Paolo Terenziani

A modular approach for representing and executing clinical guidelines

In this paper, we propose an approach for managing clinical guidelines. We outline a modular architecture, allowing us to separate two conceptually distinct aspects: the representation (and acquisition) of clinical guidelines and their execution. We propose an expressive formalism, which allows one to deal with the context-dependent character of clinical guidelines and also takes into account different temporal aspects. We also describe our tool for acquiring clinical guidelines, which provides a user-friendly interface to physicians, and automatically detects many forms of syntactic and semantic inconsistencies in the guidelines being acquired. In the second part of the paper, we describe a flexible engine for executing clinical guidelines (e.g. for clinical decision support applications, for medical education, or for integrating guidelines into the clinical practice), focusing our attention on temporal issues.

A spectrum of definitions for temporal model-based diagnosis

Abstract Model-based diagnosis (MBD) tackles the problem of troubleshooting systems starting from a description of their structure and function (or behavior). Time is a fundamental dimension in MBD: the behavior of most systems is time-dependent in one way or another. Temporal MBD, however, is a difficult task and indeed many simplifying assumptions have been adopted in the various approaches in the literature. These assumptions concern different aspects such as the type and granularity of the temporal phenomena being modeled, the definition of diagnosis, the ontology for time being adopted. Unlike the atemporal case, moreover, there is no general “theory” of temporal MBD which can be used as a knowledge-level characterization of the problem. In this paper we present a general characterization of temporal model-based diagnosis. We distinguish between different temporal phenomena that can be taken into account in diagnosis and we introduce a modeling language which can capture all such phenomena. Given a suitable logical semantics for such a modeling language, we introduce a general characterization of the notions of diagnostic problem and explanation, showing that in the temporal case these definitions involve different parameters. Different choices for the parameters lead to different approaches to temporal diagnosis. We define a framework in which different dimensions for temporal model-based diagnosis can be analyzed at the knowledge level, pointing out which are the alternatives along each dimension and showing in which cases each one of these alternatives is adequate. In the final part of the paper we show how various approaches in the literature can be classified within our framework. In this way, we propose some guidelines to choose which approach best fits a given application problem.

Integrating calendar dates and qualitative temporal constraints in the treatment of periodic events

The paper describes a framework for representing and reasoning with periodic events. In particular, it proposes a temporal formalism which deals with both (1) quantitative information concerning the frame of time (e.g., between 1990 and 1993) and the user-defined calendar-dates (e.g., on the first Mondays of April) in which periodic events are located and (2) the qualitative relations between periodic events (e.g., Sam visits the branch office XO1 before going to his office). The meaning of the temporal specifications in our formalism is described in logical terms. The paper defines the basic operations of inversion, intersection and composition of temporal specifications. These operations are correct (with respect to the logical definition of the specifications) and do not lose information. Finally, the paper also describes a correct algorithm which takes advantage of these operations for performing temporal reasoning, and analyses its complexity. An application of the temporal framework to the scheduling in an office is shown in an example.

Reconciling point-based and interval-based semantics in temporal relational databases: a treatment of the Telic/Atelic distinction

The analysis of the semantics of temporal data and queries plays a central role in the area of temporal databases. Although many different algebrae and models have been proposed, almost all of them are based on a point-based (snapshot) semantics for data. On the other hand, in the areas of linguistics, philosophy, and recently, artificial intelligence, an oft-debated issue concerns the use of an interval-based versus a point-based semantics. In this paper, we first show some problems inherent in the adoption of a point-based semantics for data, then argue that these problems arise because there is no distinction drawn in the data between telic and atelic facts. We then introduce a three-sorted temporal model and algebra including coercion functions for transforming relations of one sort into relations of the other at query time which properly copes with these issues.

Adopting model checking techniques for clinical guidelines verification

OBJECTIVES Clinical guidelines (GLs) are assuming a major role in the medical area, in order to grant the quality of the medical assistance and to optimize medical treatments within healthcare organizations. The verification of properties of the GL (e.g., the verification of GL correctness with respect to several criteria) is a demanding task, which may be enhanced through the adoption of advanced Artificial Intelligence techniques. In this paper, we propose a general and flexible approach to address such a task. METHODS AND MATERIALS Our approach to GL verification is based on the integration of a computerized GL management system with a model-checker. We propose a general methodology, and we instantiate it by loosely coupling GLARE, our system for acquiring, representing and executing GLs, with the model-checker SPIN. RESULTS We have carried out an in-depth analysis of the types of properties that can be effectively verified using our approach, and we have completed an overview of the usefulness of the verification task at the different stages of the GL life-cycle. In particular, experimentation on a GL for ischemic stroke has shown that the automatic verification of properties in the model checking approach is able to discover inconsistencies in the GL that cannot be detected in advance by hand. CONCLUSION Our approach thus represents a further step in the direction of general and flexible automated GL verification, which also meets usability requirements.

