Acquisition and Analysis of Repeating Patterns in Time-oriented Clinical Data

Objectives: (1) Creation of an expressive language for specification of temporal patterns in clinical domains, (2) Development of a graphical knowledge-acquisition tool allowing expert physicians to define meaningful domain-specific patterns, (3) Implementation of an interpreter capable of detecting such patterns in clinical databases, and (4) Evaluation of the tools in the domains of diabetes and oncology.

Methods: We describe a constraint-based language, named CAPSUL, for specification of temporal patterns. We implemented a knowledge-acquisition tool and a temporal-pattern interpreter within Résumé, a larger temporal-abstraction architecture. We evaluated the knowledge-acquisition process with the help of domain experts. In collaboration with the Rush Presbyterian/St. Luke’s Medical Center, we analyzed data of bone-marrow transplantation patients. The expert compared the detected patterns to a manual inspection of the data, with the help of an experimental information-visualization tool we are developing in a related project.

Results: The CAPSUL language was expressive enough during the knowledge-acquisition process to capture almost all of the patterns that the experts found useful. The patterns detected in the data by the pattern interpreter were all verified as correct. Completeness (whether all correct patterns were found) was difficult to assess, due to the size of the database.

Conclusions: The CAPSUL language enables medical experts to express temporal patterns involving multiple levels of abstraction of clinical data. The ability to reuse both domain-patterns and abstract constraints seems highly useful. The Résumé interpreter, augmented by the CAPSUL semantics, finds the complex patterns within a clinical time-oriented database in a sound fashion.

Change Management of Shared and Local Versions of Health-Care Terminologies

Local sites that adopt a shared health-care terminology for computer-based systems have local needs that prompt the local-terminology maintainers to make changes to the local version of the shared terminology. If the local site is motivated to conform to the shared terminology, then the burden lies with the local site to manage its own changes and to incorporate the changes of the shared version at periodic intervals. We call this process synchronization. We survey current approaches that address problems of sharing and local modification, and we present the CONCORDIA model, which supports carefully controlled divergence of a local version from a shared terminology. CONCORDIA provides the underlying design and methodology for the implementation of a synchronization-support tool.

Medical Informatics: Between Science and Engineering, Between Academia and Industry

Objective: To analyze the nature and appropriate role of the Medical Informatics research and practice area in the 21st Century, and to determine its links to academic environments versus industrial companies and health-care organizations.

Methods: A qualitative analysis of the state of the art of Medical Informatics, based on observation of current medical informatics programs and research in academic and industrial sites.

Results and Conclusions: Medical Informatics is definitely a scientific and technological area of endeavor, although somewhat ill-defined in scope. It is situated between science and engineering, but much closer to the engineering world, and its multidisciplinary nature fits well the engineering paradigm. It is better viewed as a specialization of the informatics field rather than as a basic medical science. However, there are good arguments as to why Medicine should be the first among equals to have its own informatics domain. Medical Informatics must have extensions to both academia and industry to survive. Medical informaticians, whether implicitly or explicitly, exist in three different environments: academic, clinical (user), and industrial (informatics developer); all three environments must be considered when trying to predict the future of this new multi-disciplinary area.

Commentaries on “Informatics and Medicine: From Molecules to Populations”

Objective: To discuss interdisciplinary research and education in the context of informatics and medicine by commenting on the paper of Kuhn et al. “Informatics and Medicine: From Molecules to Populations”.

Method: Inviting an international group of experts in biomedical and health informatics and related disciplines to comment on this paper.

Results and Conclusions: The commentaries include a wide range of reasoned arguments and original position statements which, while strongly endorsing the educational needs identified by Kuhn et al., also point out fundamental challenges that are very specific to the unusual combination of scientific, technological, personal and social problems characterizing biomedical informatics. They point to the ultimate objectives of managing difficult human health problems, which are unlikely to yield to technological solutions alone. The psychological, societal, and environmental components of health and disease are emphasized by several of the commentators, setting the stage for further debate and constructive suggestions.

Intelligent Interactive Visual Exploration of Temporal Associations among Multiple Time-oriented Patient Records

Objectives: To design, implement and evaluate the functionality and usability of a methodology and a tool for interactive exploration of time and value associations among multiple-patient longitudinal data and among meaningful concepts derivable from these data.

