We're told that 69 per cent of car crash fatalities took place on rural roads. Now what is the most important term in that sentence?

Is it a) 69 per cent b) rural roads c) crash d) fatalities?

I'd argue it's option d). Fatalities. Not accidents. Not collisions. But fatalities. That means that if you don't die you don't form part of the statistics.

This debate has all the hallmarks of the 'eat spinach for extra iron' argument, which persists as a piece of a dietary advice despite it being shown that the reason spinach has ten times more iron than other vegetables because a Victorian researcher put the decimal point in the wrong place.

There are two reasons that 70 per cent of fatalities occur in the countryside: the combined collision speed of the cars, which are rarely separated by a central reservation, and the distance of the collision site from the nearest hospital.

If such a collision took place in an urban area, the ambulance service would be aiming for a response time of less than eight minutes. You'd be hard-pressed to get there within half an hour in the countryside.

What are the statistics that would properly inform this debate? How about the number of accidents per passenger mile travelled. That would help decide. Any better ideas?