Martin Michalowski

Retrieving and semantically integrating heterogeneous data from the Web

Building Finder uses semantic Web technologies to integrate different data types from various online data sources. The applications use of the RDF and RDF data query language makes it usable by computer agents as well as human users. An agent would send a query, expressed in terms of its preferred ontology (schema), to a system that would then find and integrate the relevant data from multiple sources and return it using the agents ontology. We discuss about retrieving and semantically integrating heterogeneous data from the Web.

Mitigation of adverse interactions in pairs of clinical practice guidelines using constraint logic programming

We propose a new method to mitigate (identify and address) adverse interactions (drug-drug or drug-disease) that occur when a patient with comorbid diseases is managed according to two concurrently applied clinical practice guidelines (CPGs). A lack of methods to facilitate the concurrent application of CPGs severely limits their use in clinical practice and the development of such methods is one of the grand challenges for clinical decision support. The proposed method responds to this challenge. We introduce and formally define logical models of CPGs and other related concepts, and develop the mitigation algorithm that operates on these concepts. In the algorithm we combine domain knowledge encoded as interaction and revision operators using the constraint logic programming (CLP) paradigm. The operators characterize adverse interactions and describe revisions to logical models required to address these interactions, while CLP allows us to efficiently solve the logical models - a solution represents a feasible therapy that may be safely applied to a patient. The mitigation algorithm accepts two CPGs and available (likely incomplete) patient information. It reports whether mitigation has been successful or not, and on success it gives a feasible therapy and points at identified interactions (if any) together with the revisions that address them. Thus, we consider the mitigation algorithm as an alerting tool to support a physician in the concurrent application of CPGs that can be implemented as a component of a clinical decision support system. We illustrate our method in the context of two clinical scenarios involving a patient with duodenal ulcer who experiences an episode of transient ischemic attack.

Automatically utilizing secondary sources to align information across sources

XML, web services, and the semantic web have opened the door for new and exciting information-integration applications. Information sources on the web are controlled by different organizations or people, utilize different text formats, and have varying inconsistencies. Therefore, any system that integrates information from different data sources must identify common entities from these sources. Data from many data sources on the web does not contain enough information to link the records accurately using state-of-the-art record-linkage systems. However, it is possible to exploit secondary data sources on the web to improve the record-linkage process.We present an approach to accurately and automatically match entities from various data sources by utilizing a state-of-the-art record-linkage system in conjunction with a data-integration system. The data-integration system is able to automatically determine which secondary sources need to be queried when linking records from various data sources. In turn, the record-linkage system is then able to utilize this additional information to improve the accuracy of the linkage between datasets.

Using Constraint Logic Programming to Implement Iterative Actions and Numerical Measures during Mitigation of Concurrently Applied Clinical Practice Guidelines

There is a pressing need in clinical practice to mitigate (identify and address) adverse interactions that occur when a comorbid patient is managed according to multiple concurrently applied disease-specific clinical practice guidelines (CPGs). In our previous work we described an automatic algorithm for mitigating pairs of CPGs. The algorithm constructs logical models of processed CPGs and employs constraint logic programming to solve them. However, the original algorithm was unable to handle two important issues frequently occurring in CPGs – iterative actions forming a cycle and numerical measurements. Dealing with these two issues in practice relies on a physician’s knowledge and the manual analysis of CPGs. Yet for guidelines to be considered stand-alone and an easy to use clinical decision support tool this process needs to be automated. In this paper we take an additional step towards building such a tool by extending the original mitigation algorithm to handle cycles and numerical measurements present in CPGs.

Identifying inconsistencies in multiple clinical practice guidelines for a patient with co-morbidity

This paper describes a methodological approach to identifying inconsistencies when reconciling multiple clinical practice guidelines. The need to address these inconsistencies arises when a patient with co-morbidity has to be managed according to different treatment regimens. Starting with a well-known flowchart representation we discuss how to create a formal guideline model that allows for easy manipulations of its components. For this model we present how to identify conflicting actions that are manifested by treatment-treatment and treatment-disease interactions, and how to reconcile these conflicting actions.

