HTA Director Hywel Williams – Bias in Clinical Trials

HTA Director Hywel Williams – Bias in Clinical Trials


Hello, I have just come away from lecturing our medical students on the topic of biases in clinical trials,… …and I thought this would be a good topic to mention briefly in my monthly HTA Directors’ message. It applies, after all, to commission topics and researcher-led applications. And the basic message is this: If you are aware of a significant potential bias,… …when designing a study to address an important question to the NHS, then name it. Tell us about it. Say that you recognise the bias. It may be quite subtle, like… … a surgical trial, where there is a waiting list for the randomised surgical intervention… …during which time people start having other surgical procedures or physiotherapy… …or expertise bias,… …or contamination of a control group with elements of the active intervention. Or, if you anticipate significant differential drop-outs between groups. Or, some interventions, for example, are impossible to blind. So, just tell us about these. Because if you don’t, or if you just ignore them,… …I can guarantee to you, that our funding committees will spot them very quickly. There is no need to flagellate yourselves with a long string of anticipated bias,… …so that your application comes across as depressing or nihilistic,… …but just demonstrate to us that you have spotted the bias and how you will tackle it. When we are developing commissioning briefs, we sometimes realise that some of the important questions,… …have significant inherent potential challenges as a bias. And you will often see phrases like “applicants to define”. Sometimes referring to participants or interventions or control group. Now, when you see phrases like that, it is a cue to you,… …and it will give you an opportunity to, sort of, recognise ‘Oh, now I see what they are at’. ‘Now I see why they haven’t been too prescriptive in this area, … … and it is a cue for you then to demonstrate us,… …what creative solutions you might suggest to overcome such potential biases. And I guess, in addition to the more subtle form or bias that I mentioned,… …there is nothing stopping you briefly mentioning,… …how you will mitigate the risk of the traditional clinical trial biases… …such as selection performance, detection, attrician and selective reporting outcomes. So, just like a driving test, where you sort of advertise to the examiner that you are looking in the mirror,… …to see what is behind you… …it’s helpful to advertise to us as well, that you have thought about the biases,… …that you have spotted them, and you’ve told us how you may attempt to tackle them. Because, biases are inherent in all of us, as you may witness from me during the forthcoming Rugby World Cup. On that note, I shall speak to you again, next month.