Clinical Trials

Blinded trials

There are several ways in which the results of trials can be made as reliable and accurate as possible. One of these is to make the trial a ‘blind trial’. In a blind trial the participants are not told which group they are in. This is because if they knew which treatment they were getting, it might influence how they felt or reported their symptoms. Some trials are double-blind, which means that neither participants nor the doctors treating them know which people are getting which treatments. This also avoids the doctors’ hopes and expectations influencing the results of the trial.

Amanda describes a double-blind trial of pro-biotic yoghurt to control irritable bowel syndrome....

Amanda describes a double-blind trial of pro-biotic yoghurt to control irritable bowel syndrome....

Age at interview: 54
Sex: Female
SHOW TEXT VERSION
PRINT TRANSCRIPT
And then the last trial was very recently, which is, I have irritable bowel syndrome, not - it comes and goes, I mean, I can be perfectly normal for months, and then I can get attacks. And so there was a local researcher in the university I was working who was doing a trial of pro-biotic yoghurt versus plain yoghurt, not pro-biotic yoghurt, to see whether it controlled the symptoms of irritable bowel, and I was eligible and I signed up for it and it required you to drink, to eat two cartons of yoghurt a day for six months, I think it was, or - I can’t even remember. It was a long time, I remember feeling quite yoghurted-out by the end of it. And you’d just pick up your yoghurts in boxes with a number, you know, LR0911 on the top or whatever it was, and had to fill in a diary about how you felt. And I was interested in doing this trial because I’m interested, I’m actually interested - I’ve just switched jobs - I’m interested in the public running trials on themselves, you know. So people who like myself were interested in generating knowledge, we’d get together and we’d do it. And I just wanted to see how easy it was, and actually it’s, it’s quite difficult. You forget to fill in the diary and you think, “Oh, what was it last Wednesday?” or you scribble it on a piece of paper because you don’t have your book with you because you’ve gone away and, you know, you really don’t feel like a yoghurt but you have to eat it, and I was very compliant, I’m sure I was much more compliant than most people, just because there’s part of me that’s a bit of a researcher and so it makes you realise just how much noise or rubbish must be put into trials really because you’re thinking, you know, “What was it?” And then I’d think, “Oh, well, I remember” and then I’d find my bit of paper and it wasn’t true. Then I’d have to go back and change it, you know. So, but if I hadn’t found the piece of paper that misinformation would then have been put in. You had to, how many stools did you pass and what, how firm were they. There was a little chart where you had to categorise your, [laughs] your stool and this sort of thing. And actually during the whole period I was asymptomatic, and I was actually quite interested in how the research nurse dealt with me, because she, I think she was frustrated because there were no symptoms.
 
And also I was neither better nor worse and so the other half of her was wanting me to say I was better, when I didn’t necessarily feel I was better, I just felt I was unchanged. And I felt people were wanting a certain result, I really did get that impression, so I thought it’s, you know, it’s really important that staff are blinded as well. They can’t, you can’t help but want a result, that what nobody wanted to hear was, “Well, I’m just the same. You know, I haven’t noticed any change at all”, you know, so which is, was my situation. Later they told me that I had been on, on the pro-biotic yoghurt, which was a good thing because I was asymptomatic and I remained asymptomatic. So, who knows? Well, I say it was a good thing, but who knows whether that was, I haven’t seen the results of the trial yet, so….
 
Very difficult isn’t it? Particularly with those self-report kind of conditions--
 
Yeah,
 
--where there’s no sort of objective kind of, you know, like a tumour size shrinking is in a way easy to measure, isn’t it, but something like that?
 
I really think these things have to be blinded from staff as well, because you can’t help, I mean, if you’re running a trial, I want to run a trial, you can’t help wanting the result, because you, you wouldn’t be doing the trial unless you believed your intervention was effective, really, I think. Although they say you’re supposed to be in equipoise*. I don’t think you would ever begin a trial unless you actually believed there was good grounds to think that this will work.

