Analysing the trend over time of antibiotic consumption in the community: a tutorial on the detection of common change-points
Robin Bruyndonckx, Samuel Coenen, Niels Adriaenssens, Ann Versporten, Dominique L. Monnet, Herman Goossens, Geert Molenberghs, Klaus Weist, Niel Hens, the ESAC-Net study group, Reinhild Strauß, Eline Vandael, Stefana Sabtcheva, Marina Payerl-Pal, Isavella Kyriakidou, Jiřı́ Vlček, Ute Wolff Sönksen, Elviira Linask, Emmi Sarvikivi, Philippe Cavalié, Tim Eckmanns, Flora Kontopidou, Mária Matúz, Gudrun Aspelund, Karen Burns, Filomena Fortinguerra, Andis Seilis, Rolanda Valintėlienė, Marcel Bruch, Peter Zarb, Stephanie Natsch, Hege Salvesen Blix, Anna Olczak-Pieńkowska, Ana Silva, Gabriel Adrian Popescu, Tomáš Tesař, Milan Čižman, Antonio López Navas, Vendela Bergfeldt, Berit Müller‐Pebody
Abstract
OBJECTIVES: This tutorial describes and illustrates statistical methods to detect time trends possibly including abrupt changes (referred to as change-points) in the consumption of antibiotics in the community. METHODS: For the period 1997-2017, data on consumption of antibacterials for systemic use (ATC group J01) in the community, aggregated at the level of the active substance, were collected using the WHO ATC/DDD methodology and expressed in DDD (ATC/DDD index 2019) per 1000 inhabitants per day. Trends over time and presence of common change-points were studied through a set of non-linear mixed models. RESULTS: After a thorough description of the set of models used to assess the time trend and presence of common change-points herein, the methodology was applied to the consumption of antibacterials for systemic use (ATC J01) in 25 EU/European Economic Area (EEA) countries. The best fit was obtained for a model including two change-points: one in the first quarter of 2004 and one in the last quarter of 2008. CONCLUSIONS: Allowing for the inclusion of common change-points improved model fit. Individual countries investigating changes in their antibiotic consumption pattern can use this tutorial to analyse their country data.