May 30, 2013 15:21 SM12-HF301 Statistical Modelling 2013; 13(3): 253–274 Adaptive change-point mixed models applied to data on outpatient tetracycline use in Europe Girma Minalu 1 , Marc Aerts 1 , Samuel Coenen 2,3 , Ann Versporten 3 , Arno Muller 3 , Niels Adriaenssens 2,3 , Philippe Beutels 4 , Geert Molenberghs 1,5 , Herman Goossens 3 and Niel Hens 1,4 1 Interuniversity Institute for Biostatistics and statistical Bioinformatics (I-BIOSTAT), Hasselt University, Belgium 2 Vaccine & Infectious Disease Institute (VAXINFECTIO), Centre for General Practice, University of Antwerp, Belgium 3 Laboratory of Medical Microbiology, Vaccine & Infectious Diseases Institute (VAXINFECTIO), University of Antwerp, Belgium 4 Centre for Health Economics Research and Modeling Infectious Diseases (CHERMID), Centre for the Evaluation of Vaccination, Vaccine and Infectious Disease Institute (VAXINFECTIO), University of Antwerp, Belgium 5 Interuniversity Institute for Biostatistics and statistical Bioinformatics (I-BIOSTAT), Catholic University of Leuven, Belgium Abstract: In this paper, we propose a change-point mixed model to assess the change in the trend of outpatient antibiotic use in a Bayesian framework, where the change-points are unknown parameters of the model. Model selection using DIC indicates that the data supports the model with a country- specific change-point. The location of the change-points may be related to points in time where public health strategies aiming at increasing the awareness of the public to a more rational use of antibiotics or targeting to reduce overconsumption of antibiotics were initiated. Key words: amplitude; antibiotic use; change-point model; non-linear model; phase shift; seasonal variation Received January 2012; revised March 2013; accepted March 2013 1 Introduction Antibiotics are drugs that inhibit or abolish the growth of bacteria. Antibiotic resis- tance is a major European and global public health problem and international efforts Address for correspondence: Marc Aerts, Interuniversity Institute for Biostatistics and statistical Bioin- formatics (I-BIOSTAT), Hasselt University, Campus Diepenbeek, Agoralaan 1, B-3590, Diepenbeek, Belgium. E-mail: marc.aerts@uhasselt.be c  2013 SAGE Publications 10.1177/1471082X13485404 NOT FOR COMMERCIAL USE