* FOOTNOTE: ‘Equipoise’ means a situation where clinicians are uncertain which treatment is best – literally it means they are balanced equally between them.
 
Amanda explained how the yoghurt cartons were labelled with a number so no-one could tell whether they contained the pro-biotic yoghurt or a placebo version. Phil was in a study which combined a comparison of named drugs for high blood pressure and a blinded comparison of a cholesterol-lowering drug and a placebo.

Phil explains how the cholesterol-lowering drug and placebo were blinded, but he guessed he might...

Phil explains how the cholesterol-lowering drug and placebo were blinded, but he guessed he might...

Age at interview: 58
Sex: Male
SHOW TEXT VERSION
PRINT TRANSCRIPT
I knew what blood pressure pill it was. That was a particular brand of pill. The only one I didn’t know, of course, was the, the cholesterol thing. And that was, because that was like a separate, it was like the two studies in one. You know, it was on the one hand you had the blood pressure test, on the other they wanted to see what result cholesterol lowering had. So that was, that was the one you didn’t know if you were on or not.
 
How did they do that? Did they have a, sort of a blank bottle or something, with no label on, or how--?
 
Yeah, yeah, you’re just taking the pill and you’re just marked up with the, the, you know, the actual name of the trial, ASCOT trial. And it didn’t say anything. You didn’t know if you were taking sugar or a cholesterol-lowering thing. You’re just taking a tablet for, you know, several years. And it was marked up, it was coded, each one was like coded with your, I suppose your own code number and the trial number. So eventually they could, they knew at the end of the trial what, who had been taking what, but not, not during the actual trial.
 
So did the doctors send you off to another room to get that from someone else? Or did they just take the coded bottle out of the cupboard?
 
They took it out of a cupboard. While you, when you went every six months to have your tests done, while you were there they gave you the next six months’ worth of pills. There was the cholesterol, you know, the dummy pill, or whatever it was in one box, and there was the, the actual cholesterol - the blood pressure tablet.
 
But they didn’t know which?
 
They, no - well, they said they didn’t anyway [laugh]. I trust, I take it they didn’t.
 
It was hinted about halfway through, it was hinted at that I was on the cholesterol anyway. Because the, although the doctor didn’t actually, he couldn’t really tell me, he said, “I don’t know what the...” you know, he didn’t know himself which one I was on. But he said, “Whatever it is, I would keep taking it. It seems to be doing you good” you know. I mean, my cholesterol came down from about over 6 to about 3½. Which is pretty reasonable, you know. So having been told that, it was obvious that, you know, I was on the cholesterol pill. 
Kate and her husband also began to guess which arms of the trial they were in, comparing grass pollen injections with placebo to build up resistance to hay fever.

Kate and her husband guessed correctly that she was getting the active injection and he was...

Text only
Read below

Kate and her husband guessed correctly that she was getting the active injection and he was...

Age at interview: 38
Sex: Female
HIDE TEXT
PRINT TRANSCRIPT
The theory was that your body would build up a natural resistance to the grass pollen. So they start off very low dosage and gradually increase the dosage of grass pollen in your system, until your body was able to fight the reaction by itself. For me, I definitely noticed an improvement in my symptoms. Although it was a double-blind randomised controlled trial, I think it’s called, so nobody knew who was getting what, I was pretty certain from the outset that I was getting injected with grass pollen because I was getting a local reaction to every injection. My husband wasn’t getting a reaction, so we kind of guessed that he probably wasn’t on the grass pollen injection. And at the end of the study, when they announced the results to us and told us whether we were or were not on the grass pollen and it was confirmed that he was on the placebo, they did offer for him to go through the process again, to receive grass pollen injections, go through the whole process but obviously without having to keep the diary, just so he could gain benefit. Because it was shown to be of benefit.
 
Did [the research nurse] ever say anything about your suspicions that you were on it and that your husband wasn’t on it?
 
