How to halve cycle accident rates

This article was published in 2004, in Newsletter 53.

An item in the national publication Cycle Campaign News caused me to do some digging in a journal Injury Prevention. I’ve often hypothesised that as a cyclist the chance of being involved in a crash in Cambridge is lower than elsewhere, but now there is a suggestion in data from elsewhere that I might be right.

The article ‘Safety in numbers: more walkers and bicyclists, safer walking and bicycling,’ by P L Jacobsen, is a detailed analysis of data from a number of countries and sources. At first I was worried because some of the data from the UK was a time series, but much of the other data is comparing differing towns in each of California, Denmark and the Netherlands.

The analysis of these data suggests that as the volume of cyclists and pedestrians double, total injuries to these users rise by only 32%. It is not suggested that these changes are caused by differing engineering measures between these towns but simply by the fact that where there are larger numbers of ‘vulnerable users’ motorists modify their behaviour.

Some of the key points in the paper are:

  • Motorists’ behaviour evidently largely controls the likelihood of collisions with people walking and cycling.
  • Efforts to enhance pedestrian and cyclist safety, including traffic engineering and legal policies, need to be examined for their ability to modify motorists’ behaviour.
  • Policies that increase walking and bicycling appear to be an effective route to improving the safety of people walking and bicycling.

Cartoon indicating mathematical attempt to prove the above

An alternate hypothesis might be that as cyclist and pedestrian numbers rise they become more careful, but that seems extremely unlikely doesn’t it?

For those of a mathematical bent the formula is:

I = aEb

Where I is the injury measure and E is the measure of walking or cycling. If b were 1, as cycling rose, casualties would rise at the same rate, but the estimate of b given in the paper is 0.4. a is a parameter not computed.

This all came from the journal Injury Prevention 2003;9 pages 205-209. For those in the CAM domain it is one of the on-line journals available via the Cambridge University Library website and search for Injury Prevention.

So how do we halve cyclist casualty rates? Simply triple the number of cyclists!

Jim Chisholm