Showing papers presented at "Knowledge Representation for Health-Care in 2009"

Towards a possibility-theoretic approach to uncertainty in medical data interpretation for text generation

Abstract: Many real-world applications that reason about events obtained from raw data must deal with the problem of temporal uncertainty, which arises due to error or inaccuracy in data. Uncertainty also compromises reasoning where relationships between events need to be inferred. This paper discusses an approach to dealing with uncertainty in temporal and causal relations using Possibility Theory, focusing on a family of medical decision support systems that aim to generate textual summaries from raw patient data in a Neonatal Intensive Care Unit. We describe a framework to capture temporal uncertainty and to express it in generated texts by mean of linguistic modifiers. These modifiers have been chosen based on a human experiment testing the association between subjective certainty about a proposition and the participants' way of verbalising it.

Identifying disease-centric subdomains in very large medical ontologies: a case-study on breast cancer concepts in SNOMED CT. or: Finding 2500 Out of 300.000

Abstract: Modern medical vocabularies can contain up to hundreds of thousands of concepts. In any particular use-case only a small fraction of these will be needed. In this paper we first define two notions of a disease-centric subdomain of a large ontology. We then explore two methods for identifying disease-centric subdomains of such large medical vocabularies. The first method is based on lexically querying the ontology with an iteratively extended set of seed queries. The second method is based on manual mapping between concepts from a medical guideline document and ontology concepts. Both methods include concept-expansion over subsumption and equality relations. We use both methods to determine a breast-cancer-centric subdomain of the SNOMED CT ontology. Our experiments show that the two methods produce a considerable overlap, but they also yield a large degree of complementarity, with interesting differences between the sets of concepts that they return. Analysis of the results reveals strengths and weaknesses of the different methods.

Experiences in the development of electronic care plans for the management of comorbidities

Abstract: Recent studies have shown that care plans with comprehensive home interventions can be effective in the management of chronic patients. Evidence also exists about the importance of tailoring these care plans to patients, by integrating comorbidities. In this context, the development, implementation, outcome analysis, and reengineering of care plans adapted to particular patient groups earn relevance. We are concerned with the development and reengineering of electronic care plans dealing with comorbidities. Our hypothesis is that a library of reusable care plan components can facilitate these tasks. To confirm this hypothesis we have carried out an experiment consisting in developing a library of care plan components for the management of patients with COPD or CHF, and next building a care plan for stable COPD&CHF patients by (re)using these components. In this paper we report on this experiment.

Semantic web-based modeling of clinical pathways using the UML activity diagrams and OWL-S

Abstract: Clinical Pathways can be viewed as workflows, comprising an ordering of activities with associated execution constraints. Workflow models allow formal representation, analysis and execution of workflows in the Clinical Pathways. We present a semantic web-based approach where the domain knowledge and the workflow model are modeled separately as ontologies, while the Clinical Pathway and the associated workflows are modeled as the instantiations of these ontologies. Our workflow model is based on the UML Activity Diagrams and OWL-S service ontology, and the execution semantics are based on Place/Transition Petri Nets. We demonstrate our approach by capturing the workflow of the Prostate Cancer Care Pathway.

Argumentation about treatment efficacy

Abstract: The volume and complexity of knowledge produced by medical research calls for the development of technology for automated management and analysis of such knowledge. In this paper, we identify scenarios where a researcher or a clinician may wish to use automated systems for analysing knowledge from clinical trials. For this, we propose a language for encoding, capturing and synthesising knowledge from clinical trials and a framework that allows the construction of arguments from such knowledge. We develop this framework and demonstrate its use on a case study regarding chemotherapy regimens for ovarian cancer.

