Written by Alan Barker, Freelance Writer, British Science Festival

Statistics may not be everyone's idea of exciting science, but it supplies vital evidence to shape public policy. Or, as Laura Bonnett believes, it should. She’s been working to influence DVLA policy on granting licences to drivers who’ve suffered seizures. In the Rosalind Franklin Award Lecture at this year’s British Science Festival, she took us through the process. Alan Barker was hooked.

We all have an interest in keeping our roads safe. Whether we drive or cycle, catch the bus or walk, we’re all affected by driving – and the policies surrounding it.

How, for instance, do we decide whether people with long-term medical conditions should be offered driving licences? Laura Bonnett of Liverpool University has sought to answer this question for people who have had epileptic seizures.

It would surely be unfair to say that people with seizures shouldn’t be allowed to drive at all. But most of us would probably agree that a person who’s had one seizure might be at higher risk of having another – possibly while driving. Everyone needs to be protected.

So here’s the problem: how long should we prevent a person who’s had a seizure from driving? In her lecture – accompanied by two (admittedly rather passive) assistants – Laura took us through the process of solving it.

Laura Bonnett and her two assistants on stage at the British Science Festival

Until recently, the DVLA relied on expert opinion for a solution. A panel of neurologists took an educated guess, and came up with the figure of twelve months. Laura finds that decision scary – because, however expert, it’s just an opinion. If we want a decision that will satisfy everyone – the driver concerned, all the other drivers on the road, and citizens generally – then we should base it on more than expert opinion.

We should base it, she says, on evidence.

That evidence has three elements: a mass of reliable data; a risk threshold (more on that in a moment); and statistical know-how.

The data in Laura’s case comes mainly from clinical trials. Laura had access to a trial involving over 600 people. All had had a single seizure and all were 16 or over.

The risk threshold is essentially an acceptable level of risk. Imagine watching a car coming towards you. At what point do you decide not to cross the road? That point is your risk threshold. For the DVLA, the risk threshold in this instance was the socially acceptable level of risk of a driver having a second seizure. Before Laura’s intervention, they thought (or rather, their experts thought) it should be the same as the risk of a newly qualified driver having an accident in their first year behind the wheel. And that risk is 20%. Laura’s job was to assess that threshold and decide if it is, indeed, acceptable.

The last element of evidence is statistical know-how. Laura, as you might imagine, has a fair amount of that. She knows about conditional probabilities, and confidence intervals, and univariate analysis. She took us through each, with admirable clarity, showing how she was able to create a policy framework that generated personalised risk estimates for each driver.

Unfortunately, the DVLA were unwilling to go so far. It was simply too complicated. The neurologists, too, were unhappy; policy based on patient-reported information, they claimed, might be vulnerable to misreporting: drivers would have every incentive to lie, and doctors might treat people who don’t need it, simply to get them back on the road. 

Nonetheless, Laura has managed to convince the DVLA to lower the limit for drivers who have had single seizures from twelve to six months. And she’s been able to convince authorities in the EU to do the same. She’s also influenced the policy for drivers who have taken medication, and those who have decided to come off medication after a period of time. Epilepsy medication, after all, has nasty side-effects – as one audience member pointed out in the Q&A session.

And for anyone concerned that maybe the limits have become too lenient, Laura assured us that accidents caused by other medical conditions – diabetes, for example, or heart conditions – are actually 26 times more likely than accidents caused by epileptic seizures.

Laura believes that all policy should be informed by evidence. And too much, even now, is not. Laura closed by telling us that, if we think a specific policy is founded on anything less, then we should challenge it – and the policy makers themselves.