Ebook: Computer-based Support for Clinical Guidelines and Protocols

In recent years, guidelines and protocols have gained support as the vehicles for promoting best practices in clinical medicine. They offer the possibilities of reducing unwarranted practice variations, of containing cost while maintaining high quality of care, and of defining standards of care for quality assurance purposes. These promises have led to an explosion of guideline publications. Yet, studies have shown that dissemination and effective use of guidelines in clinical care remain a major bottleneck.

A number of researchers have developed different technologies for delivering computerized guidelines in clinical care. These technologies range from alerts and reminders to knowledge-based systems, information-retrieval systems, and others. The tasks to which guidelines have been applied include classic clinical decision support, workflow management, quality assurance, and resource-requirement estimates. The research has spanned several communities (information retrieval, artificial intelligence, medical informatics, software engineering, clinical medicine), but unfortunately, there has been little cross fertilization between the communities working in this area.

Following the success of the first European Workshop on Computerized Guidelines and Protocols held at Leipzig, Germany, in 2000, the Symposium on Computerized Guidelines and Protocols (CGP-2004) was organized to identify use cases for guideline-based applications in healthcare, computerized methods for supporting the guideline development process, and pressing issues and promising approaches for developing usable and maintainable vehicles for guideline delivery. It brought together researchers from different communities to examine cutting-edge approaches to guideline modelling and application development and to consider how different communities can leverage each other's strengths. The papers collected in this volume represent the best of the contributions to this symposium.

We solicited two categories of papers for this symposium: (1) long (fifteen-page) papers that present mature research results and that review focused topics; and (2) short (five-page) papers that report early results and innovative ideas and that describe practical applications. In the first category we have papers that describe the use of formal and adaptive methods in applying protocols to clinical decision support; that review the representation of guideline goals and present an empirically derived way of categorizing them; that present methods for deriving temporal abstraction and temporal action specification in guidelines; that explore interactive visualizations for medical treatment plans; that discuss the relationship between guidelines and standard terminologies, and that demonstrate improvement in health outcomes and/or cost-effectiveness ratio with guideline compliance. In the second category we have papers that describe contrasting approaches to developing, searching, and evaluating guideline knowledge bases, formal representation and reasoning methods, the possibility of translating from one formalism to another, adapting workflow to implement treatment protocols, and the use of mark-up and data-mining technologies.

The diversity of the topics belies the fact that workers in the field share a number underlying concerns. The first is representation of medical knowledge embodied in clinical guideline and protocols. Several papers presented formal, empirical, and hybrid methods for representing such knowledge, especially the temporal aspects of guideline-based data abstractions and recommendations. For computer-supported guidelines and protocols to make a difference in clinical practices, they must be integrated into clinical information and workflow systems. Papers on deployment-driven guideline encoding, integration of standard terminologies, and adaptation of workflow processes speak to these concerns. Finally, the results of deploying computerized guidelines and protocols require evaluation. Evaluation can be done in terms of the correctness of guideline information presented to clinicians and of the effects on clinicians' compliance to guideline-recommended practices, and, ultimately, on quality and cost-effectiveness of patient care.

The symposium was held at the Novotek Hotel, Prague, in the Czech Republic, as part of the International Joint Meeting EuroMISE 2004. A number of organizations, including the University of Economics (Prague), European Centre for Medical Informatics, Statistics and Epidemiology (EuroMISE), Czech Society of Cybernetics and Informatics, Guidelines International Network (G-I-N), the Austrian Society for Artificial Intelligence (ÖGAI), and Health Level 7, endorsed the symposium and encouraged their members to participate.

Dr. Kitty Rosenbrand from the Dutch Institute for Healthcare Improvement (CBO) and Dr. Gunther Schadow from the Regenstrief Institute, Indiana University School of Medicine delivered invited talks to the Symposium. Dr. Rosenbrand trained as a physician and is employed at CBO as a senior consultant on medical guidelines She is, and has been, involved in the development of several medical guidelines by medical specialists associations as well as other societies supporting the life cycle of guidelines (like, AGREE collaboration and the Guidelines International Network (G-I-N)). Dr. Schadow has led much of the development effort for HL7 version 3 Reference Information Model and has proposed innovative ideas about how to integrate of guideline recommendations and clinical data standards.

