5 Things nobody told me about running a double-blind human study
Ben Houghton writes about the randomised controlled trials and tribulations of running a double-blind study of oxytocin as part of his PhD.
“Won’t somebody think of the researchers?! After spending years carefully crafting an experiment, begging for a pittance of funding to run your obviously world changing study (spoiler alert – it won’t change the world) requiring the design to be amended so much that where it once was an all singing and dancing musical extravaganza that Hugh Jackman would be proud, it now resembled a group of toddlers bouncing feverishly to Baby Shark. On repeat. One thing has remained steadfastly unchanged – the double blind, placebo-controlled design.
For me this means recruiting 30 opioid dependent males receiving methadone or buprenorphine to receive both nasal oxytocin and placebo after being induced to crave heroin in a crossover design over 4 weeks. Oh yes, during a global pandemic where peoples’ movement is restricted on a seemingly weekly basis. After moonwalking through the ethics and approvals process and the recruitment start being stunted by the aforementioned pandemic, I finally managed to get going. Literally years of planning came down to this data collection…but I was blinded! So, for all those researchers about to start, or who are planning a double-blind study, here are 5 things nobody told me…
“Being blinded removes a large potential for my own bias. I literally do not know what I’m doing”
- The tension is unbearable (which is right, A or B??)
As the data trickles in and I keep tabs on the findings I am torn on a seemingly weekly basis. Is spray A or spray B having the effect? If it’s spray A then we could be on to something but if it’s spray B, then I have found the opposite to what I thought. As data comes in I am forced to see only the findings with no idea what I’m actually looking at. This is GREAT for critical thinking. It lets me explore why my hypothesis could be right and also why it could be wrong. For example, why does one participant have an effect that another doesn’t? What are the variables involved here? How would this impact the long-term goal? Being blinded removes a large potential for my own bias. I literally do not know what I’m doing. Of course, others on the study team DO know which spray is which and I have to look away when they have checked this for fear of me attempting to read any (non-existent) facial expressions. In fact, my biggest fear is being unblinded only to find my entire hypothesis is not only wrong but completely reversed (as could be the case with the data so far!).
- Check, check and re-check
When we ordered the oxytocin I asked ‘do I need to do anything?’. ‘No’ came the reply. When it was delivered, I asked again. ‘No’ came the reply once more, ‘It will all be delivered blinded in bottles marked A and B’. Day one of recruitment, participant arrives and I go to the fridge to get the first spray only to find the sprays were not blinded. I was staring at bottles labelled ‘oxytocin’ and bottles labelled ‘placebo’. I should have checked. Sure I asked, but myself physically checking could have saved that spike in my blood pressure. Thankfully the research and development team quickly blinded a few for me so I could proceed.
- The responsibility is huge
I’ve been a community drug worker for 10 years before this. I’ve been threatened to be set on fire by intoxicated people, had people self-harm in front of me, listened to the details of horrific trauma disclosed to me whilst being expected to have the solution for them but nothing hit me like my experience with one research participant. Part of my study is inducing craving by showing a video of someone injecting heroin to the groin. As I showed this video, the participant physically jolted and recoiled in the chair. I was confused. This was something they did themselves, to themselves, on many occasions but seeing this video triggered this kind of reaction. This emphasised the huge responsibility to me. Real people were taking part in a study that had real implications for them. Providing support is one thing but being the cause of potential distress is quite another and not something to be lost amongst the excitement of conducting research.
- The temptation to be a human guinea pig
I don’t have healthy controls for comparison in the study. The video won’t make me crave heroin but what about the other computerised tests? Maybe I could run the experiment on me, see what happens? Maybe I could see if it’s possible to tell the difference between oxytocin and placebo? I mean, both are odourless and clear but what if I feel something? I should stress that I haven’t done this but I can totally relate to scientists conducting experiments on themselves.
“Any deviation can create ‘noise’ in the data to be accounted for, consistency is key”
- Every time is the first time
I see each participant 4 times. I watch the video 3 times in total, ask the same question about craving 12 times in total and take a reading of attentional bias 10 times in total per participant. It’s fair to say I can get a bit fatigued with giving the same speech over and over but it’s so important that I do. Any deviation can create ‘noise’ in the data to be accounted for, consistency is key. It’s also important for the person in front of me. Sure, it feels like the millionth time for me but for them it is the first time and for them it is all new. As I mentioned above, human research carries a big responsibility and each participant deserves to receive the best experience.
- Expect the unexpected
Just like you weren’t expecting a sixth point in a ‘5 things nobody told me’ post, I wasn’t expecting to have an ineligible participant be so sold on the idea of an oxytocin nasal spray to reduce cravings that they decided to buy their own oxytocin spray off the internet when they found out they were excluded! They have gleefully promised to keep me updated and to use the spray immediately when they get cravings but seemingly have an utter disregard for the meteoric spike in my blood pressure produced by said text. What do I do with this information? It’s not an adverse reaction…do I give harm minimisation and explain the dose is nowhere near enough to do anything…do I offer to train them on anticipated time of effects…no, I scurry off and feverishly smash out an email to the Chief Investigator who is also a clinician at the recruiting site. They’re not recruited to the study so management of this is definitely outside the remit of my research.”
Ben Houghton is studying his PhD at St George’s University of London working at ‘Targeting Opioid Related Deaths: Prevalence, Perspectives and the Potential Role of Oxytocin’.
The opinions expressed in this post reflect the views of the author(s) and do not necessarily represent the opinions or official positions of the SSA.
The SSA does not endorse or guarantee the accuracy of the information in external sources or links and accepts no responsibility or liability for any consequences arising from the use of such information.