Argumentative reasoning and taxonomic analysis for the identification of medical errors

Abstract

International audienceTelemedicine consists of the use of information and communication technologies (ICTs) in the practice of medicine. The massive digitalisation of the society is changing the behaviour of ordinary people even in medical sectors. The impact of digitisation is also having impacts on teleexpertise, where a medical professional can remotely ask some advices through the use of ICTs to provide treatment to a patient in critical conditions in remote environment. However, sometimes the outcome of such advice obtained remotely can lead to medical errors. In these situations, it is important to determine whether the causes of the errors could have been avoidable or not for the purposes of establishing the truth and assuring justice for the victims of medical errors. The proposed work fits this perspective with the objective to formalise elements of argumentation in collaborative medical organisations using telemedicine. In other words, a technique that extends the Dung's argumentation framework in order to bring out the errors committed following a remote medical procedure has been proposed. The proposed technique is underpinned by graphical reasoning. The reasoning is represented through a directed graph in which the extended nodes specify the arguments with their source(s) and the identification of errors is done according to the Makeham's and Tempos taxonomies. To illustrate the functioning of the proposed technique or solution, an example of the practice of teleexpertise (between two French hospitals) that leads to litigation is presented

Argumentation graphs with constraint-based reasoning for collaborative expertise

“…In this section, we will recall the background of conceptual graphs and constraints drawn from [10]. The argumentation system [5] used in this work is well explained in our previous works [2][3][4]11].…”

“…In this work, the weighting of the arguments takes into consideration both competencies (competencies are modelled as levels of expectation) and the sources used to support the advice of the medical professionals involved in an act of teleexpertise [4]. Several works have been achieved in weighting competencies, we can cite for example [23][24][25].…”

“…This case study requires interaction between at least two speciality X and Y . The interaction between the different arguments is represented by a graph of attacks [2][3][4]. In a knowledge base the graph of attacks is defined as fact.…”

“…An effective integration of constraints in collaborative processes must therefore become a central feature of consistent development efforts in argumentative reasoning. In fact, in our previous works [2][3][4], we have modelled collaborative processes between using the Dung argumentation framework, itself represented in conceptual graph formalism which can support the expression of semantic constraints in order to check if there are inconsistencies or not. The solution proposed in this work is to use some laws governing remote medical collaboration that will be modelled as constraints because the laws are obligations and must be respected.…”

Data mining for decision support with uncertainty on the airplane

“…selection, prioritization and transformation of features representing the problem and its input data) is also needed to improve the proposed model. Likewise, ontology-based deep learning [65] is an emergent topic for human behavior prediction with explanations and this could contribute to ease domain-specific knowledge representation or argumentations [66][67][68][69][70] and facilitate emergency efficiency during critical medical interventions in the air transport for the safety of their passengers.…”