Assessment of management-related risk factors for paratuberculosis in Danish dairy herds using Bayesian mixture models

Transmission of Mycobacterium avium subsp. paratuberculosis (Map) to susceptible animals is primarily considered to occur via faeces and milk originating from infectious cows. However, studies of factors resulting in increased transmission of Map are difficult to perform due to a long and unpredictable incubation period and inaccurate diagnostic tests. A multi-level Bayesian mixture model has been shown to predict the infection status of an individual cow more precisely than traditional cut-off based methods used for interpretation of diagnostic test-information, thereby increasing the precision of the diagnostic information.The objective of our study was to assess management-related risk factors for within-herd transmission of Map. Management-related risk factors were recorded in 97 Danish dairy herds. Twenty-six months following that recording, the antibody status of all lactating cows (n = 7410) in the same herds was measured by the use of an ELISA. A multi-level Bayesian mixture model was used to assess the association between the probability of infection of individual cows and 41 herd-level management-related risk factors using univariable analyses. In this model, the continuous OD value was used to estimate the probability of infection, corrected for known animal covariates and laboratory factors. The statistical significance of the potential risk factors was assessed by calculating odds ratios and their 95% credibility posterior intervals.Four significant risk factors were identified: housing of cows in bed stalls compared to housing in tie stalls; low level of hygiene in the feeding area of calving areas; low amounts of straw in the bedding of the calving area; high animal density among young stock >12 months of age. Surprisingly, the hygiene level in the calving area was not found to affect the odds of infection.