The credit for the success of the symposium goes all of its participants. First and foremost are the authors who submitted papers and the presenters who gave talks, presented posters, and demonstrated their systems. A program committee consisting of more than 30 leading researchers reviewed submissions and gave constructive comments on them. Professor Vojtech Svatek from the University of Economics, Czech Republic and his colleagues provided the local logistical support. Peter Votruba at the Institute of Software Technology and Interactive Systems, Vienna University of Technology did the yeoman's work in formatting and preparing the papers for publication. Finally, we thank IOS Press for offering to issue the proceedings of the symposium as a volume in the “Studies in Health Technology and Informatics” book series.

December 2003

Katharina Kaiser

Silvia Miksch

Samson W. Tu

This paper presents an interactive visualization for medical treatment plans that are formulated in the plan representation language Asbru.

So far, most attention of the protocol-based care community was focused towards formal guideline representation and authoring partly supported by graphical tools. The intention of this work is to go the opposite way and communicate the logic of a computerized treatment plan to physicians, nursing-, and other medical personnel visually.

The visualization is based on the idea of flow-chart algorithms widely used in medical education and practice. This concept has been extended in order to cope with the powerful and expressive guideline representation language Asbru. Furthermore, a number of interactive navigational and overview extensions are used to intuitively support the understanding of the logic of plans.

The user-centered development approach applied for these interactive visualization methods has been guided by user input gathered via a user study, design reviews, and prototype evaluations as described in this document.

This paper presents the KASIMIR research project for the management of decision protocols in oncology. A decision protocol is a kind of decision tree implemented in an object-based representation formalism. A reasoner based on such a formalism and on hierarchical classification is coupled with a knowledge editor. This association provides an assistance for editing and maintenance of protocols, enabling the detection of errors and the comparison between versions of the protocol. In this way, a management of protocols takes fully advantage of the underlying knowledge representation and reasoning tools. This straightforward use of the protocol may be insufficient in some situations. Then, the protocol may have to be adapted for these situations. A study of protocol adaptation is presented. In particular a reasoner based on a combination of hierarchical classification and fuzzy logic is introduced.

Knowledge of clinical goals and the means to achieve them are either not represented in most current guideline representation systems or are encoded procedurally (e.g. as clinical algorithms, condition-action rules). There would be a number of major benefits if guideline enactment systems could reason explicitly about clinical objectives (e.g. whether a goal has been successfully achieved or not, whether it is consistent with prevailing conditions, or how the system should adapt to circumstances where a recommended action has failed to achieve the intended result). Our own guideline specification language, PROforma, includes a simple goal construct to address this need, but the interpretation is unsatisfactory in current enactment engines, and goals have yet to be included in the language semantics. This paper discusses some of the challenges involved in developing an explicit, declarative formalism for goals. As part of this, we report on a study we have undertaken which has identified over 200 goals in the routine management of breast cancer, and outline a tentative formal structure for this corpus.

This paper describes a new method for the ontologically based standardization of concepts with regard to the quality assurance of clinical trial protocols. We developed a data dictionary for medical and trial-specific terms in which concepts and relations are defined context-dependently. The data dictionary is provided to different medical research networks by means of the software tool Onto-Builder via the internet. The data dictionary is based on domain-specific ontologies and the top-level ontology of GOL. The concepts and relations described in the data dictionary are represented in natural language, semi-formally or formally according to their use.

