The Risk of Communicating Risk
“We teach our children the mathematics of certainty but not about how to handle uncertainty and risk in the world”. David Spiegelhalter, Professor of the Public Understanding of Risk, University of Cambridge, UK.
In April 2009, the UK’s Meteorology Office forecasted a 65% chance of above-average temperatures in the UK for the summer. The media grasped this statement and declared an odds-on ‘barbecue summer’. The public took action on the basis of these headlines; sales of domestic holidays went stellar. By late July, weeks of rain had caused a national public backlash against the Met Office. Few people understood or recalled that they had been advised that there was a 35% chance of a washout summer. Ironically, the temperatures had been above average despite the rain.
This episode is illustrative of the risk of communicating risk. Assume that an authority figure truly understands the probabilities surrounding a scenario. It is a substantial challenge for them to educate their audience to a degree that they understand that risk, can make a truly informed decision, and accept responsibility for that decision if the dice don’t roll their way.
As we enter an era of personalized, genomic medicine, the understanding and communication of probabilities is likely to be a stumbling block, not only for the public but for clinicians too. As genome science accelerates, scientists and clinicians will be faced with evolving risk scenarios. Particularly when looking at risk of common disease, most newly discovered variants are probabilistic rather than deterministic. Integrating these into public health policies or presenting probabilities to individual patients will require enormous skill and should stimulate debate about responsibilities for decision making. Given the complexity of genomic data there may be many scenarios in which consent can never be truly informed, risking paralysis in the system.
How to provide context, interpretation and counselling around these complex sets of probabilities is a new challenge in statistics, ethics and psychology. We could start, as Professor Spiegelhalter recommends, by teaching our children more about the mathematics of uncertainty.