From natural language descriptions in clinical guidelines to relationships in an ontology

Abstract: Knowledge Engineering allows to automate entity recognition and relation extraction from clinical texts, which in turn can be used to facilitate clinical practice guideline (CPG) modeling. This paper presents a method to recognize diagnosis and therapy entities, and to identify relationships between these entities from CPG free-text documents. Our approach applies a sequential combination of several basic methods classically used in knowledge engineering (natural language processing techniques, manually authored grammars, lexicons and ontologies), to gradually map sentences describing diagnostic and therapeutic procedures to an ontology. First, using a standardized vocabulary, our method automatically identifies guideline concepts. Next, for each sentence, it determines the patient conditions under which the descriptive knowledge of the sentence is valid. Then, it detects the central information units in the sentence, in order to match the sentence with a small set of predefined relationships. The approach enables automated extraction of relationships about findings that have manifestation in a disease, and procedures that diagnose or treat a disease.

Challenges in delivering decision support systems: the MATE experience

Abstract: Cancer Multidisciplinary Meeting (MDM) is a widely endorsed mechanism for ensuring high quality evidence-based health care. However, there are shortcomings that could ultimately result in unintended patient harm. On the other hand, clinical guidelines and clinical decision support systems (DSS) have been shown to improve decision-making in various measures. Nevertheless, their clinical use requires seamlessly interoperation with the existing electronic health record (EHR) platform to avoid the detrimental effects that duplication of data and work has in the quality of care. The aim of this work is to propose a computational framework to provide a clinical guideline-based DSS for breast cancer MDM. We discuss a range of design and implementation issues related to knowledge representation and clinical service delivery of the system, and propose a service oriented architecture based on the HL7 EHR functional model. The main result is the DSS named MATE (Multidisciplinary Assistant and Treatment sElector), which demonstrates that decision support can be effectively deployed in a real clinical setting and suggest that the technology could be generalised to other cancer MDMs.

Bridging an asbru protocol to an existing electronic patient record

Abstract: Clinical protocols can improve the quality of care if implemented in Decision Support Systems (DSS) that are used in clinical practice. For optimal user acceptance, they must use data from the existing Electronic Patient Records (EPR) and enforce only small changes in the care process and minimal extra effort for data entry. In this paper we describe how we handle the challenge of mapping a breast cancer treatment protocol encoded in Asbru to a legacy EPR which has been used by oncologists at the point of care for years. We identified different levels of integration effort ranging from readily available data in the EPR to abstractions which can only be performed by domain experts. By involving the author of the protocol in the implementation process, we were able to design a system which promises to improve the daily routine at the places of application.

A hybrid methodology for consumer-oriented healthcare knowledge acquisition

Abstract: In Consumer Healthcare Informatics it is still difficult for laypersons to find, understand and act on health information, due to the persistent communication gap between specialized medical terminology and “lay” medical terminology used by healthcare consumers. Furthermore, existing clinically-oriented terminologies cannot provide sufficient support when integrated into consumer-oriented applications, so there is a need to create consumer-friendly terminologies reflecting the different ways healthcare consumers express and think about health topics. This work suggests a methodology to acquire consumer health terminology for creating a Consumer-oriented Medical Vocabulary for Italian that mitigates this gap. This resource, aligned to a standard medical terminology, could be useful in Personal Health Records to improve users' accessibility to their healthcare data. We performed evaluation mapping on acquired data to the International Classification of Primary Care (ICPC-2) to find overlaps and the candidate “lay” terms that can be considered good synonyms for the medical ones.

Analysis of the GLARE and GPROVE approaches to clinical guidelines

Abstract: Clinical guidelines (GLs) play an important role in medical practice, and computerized support to GLs is now one of the most central areas of research in Artificial Intelligence in medicine. In recent years, many groups have developed different computer-assisted management systems of GL. Each approach has its own peculiarities and thus a comparison is necessary. Many possible aspects can be analyzed, but a first analysis has probably to consider the GL models, i.e. the representation formalisms provided. To this end, Peleg and al. [4] have analyzed and compared six different frameworks. In this paper, we analyse also GLARE and GPROVE on the basis of the same methodology. Moreover, we extend such analysis by considering the tools and the facilities that GLARE and GPROVE provide to support the use of GLs. The final goal of our analysis is to exploit the differences between these two systems and if they can be fruitfully integrated.