Guideline and protocol representation languages have reached a level of complexity where auxiliary methods are needed to support the authoring of protocols in the particular language. Several approaches and methods exist that claim high knowledge about both, the medical context and the formal requirements. Therefore, we need knowledge-based methods to facilitate the human plan designer and create the protocols of the particular language as automated as possible. We present a three-step wrapper method, called TimeWrap, to extract information, in particular temporal issues, out of semistructured data and integrate it in a formal representation. We illustrate our approach using the guideline-representation language Asbru and examples from guidelines to treat conjunctivitis.

Guidelines are often based on a mixture of evidence-based and consensus-based recommendations. It is not straightforward that providing a series of “good” recommendations result in a guideline that is easily applicable, and it is not straightforward that acting according to such recommendations leads to an effective and efficient clinical practice. In this paper we summarize our experience in evaluating both the usability and the impact of a guideline for the acute/subacute stroke management. A computerised version of the guideline has been implemented and linked to the electronic patient record. We collected data on 386 patients. Our analysis highlighted a number of non-compliances. Some of them can be easily justified, while others depend only on physician resistance to behavioural changes and on cultural biases. From our results, health outcomes and costs are related to guideline compliance: a unit increase in the number of non-compliance results in a 7% increase of mortality at six months. Patients treated according to guidelines showed a 13% increase in treatment effectiveness at discharge, and an average cost of 2929 $\euro$ vs 3694 $\euro$ for the others.

Temporal data abstraction bridges the gap between snap shot values delivered by monitoring devices and laboratory tests on one side and high-level medical concepts used in guidelines and by medical professionals on the other side. Within this field, the detection and abstraction of repeated patterns is a complex and important challenge. A repeated pattern is a combination of events or intervals which occur multiple times in a formally describable temporal relation.

While there are many approaches to detect patterns in time series without prior definition of target concepts, we describe the application of temporal data abstraction in the context of guideline execution. Here predefined concepts of temporal patterns must be compared with measurement series describing the patient state. We discuss the requirements coming from both high-frequency domains such as intensive care units and low-frequency domains such as diabetes monitoring and show our solution based on a new version of the Asgaard data abstraction unit. It interfaces the dynamically changing patient state to the guideline execution unit and features abstraction modules ranging from simple calculations to statistical measures calculated for sliding time windows.

Medical guidelines and protocols describe the optimal care for a specific group of patients and therefore, when properly applied, improve the quality of patient care. During the last decade, a large number of medical guidelines and protocols have been published. However, the work done on developing and disseminating them far outweighs the efforts on guaranteeing their quality. Indeed, anomalies like ambiguity and incompleteness are frequent in medical guidelines and protocols. An approach grounded on a formal representation, can answer these needs, as we have demonstrated in the Protocure project. The Protocure II project will aim at integrating formal methods in the life cycle of guidelines.

The integration of a computer-based system dealing with clinical guidelines with a medical ontology can provide several advantages, including standardization and knowledge sharing. Furthermore, in order to operate in the clinical practice, guideline systems must also interact with the hospital databases to retrieve patients’ data. Unfortunately, currently there seems not to be any “standard” consensus model either for the medical ontology or for (the conceptual structure of) patient databases (even if several interesting proposals have been carried out). In this paper we show how we are extending the GLARE guideline manager in order to strictly interact with both a medical ontology and a patient Database, in such a way that GLARE is not committed to any specific ontology and database (i.e., different ontologies and/or databases can be used).

One goal in modern medicine is to increase the treatment quality. A major step towards this aim is to support the execution of standardized, guideline-based clinical protocols, which are used in many medical domains, e.g., for oncological chemotherapies. Standardized chemotherapy protocols contain detailed and structured therapy plans describing the single therapy steps (e.g., examinations or drug applications). Therefore, workflow management systems offer good support for these processes. However, the treatment of a particular patient often requires modifications due to unexpected infections, toxicities, or social factors. The modifications are described in the treatment protocol but not as part of the standard process. To be able to further execute the therapy workflows in case of exceptions running workflows have to be adapted dynamically. Furthermore, the physician should be supported by automated exception detection and decision support for derivation of necessary modifications. The AdaptFlow prototype offers the required support for the field of oncological chemotherapies by enhancing a workflow system with dynamic workflow adaptation and rule based decision support for exception detection and handling.