Methods: We developed a new, user-driven, interactive knowledge-based visualization technique, called Temporal Association Charts (TACs). TACs support the investigation of temporal and statistical associations within multiple patient records among both con cepts and the temporal abstractions derived from them. The TAC methodology was implemented as part of an interactive system, called VISITORS, which supports intelligent visualization and exploration of longitudinal patient data. The TAC module was evaluated for functionality and usability by a group of ten users, five clinicians and five medical informaticians. Users were asked to answer ten questions using the VISITORS system, five of which required the use of TACs.

Results: Both types of users were able to answer the questions in reasonably short periods of time (a mean of 2.5 ± 0.27 minutes) and with high accuracy (95.3 ± 4.5 on a 0–100 scale), without a significant difference between the two groups. All five questions requiring the use of TACs were answered with similar response times and accuracy levels. Similar accuracy scores were achieved for questions requiring the use of TACs and for questions requiring the use only of general exploration operators. However, response times when using TACs were slightly longer.

Conclusions: TACs are functional and usable. Their use results in a uniform performance level, regardless of the type of clinical question or user group involved.

Effects of changes in copayment for obstetric emergency room visits on the utilization of obstetric emergency rooms

In view of the growing proportion of "non-urgent" admissions to obstetric emergency rooms (OERs) and recent changes in copayment policies for OER visits in Israel, we assessed factors contributing to OER overcrowding. The changes investigated were (a) exemption from copayment for women with birth contractions, (b) allowing phone referrals to the OER and (c) exemption from copayment during primary care clinic closing hours. We analyzed data of a large tertiary hospital with 37 deliveries per day. Counts of women discharged to home from the OER were an indicator of "non-urgent" visits. The annual number of non-urgent visits increased at a higher rate (3.4%) than the natural increase in deliveries (2.1%). Exemption from copayment for visits during non-working hours of primary care clinics was associated with increases in OER admissions (IRR=1.22) and in non-urgent OER visits (IRR=1.54). Younger and first-time mothers with medically unjustified complaints were more likely to be discharged to home. We showed that the changes in the policy for OER copayment meant to attract new clients to the HMO had an independent impact on OER utilization, and hence, added to the workload of medical personnel. The change in HMO policy regulating OER availability requires rigorous assessment of possible health system implications.

[Mining medical knowledge from time-oriented clinical data using existing clinical knowledge]

The analysis of clinicaL data accumulating over time from multipLe sources, regarding a group of patients, can lead to muLtiple insights. In particular, such an analysis enables: (a) the discovery of temporal patterns repeating above a certain threshold frequency, thus essentially creating clusters of different behaviors over time of various patient sub-groups; and (b) detection amongst the temporal patterns, some of which, together with additional patient data, such as demographic data, might be able to predict future cLinically significant outcomes, such as renal dysfunction in the case of a diabetes patient. The analysis of temporal data is even more efficient when it includes not only time-stamped, point-based raw data, but also time intervals (periods) during which certain context-sensitive, abstract interpretations of the data hold, such as a period of moderate anemia instead of a series of hemoglobin values; or a degradation in liver functions, instead of a series of different enzyme values. Such an interpretation naturally requires an explicit representation of the medical knowledge involved, in a medical knowledge base, in a manner that supports automated computational tools. In this survey, we briefly introduce several of the innovative computational methods developed in our research center, for the purpose of the multivariate analysis of time-oriented clinical data. We mainly demonstrate the use of tools for exploration and knowledge discovery which are intended for use by a cLinicaL user at the point of care, and by cLinical researchers or heaLthcare policy makers. *

Commentaries on "Informatics and medicine: from molecules to populations"

OBJECTIVE

To discuss interdisciplinary research and education in the context of informatics and medicine by commenting on the paper of Kuhn et al. "Informatics and Medicine: From Molecules to Populations".

METHOD

Inviting an international group of experts in biomedical and health informatics and related disciplines to comment on this paper.

RESULTS AND CONCLUSIONS

The commentaries include a wide range of reasoned arguments and original position statements which, while strongly endorsing the educational needs identified by Kuhn et al., also point out fundamental challenges that are very specific to the unusual combination of scientific, technological, personal and social problems characterizing biomedical informatics. They point to the ultimate objectives of managing difficult human health problems, which are unlikely to yield to technological solutions alone. The psychological, societal, and environmental components of health and disease are emphasized by several of the commentators, setting the stage for further debate and constructive suggestions.

