Survey on the importance of different traits for medical detection dogs

Abstract

Medical detection dogs are used to remotely detect a range of diseases (biodetection) and assist patients with chronic conditions. There has been little research on the behavioral traits linked to good performance for either of these roles. We developed a survey to investigate which behavioral attributes are most important in medical detection dogs and to examine how these attributes vary between dogs used for biodetection and medical alert assistance tasks. The survey was completed by 62 professionals working with medical detection dogs from 16 different countries. Participants considered 40 traits for each rating: importance during selection, the ideal level required, and the level shown by a dog with which they had recently worked. Overall, 39 respondents rated assistance dogs, and 23 rated biodetection dogs. The most important traits were “Level of motivation when working” (on a scale of 1-5 mean = 4.66 ± 0.51), “Health” (4.60 ± 0.62), “Ability to learn from being rewarded” (4.57 ± 0.57), “Concentration” (4.57 ± 0.60), and “Acuity of sense of smell” (4.5 ± 0.76). Trait importance differed significantly between the three roles: “Level of Attachment to human partner” (U = 685, = 0.001), “Confidence in different environments” (U = 604 = 0.002), and “Travel ability” (U = 576, = 0.018) were deemed significantly more important for assistance dogs. “Acuity of sense of smell” (U = 280, = 0.004), “Tendency to be distracted when working” (U = 279.5 = 0.009), and “Ability to solve problems when working” (U = 310 = 0.009) were more important for biodetection dogs. Ideal levels also differed between the roles, for example, biodetection dogs were thought to require higher levels of “Tendency to search by smell alone” (U = 277.5 = 0.006) but lower levels of “Attachment to human partner” (U = 277.5 < 0.001) than assistance dogs. When comparing ratings for the current dogs to the perceived ideal levels, there were significant deviations in several traits; for example, dogs were generally rated as having higher than desired levels of “General excitability” (mean deviation = 1.02) and “Tendency to be distracted when working” (mean deviation = 1.00). Paying particular attention to important traits and those that differ greatly from ideal levels when selecting medical detection dogs may aid improvement in task success. Differences between medical detection dog tasks must be considered during selection if subsequent performance is to be optimized.