We present a generic means of interfacing XML documents and clinical systems. The interface has been developed to allow the integration of best practice guidance information within prescribing systems.

The interface has the following characteristics: (1) integrating developers do not have to interpret the structure of the XML documents, (2) inconsistencies between integrations are reduced, (3) the structure of the XML documents can change without affecting integration and (4) specification and documentation of the NDR document interface is within the interface itself.

The Guidelines International Network (G-I-N www.g-i-n.net) is a major new international initiative involving guideline-developing organisations from around the world. G-I-N seeks to improve the quality of health care by promoting systematic development of clinical practice guidelines and their application into practice.

The Network now has over 45 international members, most of whom prepare evidence-based clinical practice guidelines, or actively promote the use of evidence in practice. One of the priorities of the organisation is to share evidence tables and adapt guidelines for local circumstances based on international evidence. In the longer term, guideline developers are planning to create ‘living guidelines’ that can be continuously updated and used by a number of different countries.

A major consideration for guideline developers is how to communicate and work with information technology scientists to develop standards and protocols for the translation of these trans-national guidelines into electronic formats. To be effective, there must be formal internationally agreed standards that allow electronic guidelines to be shared and automatically updated.

The Guidelines International Network will be taking a leading international role in working with designers and vendors of electronic decision support systems and tools to guarantee the integrity of guidelines when translated into electronic formats. This presentation by Catherine Marshall, Kitty Rosenbrand and Guenter Ollenschlaeger will:

• explore current experiences from New Zealand, Germany and the Netherlands

• identify issues from the perspective of guideline developers

• make recommendations for establishing opportunities for software designers, vendors and informatics experts to collaborate with guideline developers to ensure that up to date evidence can be easily implemented and shared throughout the world.

A major problem in the effective use of clinical guidelines is fast and accurate access at the point of care. Thus, we are developing a digital electronic guideline library (DeGeL) and a set of tools for incremental conversion of free-text guidelines into increasingly machine-comprehensible representations, which support automated application. Even if guidelines are represented in electronic fashion, care providers need to be able to quickly retrieve the guidelines that best fit the clinical situation at hand. We describe Vaidurya, a search and retrieval engine that exploits the hybrid nature of guideline representation in the DeGeL architecture. Vaidurya can use not only free-text keywords, but also multiple semantic indices along which the guidelines are classified, and the mark up of guidelines in DeGeL, using the semantic roles of one or more guideline-representation languages (ontologies). Vaidurya offers a wide variety of querying options, in order to enable different types of users to query the guideline library in a manner that is both efficient and user friendly. We describe the customizable query interface, in which each user can create their own personal query interface.

The Stepper tool was developed to assist a knowledge engineer in developing a computable version of narrative guidelines. The system is document–centric: it formalises the initial text in multiple user–definable steps corresponding to interactive XML transformations. In this paper, we report on experience obtained by applying the tool on a narrative guideline document addressing unstable angina pectoris. Possible role of the tool and associated methodology in developing a guideline–based application is also discussed.

We develop a formal framework by which clinical guidelines and protocols (CGPs) can be partially represented as a set of terminological concept definitions using standard description logics. There are two benefits in pursuing such an approach. First, it provides a foundation for logic-based CGP fusion and collision detection. Second, it allows for the checking of clinical treatment episodes from the EPR against CGPs.

ASTI is a guideline-based decision support system for therapeutic prescribing in primary care with two modes of interaction. The “critic mode” operates as a reminder system to detect non guideline-compliant physician drug orders, whereas the “guided mode” operates on demand and provides physician guidance to help her establishing best recommended drug prescriptions for the management of hypertension. A preliminary evaluation study was conducted with 10 GPs to test the complementary nature of both modes of decision support. Results tend to validate our assumption that reminder-based interaction is appropriate for simple cases and that physicians are willing to use on-demand systems as clinical situations become more complex.

