Bovine TB begins and ends with cattle, with badgers playing at most a minor role. Photo: Will Fisher via Flickr (CC BY-SA).
Dodgy data, bad science, rotten politics: why the badger cull is wrong and stupid
28th April 2016
If we are ever to bring bovine TB under control in Britain's cattle herd, we must begin with the main disease reservoir, writes Tom Langton: the cattle themselves. The insistence on culling badgers has little to do with disease control, and everything to do with the short term economics of the beef and dairy industries, unwilling to sacrifice an iota of production in the interests of a real solution.
We really must accept that the only known way to tackle bTB is to test cattle effectively and remove infected animals. In other words, to stop side-lining disease control in favour of short term economic output.
Can it be ten years since the RBCT began winding up its field work and an Independent Specialist Group (ISG) began preparing for its 2007 final report?
I have to admit to not reading its many pages when badger culling first looked likely in 2012.
At that time, incredulity for me sprang from how it could be possible to reliably quantify 70% of any given badger population, and what the future impacts of badger removal on natural communities would be.
As it turned out, the economic crash of 2007-10 would play a more important role than the RBCT in formulating UK policy on bovine tuberculosis (bTB) and badger culling.
Last year, once the invitation to cull badgers over a wider area was announced, it was time to give the 300 or so pages of ISG report and scientific papers based upon it a proper read.
Science led policy? In our dreams ...
For those petitioning or signing a full page complaint in the Observer newspaper in October 2013, the government was just plain wrong.
It was simple. Any bTB reduction from culling badgers is all but erased as the result of something called the perturbation effect. Frightened and fleeing badgers, raising levels of bTB in cattle at the edge of the cull zones. It all made sense.
The celebrities, glossy wildlife mags and occasional expert or circumspect grandee, agreed with the ISG findings and repeated that there would be 'no meaningful contribution' from badger culling to resolving the bovine tuberculosis crisis.
Surely the government would concede this in the face of the overwhelming scientific argument? Sadly, no. Government policy, formulated around 2010, is what chief scientist Ian Boyd articulated as being to; "Achieve Officially TB free status for England by 2038 whilst maintaining a sustainable livestock industry."
However, unlike any other country in the world tackling bTB successfully, the UK, leaned on by the NFU, had refused to adequately test for bTB or to depopulate (slaughter) infected herds, causing the return of bTB to half of our countryside since the 1990s.
The 'sustainable industry' bit of the policy, means keeping infected herds and the beef and dairy exports to China alive, rather than tackling the disease head on, as in the past. The objective seems largely to try to chip away lightly at its spread, pretending the problem will resolve.
The overall aim? To try to recover the UK export markets after ruination by Foot and Mouth Disease (FMD) mismanagement in 2001. It is a long time surely, since established epidemiological advice on a terrible disease epidemic like bTB was ignored in favour of some kind of 'win now - pay later' economic plan.
What the figures show: the minor role of badgers in infecting cattle
However, I needed to focus on the ISG report and the papers. It was, in fact, fairly simple. The ISG said that badgers pass bTB to cattle at significant levels. It attempted to measure the effects of culling badgers on the frequency of new herds 'breaking down' with bTB.
Two types of badger killing were conducted by the RBCT in separate 'treatment' areas, each of about 100 sq.km in size.
- 'Reactive' culling removed badgers in response to, and at and around the locations of confirmed herd breakdowns, but was scrapped prematurely after a year.
- 'Proactive' culling removed badgers randomly across a separate 'treatment' area. Both approaches removed a high proportion of badgers, and control study areas were set up for comparison. Edge areas were also monitored.
The analysis in 2007 had also been very simple. Twenty areas in a line from Lands' End to Staffordshire, either culled or left alone. Ten paired comparisons. The ISG report offered tables that were hard to compare by eye, but with an evening or two of simple number crunching it was plain to see the trends.
