ORIGINAL ARTICLE Medical equipment classification: method and decision-making support based on paraconsistent annotated logic Nata ´lia F. Oshiyama • Rosana A. Bassani • Itala M. L. D’Ottaviano • Jose ´ W. M. Bassani Received: 29 July 2011 / Accepted: 26 February 2012 / Published online: 11 March 2012 Ó International Federation for Medical and Biological Engineering 2012 Abstract As technology evolves, the role of medical equipment in the healthcare system, as well as technology management, becomes more important. Although the exis- tence of large databases containing management information is currently common, extracting useful information from them is still difficult. A useful tool for identification of fre- quently failing equipment, which increases maintenance cost and downtime, would be the classification according to the corrective maintenance data. Nevertheless, establishment of classes may create inconsistencies, since an item may be close to two classes by the same extent. Paraconsistent logic might help solve this problem, as it allows the existence of inconsistent (contradictory) information without trivializa- tion. In this paper, a methodology for medical equipment classification based on the ABC analysis of corrective maintenance data is presented, and complemented with a paraconsistent annotated logic analysis, which may enable the decision maker to take into consideration alerts created by the identification of inconsistencies and indeterminacies in the classification. Keywords Clinical engineering  Medical technology—classification  Database  Decision making 1 Introduction The constant technology advances improve the quality of the health service provided and also favor cost reduction and access to the healthcare system [6]. In the USA, Eur- ope and Japan, 5–6.2 % of the total health expenditure is destined to medical devices [13]. It is estimated that in the USA this percentage was equivalent to US$365 per capita in 2008 [17]. This considerable cost and the key role of medical devices in the health care systems emphasize the need of efficient management of health technology. Nowadays, computerized maintenance management systems provide the means for storing large amounts of maintenance data [3] that allows working on many of the medical equipment lifecycle areas [11]. However, in most of the times, biomedical engineers cannot transform the available maintenance data in concrete tools for inventory management, so that decisions are frequently based on intuition, rather than on objective information [14]. That is why data analysis tools, e.g. classification (used to elabo- rate models that may reveal interesting patterns present in the database), are required [12]. One way of grouping elements into classes is the ABC (always better control) analysis, which is based on the assumption that approximately 70 % of the total cost are due to only 10 % (‘‘vital few’’) of the total number of items, whereas 70 % of the items (‘‘trivial many’’) would account for only approximately 10 % of the cost [10]. This pattern also applies to medical equipment, since greater incidence of failure, cost and time of repair are attributable N. F. Oshiyama  J. W. M. Bassani (&) Department of Biomedical Engineering, School of Electrical and Computer Engineering, University of Campinas, Campinas, SP, Brazil e-mail: bassani@ceb.unicamp.br R. A. Bassani  J. W. M. Bassani Center for Biomedical Engineering, University of Campinas, Campinas, SP, Brazil I. M. L. D’Ottaviano Center for Logic, Epistemology and the History of Science, University of Campinas, Campinas, SP, Brazil 123 Med Biol Eng Comput (2012) 50:395–402 DOI 10.1007/s11517-012-0888-6