Stéphanie Bernonville

Data Mining to Generate Adverse Drug Events Detection Rules

Adverse drug events (ADEs) are a public health is sue. Their detection usually relies on voluntary reporting or medical chart reviews. The objective of this paper is to automatically detect cases of ADEs by data mining. 115 447 complete past hospital stays are extracted from six French, Danish, and Bulgarian hospitals using a common data model including diagnoses, drug administrations, laboratory results, and free-text records. Different kinds of outcomes are traced, and supervised rule induction methods (decision trees and association rules) are used to discover ADE detection rules, with respect to time constraints. The rules are then filtered, validated, and reorganized by a committee of experts. The rules are described in a rule repository, and several statistics are automatically computed in every medical department, such as the confidence, relative risk, and median delay of outcome appearance. 236 validated ADE-detection rules are discovered; they enable to detect 27 different kinds of outcomes. The rules use a various number of conditions related to laboratory results, diseases, drug administration, and demographics. Some rules involve innovative conditions, such as drug discontinuations.

Integrating the SE and HCI models in the human factors engineering cycle for re-engineering Computerized Physician Order Entry systems for medications: Basic principles illustrated by a case study

OBJECTIVES The integration of Human Factors is still insufficient in the design and implementation phases of complex interactive systems such as Computerized Physician Order Entry (CPOE) systems. One of the problems is that human factors specialists have difficulties to communicate their data and to have them properly understood by the computer scientists in the design and implementation phases. This paper presents a solution to this problem based on the creation of common documentation supports using Software Engineering (SE) and Human-Computer Interaction (HCI) methods. METHOD The integration of SE and HCI methods and models is an interesting means for modelling an organizations activities, with software applications being part of these activities. Integrating these SE and HCI methods and models allows case studies to be seen from the technical, organizational and ergonomic perspectives, and also makes it easier to compare current and future work situations. RESULTS The exploitation of these techniques allows the creation of common work supports that can be easily understandable by computer scientists and relevant for re-engineering or design. In this paper, the basic principles behind such communication supports are described and illustrated by a real case study.

Contextual data entry system design in the healthcare domain

To create an adapted and effective data entry system, we need to answer to a number of questions. Who is the system user? What tasks can be completed with the system? Where is the system used? What are the tools that can help the user during data entry? The answers to these questions represent the data entry context and the needs of the user. This paper concerns data entry assistance for data coding in the healthcare domain. We propose a contextual data entry system dedicated to healthcare professionals. The name of our system is Hermes. The aim of Hermes is to manage a set of data entry tools. A data entry tool is selected according to the context of the users data entry. Finally, we test the performance of our system by simulating scenarios. The results are presented and discussed in this paper.

First experimentation of the ErgoPNets method using dynamic modeling to communicate usability evaluation results

When a computer application is being designed or re-engineered, especially a user-centred application, communication between ergonomists and computer scientists is very important. However, the formalisms used to describe ergonomic problems and recommendations are often based on natural language. Consequently, the results of ergonomic evaluation can be poorly understood or interpreted by computer scientists. To remedy this problem, we propose a method, called ErgoPNets. The method creates common work support for both the ergonomists and the computer scientists working on the same project. Comprehensible for everyone, this support must provide an efficient tool that can be used by each person involved. ErgoPNets uses Petri nets to model Human-Computer Interaction (HCI) procedures and ergonomic criteria to model the ergonomic analysis. A first experimentation has been performed with designers/developers and academic researchers.

Towards an assistance for selecting methods and models for interactive software design or re-engineering within complex organisation: Application case of a CPOE software

The hospital activities display all of the characteristics of a very complex sociotechnic organisation. Consequently, the design of Computerized Physician Order Entry (CPOE) softwares which support this type of activity is very difficult. The Study of the procedure of these projects shows communication information problems between project partners. This article proposes the principles of an approach, based on Software Engineering (SE) and Human-Computer Interaction (HCI) methods and models, allowing improving the transmission of information during design or re-engineering projects within complex organisation.

Développement et exploitation d’une taxonomie visant l’aide à la conception d’un système d’aide à la décision médicamenteuse contextualisé

Clinical Decision Support Systems (CDSS) are implemented in hospital settings to improve the reliability of drugs ordering. Nowadays, CDSS have limited effects due to their tendency to over-alert. In order to help healthcare professionals to take into account alerts, it is necessary to adapt the CDSS to their activity. Thus, we have to consider contextualisation aspects in the design of the system. In this article, we propose a taxonomy embedding elements of contextualisation extracted from an activity analysis to guide the design of a contextualised CDSS. This taxonomy has been developed within the framework of the European project PSIP (Patient Safety through Intelligent Procedures in medication) aiming at identifying and preventing Adverse Drug Events.