I think it’s very difficult to sort of avoid it, because to me it was so obvious that I was receiving the grass pollen just because of the reaction I was getting. But he always said, “Of course I couldn’t possibly comment.” Sort of made a joke of it. But I think we both knew that I was clearly on it.
Kate and her husband were proved right about which arm of the trial they were in. However, it is possible that someone who is taking a placebo may notice some improvements in their health. This is known as the placebo effect or placebo response. It is also possible that someone taking the active ingredient may notice no change.
 
As Kate explained, her husband was offered a course of the active injections after the trial ended, because it had shown they did improve hay fever symptoms. But she also noted that one problem with guessing which treatment you are on is that it may make some people want to stop taking part. Her husband did in fact continue, but it was a big commitment, knowing he was on the placebo. (See also ‘Feeling about being in a placebo-controlled trial’ and ‘Thinking about withdrawing from a trial’).
 
Normally blinding requires that the active treatment and the placebo should look as similar as possible. Amanda tried to set up an ‘n of 1’ or ‘number of 1’ trial on treating her rheumatoid arthritis. She would have been the only participant and would have switched between a placebo and a drug she had not taken before, without knowing which she was taking. (This is known as a crossover trial design). In the end the trial did not go ahead, but she was struck by the fact that the placebo did not look like the active drug.

Amanda wanted to run a trial on herself to see if a change of drug made any real difference to...

Amanda wanted to run a trial on herself to see if a change of drug made any real difference to...

Age at interview: 54
Sex: Female
SHOW TEXT VERSION
PRINT TRANSCRIPT
I had been on gold injections for many years for my arthritis and, you know, if I got iller they put the dose up and then as I got better they would lower the dose and, and I was just thinking, “Well, you know, this is crazy. Of course I’m going to get better. If I’m ill I’m going to get better, because it’s a disease that cycles.” And I had asked them - this was long before I was a doctor - if they would just give me water sometimes instead of gold and not tell me about it, at which they laughed and said “Oh no, we can’t do that.”
 
Later on I was told I was being withdrawn from the gold, because I had beginnings of kidney damage and, you know, they would try something else, azathioprine, which is quite a toxic drug. So this time I asked if it could be introduced experimentally without me, you know, being told what the drug was, because they said the gold had probably not been doing me much good even though I’d had it for years and years and years. So it’s known as an ‘n of 1’ trial - a one person trial where you have the drug, then you have something that looks like the drug that isn’t the drug, and you don’t know when it’s starting and ending, to see whether actually it is related to your symptoms, because in a disease like rheumatoid arthritis you can be very ill and then you can be better just because the natural variability of the disease.
 
And my rheumatologist, she was very keen to do it with me and we looked up how these trials were done. We found a pharmacist who is experienced in n of 1 trials, we got the drugs from the [laughs] pharmaceutical industry, and we were going to do the trial. But then, interestingly enough - which was simply giving me the drug - we were then told we couldn’t do it because it would have to go to an ethics committee because it was an experiment, which I found ridiculous. You know, you can give me the drug, you don’t know whether it’s doing me any good, it’s toxic, but to give me it in a way that I could actually find out whether it’s doing me good suddenly we weren’t allowed to do it. And the other thing that was very interesting about that trial was the placebo was different from the drug, so you could tell which was which. And we contacted the pharmaceutical company concerned and they said yes, it was always like that, and it was like that in the trial, which made me very suspicious now about blinding in trials. You just think, “God”. I could distinguish those. So then we had to look up things like, we were going to crush them or put them in capsules, but because it’s toxic, that breached health & safety regulations. Then we were thinking about covering [laughs] them in chocolate. But by the time we’d done this I was getting very ill. My consultant just started me on the drugs, she said, “We just can’t wait to set this up”, and it didn’t work. I actually got to quite a toxic dose. She said “This is clearly not working” We were upping the dose and then we just stopped it, so we never got through the process of doing the trial, because it was blocked by the powers that be saying it needed to go to an ethics committee, and that was just taking way too long for me as an individual patient.
 
Were you trying to go through the ethical process?
 
