This study used probabilistic elicitation and a Bayesian framework to quantitatively explore how logically practitioners' update their clinical beliefs after exposure to new data. The clinical context was the efficacy of antibiotics versus teat sealants for preventing mammary infections during the dry period. While most practitioners updated their clinical expectations logically, the majority failed to draw sufficient strength from the new data so that their clinical confidence afterwards was lower than merited. This study provides quantitative insight into how practitioners' update their beliefs. We discuss some of the psychological issues that may be faced by practitioners when interpreting new data. The results have important implications for evidence-based practice and clinical research in terms of the impact that new data may bring to the clinical community.
Keywords: Antimicrobial resistance; Bayesian analysis; Belief updating; Clinical priors; Evidence-based veterinary medicine; Probabilistic elicitation.
Copyright © 2016 Elsevier B.V. All rights reserved.