Hybrid specification, storage, retrieval and runtime application of clinical guidelines

Clinical guidelines are a major tool in improving the quality of medical care. However, most guidelines are in free text, are not machine-comprehensible and are not easily accessible to clinicians at the point of care. We have designed and implemented a web-based, modular, distributed architecture, the Digital Electronic Guideline Library (DeGeL), which facilitates gradual conversion of clinical guidelines from text to a formal representation in the chosen target guideline ontology. The architecture supports guideline classification, semantic markup, context-sensitive search, browsing, run-time application and retrospective quality assessment. The DeGeL hybrid meta-ontology includes elements common to all guideline ontologies, such as semantic classification and domain knowledge; it also includes four content-representation formats: free text, semi-structured text, semi-formal representation and a formal representation. These formats support increasingly sophisticated computational tasks. Guidelines can thus be in a hybrid representation in which guidelines, and even parts of the same guideline, might exist at different formalisation levels. We have also developed and rigorously evaluated a methodology and an associated web-based tool, Uruz, for gradually structuring and semi-formalising free-text clinical guidelines. Finally, we have designed, implemented and evaluated a new approach, the hybrid runtime application model, for supporting runtime application of clinical guidelines that are not necessarily in a machine-comprehensible format; in particular, when the guideline is in a semi-formal representation and the patient's data are either in an electronic medical record or in a paper format. The tool implementing this new approach, the Spock module, is customised at this point to the Asbru guideline specification language and exploits the hybrid structure of guidelines in DeGeL. The Spock module also exploits our temporal-abstraction mediator to the patient record, IDAN, and our interactive intelligent-visualisation tool, KNAVE-II.

Medical informatics: between science and engineering, between academia and industry

OBJECTIVE

To analyze the nature and appropriate role of the Medical Informatics research and practice area in the 21st Century, and to determine its links to academic environments versus industrial companies and health-care organizations.

METHODS

A qualitative analysis of the state of the art of Medical Informatics, based on observation of current medical informatics programs and research in academic and industrial sites.

RESULTS AND CONCLUSIONS

Medical Informatics is definitely a scientific and technological area of endeavor, although somewhat ill-defined in scope. It is situated between science and engineering, but much closer to the engineering world, and its multidisciplinary nature fits well the engineering paradigm. It is better viewed as a specialization of the informatics field rather than as a basic medical science. However, there are good arguments as to why Medicine should be the first among equals to have its own informatics domain. Medical Informatics must have extensions to both academia and industry to survive. Medical informaticians, whether implicitly or explicitly, exist in three different environments: academic, clinical (user), and industrial (informatics developer); all three environments must be considered when trying to predict the future of this new multidisciplinary area.

Acquisition and analysis of repeating patterns in time-oriented clinical data

OBJECTIVES

(1) Creation of an expressive language for specification of temporal patterns in clinical domains, (2) Development of a graphical knowledge-acquisition tool allowing expert physicians to define meaningful domain-specific patterns, (3) Implementation of an interpreter capable of detecting such patterns in clinical databases, and (4) Evaluation of the tools in the domains of diabetes and oncology.

METHODS

We describe a constraint-based language, named CAPSUL, for specification of temporal patterns. We implemented a knowledge-acquisition tool and a temporal-pattern interpreter within Résumé, a larger temporal-abstraction architecture. We evaluated the knowledge-acquisition process with the help of domain experts. In collaboration with the Rush Presbyterian/St. Luke's Medical Center, we analyzed data of bone-marrow transplantation patients. The expert compared the detected patterns to a manual inspection of the data, with the help of an experimental information-visualization tool we are developing in a related project.

RESULTS

The CAPSUL language was expressive enough during the knowledge-acquisition process to capture almost all of the patterns that the experts found useful. The patterns detected in the data by the pattern interpreter were all verified as correct. Completeness (whether all correct patterns were found) was difficult to assess, due to the size of the database.

CONCLUSIONS

The CAPSUL language enables medical experts to express temporal patterns involving multiple levels of abstraction of clinical data. The ability to reuse both domain-patterns and abstract constraints seems highly useful. The Résumé interpreter, augmented by the CAPSUL semantics, finds the complex patterns within a clinical time-oriented database in a sound fashion.