Later: managing temporal information efficiently

The Later (LAyered TEmporal Reasoner) system is a general-purpose manager of temporal information which provides a high-level interface language. Later works as a specialized temporal module loosely coupled with other modules to deal with time in different tasks. The Later knowledge server operates as a cooperative agent for use by various problem solvers (or applications) that need to deal with time. Its clear, easy-to-use interface language lets users easily manipulate and query a temporal knowledge base. Laters predictable behavior means that temporal reasoning is correct and complete, and that reasoning is computationally tractable. Its query processing is efficient; in fact, query processing is the basis of the integration with other reasoning tasks.

On the computational complexity of querying bounds on differences constraints

Abstract Given a consistent knowledge base formed by a set of constraints, efficient query answering (e.g., checking whether a set of constraints is consistent with the knowledge base or necessarily true in it) is practically very important. In the paper we consider bounds on differences (which are an important class of constraints based on linear inequalities) and we analyze the computational complexity of query answering. More specifically, we consider various common types of queries and we prove that if the minimal network produced by constraint satisfaction algorithms (and characterizing the solutions to a set of constraints) is maintained, then the complexity of answering a query depends only on the dimension of the query and not on the dimension of the knowledge base (which is usually much larger than the query). We also analyse how the approach can be used to deal efficiently with a class of updates to the knowledge base. Some applications of the results are sketched in the conclusion.

Applying Artificial Intelligence to Clinical Guidelines: The GLARE Approach.

In this paper, we present GLARE, a domain-independent system for acquiring, representing and executing clinical guidelines. GLARE is characterized by the adoption of Artificial Intelligence (AI) techniques at different levels in the definition and implementation of the system. First of all, a high-level and user-friendly knowledge representation language has been designed, providing a set of representation primitives. Second, a user-friendly acquisition tool has been designed and implemented, on the basis of the knowledge representation formalism. The acquisition tool provides various forms of help for the expert physicians, including different levels of syntactic and semantic tests in order to check the “well-formedness” of the guidelines being acquired. Third, a tool for executing guidelines on a specific patient has been made available. The execution module provides a hypothetical reasoning facility, to support physicians in the comparison of alternative diagnostic and/or therapeutic strategies. Moreover, advanced and extended AI techniques for temporal reasoning and temporal consistency checking are used both in the acquisition and in the execution phase. The GLARE approach has been successfully tested on clinical guidelines in different domains, including bladder cancer, reflux esophagitis, and heart failure.

Extending Temporal Relational Databases to Deal with Imprecise and Qualitative Temporal Information

In several application domains, there is a need for historical databases supporting also imprecise and qualitative temporal information. We propose a formal theory for handling imprecise and qualitative temporal information in a relational temporal database. In particular, we introduce a relational algebra in which the traditional operators are extended in order to deal with imprecise and qualitative time; we examine the properties of the operators, showing that the algebra we introduce is a consistent extension of the snapshot algebra. Finally, we show how a temporal relational database based on this theory can be built on top of a conventional relational database.

TIME, ACTION-TYPES, AND CAUSATION: AN INTEGRATED ANALYSIS

In this paper we focus on the temporal constraints between causes and effects of causal relations, and, to deal correctly with such relations, we stress the importance of analyzing the action‐types (aspectual category) of causes and effects. In particular, we propose a domain‐independent ontology in which the distinctions between action‐types (e.g., the distinction between durative and punctual situations) are dealt with, and different types of causal relations are distinguished, on the basis of (i) the temporal constraints they impose between causes and effects (these constraints are expressed in a temporal formalism that extends Vilains point interval algebra) and (ii) the action‐types of their causes and effects. Our ontology allows one to capture precisely the temporal constraints imposed by causation and the action‐types of the related situations. Moreover, in case the user has no accurate knowledge about the action‐types of some situations and/or the types of some causal connections to be dealt with, our formalism allows the user to leave the descriptions underspecified, and more specific pieces of information may be inferred. Inferences provide a flow of information about action‐types to information about temporal constraints in causation and vice versa, and demonstrate that a deep integration of time and causation is provided also at the inferential level.