Yeah, we, at that stage we were going to go through the ethical process, we’d drawn up a protocol, we had everything written and we were still exploring methods of making the drug truly blinded so I wouldn’t know what I was on and that - so we, we hadn’t actually submitted it to an ethics committee because we were still, hadn’t – we’d just got to the chocolate-coated tablet idea when, by which time [laughs] we she’d decided it was toxic and was not going to do any good anyway.
 
So then I stopped that.
She said “It’s made me much more sceptical about trials… I’m also sceptical about trial results because people want things to happen.” On the other hand Danny noted a rather different problem. She and her husband both started the same trial for high blood pressure medication. In the end he was ineligible and she dropped out later because of side effects. But she wondered what would have happened if they had ever got their tablets mixed up. “We were both given a packet and there was no way of differentiating his packet and mine.”
 
Sometimes it is not possible to conceal what treatment someone is getting if there are very obvious differences. For example, the trial Jenny was in was comparing different medications for heavy periods with fitting a coil or intrauterine device (IUD). In Judith’s trial, patients were randomised to have chemotherapy every two or every three weeks. Even if treatments cannot be blinded from the doctors or patients, one possible solution is to send the results away for analysis, so the person analysing them does not know the patients or what treatment they have had.

Jayne explains that blinding to patients and staff is not always possible, but you can still make...

Text only
Read below

Jayne explains that blinding to patients and staff is not always possible, but you can still make...

Age at interview: 47
Sex: Female
HIDE TEXT
PRINT TRANSCRIPT
Blinding is so that you – well, if possible - you don’t know whether you’re getting Treatment A or Treatment B, but in a lot of trials that’s not possible because there are differences in Treatment A and Treatment B. In the trial that I was in, one was a tablet and one was an injection, so it’s pretty obvious. So it’s not always possible for the patient to be blind, and it’s not always possible for the clinician to be blind because they know whether they’re giving an injection or whether they’re writing a prescription, or in our case whether they’re fitting the different appliances. But what you can do is to get an independent person to look at the outcomes, and that’s probably the - I mean, if everybody’s blind to it that’s great, but if an independent person looks at the outcomes they don’t know what treatment the patients had and that’s how we get round it in our discipline [dentistry research] because we, a lot of our outcomes are assessed from photographs, models or x-rays, and you can blind them to the person that’s looking at them or measuring them.
 
So they have no idea what happened?
 
No, which – yes, because again people have in-built prejudices and they think that Treatment A is going to be better, so they sort of want it to win, so they might measure it slightly more favourably.
A lot of the terms used in trials are quite complex and closely related to each other. It was evident in our interviews that people sometimes got them mixed up, even including some well-informed people who had worked in health care. For example, people asked about their understanding of why randomisation was used might give a good explanation of why a control group was needed. Sometimes people were confused about the difference between placebos and controls, or between randomisation and blinding. Below we give some brief definitions as a reminder of how they fit together (See also ‘What are clinical trials and why do we need them?’ and Resources and Information).
 
Control group – a comparison group in which people often get a standard treatment. If no standard treatment exists the control group receives no specific treatment or a placebo. Comparing the results of people in the control group with people in the experimental group helps assess whether there are differences between the new treatment and the control treatment. (See also ‘Feelings about being allocated (randomised) to a treatment group').

Randomisation – allocating people at random to one group or another, so that each group contains a similar mix of people. Random allocation helps ensure we are comparing two very similar groups of patients, so if one group does better than another, it is very likely to be because the treatments being compared have different effects, and not because of differences between the people in the groups. (See also ‘Feelings about being allocated (randomised) to a treatment group').

Placebo – a treatment with no active ingredient which is designed to appear very similar to the treatment being tested. By comparing people’s responses to the placebo and to the treatment being tested, researchers can tell whether the treatment provides any additional benefit. (See also ‘Feelings about being in a placebo-controlled trial’).

Double-blinding – making sure neither patients nor doctors know which treatment each person is getting, so that this knowledge does not influence how patients feel or how doctors interact with their patients or interpret the results.

Last reviewed September 2018.
Last updated July 2011

Copyright © 2024 University of Oxford. All rights reserved.