Age pattern of mortality in eight breeds of insured dogs in Sweden.

摘要: Abstract The objective of this study was to use several methods describe the age patterns for risk death in selected breeds dogs insured life a Swedish animal-insurance company 1996. Data on eight were analyzed at (including euthanasia). If left insurance reasons other than death, they regarded as censored. Dogs only up 10 years age. Four analytical approaches used. First, descriptive statistics distributions (e.g. breed-specific median ages breed- and age-specific mortality risks) computed. Second, estimates survival calculated using formula: survival=(1−risk )(1−risk age 1 )⋯(1−risk age 9 ). Third, Cox regression (proportional-hazards model) used estimate hazard functions. Finally, hierarchically coded Poisson determine cut-points death. hazards from incidence-density rates models transformed risk: risk=1−exp{−(hazard)} or 1−exp{−(incidence-density rate)}. studied Beagle, Bernese mountain dog, Boxer, Cavalier King Charles spaniel, Drever, German shepherd Mongrel Poodle, together representing over 50 000 each year. yearly varied between 1.7% (Poodle) 6.5% (Bernese dog). In all breeds, increased with but pattern by breed. probability 5 94% (Cavalier spaniel Poodle) 83% dog) 30% those derived combined-risk formula similar. cut-point provided simplified picture when changed significantly compared previous categories. As anticipated, differed widely there marked differences mortality. implications these findings should be considered multivariable analyses, where confounding effect is often controlled single variable common breeds.