Medical quality assessment by scoring adherence to guideline intentions

Quality assessment of clinician actions and patient outcomes is a central problem in guideline- or standards-based medical care. In this paper we describe an approach for evaluating and consistently scoring clinician adherence to medical guidelines using the intentions of guideline authors. We present the Quality Indicator Language (QUIL) that may be used to formally specify quality constraints on physician behavior and patient outcomes derived from medical guidelines. We present a modeling and scoring methodology for consistently evaluating multi-step and multi-choice guideline plans based on guideline intentions and their revisions.

Change management of shared and local versions of health-care terminologies

Local sites that adopt a shared health-care terminology for computer-based systems have local needs that prompt the local-terminology maintainers to make changes to the local version of the shared terminology. If the local site is motivated to conform to the shared terminology, then the burden lies with the local site to manage its own changes and to incorporate the changes of the shared version at periodic intervals. We call this process synchronization. We survey current approaches that address problems of sharing and local modification, and we present the CONCORDIA model, which supports carefully controlled divergence of a local version from a shared terminology. CONCORDIA provides the underlying design and methodology for the implementation of a synchronization-support tool.

Dimensions of time in illness: an objective view

It is almost impossible to try to represent and analyze clinical data in the absence of a temporal dimension. The temporal aspect is especially important when automated decision support is used for patient care over substantial periods. This paper emphasizes the crucial role that tasks of temporal reasoning and temporal maintenance play in modern medical information and decision support systems; it also discusses the implications of providing automated support to clinicians who must perform such tasks as part of broader clinical tasks (for example, monitoring and therapy). Temporal reasoning tasks mainly involve intelligent analysis of time-oriented clinical data, and temporal maintenance tasks focus on effective storage and retrieval of these data. Both types of tasks, however, are highly relevant for such applications as patient monitoring, proper use of therapeutic guidelines, assessment of the quality of guideline use, and visualization and exploration of time-oriented biomedical data. Scientists in medical informatics should view the integration of these two areas as a major research and development goal. This paper demonstrates one approach to integration by presenting the concept of a temporal mediator, which combines temporal reasoning and temporal maintenance. Use of the temporal mediator in several clinical tasks is also presented and discussed.

Intelligent visualization and exploration of time-oriented clinical data

Physicians and other care providers often need to quickly browse and interpret large numbers of time-oriented clinical data. Reducing the information overload involving such tasks is a major goal for medical information systems. We describe a conceptual architecture and software implementation specific to the task of interpretation, summarization, visualization, explanation, and interactive exploration of time-oriented clinical data and the multiple levels of meaningful concepts that can be derived from these data. We build on our work on abstraction of time-oriented clinical data using a knowledge base, acquired from expert physicians, of temporal properties of the data. The core module of the new framework is called KNAVE (Knowledge-based Navigation of Abstractions for Visualization and Explanation). Health care providers can manipulate the display though several visualization and exploration operators. These operators have semantics that are domain independent but that are customized automatically for the application by access to the domain-specific knowledge base. The display, which reflects data and derived interpretations in the patient's database, changes when the user explores key relations (e.g., the dependency hierarchy) in the knowledge base of the relevant clinical domain. Preliminary assessment of the initial prototype with several clinical users has been encouraging. The KNAVE methodology has broad ramifications for reducing the load that large numbers of time-oriented clinical data put on care providers.

Semi-automated entry of clinical temporal-abstraction knowledge

OBJECTIVES

The authors discuss the usability of an automated tool that supports entry, by clinical experts, of the knowledge necessary for forming high-level concepts and patterns from raw time-oriented clinical data.

DESIGN

Based on their previous work on the RESUME system for forming high-level concepts from raw time-oriented clinical data, the authors designed a graphical knowledge acquisition (KA) tool that acquires the knowledge required by RESUME. This tool was designed using Protégé, a general framework and set of tools for the construction of knowledge-based systems. The usability of the KA tool was evaluated by three expert physicians and three knowledge engineers in three domains-the monitoring of children's growth, the care of patients with diabetes, and protocol-based care in oncology and in experimental therapy for AIDS. The study evaluated the usability of the KA tool for the entry of previously elicited knowledge.

MEASUREMENTS

The authors recorded the time required to understand the methodology and the KA tool and to enter the knowledge; they examined the subjects' qualitative comments; and they compared the output abstractions with benchmark abstractions computed from the same data and a version of the same knowledge entered manually by RESUME experts.

