Sticks and stones may break bones, but this could really hurt us

Animal welfare should be of paramount concern to all pest management professionals. It is not just a case of what we use, but how we use it which distinguishes a pest control program.

Sticks and stones may break bones, but this could really hurt us.

Animal welfare should be of paramount concern to all pest management professionals.  It is not just a case of what we use, but how we use it which distinguishes a pest control program. 

This is a critical review of information presented in the paper, An assessment of animal welfare impacts in wild Norway rat (Rattus norvegicus) management.  The paper will be broken into its salient points and then discussed as to how the data were collected, how they have been interpreted and how they have been presented in order to provide an objective review as to how the authors came to the conclusions that they did.

It ain’t what you don’t know that gets you into trouble.  It’s what you know for sure that just ain’t so.

Mark Twain.


The Framework of the study.

Welfare Assessment model.

In order to establish a metric for animal welfare this paper needed to establish a model which had definitive standards, criteria and parameters.  To achieve this task, the authors decided to use a model which was based on Mellor and Reids more general Five Domains Model (1994), which was developed from the UK Farm Animal Welfare Councils Five Freedoms (FAWC 2009).

This model in and of itself is a robust tool used throughout the world to measure the impact on livestock animals.  The model looks at five key components, Nutrition, Environment, Health, Behaviour and Mental State with the last component being an interpretation of the previous four components.

Nutrition is based on the access to food, water and a freedom from malnutrition.  Environment is based on the challenge of the environment on that animal.  Health is measured in the freedom from diseases, injury and functional impairment.  Behaviour is the ability to exhibit natural behaviours or to interaction normally with its environment.  Finally the animals mental state is then derived from observing the animal exhibiting pain, sickness, weakness etc with the inference being that some factor presented in the previous four components (domains) is the culprit.

Whilst this model is a fantastic tool for understanding animal welfare in livestock animals, it is largely misplaced in measuring the impacts on animal welfare in animals over whom we are not in total control of their care.

Domains 1 (nutrition), 2 (Environment) and 4 (Behaviour) may well be impacted by a pest management program, but the degree of impact is largely unmeasurable.  Changes in food abundance may well be affected by a Pest Management Professional (PMP) undertaking a program, but equally food abundance might be affected by the exhaustion of a spilt food, the removal of that foodstuff by adverse weather or other animals.  It is unreasonable therefore to assume that the actions of a Pest Management Professional would constitute a serious impact on the welfare of that animal beyond which it is naturally challenged within the scope of its existence.

Equally the measures of environmental challenge and behaviour freedom will be impacted by the actions of a PMP, but it is reductive thinking to assume that all challenges to these freedoms are to be placed at the hands of PMPs exclusively.

Therefore, it could be strongly argued that whilst this model is reliable for the provision of livestock, it is redundant in terms of measuring the impacts on wild animals in anything other than the interaction between Domain 5 (Mental State) and Domain 3 (Health).

These data are decidedly more appropriate to the welfare of animals managed in pest control.

Impact Criteria.

Part A measures the impact that the non-lethal aspects of Pest Management has on rodents by using a matrix of severity over impact.  The paper further divides the impacts of these into effects occurring prior to treatment, e.g. the clearing of spilt feed prior to applying rodenticides; and the effects occurring as a direct result of or as a result of the treatment, e.g. the restraint of an animal on a glue board.

The tools for measuring non-lethal impacts prior to treatments are largely superfluous in terms of wild animals, however there is a serious case to be made for the non-lethal impacts these tools have after deployment.  Especially in relation to glue boards and cage trapping.

Part B focuses on the welfare impacts pertaining to lethal control.  This too is based on a matrix of severity over time.  The paper defines the measurement of these criteria as starting from the first instance of impact through to the point of irreversible unconsciousness.

However there is a slight discrepancy with the maths with such matrices.  Hunger and thirst for example will increase with intensity depending on the duration of depravation.  So how does one measure the impacts of hunger on an animal restrained only for 5 minutes over traps which have been left overnight?  The paper also argues that if Domain 5, which covers thirst and hunger, cannot be adequately quantified then the overall impact of the cumulative values of Domains 1-4 should be used.  This means that it is an animals availability to food which is measured, rather than the physical impacts that hunger might impose on an animals mental state.

This discrepancy becomes all the greater when considering the overlapping physical symptoms presented with chemical control methods which range from nuisance through to debilitating.


Results & Discussions.

Pre-deployment values (Part A).

All pre-deployment values for all tools ranked similar impacts.  All tools (with the exception of glue boards) were ranked to be of a mild impact which lasted for days giving each tool a total score of 5 (based on a rank ranging from 0-8).

To further elucidate this point the interpretation of the model claims that the placement of a snap trap in an area where there are rodents present earns this rank as a result of a mild behavioural impact in D4 because rats are likely to experience opposing drives to both avoid and explore novel objects.  This would be expected to lead to mild anxiety in D5.

