Scientists have spoken out for and against the 'evidence-based' policy for badger culling in England for the control of cattle tuberculosis (TB) (see M. Woolhouse and J. Wood, Nature 498, 434; 2013 and go.nature.com/nem9ua). Each faction emphasizes different statistics from the Randomised Badger Culling Trial (RBCT) on the impact of culling.

Appreciable uncertainty surrounds Woolhouse and Wood's statement that widespread badger culls “roughly halved” the incidence of cattle TB. This 54% reduction occurred inside culling areas only after five years of annual culls, and the benefits diminished after just 18 months (95% confidence interval: 38–66%; H. E. Jenkins et al. Int. J. Infect. Dis. 12, 457–465; 2008). In my view, this maximal risk reduction is relevant in setting stakeholder and policy-maker expectations for culling only if it can be sustained beyond 18 months (H. E. Jenkins et al. PLoS ONE 5, e9090; 2010).

'On–off culling', in which annual widespread culling resumes when cattle TB rates increase, might in principle sustain such a risk reduction, but the RBCT did not test this approach. Careful epidemiological and ecological modelling and cost analysis would be required to predict the impacts of on–off culling. It might trigger the reappearance of the transient increases in TB that were observed early in the RBCT outside culling areas, attributed to increased badger movements.