RESULTS

Understanding RESUME required 6 to 20 hours (median, 15 to 20 hours); learning to use the KA tool required 2 to 6 hours (median, 3 to 4 hours). Entry times for physicians varied by domain-2 to 20 hours for growth monitoring (median, 3 hours), 6 and 12 hours for diabetes care, and 5 to 60 hours for protocol-based care (median, 10 hours). An increase in speed of up to 25 times (median, 3 times) was demonstrated for all participants when the KA process was repeated. On their first attempt at using the tool to enter the knowledge, the knowledge engineers recorded entry times similar to those of the expert physicians' second attempt at entering the same knowledge. In all cases RESUME, using knowledge entered by means of the KA tool, generated abstractions that were almost identical to those generated using the same knowledge entered manually.

CONCLUSION

The authors demonstrate that the KA tool is usable and effective for expert physicians and knowledge engineers to enter clinical temporal-abstraction knowledge and that the resulting knowledge bases are as valid as those produced by manual entry.

Representation of change in controlled medical terminologies

Computer-based systems that support health care require large controlled terminologies to manage names and meanings of data elements. These terminologies are not static, because change in health care is inevitable. To share data and applications in health care, we need standards not only for terminologies and concept representation, but also for representing change. To develop a principled approach to managing change, we analyze the requirements of controlled medical terminologies and consider features that frame knowledge-representation systems have to offer. Based on our analysis, we present a concept model, a set of change operations, and a change-documentation model that may be appropriate for controlled terminologies in health care. We are currently implementing our modeling approach within a computational architecture.

The Asgaard project: a task-specific framework for the application and critiquing of time-oriented clinical guidelines

Clinical guidelines can be viewed as generic skeletal-plan schemata that represent clinical procedural knowledge and that are instantiated and refined dynamically by care providers over significant time periods. In the Asgaard project, we are investigating a set of tasks that support the application of clinical guidelines by a care provider other than the guideline's designer. We are focusing on the application of the guideline, recognition of care providers' intentions from their actions, and critique of care providers' actions given the guideline and the patient's medical record. We are developing methods that perform these tasks in multiple clinical domains, given an instance of a properly represented clinical guideline and an electronic medical patient record. In this paper, we point out the precise domain-specific knowledge required by each method, such as the explicit intentions of the guideline designer (represented as temporal patterns to be achieved or avoided). We present a machine-readable language, called Asbru, to represent and to annotate guidelines based on the task-specific ontology. We also introduce an automated tool for the acquisition of clinical guidelines based on the same ontology, developed using the PROTEGE-II framework.

Timing is everything. Time-oriented clinical information systems

Time is important in clinical information systems. Representing, maintaining, querying, and reasoning about time-oriented clinical data is a major theoretical and practical research area in medical informatics. In this nonexhaustive overview, we present a brief synopsis of research efforts in designing and developing time-oriented information systems in medicine. These efforts can be viewed from either an application point of view, distinguishing between different clinical tasks (such as diagnosis versus therapy) and clinical areas (such as infectious diseases versus oncology), or a methodological point of view, distinguishing between different theoretical approaches. We also explore the two primary methodological and theoretical paths research has taken in the past decade: temporal reasoning and temporal data maintenance. Both of these research areas include efforts to model time, temporal entities, and temporal queries. Collaboration between the two areas is possible, through tasks such as the abstraction of raw time-oriented clinical data into higher-level meaningful clinical concepts and the management of different levels of temporal granularity. Such collaboration could provide a common ground and useful areas for future research and development. We conclude with our view of future research directions.

Knowledge-based visualization of time-oriented clinical data

We describe a domain-independent framework (KNAVE) specific to the task of interpretation, summarization, visualization, explanation, and interactive exploration in a context-sensitive manner through time-oriented raw clinical data and the multiple levels of higher-level, interval-based concepts that can be abstracted from these data. The KNAVE exploration operators, which are independent of any particular clinical domain, access a knowledge base of temporal properties of measured data and interventions that is specific to the clinical domain. Thus, domain-specific knowledge underlies the domain-independent semantics of the interpretation, visualization, and exploration processes. Initial evaluation of the KNAVE prototype by a small number of users with variable clinical and informatics training has been encouraging.

