scientists who know what they're doing. They've been validated by some sort of scientific community, what we call scientific review committees by the National Institutes of Health or some other organ that's responsible or body that's responsible for designing and implementing these studies.

What's in between? In between could be things like case reports. I, as a doctor can tell you, I have four patients and in 20 years, and I did this for them and it worked, for example. So, that's what we call case series. Or four patients based on my experience. Well guess what? There could be another physician in Florida that treated the same four patients and did not get an outcome. That's what we call case reports. Then there is something called case control studies. For example, studies that I compare my results that I had in 2019, I treated four patients with drug A and I compare them to the same potential patients or very similar that I treated in 2016 with a similar drug or a different drug. And I can say, oh, my 2019 results are better than my 2016 results. So, I'm comparing a group of patients to another group of patients. These are not what we call randomized. These are over different times. They're not the same population. They may be different in age; they may be different in gender, but I'm come making this comparison that, although I feel good about, it is not necessarily sexually scientifically valid.

And if we go a little bit further down on the pyramid, so the wider the bottom, the more validated and the more trustworthy the information is. So, if we get to the middle of the pyramid, there could be studies of like 30, 40, 50 patients in a group that's not randomized. Everyone gets the drug, and we see results. And that's has certain validity to it. If I treat 40 patients with drug A and 20 of them respond, well, that's pretty impressive, you would think, right? Fifty percent of my patients responded to the drug. The problem with that interpretation is it may be 50 percent for me, but if my colleague does this in another city, they may get 20 percent. So, I cannot tell my patient is the chance of responding to this drug 20 or 50 percent? Are they to believe me or they to believe my colleague in Pittsburgh? The truth is that these are all statistical data that needs to be interpreted. The truth is somewhere between 20 to 50 percent probably. So, interpreting clinical trials or data online is very difficult. You have to know the source. You have to know the type of study that was done. You have to know how many patients were on it. And as a physician, it can be very difficult for us to interpret clinical trial sometimes. You know, one day eggs are good for you; the next day they're not good for you, right? They keep changing back and forth. So, I would strongly recommend to everyone to discuss what their findings are with a physician who's qualified in this field and see, is this a study that should be interpreted in this way or in that way?

Macie Jepson

You've just laid it out so beautifully. How can the general person know the difference between the top of the pyramid and the bottom?

Dr. Afshin Dowlati

So, unfortunately it can be difficult because it requires a degree of training, statistical training, clinical trials training. Physicians have been trained to be able to interpret clinical trials correctly. That's part of our medical school curriculum.

You know, what type of trials do we believe in, what type of trials we don't believe in or we need more information on before we can make a judgment for our patient. So, you always have to discuss this with the patient, with the physician. But let me just set some guidelines for you. And the guidelines are that there's three different types of studies. There's something which we call phase one clinical trials. And these phase one clinical trials are trials that are really at the, when they, at their infancy, meaning that we don't know what the dose of the drug is, we don't know how to give it to the person twice a day, once a day. What dose? 40 milligrams, 80 milligrams. We don't know much about the drug, and we're just trying to figure that out. These dose-finding studies, in other words, phase one, clinical trials are really not developed to be able to answer if this drug helps you or not.

Now, every once in a while, the results can be incredible. You put 10, 20, 30 patients on a clinical trial and guess what? Out of the 30, 28 had an incredible response. Now, those are far in-between, and that's not difficult to interpret. If 30 people had 20, out of 30, 28 had an incredible response to whatever drug they were given. But in general, it's not like that. They're very difficult to interpret because we're just trying to figure out the dose, and we're trying to figure out the schedule. What is safe for the patient? That's what we call a phase one study. Those studies generally have very few patients on them. Anything from 10 to 20 to 80, every once in a while, it can be more than that, but generally not. In my experience, they average around 40 or so.

The next phase of trials are what we call phase two clinical trials. Now, phase two clinical trials, we know the dose, we know the schedule. We know in general how safe a drug may be or have some sort of idea about it, and we're putting on anything from 20 to 100, 120 patients. Those are just some numbers to think about, but generally more than a phase one clinical trial. And we're zeroing in on a very specific population. Let me give you an example. So, let's think about breast cancer, and we're trying to figure out if this drug works in patients that have metastatic breast cancer, and they have a very particular type of breast cancer. So, we put 40, 60 patients on this particular trial. These are designed by statisticians. They'll tell us if we need 40 or if we need 60. And we see a certain amount of good responses. Let's say out of 40 patients we see 20. That would give us a 50 percent what we call response rate.

So, these studies are, of course, much, are more easier to interpret in terms of a drug working than a phase one clinical trial. The population is homogenous. We know who the patients are. We're giving the same dose, the same schedule. And if we get a 50 percent response in these patients, we tell them, you know, this drug may be working. But unfortunately, even interpreting these can sometimes be difficult, right? Because a statistician will tell you that if 20 out of the 40 patients responded, guess what? If you repeat this study over and over again, you're not going to get 50 percent each time. One time it may be 20. One time, it may be 70. We just don't know the exact number, but it is enough to make both the patient excited and the physician excited that, listen, something's coming out of this particular trial. We're getting excited.

The next step goes to what we call a phase three clinical trial, which is generally hundreds and sometimes thousands of patients on trial where we're comparing the new drug to what we call the standard of care in a randomized fashion. And if that study comes out positive, meaning your drug is better than the standard, we have a drug that's really applicable. It needs to get out to the community sooner rather than later.

Pete Kenworthy

That's a great way to put it all. So, phase one, phase two, phase three, the higher the number gets, the more reliable the data.

Dr. Afshin Dowlati

Absolutely. There's no doubt about it.

