This is in relation to a recent news item, where scientists have suggested culling sick Koalas to help populations. Read more here.

The statistician George Box coined a phrase in 1978: “All models are wrong but some are useful.”  His point was that the real world is too complex to recreate exactly in a model, but simple models that approximate the real world may provide helpful insights. 

The flipside, as a recent paper (Wilson et al. 2015.  “The paradox of euthanizing Koalas to save populations from elimination”) demonstrates, is that not only are some models not useful, they can be very dangerous.

In brief, the paper proposes selective culling and treatment of Koalas could eliminate disease from the population, perhaps within 4 years. 

If researchers are going to propose a course of action as controversial as culling a Federally-listed threatened species such as the Koala, it is beholden upon them to make sure their work is correct.

In this case we argue there are some fundamental flaws that undermine the described model.  For example:

1)      The model assumes that Koalas either die due to disease, euthanasia, or natural causes.

Mortality rates due to natural causes are taken from a study of Koalas living on French Island in Victoria that was published back in 1990.  Is the use of these mortality rates suitable for modelling Koala populations in the heavily urbanised southeast Queensland in 2015?  We certainly don’t think so.

In effect, the model ignores additional mortality due to habitat loss, and associated impacts such as vehicle strike and dog attack

Some authors on this study have alrady seen the devastating impact of habitat loss, dogs and cars on Koalas in their own study populations. From May 2013 to August 2015, 453 Koalas were been examined as part of the Moreton Bay Koala Tagging and Monitoring Program.  More than half (251) of these animals are now dead.  Only 64 of these animals died of disease and illness.  As many as 157 have been killed by dogs. 

The failure to include this mortality into modelling is inexcusable.  Koalas in southeast Queensland are dying – from habitat loss, dogs and cars – in greater numbers than the model allows for.  So the addition of more Koala deaths from culling means that, rather than eliminating chlamydia, there is a real chance that it is the Koala that is eliminated.

2)     The model makes some unjustified assumptions about Chlamydia in Koalas; that Koalas can clear an infection (i.e. cure themselves, without treatment) within 10-18 months, and, having cleared the infection, they have a 20-60 per cent chance of avoiding reinfection for 6 to 12 months.  Where did these figures come from?  Certainly there is no published research to support these assumptions and numbers.

The paper claims they have been “assumed based on expert opinion or inferred through calibration to provide realistic model projections which match observed epidemiological trends.”  Truly shocking comments, a throw-away line that doesn’t really explain anything. 

AKF is not sure whose expert opinion was used, or which observed epidemiological trends were used to get these figures. 

Certainly, authors on the study have previously claimed that “Chlamydial infections in most hosts are not uncommon, and usually chronic, often staying with the host for many years, if not for life.” 

And while there is no published information on chlamydia clearance in Koalas, researchers looking at human chlamydia infections describe case studies where individuals with untreated chlamydia can still be infected (and symptom free) for up to 3.5 years.

If these assumptions are removed from the model, the headline story – that culling and treatment could eliminate chlamydia within four years – no longer holds true.

Why are these issues important? 

Koalas are a National Icon that has been recognised as a threatened species.  Any proposal to cull a threatened species needs the most robust support imaginable.  Frankly the current study is not up to the task.

This study is just a thought bubble.

It is scary to see how quickly this idea has gained traction.  Very few would have read the paper beyond the headlines.  The entire premise is based upon some quite flimsy assumptions and poor knowledge of both chlamydia and Koala ecology (particularly in southeast Queensland).  It simply does not bear up to scrutiny. 

AKF would never allow a cull on the strength of a weird thought bubble model.   In AKF’s opinion, any management strategy based on this model will be fundamentally flawed.

This is a dangerous idea that must not take hold.