Frontiers in Veterinary Science (Jan 2020)
Developing Diagnostic Frameworks in Veterinary Behavioral Medicine: Disambiguating Separation Related Problems in Dogs
Abstract
Diagnoses are widely used in both human and veterinary medicine to describe the nature of a condition; by contrast, syndromes are collections of signs that consistently occur together to form a characteristic presentation. Treatment of syndromes, due to either their lack of a clear biological cause or multiple causes, necessarily remains non-specific. However, the discovery of interventions may help refine the definition of a syndrome into a diagnosis. Within the field of veterinary behavioral medicine, separation related problems (SRPs) provide a good example of a syndrome. We describe here a comprehensive process to develop a diagnostic framework (including quality control assessments), for disambiguating the signs of SRPs as an example of a heterogeneous behavioral syndrome in non-human animals requiring greater diagnostic and treatment precision. To do this we developed an online questionnaire (243 items) that covered the full spectrum of theoretical bases to the syndrome and undertook a large-scale survey of the presenting signs of dogs with one or more of the signs of SRPs (n = 2,757). Principal components analysis (n1 = 345), replicated in a second sample (n2 = 417; total n = 762), was used to define the structure of variation in behavioral presentation, while hierarchical agglomerative cluster analysis cross checked with the partitioned around medoids method was used to determine sub-populations. A total of 54 signs were of value in defining a latent structure consisting of seven principal components (termed “exit frustration,” “social panic,” “elimination,” “redirected frustration,” “reactive communication,” “immediate frustration,” “noise sensitivity”), which divided the population in four clusters (termed “exit frustration,” “redirected reactive,” “reactive inhibited” and “boredom” related SRPs) with 11 sub-clusters (3, 3, 3, and 2, respectively). We used a bottom-up data-driven approach with numerous quality checks for the definition of robust clusters to provide a robust methodology for nosological studies in veterinary behavioral medicine, that can extend our understanding of the nature of problems beyond SRPs. This provides a solid foundation for future work examining aetiological, and differential treatment outcomes, that will allow both more effective treatment and prevention programmes, based on a fully appreciation of the nature of the problem of concern.
Keywords
