Abstract Chapter 17

Chapter 17 aims at assessing the performance of procurement policies for drugs in Italy in terms of i) the level of competition among sellers ii) the efficiency of purchasers.  To describe this phenomenon, we relied on IHS (Information Hospital System) database (data on drugs tenders 2005-2017) in order to map 1,951 tenders and 310,989 lots/sub-lots and test the level of competition through a linear regression, with the number of offers received per lot/sub-lot and discount over list prices as dependent variables. Efficiency was tested through a probit regression with the likelihood of failure as dependent variable (lot/sub-lot not awarded). Explanatory variables were the aggregation level (hospital/hospitals’ network/region), the procedure (e.g. open procedures, invited negotiations), medicines included in the lot (same vs more formulations and dosages), dimension of the lot (volumes procured), presence of generic/biosimilars, awarded product. We controlled for time, region, ATC-class and, where needed, active ingredient. Tenders on biological products and drugs from chemical synthesis were considered separately. The chapter underlines that the number of offers and the discount over list prices are higher if the awarded product is a generic and if generics/biosimilars join the same lots of the relevant originator. More controversial (or less important) is the role played by the procedure, aggregation level and volumes. The likelihood that the lot/sub-lot is awarded is higher when a generic is available, and lots include more dosage and formulation. The presence of biosimilars and the originator in same tender is associated to lower odds of failure too. Our empirical evidence was partially expected (e.g. on the role played by generics and biosimilars), do not support other common hypotheses (e.g. that buyers aggregation provides for a higher level of competition) and provides further insights (e.g. biosimilars and the originator in the same tender) to make procurement policies more efficient.