SEARCH · Search PubMed
Search indexed PubMed citations on genomics, clinical trials, systematic reviews and public health. Explore titles, authors and supplied subject terms, then open the PubMed record.
Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.
We have presented the management of a case of benign lipoblastomatosis in the face. Knowledge of this rare condition will enable the surgeon to make intelligent intraoperative decisions when confronted with similar findings.
1. The Occupational Safety and Health Administration (OSHA) gives the OR nurse the responsibility of selecting the appropriate product required, based on the nurse's assessment of "the task and degree of exposure anticipated." 2. Currently no reliable standards exist for judging materials/gowns. 3. Nurses should be aware of and dispel misconceptions regarding manufacturers' claims, so that they can make an intelligent, informed decision in selecting and/or assessing the use of an appropriate product.
Explore the source record for details and available documents.
PRETRIEVE is a belief-network-based, unsolicited information-retrieval system that performs machine learning based on user feedback. We report here on the document-ordering and document-retrieval performance of PRETRIEVE. We developed a test collection of 410 judgments of document utility in a simulated medical order-entry context. We characterized the validity of these judgments, which were elicited from domain experts, by measuring interrater and intrarater reproducibility. We developed a measure of the quality of document orderings similar to the ROC-curve analysis used to evaluate document-retrieval systems. We found that the ordering performance of the PRETRIEVE system was (1) substantially better than random, (2) somewhat less than ideal, and (3) superior to that of versions of the PRETRIEVE system that used relevance feedback instead of utility feedback. Under a set of assumptions, which we make explicit, we found that the documents retrieved by a version of PRETRIEVE that modeled time cost were of higher utility than those retrieved by a similar rule-based system.
This paper describes an integrated system in Prolog that permits the creation of a personal Knowledge Base to express and formalize specialist knowledge in medicine. Formalisms used are production rules and frames. The integrated system is able to manage data and knowledge stored in a database built in M Technology (MUMPS).
We have created a tool that allows users unfamiliar with the Arden Syntax and our underlying database to create Medical Logic Modules (MLMs). In a study of this tool (N 16), subjects found it easy to use (mean score - 4.69 on a scale of 1-5, 5 being best). Each subject created 3 MLMs of varying complexity following a protocol. On average, subjects required 312, 308 and 318 seconds, respectively, to complete each MLM. Comparison of clinicians to non-clinicians and those with to those without knowledge of Arden showed no significant difference. Of the 48 MLMs, 47 compiled and executed with appropriate output. Independent manual review of the MLM correlated well and found few errors. We conclude that our tool is easily used by inexperienced persons to write MLMs in the Arden Syntax.
This article describes a system for the optimization of the prehospital assessment of emergency, in cases involving thoraco-abdominal complaints and consciousness problems. This assessment is performed by nurses on the basis of a telephone interview at ambulance dispatch centers. The system has a body of biomedical and policy knowledge available to guide the interview and to provide advice.
Each science is identified with the questions it raises with respect to its object of study. This paper discussed the formulation of the basic questions of medical informatics. From a historical point of view, it first dealt with the problems of medical computing. Thereby, three classical questions arose: How can existing computers and information technologies assist in medical activities? Which components of the mathematical apparatus of informatics can be used for solving medical problems, and how and what activities of a physician are subject to algoritmization? The present time raises a new circle of questions centered around the basic one: How is the information system of the human organism structured and how does it function? This question and others form the basis of a new trend in medical informatics.
One way to fulfill point-of-care knowledge needs is to present caregivers with a visual representation of the available "answers". Using such a representation, caregivers can recognize what they want, rather than have to recall what they need, and then navigate to an appropriate answer. Given selected pieces of information from a computer-based patient record, an interface can anticipate certain knowledge needs by initializing caregiver navigation in a semantic neighborhood of answers likely to be relevant to the patient at hand. These notions draw heavily on two collaborative projects--the U.S. National Library of Medicine Unified Medical Language System and the U.S. National Cancer Institute Knowledge Server. Both of these projects support navigation because they make the structure of medical knowledge explicit in a way that can be exploited by human interfaces.
