Research published last year suggested that around 110 lives a year could be saved in the UK if people did not take certain types of drug.

The same study found that people on some of the most common tranquillisers could be as much as 50% more likely to have accidents than those not on the drugs.

Previously, a European Commission report suggested that at least 10% of those injured in traffic accidents had taken medicines that affected the brain and central nervous system.

Some doctors are concerned that newer drugs may not have the same effect on reaction times, but the old drugs are still being prescribed.

Quantifying the risk

The new study is being carried out at the University of Surrey, and aims to determine exactly what effect the drugs have on reaction times.

Fran Ridout, an experimental officer involved in the research, said: "Many accidents are a direct consequence of slow reaction time in a critical situation.

"Some of the older medicines, particularly tricyclic antidepressants and benzodiazepines, have a detrimental effect of driving performance that we can now quantify."

The researchers will use a test car on a special circuit to measure patients' reaction times.

The car has a red light fitted on the bonnet to simulate the braking lights on a car in front.

Braking distance

Subjects will drive around the circuit at 30mph and, when the red light comes on, will have to slow to put it out.

The researchers say the test conditions will allow them to measure reaction times to the millisecond.

Ultimately they aim to develop a database of medicines that are safe to take before driving.

Researchers from the Medicines Monitoring Unit at Ninewells Medical School in Dundee carried out last year's study.