Struggling to come to terms with the new realty? Me too! So I thought I’d try to at least help people come to terms with drug trial and terms! Open-label, double-blind, what do investigators have in mind?! Control arm, study size, placebo – these are just a few of the terms that you should know (and which you’ve likely been hearing and/or will be hearing a lot about these days).
Today’s post (was) going to be shorter than usual – I’m spending most of the day working on continuing to format the translated versions of my Covid-19 testing infographics. Huge thanks to all of the amazing people who have offered to translate into over a dozen languages! Sorry for the delays in getting them posted – it takes quite some time to get all the text nicely copy-pasted-formatted and proofed. But I’m trying my best! more added continuously here: https://bit.ly/covid19bbresources
I only speak a single language, but scientific and medical terms can be so jargonny they can sometimes seem like a whole ‘nother languages so I thought, especially as a follow-up to yesterday’s post on potential treatments for Covid-19, I’d attempt to “translate” some of the jargon around drug testing into more accessible language. more on those potential treatments here: https://bit.ly/2Uaa3Gg
First off, the least scientific type of “evidence” – anecdotal evidence. This can be something like “dude, I rubbed a tomato on my scalp and my hair grew back!” or, more common these days, a Facebook post from a friend of a friend about how their cousin’s fiancé smoked some weed and it helped with their insomnia. Anecdotal evidence isn’t inherently wrong – and it might really have some validity – but we can’t know for sure unless we test things under more controlled settings.
Say you have a group of people with a potentially deadly virus. If you do nothing, some people will get better, some people will die, and some people will eventually get better but experience some complications along the way. That’s if you do nothing. What if you give them all some experimental drug? It’s likely that some people will get better, some people will die, and some people will eventually get better but experience some complications along the way.
Yes, I purposely just repeated that because the same things can happen with or without the drug, so in order to know whether those things are related to the drug, you need to be able to directly compare.
Anecdotal evidence can often gain traction because of something called “regression to the mean.” Say you bang your knee and get a bruise. And the bruise gets really grody looking. And your friend says to rub some horseradish on it – so you do. And – voila! – your bruise starts going away. But the bruise would have gone away anyways – you often try something when things are at their worst, so, when things inevitably get better, you attribute the getting better to the thing you tried. Even if that thing had nothing to do with it.
And it can work the other way too – attributing negative occurrences (adverse events) to some unrelated thing that happened to precede the event. For example, when I was a toddler I had a couple of febrile seizures – I had these kinda random seizures accompanied by a high fever. Thankfully, I was fine and as far as I know, suffered no ill effects (the only consequence is that I have to check the box on forms asking if you’ve ever had a seizure…). And, also thankfully, that seizure didn’t come right after I had received a vaccination. If it had, my parents might have mistakenly, but understandably & well-meaning-ly attributed the seizure to the vaccine I had.
Similarly, kids often get vaccines around the same time that signs of autism begin appearing, so autism can FALSELY get attributed to vaccination, even though those signs would have begun appearing even if the child hadn’t had the vaccine. sidenote: VACCINATE YOUR CHILDREN – PLEASE!
Moral of these stories – it’s really important that doctors carry out controlled clinical trials where they monitor patients with and without some new treatment to figure out what is and isn’t treatment-related. The groups of patients receiving different treatments are sometimes referred to as “arms” – e.g. a group of patients receiving the drug is referred to as a “therapeutic arm” and they’re compared to a “control arm” where patients don’t get the drug, or get the common “standard of care” (more on this below).
Some trials have multiple therapeutic arms, so they can compare multiple treatment strategies (different drugs, different doses, etc.). It’d be great to have a lot of arms, but it’s difficult in part because you need to keep the study size high – you need a large “n” (number of patients) in each group in order for any result to be statistically significant. If you have only 2 patients in each group, 1 dies in the control group and both live in the treatment group, you really don’t have enough evidence to say the treatment’s helpful. But if you have 200 patients in each group, half of the control group patients die and all of the treatment group patients live, now you can have more confidence. Assuming the trial is well-controlled that is…
Trials are “controlled” in the sense that you want to keep as much as possible similar between people who do and don’t get the treatment. Potential differences include things about the patient which are unrelated to the disease (age, sex, BMI, pre-existing health problems (so-called co-morbidities like asthma or heart disease) etc.) as well as things related to the disease (time since disease onset, severity of symptoms, etc.)
We call these “independent variables” since they are *independent of the treatment* – but they *can* influence the effects the treatment has, or appears to have, and thus can confuse the results of a study if you don’t control for them. For example, if all of the patients in your control group are old with co-morbidities, while all the patients in your treatment group are young and otherwise healthy, the patients in the treatment group are likely to have much better outcomes than those in your control group for reasons unrelated to the treatment.