Intention-based critiquing of guideline-oriented medical care

We present a methodology and tool for providing retrospective review and critiquing of guideline-based medical care given to patients. We show how our guideline representation language, Asbru, which supports the use of physicians intentions in addition to physician's actions, allows us to compare the care given to a patient at the level of the intention to treat in addition to the more detailed plan carried out. We have developed an algorithm based on this representation for retrospective quality assessment of guideline-based care. Our method takes the physician's and institution's preferences and policies into account in explaining or justifying physician deviations from the recommendations of a guideline.

Temporal reasoning and temporal data maintenance in medicine: issues and challenges

We present a brief, nonexhaustive overview of research efforts in designing and developing time-oriented systems in medicine. The growing volume of research on time-oriented systems in medicine can be viewed from either an application point of view, focusing on different generic tasks (e.g. diagnosis) and clinical areas (e.g. cardiology), or from a methodological point of view, distinguishing between different theoretical approaches. In this overview, we focus on highlighting methodological and theoretical choices, and conclude with suggestions for new research directions. Two main research directions can be noted: temporal reasoning, which supports various temporal inference tasks (e.g. temporal abstraction, time-oriented decision support, forecasting, data validation), and temporal data maintenance, which deals with storage and retrieval of data that have heterogeneous temporal dimensions. Efforts common to both research areas include the modeling of time, of temporal entities, and of temporal queries. We suggest that tasks such as abstraction of time-oriented data and the handling of different temporal-granularity levels should provide common ground for collaboration between the two research directions and fruitful areas for future research.

Development of a change model for a controlled medical vocabulary

Managing change in controlled medical vocabularies is labor intensive and costly, but change is inevitable if vocabularies are to be kept up to date. The changes that are appropriate for a controlled medical vocabulary depend on the data stored for that vocabulary, and those data in turn depend on the needs of users. The set of change operations is the change model; the data stored about concepts comprise the concept model. Because the change model depends directly on the concept model, a discussion of the former necessitates a discussion of the latter. In this paper, we first present a set of tasks that we believe controlled medical vocabularies should handle. Next, we describe our concept model for a controlled medical vocabulary. Then, we review the literature on changes in existing vocabulary systems. Finally, we present our change model. We call our system, which incorporates the concept model and change model, the General Online Dictionary of Medicine (GOLDMINE).

A temporal database mediator for protocol-based decision support

To meet the data-processing requirements for protocol-based decision support, a clinical data-management system must be capable of creating high-level summaries of time-oriented patient data, and of retrieving those summaries in a temporally meaningful fashion. We previously described a temporal-abstraction module (RESUME) and a temporal-querying module (Chronus) that can be used together to perform these tasks. These modules had to be coordinated by individual applications, however, to resolve the temporal queries of protocol planners. In this paper, we present a new module that integrates the previous two modules and that provides for their coordination automatically. The new module can be used as a standalone system for retrieving both primitive and abstracted time-oriented data, or can be embedded in a larger computational framework for protocol-based reasoning.

Knowledge-based temporal abstraction in clinical domains

We have defined a knowledge-based framework for the creation of abstract, interval-based concepts from time-stamped clinical data, the knowledge-based temporal-abstraction (KBTA) method. The KBTA method decomposes its task into five subtasks; for each subtask we propose a formal solving mechanism. Our framework emphasizes explicit representation of knowledge required for abstraction of time-oriented clinical data, and facilitates its acquisition, maintenance, reuse and sharing. The RESUME system implements the KBTA method. We tested RESUME in several clinical-monitoring domains, including the domain of monitoring patients who have insulin-dependent diabetes. We acquired from a diabetes-therapy expert diabetes-therapy temporal-abstraction knowledge. Two diabetes-therapy experts (including the first one) created temporal abstractions from about 800 points of diabetic-patients' data. RESUME generated about 80% of the abstractions agreed by both experts; about 97% of the generated abstractions were valid. We discuss the advantages and limitations of the current architecture.

An intention-based language for representing clinical guidelines

Automated support for guideline-based care would be enhanced considerably by a standard representation of clinical guidelines. To faciliate use and reuse, we suggest a representation that includes the explicit intentions of the guideline's author. These intentions include the desirable actions of the care provider and the patient states to be achieved before, during, and after the administration of the guideline. Intentions are temporal patterns of provider actions or patient states to be maintained, achieved, or avoided. We view automated support as a collaborative effort of the health-care provider and an automated assistant and involves several different tasks. We defined the syntax and, the semantics of a text-based language (ASBRU) for representation and annotation of clinical guidelines. The language supports maintenance of the automated assistant's knowledge base and could improve the quality and flexibility of the automated assistant's recommendations. In the ASGAARD project, we are developing reasoning mechanisms that use the ASBRU language for execution and critiquing tasks in conjunction with online electronic patient medical records.

