Interpretable associations over DataCubes: application to hospital managerial decision making.

The world concern about the costs of the health care systems has raised the importance of counting on precise and interpretable tools, that help the health care institution's managers to make decisions to optimize the use of health resources. In this paper we propose a new Classification based on Association Rules (CAR) algorithm that improves the interpretability of the results, making it specially useful for decision making. Changing the usual way to obtain the rules we follow four goals: first to improve the interpretability of the result by obtaining rules meaningful and interpretable by themselves, secondly to reduce the complexity of the result obtaining a lower number of rules; thirdly, to obtain simpler rules, with less size in number of antecedents; and finally to avoid the usual over-fitting problem of the classification methods by obtaining a generic final result set, where specific rules for specific cases are avoided unless they are necessary. To prove the utility of our proposal we have used it in an example of decision support regarding the planning of the surgery rooms.

Improving hospital decision making with interpretable associations over datacubes.

In this paper we propose a new Classification based on Association Rules (CAR) algorithm that improves the interpretability of the results, works over real data from the electronic health records (EHRs), and allows the study of the patient as a whole. It enables tasks such as the discovery of relationships between diseases, or offering several alternative and reasoned diagnoses for the cases of patients with several diseases that analysed separately could lead to mistaken diagnosis. We aim to achieve several goals: to discover hidden relationships; to improve the interpretability and reduce the complexity of the result; to obtain more reliable diagnosis (getting alternative reasoned diagnoses and higher robustness to noisy rules), and to improve the quality of the classifier avoiding the usual over-fitting problem. To this purpose, we define and exploit hierarchies defined over datacubes dimensions, and change the way the association rules are obtained, and their evaluation at the classification process. To prove the utility of our proposal we have used it in an example of cancer discrimination.

Contextualized access to electronical health records in cardiology.

In this paper, we propose a new approach for accessing the electronical health records (EHR), and we apply it to the cardiology medical specialty. Though the use of EHR improves the storage and access to the information in it regarding the previous health records in papers, it entails the risk of having the same problems of huge size and of becoming inoperative and really difficult to handle, especially if the user is looking for a specific data item. Our proposal is based on the contextualization of the access, providing the user with the most important information for the assistance act in which he/she is involved. To do this, we define the set of possible contexts and consider different aspects of the pertinence of the documents to each context. We do it by using fuzzy logic and pay special attention to the efficiency, due to the huge size of the involved databases. Our proposal does not limit the access to the EHR, but establishes a prioritization based on the access needs, which provides the system with an additional advantage, easily enabling the use of new terminals and devices like tablet PCs and PDAs, which have great limitations in the interfaces.