Quantitative assessment of TB freedom in possums

Sometimes, inspired ideas have to wait before they can be implemented. For example, the value of having a quantitative tool for assessing the probability that bovine tuberculosis (TB) had been eliminated from possums in a local area was recognised 8 years before that tool was first used in 2012. This article summarises the key roles current and former Landcare Research staff have played in developing that tool, in collaboration with key TBfree New Zealand staff (Mark Bosson and Paul Livingstone) and overseas researchers (Tony Martin, Western Australia).

Declarations of ‘TB freedom’ are largely based on a calculated probability that there is a less than 5% chance the possums are still infected (i.e. PFree >0.95). These calculations are made within a ‘Proof of Freedom’ data model developed by Dean Anderson and others in about 2010. That model is initialised with predictions from a Possum-TB model developed by Dave Ramsey and Murray Efford and refined by Mandy Barron. In essence, the Possum-TB model provides a quantitative statement of belief that previous possum control has been sufficient to eradicate TB from an area (Fig. 1).

Once the possum-model-predicted PFree reaches an agreed trigger level (currently ≥80%), surveillance is initiated to provide empirical evidence that TB is indeed absent from a possum population. The greater the survey effort undertaken without finding TB, the greater the confidence an area is TB-free. Surveillance data can be derived not only directly from possums, but also from surveys of ‘sentinel’ species such as deer, pigs, and ferrets.

The underpinning data and theory span Bayesian probability and classical sampling theory, TB epidemiology, population dynamics, home range utilisation and movement patterns of possums and sentinel species, and, for possums, their trappability or detectability. Most of the underpinning research was conducted without Proof of Freedom in mind, highlighting how current research can have unforeseen additional uses.

Qualitative ‘stopping rules’, developed by TBfree New Zealand to decide when they could stop local possum control, underpinned the development of a quantitative Proof of Freedom model beginning in 2004 with a workshop between TB managers, Landcare Research, AgResearch, Environment Waikato and others. Some key points discussed were:

The desirability of using actual surveillance to quantify confidence that TB was absent.
The difficulty of surveying a large sample of possums given that in well-managed areas possum densities were very low.
Whether simply knowing that enough possum control had been done would suffice; e.g. would achieving >95% kills on three occasions over 10 years guarantee local eradication of TB from possums?

By 2006, Graham Nugent, Dave Ramsey and Peter Caley had produced a discussion document that showed how all three issues could be combined into a workable tool. This included the development of a new concept of spatially explicit surveillance based on ‘TB-detection kernels’ for sentinels (Fig. 2). The kernels from different species could be added together to predict where in the landscape TB was least likely to be.

Over the next few years, a major review was undertaken by the funders of the TB programme to decide whether the objectives of TB management should be to control the disease at some low level or push for eradication. Agreement was eventually reached that eradication would indeed be a key objective in the National Pest Management Plan (NPMP) for TB, and the development of a Proof of Freedom tool by researchers began in earnest in 2009 – leading to its first formal application in early 2012. As at June 2014, TBfree New Zealand had declared 830,000 ha, or 33% of the 2.5 million ha proposed for TB eradication by 2026 in their NPMP, as TB-free. This means that 8% of the 10.5 million ha that was defined as containing tuberculous wildlife at June 2011 is now free of infected possums.

In summary, development and adoption of the Proof of Freedom framework as a major new paradigm in TB management took almost a decade, partly because it inevitably takes time to ‘sell’ new concepts to users, but mainly because the need for such an approach only became a priority when local eradication of TB became a formal management goal. There is now a strong focus on refinement and extension of the framework and associated software. An example is the development of a new theory by Dean Anderson that would allow inclusion of data from livestock TB testing and slaughterhouse inspections to help increase confidence that TB is absent from possums in farmed areas. Currently that information is used only qualitatively and does not affect the calculated probabilities of TB freedom.

By providing an objective measure for comparing progress between areas, the Proof of Freedom tool has enabled a shift away from focusing on possum control toward the true objective of TB management – the elimination of TB from possums. In doing so, it has led Andrew Gormley and Graham Nugent to explore new strategic concepts about when to start TB surveillance. These concepts have the potential to greatly shorten the cost and duration of possum control required for TB freedom.

This work was funded by TBfree New Zealand.