Knowledge acquisition for temporal abstraction

Temporal abstraction is the task of detecting relevant patterns in data over time. The knowledge-based temporal-abstraction method uses knowledge about a clinical domain's contexts, external events, and parameters to create meaningful interval-based abstractions from raw time-stamped clinical data. In this paper, we describe the acquisition and maintenance of domain-specific temporal-abstraction knowledge. Using the PROTEGE-II framework, we have designed a graphical tool for acquiring temporal knowledge directly from expert physicians, maintaining the knowledge in a sharable form, and converting the knowledge into a suitable format for use by an appropriate problem-solving method. In initial tests, the tool offered significant gains in our ability to rapidly acquire temporal knowledge and to use that knowledge to perform automated temporal reasoning.

EON: a component-based approach to automation of protocol-directed therapy

Provision of automated support for planning protocol-directed therapy requires a computer program to take as input clinical data stored in an electronic patient-record system and to generate as output recommendations for therapeutic interventions and laboratory testing that are defined by applicable protocols. This paper presents a synthesis of research carried out at Stanford University to model the therapy-planning task and to demonstrate a component-based architecture for building protocol-based decision-support systems. We have constructed general-purpose software components that (1) interpret abstract protocol specifications to construct appropriate patient-specific treatment plans; (2) infer from time-stamped patient data higher-level, interval-based, abstract concepts; (3) perform time-oriented queries on a time-oriented patient database; and (4) allow acquisition and maintenance of protocol knowledge in a manner that facilitates efficient processing both by humans and by computers. We have implemented these components in a computer system known as EON. Each of the components has been developed, evaluated, and reported independently. We have evaluated the integration of the components as a composite architecture by implementing T-HELPER, a computer-based patient-record system that uses EON to offer advice regarding the management of patients who are following clinical trial protocols for AIDS or HIV infection. A test of the reuse of the software components in a different clinical domain demonstrated rapid development of a prototype application to support protocol-based care of patients who have breast cancer.

Ontology-based configuration of problem-solving methods and generation of knowledge-acquisition tools: application of PROTEGE-II to protocol-based decision support

PROTEGE-II is a suite of tools and a methodology for building knowledge-based systems and domain-specific knowledge-acquisition tools. In this paper, we show how PROTEGE-II can be applied to the task of providing protocol-based decision support in the domain of treating HIV-infected patients. To apply PROTEGE-II, (1) we construct a decomposable problem-solving method called episodic skeletal-plan refinement, (2) we build an application ontology that consists of the terms and relations in the domain, and of method-specific distinctions not already captured in the domain terms, and (3) we specify mapping relations that link terms from the application ontology to the domain-independent terms used in the problem-solving method. From the application ontology, we automatically generate a domain-specific knowledge-acquisition tool that is custom-tailored for the application. The knowledge-acquisition tool is used for the creation and maintenance of domain knowledge used by the problem-solving method. The general goal of the PROTEGE-II approach is to produce systems and components that are reusable and easily maintained. This is the rationale for constructing ontologies and problem-solving methods that can be composed from a set of smaller-grained methods and mechanisms. This is also why we tightly couple the knowledge-acquisition tools to the application ontology that specifies the domain terms used in the problem-solving systems. Although our evaluation is still preliminary, for the application task of providing protocol-based decision support, we show that these goals of reusability and easy maintenance can be achieved. We discuss design decisions and the tradeoffs that have to be made in the development of the system.

A temporal-abstraction mediator for protocol-based decision-support systems

The inability of many clinical decision-support applications to integrate with existing databases limits the wide-scale deployment of such systems. To overcome this obstacle, we have designed a data-interpretation module that can be embedded in a general architecture for protocol-based reasoning and that can support the fundamental task of detecting temporal abstractions. We have developed this software module by coupling two existing systems--RESUME and Chronus--that provide complementary temporal-abstraction techniques at the application and the database levels, respectively. Their encapsulation into a single module thus can resolve the temporal queries of protocol planners with the domain-specific knowledge needed for the temporal-abstraction task and with primary time-stamped data stored in autonomous clinical databases. We show that other computer methods for the detection of temporal abstractions do not scale up to the data- and knowledge-intensive environments of protocol-based decision-support systems.

