A Rescue That Cannot Say When It Would Euthanize Has Not Made a Decision

The Association of Shelter Veterinarians is unambiguous about the process, if not the answer: "Euthanasia decision-making must occur through a transparent process that lessens the decision-making burden on any one individual."¹ A process that has never been written down cannot lessen anyone's burden. It relocates the whole thing onto whoever is holding the leash at 9pm on a Friday.

Rescues do euthanize. US sheltering organizations euthanized an estimated 597,000 dogs and cats in 2025, and rescues without a government contract accounted for 9% of the 757,000 non-live outcomes.² The question was never whether. It is whether criteria exist before the case does.

Three Questions Wearing One Word

"Should we euthanize this animal" is three questions with three different kinds of evidence behind them, and collapsing them is why the conversations go badly.

Is this animal's day worth having? A welfare question, answerable by observation.

What is the probability and severity of harm to a person or another animal, and who will be standing next to it? A risk question, answerable badly.

What is continued care for this animal costing the animals you are not admitting? A capacity question, answerable arithmetically, almost never asked.

Most rescue people accept the first, argue about the second, and deny the third is a question at all. My position is that it is one, and that refusing to name it only moves the cost out of the meeting and onto one exhausted coordinator who quietly stops answering the intake line.

The Medical Threshold Describes a Day, Not a Diagnosis

No national body will write this criterion for you. The AVMA Guidelines for the Euthanasia of Animals govern method and agent, and the ASV binds shelters to them: agents "deemed unacceptable in the AVMA Guidelines for the Euthanasia of Animals are unacceptable to use in shelters."³¹ Neither tells you when. That gap is yours.

The available tools are quality-of-life scales. The HHHHHMM instrument scores seven domains — hurt, hunger, hydration, hygiene, happiness, mobility, more good days than bad — and a 2023 study in an Italian owner population found it useful, with the expected correlations to age, attachment and bereavement.⁴ That is a weaker claim than "validated" usually implies in rescue conversation, and the validation population matters: owners assessing their own long-known pet.

Here I am stating a judgement, not a finding. A foster animal in week two has no such observer. Nobody knows what it was like when it was well, so the scale's most useful axis — change from baseline — is unavailable. What is available is a harder criterion: can the people who actually have this animal deliver the treatment, at the frequency it requires, for as long as it requires? A protocol that works on a teaching-hospital ward and not in a foster's kitchen at 3am is not a treatment option you have. Asking a foster to attempt it anyway is how a slow medical death gets recorded as a rescue trying everything.

The Behavioral Threshold Is the Injury Already Inflicted

The best available data on behavioral euthanasia is Hitchcock and colleagues' 2024 survey of US owners, and it deserves to be read for what it actually establishes.⁵

Of 800 qualifying respondents, 729 reported on their dogs, and aggression toward people was ranked the primary reason by 57.4%. Among the 476 answering the human-aggression section, 77.5% said the dog had bitten and broken skin; among those 369 dogs the median was three skin-breaking bites, 36.9% had bitten four or more times, 12.7% ten or more. Taking each dog's worst bite, 37.6% fell in the authors' most-severe category and 13.8% required more than ten stitches, surgery, or hospital admission. The median dog was three years old, 60.4% had shown the behavior for at least a year, and 77.9% described a final incident rather than a single one. The authors' reading: owners "made these decisions over time and with extensive consideration for safety."

Now what it does not show. It is a convenience sample recruited largely from behavioral-euthanasia social media groups, 94.5% women and 91.3% white non-Hispanic. There is no comparison group of dogs with the same behavior that were not euthanized, so nothing here tells you the probability that any given biting dog gets euthanized. The paper does not report what interventions owners tried, and rehoming was not measured at all. The authors: "This study is not generalizable... individual criteria should not be interpreted to predict behavioral euthanasia decisions for individual pet dogs."⁵

So it is not a threshold. It is a description of how long people wait.

The criterion I would defend does not need the study: decide on the injury the dog has already inflicted and the predictability of the trigger, not on whether a theoretically correct home exists somewhere. In that same sample, only 42% of owners could predict aggressive reactions "most of the time" or "always," and only 30.9% said the dog gave clear warning signals that reliably. A dog that does not signal cannot be managed by a signal-based safety plan, and "experienced adopter, no children, no other animals, no visitors, muzzle in public" is not a placement plan. It describes a household that is rare, that you cannot verify a year after the adoption, and whose failure mode is someone else's face.

