Hi all – I wrote this to coincide with Clinical Trials Day, and then managed to miss it completely. Smooth move, Knowles! Nevertheless, I still think it’s an interesting topic and one which I’d really love to hear peoples’ views on, so here it is better late than never!
The inspiration for this post is that I’ve often observed negative attitudes toward trials due to people being putting in the ‘control group’, meaning that half of all the people in the trial have the new treatment withheld from them. When I’ve recruited participants into trials for example they often look disappointed if they’re allocated to the control group. I wanted to dig into that a bit, and try to explain why being in the control group is no bad thing, and perhaps is the most important thing anyone can do to help us make progress in health research. So, here are 3 points on why it’s ok to be in the control group in a trial:
1. “Control” in complex trials will rarely mean “you’ll be prevented from receiving treatment during the trial.” New treatments don’t have to be better than nothing these days- we have lots of somethings. They have to be better than the treatment you would have gotten anyway.
This is why ” control” arms are increasingly “treatment as usual” or “TAU” arms. In a pragmatic trial (one where you try to minimise the interference the trial has with things-as-they-are-usually, so you’re leaving situations and decisions to run largely as they would if the trial wasn’t there), you will typically avoid placing any restrictions on which treatments the TAU participants get. For example, in a trial we did looking at computerised therapy, we didn’t restrict access to any other therapies or medication on offer. The reason is that we’re not aiming to get a pure sense of what our trial therapy adds to people who get no other treatment – we want to look at people getting whatever treatments the NHS currently has to offer, and see if adding computerised therapy into that has any additional impact. This is why the name “control” is a bit misleading these days – we often only ‘control’ whether you access the new treatment as well, rather than controlling anything else.
2. New does not mean better.Only just over half of trials find that the ‘new’ treatment is better than the existing one.
What I find interesting in the discussions on trials is that often it’s framed as the ethical dilemma being that you deliberately withhold an effective treatment from the control patients. This doesn’t follow though, if you accept that you don’t know if a treatment is effective or not until you’ve done the trial. And as the statistic above shows, a lot of the time that ‘new’ treatment wasn’t any better anyway. To look at this another way, I’ve seen arguments that every trial with a control arm should promise to provide the new treatment to all the control patients once the trial is finished. But this again assumes that the new treatment works. If we find out it doesn’t work, what’s the equivalent for ‘correcting’ the fact that we gave everyone in the intervention arm a dud? We can’t go back in time and take it off them again. The point is we don’t know whether the treatment is better or not. That’s what the trial is for in the first place.
3.If anything, the control arm is the most important.It’s the only way we really find out if the new treatment is worth using.
I’ve had participants in the past who, upon hearing they’ve been randomised to the control arm, assume they’re effectively out of the study and won’t be hearing from us again. This can be problematic as it’s actually critical that we hear about the progress of people in the control arm.
This is because quite often people tend to get better (or their symptoms change) over time anyway. If we only looked at people who got a new treatment, and they get better, we might automatically think it’s because of that treatment. If we can compare them to people who didn’t have the treatment though, and it turns out everyone got better, then this shows that the new treatment is a bit of a waste of time and/or money, as it didn’t actually improve things for the people who received it in comparison to what would have happened anyway. So, the only way we know if the treatment actually added anything and made a difference is by looking at the people in the control arm. They’re absolutely essential to finding out if something works – or not.
So, that’s my 3 point list for why being in the control group isn’t as bad as I think it’s assumed to be, and why actually it’s probably the most important position in the trial. Let me know in the comments if you agree, disagree, or think I’ve missed anything out!