Politicians should not get involved in the detail of clinical criteria and shouldn't be arguing with professors and consultants over whether there is one standard deviation or two standard deviations.