Basically, in two of the control areas in one part of England, there had been large spikes of breakdown, elsewhere everything else looked pretty similar between treatment and control. Two other things stood out:
Firstly, and you had to go back to a previous ISG report to find it, the study areas had been chosen, not as random bits of countryside with bTB, but as the very centres of the most intense bTB outbreaks. One with nearly half the herds infected from the start. Not surprising then, that over five years, the rate of breakdown slowed from the centre, as study areas ran out of new herds to infect, and bTB accelerated at its edges.
It's a pattern that might mistakenly be interpreted as a perturbation effect. Some areas had over 80% of herds infected by the end of the trials, and in one, all herds had broken down.
Second, a 2013 paper derived from the RBCT had suggested that the new breakdowns caused by badgers were not around 1 in 20 as first thought, but 1 in 27. This means that of the total 472 breakdowns in the studies, around 18 were due to badgers. That is just about a third of a new breakdown per year, per study area. Or one new herd infection over around three years.
The Foot & Mouth Disease outbreak should have ended the trials
So subtle were the effects in fact, and so vast the confidence intervals, that I was beginning to have some sympathy with the amateur maths bloggers who have derided both Imperial Colleges' statistics and the reliability of the peer reviewers. A few breakdowns, from some unknown, uncertain or misconstrued variable, crashes the analysis.
Checking a range of papers since 2006, it seemed that the significance of findings has relied on 'data adjustments'. One of these papers had dropped three of the ten comparison pairs due to the impact that Foot & Mouth disease (FMD) had on the calculations, taking the overall analytical power below the critical threshold.
In the middle of the RBCT, FMD resulted in the slaughter on a massive scale, of sheep, pigs and cattle, based on what other published criticism later called "unvalidated predictive models".
What happened during the FMD epidemic has been described by one group of published statistical experts as resulting from the "abuse of mathematical models". As the UK authorities culled herds in the proximity of outbreaks, Ireland used FMD vaccination and avoided culling on the UK scale.
The additional expenditure from the UK culling has been estimated to have resulted in the over-expenditure of more than £3 billion. Problems were apparently attributed shortly after the events to "a lack of mutual understanding between veterinarians and modellers." The FMD outbreak messed up and should have ended the RBCT.
A full re-examination of the RBCT is needed
What was described to me as a 'pro-cull' website is also commenting on ISG analysis, having obtained raw RBCT data from DEFRA in March 2016. It is offering graphs showing why the concept of perturbation may be wrong. The data 'adjustments' create the significance, or rather - the perturbation effect is a function of the adjustments made. Something that is clear from the raw data.
This is actually quite helpful. There are many uncontrolled variables in the RBCT; bTB testing and herd numbers varied before and during the trials, pre-RBCT culling of badgers was high and uneven, and cattle testing regimes variable in type and frequency and so on. It is hard to square this with the ISG comment "very few such interactions [variables] were uncovered". They were everywhere.
What next? Well, with unsafe controls, do badgers contribute 1 in 27 new breakdowns, 1 in 2,700 or 1 in 27,000? Impossible to say. Those closer to the folk involved in the study say (in protection of reputations and funding), that if there is contra-evidence, then publish it. A re-examination of the RBCT assumptions and methods is needed.
So was £50 million really wasted? Well not completely. We did learn about badger behaviour along the way, and that killing badgers may cause a carnivore release effect on European designated sites and species - and all nature reserves and countryside actually. This, the government now accepts, requires proper investigation before considering more badger killings.
The post-RBCT period has been peppered with new offshoots and arguments. Some big sums have been allocated to vaccine development for badger and cattle in an unenthusiastic way.
But the 'elephant in the room' is what EC officials pointed out to Environment, Food and Rural Affairs Committee members in 2013: that we really must accept that the only known way to tackle bTB is to test cattle effectively and remove infected animals. In other words, to stop side-lining disease control in favour of short term economic output.
Conclusion; badgers have been more than unlucky. They have had the bad end of equivocal science, and the bad end of bad government. The question of whether or not to kill badgers at public expense, without good evidence, is surely facing an inevitable reality check.
Tom Langton has been a consulting ecologist to government, business and industry and a voluntary sector volunteer, more recently working on assisting small pressure groups in legal opposition to destruction of species and habitats in Europe.
Using this website means you agree to us using simple cookies.