Whilst this may arguably be the case, one could posit that ANY form of environmental manipulation could elicit the same response.

Impact values (Part B).

For the sake of brevity, this review will focus only on snap traps, glue boards and anticoagulant poisoning (but it is HIGHLY recommended that you read the whole paper in full).

Snap Traps

The data on snap traps proved to be the most variable with the least confidence in the data presented.  This is largely due to the high degree of perceived erraticism in snap trap construction and design in all traps available within the UK (rather than those sold exclusively to Pest Management Operators).

Snap Traps produced no impacts to extreme impacts, causing unconsciousness either immediately or within seconds to minutes, scoring welfare impact scores of AF Part B, STs scored AF, because a trap taking up to 5 min to produce irreversible unconsciousness could cause suffering ranging from none through to extreme impact.

Whilst this range of scores is worrying, and likely accurate considering the quality of traps available to the general public.  It is likely that if traps were chosen exclusively from the suppliers of professional products to PMPs these scores would be significantly lower with and significantly higher confidence


Glue Boards

Glue boards were viewed with the same lens as cage trapping in that they had an additional consideration under Part A, live capture.  In this the Glue trap was considered to have an extreme impact that lasted for hours, scoring 7.  However, due to the animal already being restrained the concussive killing of the rat on the Glue Trap was more humane than the process of dispatching a rodent in a cage trap, producing mild to moderate impacts, for seconds to minutes, scoring BD.

Compared to the previous control method, cage trapping, which scored 5-6 on Part A: Live Capture, and B-D on Part B: Killing, almost producing an overlap between the two methods.

This brings an uncomfortable question to the fore, this paper postulates that the difference in welfare between cage traps (often sold to the public as humane traps) and Glue Traps (currently under legal review in the House of Lords) is minor. 


Anticoagulant Baiting.

Anticoagulants caused a severe to extreme impact for days, scoring GH however the confidence in this data was only moderate to high.

This lack of confidence with the authors is as a result of the data pertaining to the severity and chronology of anticoagulant poisoning is either sparse, based on assumptions or at times contradictory.  Many of the papers used to determine the impacts on anticoagulant poisoning are cyclical.  Ultimately cited papers often come to the same conclusion, there is almost certainly a negative impact on animal welfare because of anticoagulant toxicity, however the severity of this impact is largely inferred not from impacts to Domains 3 or 5, but from domains 1,2 and 4.

Some of the contributing factors to this determination come from statements such as the following, poisoned rodents sometimes remain above ground in exposed positions (Fisher et al 2010), potentially resulting in environmental impacts (D2).

Some of the more rational arguments for Part B impacts are the pathology of haemorrhaging and the speed of blood loss.  What the paper does not recognise however is that the major symptoms which are likely to be most painful will only persist for very short periods as these symptoms are almost always lethal.  Whilst the more mild symptoms which provide the lowest level of discomfort may well persist for several days in the lead to the animals death.  The paper fails to make this distinction, happily lumping acute symptoms with chronic.  This consolidation of symptoms gives the reader the erroneous notion that symptoms are of the highest severity from day one and persist until the animals demise many days later, rather than reflecting the reality that there is a progressive scale of symptom, with the most sever of these potentially only occurring immediately prior to death.


Up unto this point this paper, although flawed in places, had shown drawn inferences from objective reasoning.  However, as the discourse falls to the discussion the paper rapidly loses all objectivity, veering to the subjective and often outright sensational. 

If welfare impacts are a function of the degree to which an individual animal suffers, multiplied by the number of individuals affected, then rodent control using these methods must be among the most significant of deliberate human activities affecting animal welfare.

The sensationalism of such a statement has no place in this article.  In riposte to such inflammatory claims one might also consider the wanton destruction of the Amazonian rainforest or the over fishing of coastal seas.  Such activities, and more, have significant impacts on animal welfare.  Yet unlike public health pest management ALL the above are motivated in terms of commercial gain, rather than in an effort to prevent the damage to buildings and stored food, prevent the transmission of disease and to protect the safety and mental wellbeing of society.

Closing Proposal.

The core concept of this paper is sound, there does need to be a review into the use, misuse and abuse of pest control products.  However, the model used is not appropriate to the arena these tools are being used in.  Additionally, the welfare impacts of these tools themselves varies wildly depending on their implementation and the impetus behind their use.

Therefore, it is proposed that this paper is revised with a new model, one that addresses the impacts of these tools under a lens specific to Pest Management rather than Livestock Management, and in addition looks at the distinction between when these tools are used by skilled and trained professionals and the impacts when used by unskilled persons or people who may (un)intentionally misuse or abuse these tools.

To Read this article in print; click here.