Using First-Order Logic to Represent Clinical Practice Guidelines and to Mitigate Adverse Interactions

Clinical practice guidelines (CPGs) were originally designed to help with evidence-based management of a single disease and such single disease focus has impacted research on CPG computerization. This computerization is mostly concerned with supporting different representation formats and identifying potential inconsistencies in the definitions of CPGs. However, one of the biggest challenges facing physicians is the application of multiple CPGs to comorbid patients. While various research initiatives propose ways of mitigating adverse interactions in concurrently applied CPGs, there are no attempts to develop a generalized framework for mitigation that captures generic characteristics of the problem, while handling nuances such as precedence relationships. In this paper we present our research towards developing a mitigation framework that relies on a first-order logic-based representation and related theorem proving and model finding techniques. The application of the proposed framework is illustrated with a simple clinical example.

A constraint logic programming approach to identifying inconsistencies in clinical practice guidelines for patients with comorbidity

This paper describes a novel methodological approach to identifying inconsistencies when concurrently using multiple clinical practice guidelines. We discuss how to construct a formal guideline model using Constraint Logic Programming, chosen for its ability to handle relationships between patient information, diagnoses, and treatment suggestions. We present methods to identify inconsistencies that are manifested by treatment-treatment and treatment-disease interactions associated with comorbidity. Using an open source constraint programming system (ECLiPSe), we demonstrate the ability of our approach to find treatment given incomplete patient data and to identify possible inconsistencies.

Reformulating constraint satisfaction problems to improve scalability

Constraint Programming is a powerful approach for modeling and solving many combinatorial problems, scalability, however, remains an issue in practice. Abstraction and reformulation techniques are often sought to overcome the complexity barrier. In this paper we introduce four reformulation techniques that operate on the various components of a Constraint Satisfaction Problem (CSP) in order to reduce the cost of problem solving and facilitate scalability. Our reformulations modify one or more component of the CSP (i.e., the query, variables domains, constraints) and detect symmetrical solutions to avoid generating them. We describe each of these reformulations in the context of CSPs, then evaluate their performance and effects in on the building identification problem introduced by Michalowski and Knoblock.

Reformulating CSPs for scalability with application to geospatial reasoning

While many real-world combinatorial problems can be advantageously modeled and solved usingConstraint Programming, scalability remains a major issue in practice. Constraint models that accurately reflect the inherent structure of a problem, solvers that exploit the properties of this structure, and reformulation techniques that modify the problem encoding to reduce the cost of problem solving are typically used to overcome the complexity barrier. In this paper, we investigate such approaches in a geospatial reasoning task, the building-identification problem (BID), introduced and modeled as a Constraint Satisfaction Problem by Michalowski and Knoblock [1]. We introduce an improved constraint model, a custom solver for this problem, and a number of reformulation techniques that modify various aspects of the problem encoding to improve scalability. We show how interleaving these reformulations with the various stages of the solver allows us to solve much larger BID problems than was previously possible. Importantly, we describe the usefulness of our reformulations techniques for general Constraint Satisfaction Problems, beyond the BID application.

Exploiting automatically inferred constraint-models for building identification in satellite imagery

The building identification (BID) problem is based on a process that uses publicly available information to automatically assign addresses to buildings in satellite imagery. In previous work, we have shown the advantages of casting the BID problem as a Constraint Satisfaction Problem (CSP) using the same generic constraint-model to represent all problem instances. However, a generic model is unable to represent with the necessary precision the addressing variations throughout the world, limiting the applicability of our previous approach. In this paper, we describe the end-to-end process used to solve the BID with a new model-generation technique that uses instance-specific information to automatically infer a representative constraint model of the BID. This inferred model is used by our custom constraint solver to identify buildings in satellite imagery more efficiently and with higher precision than using a single model. We evaluate our approach on El Segundo California, and empirically demonstrate its effectiveness for geographic areas larger than previously tested. We conclude with a discussion of the generality of our approach, and present directions for future work.