bTB Evidence Project

A Restatement of the Natural Science Evidence Base Relevant to the Control of Bovine Tuberculosis in Great Britain

Bovine TB is a very important disease of cattle that has major costs to the government and to the farming community. It is also found in wildlife, especially badgers, and whether culling badgers should be part of bTB control is very controversial.

This project aims to provide a succinct summary of the natural science evidence base underlying bovine tuberculosis policy in the UK.  It has been led by Charles Godfray and Angela McLean from the Oxford Martin School and also involves Christl Donnelly (Imperial College), Rowland Kao (Glasgow University), David Macdonald and Gillian Petrokofsky (Oxford University), James Wood (Cambridge University), Rosie Woodroffe (Institute of Zoology), Douglas Young (MRC National Institute of Medical Research) and Robbie McDonald (University of Exeter).

A brief announcement of the project’s conclusion is here

The evidence summary can be found as the Appendix to a paper published (August 7th 2013) in the Proceedings of the Royal Society, Biological Sciences with an annotated bibliography as supplementary material online. The open access article can be accessed here and a version as a single pdf can be downloaded here.

We welcome comments at

Though it is not our intention to update the summary as new evidence becomes available, we will add any clarifications, corrections and updates to papers we mention as in review, on this webpage.

  • 7 August 2013. The published version differs from the preprint that was circulated in some minor corrections to the statistics on bTB testing in paragraphs 23 and 24.
  • 8 August 2013. The top figure in Box 1 shows seasonal variation in bTB cases – we should have added that much of this is driven by seasonal changes in the number of tests carried out.
  • 8 August 2013. The scope of the review does not include social sciences and economics though in the Annotated Bibliography we provide an entry into this literature.  We did not attempt to be comprehensive but it has been suggested that the following references may also be useful: (I) Bennett RM and Cooke RJ. (2006).  Costs to farmers of a tuberculosis breakdown. Veterinary Record 158: 429-432. (ii) Bennett RM  (2009).  Farm costs associated with premovement testing for bovine tuberculosis.   Veterinary Record 164: 77-79. (iii) Temple, M, & Tuer, SM. (2000). The Cost At Farm Level of Consequential Losses from Tuberculosis Control Measures. MAFF, Bovine TB and Zoonoses Division
  • 9 October 2013. The July 2013 edition of Epidemiology and Infection  (volume 141, special issue 07) is devoted to “Dealing with TB in wildlife”; the table of contents can be found here.
  • 14 October 2013. In paragraph 14 the restatement says “In the RBCT proactive cull areas (see paragraph 30), it has been estimated that 50% of confirmed herd breakdowns in the year before culling began were because of badgers, though this figure has very broad confidence limits”, and in the Annotated Bibliography that these estimates are currently in review.  The 95% confidence limits (9.1-100%) are now published:
    Donnelly CA, Nouvellet P. The contribution of badgers to confirmed tuberculosis in cattle in high-incidence areas in England. PLOS Currents Outbreaks. 2013 Oct 10. Edition 1. doi: 10.1371/currents.outbreaks.097a904d3f3619db2fe78d24bc776098.

    The paper uses a mathematical model of the dynamics of herd infection and bTB detection to estimate the difference between the observed number of confirmed herd breakdowns in the proactive RBCT sites prior to culling and the number that would be expected in the absence of infection in badgers.  This difference includes the direct effects of cattle infected by badgers, but also the secondary cases that result from subsequent cattle to cattle transmission.  The study estimates that in the absence of infection in badgers there would be 52% fewer confirmed herd breakdowns though the 95% confidence limits range from 9.1-100%.  Of all confirmed herd breakdowns, 5.7% (95% confidence limits: 0.9-25%) are the direct results of badger-to-cattle transmission (the additional breakdowns occurring because of the subsequent cattle-to-cattle transmission).  Other information about the role of badgers can be obtained by looking at the maximum reduction in confirmed herd breakdowns that occurred in the RBCT sites once culling began.  This was estimated to be 54% (95% confidence limits 38-66%).  The authors propose that 38%, the lower confidence interval, is a robust estimate of the minimum effect of infection in badgers (both directly and through subsequent secondary cases).  It is robust because it takes the lowest value from the confidence interval (the data suggest that there is just a one in forty chance of the true value being lower) and because in none of the RBCT sites were all badgers culled (and hence the risk of badger-to-cattle transmission completely eliminated).  The results provide important information about the consequences of reducing badger-to-cattle and cattle-to-cattle transmission though, as the authors emphasise, “do not point to a single way forward”.