We propose to present a poster (and potentially also a demonstration of the implemented system) summarizing the current state of our work on a hybrid, multiple-format representation of clinical guidelines that facilitates conversion of guidelines from free text to a formal representation. We describe a distributed Web-based architecture (DeGeL) and a set of tools using the hybrid representation. The tools enable performing tasks such as guideline specification, semantic markup, search, retrieval, visualization, eligibility determination, run-time application and retrospective quality assessment. The representation includes four parallel formats: Free text (one or more original sources); semi-structured text (labeled by the target guideline-ontology semantic labels); semi-formal text (which includes some control specification); and a formal, machine-executable representation. The specification, indexing, search, retrieval, and browsing tools are essentially independent of the ontology chosen for guideline representation, but editing the semi-formal and formal formats requires ontology-specific tools, which we have developed in the case of the Asbru guideline-specification language. The four formats support increasingly sophisticated computational tasks. The hybrid guidelines are stored in a Web-based library. All tools, such as for runtime guideline application or retrospective quality assessment, are designed to operate on all representations. We demonstrate the hybrid framework by providing examples from the semantic markup and search tools.

Knowledge acquisition for the design of clinical decision support systems can be facilitated when clinical practice guidelines serve as a knowledge source. We describe application of the Guideline Elements Model (GEM) in the design of a decision support system to promote smoking cessation. Following selection of relevant recommendations and markup of knowledge components with the GEM Cutter editor, the Extractor stylesheet was used to create a list of decision variables and actions for further processing. Decision variables and actions that reflect similar concepts were consolidated. Action types were identified. Extracting the critical concepts from the narrative text facilitates clarification of necessary content. The guideline-centric approach promotes accurate translation of guideline knowledge.

While guideline–based decision support is safety–critical and typically requires human interaction, offline analysis of guideline compliance can be performed to large extent automatically. We examine the possibility of automatic detection of potential non-compliance followed up with (statistical) association mining. Only frequent associations of non-compliance patterns with various patient data are submitted to medical expert for interpretation. The initial experiment was carried out in the domain of hypertension management.

In this paper, we present GLARE, a domain-independent prototypical system for acquiring, representing and executing clinical guidelines. GLARE has been built within a 7-year project with Azienda Ospedaliera San Giovanni Battista in Turin (one of the largest hospitals in Italy) and has been successfully tested on clinical guidelines in different domains, including bladder cancer, reflux esophagitis, and heart failure. GLARE is characterized by the adoption of advanced Artificial Intelligence (AI) techniques, to support medical decision making and to manage temporal knowledge.

The SAGE (Standards-Based Sharable Active Guideline Environment) project is a collaboration among research groups at six institutions in the US. The ultimate goal of the project is to create an infrastructure that will allow execution of standards-based clinical practice guidelines across heterogeneous clinical information systems. This paper describes the design goals of the SAGE guideline model in the context of the technological infrastructure and guideline modeling methodology that the project is developing.

This paper presents a new guideline authoring tool, called Guideline Markup Tool (GMT). It proposes two useful features, which are missing in existing tools. First, it facilitates the translation of a free-text guideline into a formal representation, providing special XML macros. Second, it can be used to create links between the original guideline and its formal representation. Therefore, the GMT eases the implementation of clinical guidelines in a formal representation, which can be used in monitoring and therapy planning systems.

To re-examine the validity of the medical knowledge that are embedded in the legacy system, we translated a Medical Logic Module (MLM) for hyperkalemia patient screening into the GuideLine Interchange Format (GLIF). We used a set of guiding principles to direct the translation. In addition, we used the GLIF3 Guideline Execution Engine (GLEE) as a testing tool to validate the encoded GLIF guideline by applying it to 5 simulated patient cases. The result has shown that it is possible to translate Arden MLMs into GLIF guidelines. However, significant efforts are necessary to handle the problems arose during the translation process. Automatic translation could be a more generalizable approach for future work.