In this paper, we perform a cognitive analysis of knowledge discovery processes. As a result of this analysis, the construction-integration theory is proposed as a general framework for developing cooperative knowledge evolution systems. We thus suggest that for the acquisition of new domain knowledge in medicine, one should first construct pluralistic views on a given topic which may contain inconsistencies as well as redundancies. Only thereafter does this knowledge become consolidated into a situation-specific circumscription and the early inconsistencies become eliminated. As a proof for the viability of such knowledge acquisition processes in medicine, we present the IDEAS system, which can be used for the intelligent documentation of adverse events in clinical studies. This system provides a better documentation of the side-effects of medical drugs. Thereby, knowledge evolution occurs by achieving consistent explanations in increasingly larger contexts (i.e., more cases and more pharmaceutical substrates). Finally, it is shown how prototypes, model-based approaches and cooperative knowledge evolution systems can be distinguished as different classes of knowledge-based systems.
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.
The introduction of intelligent robots, expert systems and other forms of intelligent automatization in the current practice of medicine seems to be inevitable. It appears interesting to look back to the efforts that have been done, since the former steps, about three decades ago and consider the prospects in this field for both short and long term. Simultaneously it is interesting to reckon the new aspects which are raised with the evolution of these methodologies such as the responsibility of decisions taken by intelligent systems, the probable advantages, at the present stage, of the interactive systems and the risk of self-learning systems. Some efforts carried out in our department in this field are described.
This paper partly addresses the question "What artificial intelligence (AI) tools are appropriate for which parts of pathology?" by analyzing the structure and components of knowledge in pathology (e.g., observations plus archival and reference data) and which aspects of that knowledge should be expressible in an AI consulting system. The different aspects of uncertainty (observational, prevalence and validity) play an important role in both human and computer-based decision-making processes, as do relationships between the components of knowledge. The design of an AI consultant system is discussed in terms of the way uncertainty is expressed and in how many parameters, the way uncertainty is propagated (Bayes, certainty factors, Dempster-Schafer, logic or Pathfinder heuristic methods), whether the system reasons from data to a conclusion or vice versa and what the aim of the system is. The suitability of an AI tool is determined by the knowable facts of the pathology subfield, by the match with its knowledge structure and by its requirements. While the success of an AI tool will partly depend on an appropriate definition of its scope, the appropriate combinatoric also depends on the expertise of the user.
Clinical decision analysis seeks to identify the optimal management strategy by modelling the uncertainty and risks entailed in the diagnosis, natural history, and treatment of a particular problem or disorder. Decision trees are the most frequently used model in clinical decision analysis, but can be tedious to construct, cumbersome to use, and computationally prohibitive, especially with large, complex decision problems. We present a new method for clinical decision analysis that combines the techniques of decision theory and artificial intelligence. Our model uses a modular representation of knowledge that simplifies model building and enables more fully automated decision making. Moreover, the model exploits problem structures to yield better computational efficiency. As an example we apply our techniques to the problem of management of acute deep venous thrombosis.
We developed an artificial intelligence program from a large computerized database of hemodynamic and oxygen transport measurements together with prior studies defining survivors' values, outcome predictors, and a branched-chain decision tree. The artificial intelligence program was then tested on the data of 100 survivors and 100 nonsurvivors not used for the development of the program or other analyses. Using the predictor as a surrogate outcome measure, the therapy recommended by the program improved the predicted outcome 3.16% per therapeutic intervention while the actual therapy given increased outcome 1.86% in surviving patients; the artificial intelligence-recommended therapy improved outcome 7.9% in nonsurvivors, while the actual therapy given increased predicted outcome -0.29% in nonsurvivors (p < .05). There were fewer patients whose predicted outcome decreased after recommended treatment (14%) than after the actual therapy given (37%). Review of therapy recommended by the program did not reveal instances of inappropriate or potentially harmful recommendations.
Nine-year-old Korean children (N = 299) were tested for reaction time (RT) and intelligence measured by Raven's Standard Progressive Matrices. Reaction times were broken into decision times and movement times and into three degrees of complexity. The results showed low but generally statistically significant correlations between decision times and intelligence. Generally, no significant differences existed between movement times and intelligence. Boys showed significantly faster movement times than girls did.
A sample of 150 children referred to Student Study Teams was assessed with a psychometric battery. Behavioral and academic ratings were obtained from teachers. Forty-three children scored at or below 75 on the WISC-III. We examined schools' subsequent classification decisions to ascertain how schools dealt with low-IQ students with academic and behavioral problems. Schools reached decisions regarding 35 of the children: Only 6 were classified as having mental retardation, and 18, as having learning disabilities. Findings were discussed in terms of (a) decline in identification rates of mild mental retardation and (b) the extent to which school decisions adhere to the research criteria.
Explore the source record for details and available documents.