Sharable appropriateness criteria in GLIF3 using standards and the knowledge-data ontology mapper

Abstract: Creating computer-interpretable guidelines (CIGs) requires much effort. This effort would be leveraged by sharing CIGs with more than one implementing institution. Sharing necessitates mapping the CIG's data items to institutional EMRs. Sharing can be enhanced by using standard formats and a Global-as-view approach to data integration, where a common data model is used to generate standard views of proprietary EMRs. In this paper we demonstrate how generic guideline expressions could be encoded in the GELLO standard using HL7-RIM-based views. We also explain how the Knowledge-Data Ontology Mapper (KDOM) can be used to simplify GELLO expressions. We are aiming to use this approach for computerizing radiology appropriateness criteria and linking them with EMR data from Stanford Hospital. We discuss our initial study to assess whether such computerization would be possible and beneficial.

Technical solutions for integrating clinical practice guidelines with electronic patient records

Abstract: The success of a decision support system based on clinical practice guidelines does not only depend on the quality of the decision model used to represent and execute guideline recommendations, but also on the design of interactions of the system with the end-user interface and the electronic patient record This paper describes technical solutions adopted to add decision support functionalities to two existing information systems for stroke patients Despite the specific medical application, the approach is quite general, relying on two main functionalities: a real-time decision support system based on workflow technology (careflow) and an off-line tool for non-compliance detection, called “Reasoning on Medical Action” (RoMA) The integration has been developed maintaining independence between data management and knowledge management, and minimizing changes to existing user's interfaces The paper illustrates in particular the middleware layer created to allow communication between the evidence-based system and the electronic patient record

A knowledge-management architecture to integrate and to share medical and clinical data, information, and knowledge

Abstract: Data, information, and knowledge in medicine is varied, changing, interrelated, and for diverse purposes. Medical and Clinical care depends on the correct and efficient combined application of these elements to concrete health care situations as prophylactics, screening, diagnosis, therapy, and prognosis. In this paper, we propose a Knowledge Management Architecture (KMA) to allow the integration of medical and clinical data, information and knowledge in a consistent and incremental way. The components of KMA are described and the already implemented parts are provided with references to papers where they are explained in more detail. For the first time, we present the conceptual integration of the isolated works performed in the research group of artificial intelligence of the Rovira i Virgili University and in collaboration with the Clinical Hospital of Barcelona, and the SAGESSA health care organization.

Extracting qualitative knowledge from medical guidelines for clinical decision-support systems

Abstract: Medical guidelines provide knowledge about processes that is not directly suitable for building clinical decision-support systems We discuss a two-step approach where knowledge from a guideline on COPD is translated into temporal logic, and augmented with physiological background knowledge This allows capturing the dynamics of the processes using qualitative knowledge, while maintaining the temporal nature of the processes As a second step, this represented clinical knowledge is translated into a decision-theoretic framework We thus present a representation that can act as a basis for the construction of a decision-support system concerning monitoring of COPD

Creating topic hierarchies for large medical libraries

Abstract: Web-based medical digital libraries contain a huge amount of valuable, up-to-date health care information. However, their size, their keyword-based access methods and their lack of semantic structure make it difficult to find the desired information. In this paper we present an automatic, unsupervised and domain-independent approach for structuring the resources available in an electronic repository. The system automatically detects and extracts the main topics related to a given domain, building a taxonomical structure. Our Web-based system is integrated smoothly with the digital library's search engine, offering a tool for accessing the library's resources by hierarchically browsing domain topics in a comprehensive and natural way. The system has been tested over the well-known PubMed medical library, obtaining better topic hierarchies than those generated by widely-used taxonomic search engines employing clustering techniques.