Experienced trainers will disagree, and they have a case: structurally managed dogs do live long, safe lives. My claim is about base rates and who carries the risk, not about whether a particular adopter could do it.

Transfer Is a Location, Not an Outcome

The accounting invites the worst version of this decision. Under the Asilomar formula, live outcomes are defined to include "transfer to another organization for adoption."⁶ A transferred animal is a saved animal, permanently, in your numbers, whatever happens next. In 2025, 906,000 dogs and cats were transferred out of US sheltering organizations.² So a rescue can hold a spotless live release rate while sending onward every animal it would not place itself — which is why the ASV insists that "live release rates or save rates must be evaluated in the context of animal welfare and cannot be used alone as a measure of success."¹

Two states have made the transfer point legally explicit. Virginia requires that on release for adoption, return to owner, or transfer to another agency, the agency "shall disclose, if known, that the dog or cat has bitten a person or other animal and the circumstances and date of such bite" — a Class 3 misdemeanor if it does not.⁷ California's statute reaches rescue groups by name and requires written disclosure of a known bite history plus a signed acknowledgment before release.⁸ Both also say a bite history does not automatically disqualify a dog. The obligation is candour, not refusal.

Stated plainly, and this is the sentence most likely to get argued with: transferring a dog with a serious bite history to a group you have not vetted, because you cannot bring yourself to euthanize it, is a worse act than euthanizing it. You have kept your number and moved the risk to a stranger. If the receiving group would not take the dog with the full bite record in front of them, you did not find it an outcome. You found it a location.

Capacity Belongs in the Decision

The ASV writes the intake constraint as a must: "An animal must only be admitted if the shelter can provide the care they require."¹ For a foster-based rescue, capacity is not kennels. It is foster-weeks, and the supply is fixed by how many people answered the last recruitment post.

Which produces the criterion nobody wants in the minutes: an animal in month nine of a foster home, not improving, with a plan that amounts to keep waiting, is consuming a slot with a queue behind it. That is a cost, denominated in the intakes you decline, and the anonymity of those animals does not make them hypothetical.

Two guards, because the claim is easy to abuse. This is not an argument for space euthanasia, and it does not apply to an animal that is progressing or simply taking a while to be adopted; it applies where the honest answer to "what is the plan" is that there isn't one. And I cannot cite a number for how many animals a rescue turns away per long-stay case, because nobody publishes it. That figure would settle the argument and it does not exist.

Nobody Should Carry This Alone

The ASV's process requirement survives every disagreement above.¹ Concretely — criteria agreed in writing when nothing is on fire; a named second person, a veterinarian for medical cases, who has to concur; a rule that whoever raised the case cannot be the only one who can approve it; a fixed pause before anything happens.

The mechanical safeguards fail under exactly this pressure, so name them too: personnel credentialed where state or local rules require it, a microchip scan immediately beforehand, legal eligibility verified, and confirmation that the animal is the individual the organization intends to euthanize.¹ Every one is an ASV "must," and every one gets skipped on bad days.

What Would Make This Framework Wrong

If foster capacity were elastic, the third question collapses. If a rescue holding a long-stay case simply recruits another foster rather than declining an intake, no animal is displaced and the opportunity-cost argument is an accounting fiction. I do not believe capacity behaves that way, but I cannot prove it — declined intakes are not a reported field anywhere.

The behavioral base rates are guesswork. Hitchcock's authors flag the underlying problem: there is no centralized reporting of causes of death in veterinary practice comparable to Shelter Animals Count.⁵ Anyone claiming to know what share of seriously biting dogs are safely placed is extrapolating from their own caseload.

Written criteria get applied. A rescue that writes down a threshold will euthanize some animals it would otherwise have muddled through and placed successfully. That is a real cost. I think it is smaller than the cost of deciding under pressure with no criteria at all, but someone who has watched a policy applied mechanically by a tired manager has grounds to disagree.

A rescue is not a shelter. The ASV guidelines were written with open-admission facilities in mind, and a small foster group can reasonably argue that its purpose is to be the place that keeps trying after the shelter has stopped. That is coherent. It stops being coherent when the same group also declines to say what would ever be enough.

None of this requires a rescue to euthanize more animals. It requires the organization to be able to finish the sentence "we would euthanize if ___" before someone hands it a case that finishes the sentence for them.