Knowledge-based temporal abstraction for diabetic monitoring

We have developed a general method that solves the task of creating abstract, interval-based concepts from time-stamped clinical data. We refer to this method as knowledge-based temporal-abstraction (KBTA). In this paper, we focus on the knowledge representation, acquisition, maintenance, reuse and sharing aspects of the KBTA method. We describe five problem-solving mechanisms that solve the five subtasks into which the KBTA method decomposes its task, and four types of knowledge necessary for instantiating these mechanisms in a particular domain. We present an example of instantiating the KBTA method in the clinical area of monitoring insulin-dependent-diabetes patients.

RESUME: a temporal-abstraction system for patient monitoring

RESUME is a system that performs temporal abstraction of time-stamped data. The temporal-abstraction task is crucial for planning treatment, for executing treatment plans, for identifying clinical problems, and for revising treatment plans. The RESUME system is based on a model of three basic temporal-abstraction mechanisms: point temporal abstraction, a mechanism for abstracting the values of several parameters into a value of another parameter; temporal inference, a mechanism for inferring sound logical conclusions over a single interval or two meeting intervals; and temporal interpolation, a mechanism for bridging nonmeeting temporal intervals. Making explicit the knowledge required for temporal abstraction supports the acquisition and the sharing of that knowledge. We have implemented the RESUME system using the CLIPS knowledge-representation shell. The RESUME system emphasizes the need for explicit representation of temporal-abstraction knowledge, and the advantages of modular, task-specific but domain-independent architectures for building medical knowledge-based systems.

Knowledge reuse: temporal-abstraction mechanisms for the assessment of children's growth

Currently, many workers in the field of medical informatics realize the importance of knowledge reuse. The PROTEGE-II project seeks to develop and implement a domain-independent framework that allows system builders to create custom-tailored role-limiting methods from generic reusable components. These new role-limiting methods are used to create domain- and task-specific knowledge-acquisition tools with which an application expert can generate domain- and task-specific decision-support systems. One required set of reusable components embodies the problem-solving knowledge to generate temporal abstractions. Previously, members of the PROTEGE-II project have used these temporal-abstraction mechanisms to infer the presence of myelotoxicity in patients with AIDS. In this paper, we show that these mechanisms are reusable in the domain of assessment of children's growth.

A temporal-abstraction system for patient monitoring

RESUME is a system that performs temporal abstraction of time-stamped data. RESUME is based on a model of three temporal-abstraction mechanisms: point temporal abstraction (a mechanism for abstracting values of several parameters into a value of another parameter); temporal inference (a mechanism for inferring sound logical conclusions over a single interval or two meeting intervals); and temporal interpolation (a mechanism for bridging nonmeeting temporal intervals). Making explicit the knowledge required for temporal abstraction supports the acquisition of that knowledge.

Changes in smoking patterns in young military recruits in relationship to psychosocial characteristics

We followed 295 young infantry recruits during their first 14 weeks of basic training. The prevalence of smoking increased by 50%. About half of this increase was accounted for by ex-smokers, 57% of whom had resumed the habit. Average education and military psychometric measures of both the baseline smokers and the new smokers were significantly lower than those of the abstaining never-smokers. Asian and North African origin and a lower peer group evaluation score were also risk factors. These relationships were not demonstrated among resuming ex-smokers. The rise in the smoking rate accounts for most of the known rise during full military service. We suggest early preventive measures, especially for the two groups at risk.

Temporal-abstraction mechanisms in management of clinical protocols

We have identified several general temporal-abstraction mechanisms needed for reasoning about time-stamped data, such as are needed in management of patients being treated on clinical protocols: simple temporal abstraction (a mechanism for abstracting several parameter values into one class), temporal inference (a mechanism for inferring sound logical conclusions over a single interval or two meeting intervals), and temporal interpolation (a mechanism for bridging non-meeting temporal intervals). Making explicit the knowledge required for temporal abstractions supports the acquisition of planning knowledge, the identification of clinical problems, and the formulation of clinical-management-plan revisions.