New research published by the British Medical Journal this week has the potential to revolutionise global multi-billion pound research activity into the delivery of health services.
The research was conducted by Dr Richard Hooper, Queen Mary University of London, and Dr Liam Bourke, Sheffield Hallam University. Their work outlines a novel approach to research design which can drastically reduce the number of people involved in some clinical trials, therefore making these trials cheaper and easier to run.
The new technique could be particularly beneficial to trails that involve health interventions at an organisational or institutional level such as in GP's surgeries or schools.
The new trial designs are a departure from the classic "parallel group" approach to running a clinical trial, in which a group of "control" participants are followed in parallel with a group who receive an experimental treatment. Instead they involve repeated samples of individuals from the same general practice or school, say, with some assessed before the new treatment is introduced and some after.
This allows researchers to compare individuals' outcomes within an organisational unit as well as between parallel groups of units. If the samples are taken in the right pattern the number of individual participants required is reduced.
In the last few decades there have been few radical innovations in the overall plan for running a trial. As well as having an important role to play in meeting ethical and financial requirements, techniques for reducing the number of people involved in a trial could create new opportunities for research in small populations or rare conditions.
"This new technique could prove to be a valuable new tool for researchers working in public health and health service delivery" said Dr Liam Bourke, Principle Research Fellow at Sheffield Hallam University. "One of the main barriers to carrying out evidence based medicine and research is often cost. So a method that reduces the number of people required for a trial, whilst maintaining non-biased results, could potentially be a game-changer".
"Something as unusual-looking as our design shouldn't work: but it does," said Dr Richard Hooper, Senior Lecturer in Medical Statistics at Queen Mary University of London. "It's a beautiful mathematical idea that has real, practical use in health research which ultimately benefits the users of health services."
Dr Bourke added "Whilst these new designs are not going to be applicable in every case, the new technique could be the most practical way to conduct some research. The next stage is for the design to 'go-live' and be put into practice."
The report Cluster randomised trials with repeated cross sections: alternatives to parallel group designs, can be viewed at -http://www.bmj.com/content/350/bmj.h2925.full?ijkey=BA5vNtUNciEzflc&keytype=ref
The potential to optimize laboratory testing and operations with automation and AI