Pete Kenworthy

All right. So, you covered quite a bit, and that's a lot of really valuable, great information for everyone. A couple of the things I want to hit on before we let you go though, because many studies when they start, and maybe this is even before phase one for all I know, are done in animals, right? Instead of humans. How do we interpret that data or does it need to get to humans for it to be valuable to us?

Dr. Afshin Dowlati

No one outside of a scientist should be interpreting data in, in animals. Because I can tell you for someone who's been doing this for 20 years, we've cured a lot of animals with a lot of diseases, but those do not necessarily apply to humans. There's many differences between animals and humans. The disease models, meaning, for example, giving an animal lung cancer: that lung cancer is not necessarily the same as a human lung cancer. Drugs that work in animals may not necessarily work in human beings. Treating animals, for example, what we define as something working is simply, well, maybe the tumor is smaller in 30 days comparing 10 groups of mice to the other 10 group. So, maybe if you give a drug and the tumor in the treated mice, 10 mice, is five inches and compare it to the other group, which is 10 inches, you're going to say, you know, this drug works. So, interpretation of efficacy in animals is a very tricky thing. But also relating that to what's going to happen in humans is extremely difficult. So, most drugs that end up working in animals do not necessarily work in humans. And there's been many, many failures over the years. So, I don't think outside of a scientist who deals with that specific drug in that specific disease, anyone should be interpreting data in animals. Very tough.

Pete Kenworthy,

Along with that, we as humans aren't even like other humans. Right? So, if a study works in someone or a population that I see as, oh, that's me. That's what I, that's my sickness. That's what I have. That doesn't necessarily mean that's going to work for me either. Right? Because just because I have that breast cancer that you described that that population that was working on and I see this trial going on and oh my gosh, it's working for them, there's no guarantee that's going to work for me. Right?

Dr. Afshin Dowlati

Absolutely. So, what we know about human diseases as a whole is that we used to be lumpers, meaning that as physicians, as scientists, we used to tell you got breast cancer, you have lung cancer, you have prostate cancer. And what we now know is that being a lumper is not a good thing. We're actually now dividers. And we now know that there isn't one lung cancer, but there's dozens and dozens of different lung cancers. So, what works in one form of lung cancer may not work in another form of lung cancer. Let me give you an example. About 1 percent of lung cancers have a very specific genetic problem, and you give them drug A and it works incredibly well in that 1 percent. Eighty percent of those patients have a incredible shrinkage of their cancer. Now if you give that drug to the other 99 percent, it doesn't work at all.

So, let's think about a trial where you put 100 patients on the clinical trial, 100 lung cancer patients, and only one of them respond. What are you going to tell me about that drug? You're going to tell me it's a lousy drug. It doesn't work. It only works in 1 percent. But guess what? You're wrong. It works in 100 percent of those patients that have that very specific genetic abnormality. So, we have to be very careful when we lump people together. In the sciences now, moving forward to dividing patients into subgroups that may drug respond to drug A versus drug B, drug C, that's where the science is going now.

Macie Jepson

So, we're talking a lot about clinical trials and we began this discussion talking about what's already out there, what's published, and some of it not being quite as scientifically rich as other things. So, what is your advice to someone who's been given a diagnosis who wants to go out there and start looking for the best course of action? Where do they start?

Dr. Afshin Dowlati

Now, that's a great question. So, my general advice to my patients are, is that you should consult websites and organizations that help with the interpretation of clinical trials. It's very difficult for a single person without a statistical knowledge, without a scientific knowledge to interpret most clinical trials. But guess what? There are people out there who do it for us, right? These are people who actually have patient advocates on their staff. These are organizations that have scientists on their staff, clinical researchers, clinicians on their staff. As an example, the American Cancer Society. The American Cancer Society is an organization that has researchers, patient advocates, gives grants out, supports research, and they have a website that talks about various cancers and what is out there? What is going on? What is standard of care? What is new in what can help a specific patient with a specific disease?

So, I tend to encourage my patients to look at these websites that have done a lot of the interpretation for us. It takes a lot of the guessing out. And I'm sure that such websites exist for most diseases that are out there that we can consult. The National Cancer Institute. We have, in cancer, we have something called the National Comprehensive Cancer Network or NCCN. And in there there's a link for patient information about what's new. And they are very good at updating this on a almost monthly basis, if not several times a year. So, that's how I tell my patients to look at the clinical trials. Trying to interpret a clinical trial simply on based what journal it's published can be tricky at times. Obviously, the most important studies tend to be published in the most reputable journals, journals that have a great track record in publishing important studies such as The New England Journal of Medicine, such as some, the journal The Lancet. But we have to be aware that not all good clinical trials are just published in those journals. They can sometimes be published in journals that are not at the level they are. It is true that the most important studies tend to be published in what we call the higher impact journals. But it's not always the case. So, that's why I think it's very important to look at the professional organizations and patient advocacy groups that interpret these clinical trials for us.

Pete Kenworthy

And of course, talk to your physician before deciding anything.

Dr. Afshin Dowlati

Absolutely. So, nothing is going to beat talking to your physician, and make sure you do talk to them. I find sometimes patients are reluctant to bring up information they found on the internet because the physician doesn't have the time or they're afraid they're bugging the physician. I strongly encourage my patients to bring up anything they find on the internet and discuss it with me because the best patient, the best outcome is always a good physician/patient relationship and an informed patient. I find that extremely important.

Macie Jepson

Dr. Dowlati, thank you. We so appreciate you breaking this down for us and explaining it. It really did clear some things up for us. Thank you.

Pete Kenworthy

Remember you can find and subscribe to this podcast on iTunes, GooglePlay, Stitcher or wherever you get your podcasts. Search University Hospitals or Healthy@UH, depending on where you subscribe.

Macie Jepson

And for more health news, advice from our medical experts and Healthy@UH podcasts, go to UHHospitals.org/blog.