Clinical trials are pretty amazing. For 60 years, not only have they provided the gold standard of evidence for new medical treatments, the basic design of the first modern clinical trial was so good it has remained essentially unchanged for decades.
This design, known as a parallel groups design, involves one group of people being given an experimental treatment whilst another group are given a 'control' treatment (e.g. a placebo or routine care). The two groups are followed up and assessed concurrently so the impact of the experimental treatment can be fully measured.
Perhaps the most common variation on the parallel groups design is the crossover design, in which one group of people has the control treatment followed by the experimental treatment (and is assessed after each), and another group has the experimental treatment followed by the control.
Developments to trial design, such as the crossover, have an important advantage in that they require fewer people to take part in the research. This is beneficial because all clinical trials must be carried out according to strict ethical principles – one of the key principles being to have enough trial participants to provide a reliable answer to the research question, but not to involve more participants than are necessary. i.e., not to expose more people than necessary to medical research. Trials involving fewer people are therefore a good thing, and they may also be cheaper and quicker to run.
But… crossover trials can't always be done. They assume that when you withdraw the experimental treatment, people go back to the 'blank canvas' they were before, allowing you then to evaluate the 'control' treatment. But often treatments have lasting effects – even when you withdraw them, people don't go back to being the exactly as they were before they had experimental treatment.
Enter the dog-leg design. This alternative to parallel group design is a brand new, clever yet simple method of trial design which not only reduces the number of people taking part in a clinical trial, it doesn't assume (as the crossover design does) that people revert back to neutral when you withdraw the experimental treatment.
The design works like this. There are three groups of participants: one gets the control followed by the experimental treatment and is assessed after each; the other two groups are each assessed only once, but timed and treated so as to allow a treatment-control comparison at each of the two assessment times. Crucially, in many situations a dog-leg trial needs fewer people and fewer resources to achieve the same statistical power to detect a treatment effect as a parallel groups trial.
This is the strange yet brilliant thing about the design: it uses more groups but requires fewer people. It shouldn't work but it does – when you do the maths you find that the dog-leg method is a surprisingly efficient way to run a trial.
The dog-leg method, which has been reported in the International Journal of Epidemiology (the leading journal on methods in medical research) takes mathematical and statistical ideas and applies them to real problems in medical research, with real impact on patients and the NHS. It will be particularly useful for research looking into chronic diseases, and for so-called 'pragmatic' trials (looking at how treatments will actually be delivered by health services.)
Though similar to a crossover design, the dog-leg design owes its inspiration to another variant on trial design called a stepped wedge. In a stepped wedge trial all participants end up getting the experimental treatment, but the treatment is staggered over a (sometimes large) number of groups and time intervals.
Stepped wedge designs can involve a heavy burden of repeated assessments, so we wanted to try and write down the simplest, most pared down stepped wedge design that we could. Writing 'A' for 'after treatment' and 'B' for 'before treatment' we came up with this crooked arrangement of the letters ABAB, and realised we were looking at something new and potentially very powerful.
It is the crooked, offset arrangement of assessments in the dog-leg design that gives it its name. But do we wish we'd chosen something less awkward sounding? Well truth be told we like awkward, but we'd overestimated people's familiarity with the term 'dog-leg'. One reviewer of our paper asked us what canine anatomy had to do with clinical trial design. Answers on a postcard, please.
Hooper R & Bourke "L. The dog-leg: an alternative to a cross-over design for pragmatic clinical trials in relatively stable populations." International Journal of Epidemiology 2014; DOI: 10.1093/ije/dyt281