Vet. Res. 36 (2005) 199-211
DOI: 10.1051/vetres:2004064

A pharmaco-epidemiological analysis of factors associated with antimicrobial consumption level in turkey broiler flocks

Claire Chauvin^a, Isabelle Bouvarel^b, Pierre-Alexandre Beloeil^a, Jean-Pierre Orand^c, Didier Guillemot^d and Pascal Sanders<sup>e

a</sup>  AFSSA, French Agency for Food Safety, Pig and Poultry Research Laboratory, Epidemiology and Quality Assurance Unit, Zoopôle, BP 53, 22440 Ploufragan, France
^b  ITAVI, Technical Institute for Poultry Production, 28 rue du Rocher, 75008 Paris, France
^c  French Ministry of Agriculture, Food Fisheries and Rural Affairs, 251 rue de Vaugirard, 75732 Paris Cedex 15, France
^d  Institut Pasteur, Resource Center for Biostatistics, Epidemiology and Pharmacoepidemiology applied to Infectious Diseases and INSERM UMR 657, 25-28 rue du Dr Roux, 75015 Paris, France
^e  AFSSA, French Agency for Food Safety, Laboratory for the Research and Study of Veterinary Medicinal Products and Disinfectants, La Haute Marche Javené, 35302 Fougères, France

(Received 26 May 2004; accepted 12 October 2004)

Abstract - An on-farm pharmaco-epidemiological survey of 246 turkey broiler flocks from 131 farms was carried out to assess the homogeneity of antimicrobial use between flocks on the same farm and to explore the possible relationships between farm and farmer characteristics and the level of antimicrobial use. The antimicrobial use in each flock was quantified by an invoice study, expressed as the number of national animal daily doses (ADD)/turkey broiler and characterised as "high", "medium" or "low" according to the tertiles of the resulting distribution. Antimicrobial use was then correlated with variables collected from the farmer by means of an alternating logistic regression method which calculates the pairwise odds ratio (PWOR) for within-farm clustering. Two independent models were fitted: (1) "low" versus "medium" + "high" antimicrobial consumption and (2) "high" versus "medium" + "low" antimicrobial consumption. PWOR from the null models were significant (P < 0.005), but only remained significant in the first final model (P = 0.002). Four explanatory variables were retained for both models. Prophylactic antimicrobial administration and veterinarian antimicrobial prescription attaining the farm technical staff's expectation were associated with a higher antimicrobial consumption level. Administration of competitive exclusion flora and compliance with biosecurity rules of changing clothes and shoes before entering the facilities, were associated with a lower antimicrobial consumption level. In the first model, the number of full-time jobs devoted to the turkey production unit (1 versus more than 1) was also found to be associated with the antimicrobial consumption level. The study tends to confirm the feasibility of the adopted approach to quantify antimicrobial use and to determine the factors likely to influence antimicrobial consumption.

Key words: antimicrobial / pharmaco-epidemiology / alternating logistic regression (ALR) / turkey broiler / animal daily dose (ADD)

Corresponding author: Claire Chauvin c.chauvin@ploufragan.afssa.fr

© INRA, EDP Sciences 2005