Factors Associated with Shelter Dog Euthanasia versus Live Release by Adoption or Transfer in the United States
Level 3 - non-randomized controlled study
Retrospective cohort study analyzed via time-to-event modeling (graded by design analogy for non-clinical/veterinary research)
OpenAlex W3148245838 · doi:10.3390/ani11040927
What was done
Researchers analyzed electronic shelter intake and outcome records from 2017 across 17 municipally funded shelters in five US states (Mississippi, Pennsylvania, Michigan, Colorado, and Oklahoma). From 25,047 unique dogs, 19,514 adoptable dogs were identified, and a simple random sample of 6,200 dogs was modeled. Variables describing coat length, estimated adult size, and skull type were imputed from breed descriptions. A Cox proportional hazards model with a random effect for shelter was developed to assess predictors of euthanasia versus live release (adoption or transfer).
What was found
Specific effect sizes, hazard ratios, and direction of effects were not reported in the abstract. The model found that a size by geographic region interaction (p = 0.0204), age group (p < 0.0001), and skull type (p < 0.0001) were significantly associated with the hazard of euthanasia.
Why it matters
Identifying physical and demographic traits associated with euthanasia risk can help animal shelters target adoption campaigns, foster care, and retention strategies toward vulnerable canine populations.
Limits
The abstract reports p-values but no effect sizes, hazard ratios, or directions of effect. Phenotypic traits were imputed from breed designations rather than directly measured, which can introduce misclassification. Clinical health status, behavioral evaluations, and shelter capacity constraints were not detailed in the abstract.
Cited by
- context Millions of dogs die each year in the United States because of an overpopulation problem.