Human-related variables are hard to control for, but scientists try to make sure things are relatively even among all the patients in a trial by setting strict eligibility guidelines regarding age, time since disease onset, etc. They often exclude super-sick people from these types of trials, but please don’t get the wrong impression – know that doctors don’t exclude people due to lack of compassion – it’s just really important that they figure out if the drug is effective – or even harmful?! – before they just go out and give it to everyone- and if there are too many confounding factors, it’s impossible to get interpretable results.
But doctors really do care, so patients who are ineligible for treatment might be able to receive the treatment in other types of studies and/or through compassionate use (aka expanded access) which is a process that allows doctors to give not-yet-approved experimental drugs to patients if the patient meets certain criteria and there aren’t other treatment options available to them.
The variables that scientists have to control for aren’t just physical ones – they also have to take into account psychological variables. In addition to the whole regression to the mean thing, patients will often get hope just knowing they’re on a drug and this positive attitude can make them feel better (placebo effect). Conversely, if a patient has one of those “adverse events” like a seizure during the trial, and they think they’re on a drug, even if they aren’t, they might blame the drug, especially if they’re told the drug could cause seizures. This is referred to as the nocebo effect.
Enter the placebo – this can be a “sugar pill” or some other harmless “fake treatment” that the “control group” patients who receive instead of the real treatment so that the patients don’t know which group they’re in. If the patients don’t know, but the doctors do, we call a study SINGLY-BLINDED.
But, doctors are also vulnerable to these tricks of the mind – for example, if they know a patient is on the drug, they might, subconsciously, more positively interpret effects they see, or even treat the patient differently. To prevent this, the doctor shouldn’t know who is on the drug vs the placebo either, so the gold standard for clinical trials is ones that are DOUBLE-BLIND (neither doctor nor patient knows who’s getting what – of course, someone knows, but not the people who could potentially affect the results.) If both the doctor and the patient know, a study is referred to as OPEN-LABEL.
To further make sure that all the groups are as similar as possible, the best trials are RANDOMIZED – e.g. patients are randomly chosen to receive the drug or the placebo.
I’ve been talking in terms of a placebo, because that’s the easiest to understand – but often in drug trials, there’s some drug that’s already being used for that condition which is somewhat effective. It would be unethical to give patients a sugar pill when there’s something that is already known to be helpful. So, instead of receiving a fake drug, the control group patients receive whatever the currently-used “standard of care” treatment is.
There are different types of drug trials and, in order to become FDA-approved, a drug goes through a series of phases. There’s some variability depending on the nature of the condition, the treatments currently available, etc. but here’s the general gist:
Phase 1 trials test for safety – the treatment’s given to a small group of people (typically 20-80 people) to check for potential side effects. Sometimes the people are healthy volunteers and other times they have the disease
Phase 2 trials – now you start to look at whether the treatment is actually helpful! In Phase 2 trials, more people (~100-300) are given the treatment with an eye on whether it’s effective (and further make sure it’s safe).
If that all goes well, it’s on to Phase 3 – this involves a major scale-up (typically 1000-3000 participants) and if that goes well, a drug can get FDA-approved.
Phase 4 trials – these are *after* a treatment is FDA approved – scientists are just trying to figure out some more about it
How do you determine if a trial is “successful”? When setting up a trial, scientists set out particular “endpoints” – which are the things that they measure in order to see if they meet the “objectives” the study sets out. For example, you might measure the endpoint of days of required hospitalization to see if you meet an objective of a 50% reduction in hospitalization time in the treatment group compared to the control group.
“Primary endpoint”s are the main things you’re measuring – in early phase studies, the primary endpoints involve safety, but since you’re giving patients the drug you might as well check to see if it’s doing what it’s designed to do, right? So, often “secondary endpoints” are monitored as well.
This all might sound kinda cold, especially when people are suffering. But doctors can’t know if a drug is helpful without such clinical studies. For all they know, it could cause more harm than good! When it comes to pharmaceutical drugs, it really is NOT a “nothing to lose” situation. Take chloroquine for example…. well, don’t actually take it – at least not without an “in the know” doctor telling you to – it can cause serious side effects, and there isn’t even strong evidence it’s helpful for treating Covid-19, so it’s irresponsible to advocate for just giving it to all patients willy-nilly.
more on topics mentioned (& others) #365DaysOfScience All (with topics listed) 👉 http://bit.ly/2OllAB0