Time components of circulatory transport from the lungs to a peripheral artery in humans.

Blood gas changes occurring in the lung undergo delay and damping on their way to a peripheral artery sampling site. Knowledge of the time components of circulatory transfer is important for the understanding of respiratory control and cardiovascular reflexes in response to blood gas transients. Providing steady state with regard to VA/Q distribution, cardiac output and peripheral blood flow, the relationship between the time courses of small end-tidal and peripheral PO2 changes is determined by the transfer function of the interposed vascular segment. This transfer function, expressed as delay time TD and mean transit time (MTT), was measured in six well-trained subjects, allowing the calculation of arterial time-courses from end-tidal to the reverse. They were studied at rest and during four different dynamic leg exercise intensities in the supine posture. TD and MTT amounted to 15.8 +/- 1.7 (mean +/- SEM) and 18.3 +/- 2.1 s at rest and were shortened to 7.7 +/- 0.6 and 11.5 +/- 1.8 s during exercise at 170 W. The shortening of TD and MTT did not appear to be simply an inverse function of cardiac output, suggesting that the shortening occurs in the central circulatory segment but not in the arm segment.

A study of the usage of a decision-support system for infective endocarditis.

The objective of this study was to examine a design for a World Wide Web-based decision-support system in use by clinically active physicians. A prototype implementation of the design concerned management of infective endocarditis patient cases. The design was based on an integration of hypertext and rule-based knowledge. In the study sessions, physicians in the field of internal medicine worked on managing authentic patient cases in a laboratory setting. Data was collected from interviews with the physicians using video recordings and stimulated recall technique. The qualitative data was analysed according to the constant comparative method in order to develop a model of the physicians' usage of the system. The resulting model describes perceived contributions and criteria for usefulness of the system. The ways the physicians used the system showed that it was able to provide patient-specific support for confirming clinical decisions, for higher-level patient management, and for preparing for and initiating expert consultations. Users also stated that new medical knowledge could be gained as a side effect of using the system.

Assessing an AI knowledge-base for asymptomatic liver diseases.

Discovering not yet seen knowledge from clinical data is of importance in the field of asymptomatic liver diseases. Avoidance of liver biopsy which is used as the ultimate confirmation of diagnosis by making the decision based on relevant laboratory findings only, would be considered an essential support. The system based on Quinlan's ID3 algorithm was simple and efficient in extracting the sought knowledge. Basic principles of applying the AI systems are therefore described and complemented with medical evaluation. Some of the diagnostic rules were found to be useful as decision algorithms i.e. they could be directly applied in clinical work and made a part of the knowledge-base of the Liver Guide, an automated decision support system.

Design and application of a terminology management system.

A Swedish data model for handling terminology, Spriterm, is presented in this paper. A prototype terminology management system, using the Spriterm data model in also described. This prototype is implemented is Microsoft ACCESS. Furthermore, two other applications using this prototype as a base are introduced. One World Wide Web based application, and a data dictionary.

Semantic modeling of a traditional classification: results and implications.

A primary health care version of the International Statistical Classification of Diseases and Related Health Problems, Tenth Revision (ICD-10), together with a three-dimensional model for classification of diseases according to location, origin, and type has been semantically represented. The resulting computer-based version is made available via the World Wide Web.

Individuals living in areas with high background radon: a GIS method to identify populations at risk.

OBJECTIVE: to identify and link populations and individuals that live within high risk areas. DESIGN: census registers and disease registers which contain data on individuals can only give aggregate statistics relating to postal code districts, town, county or state boundaries. However environmental risk factors rarely, if ever, respect these man-made boundaries. What is needed is a method to rapidly identify individuals who may live within a described area or region and to further identify the disease(s) occurring among these individuals and/or in these areas. METHOD: this paper describes a method for linking the standard registers available in Sweden, notably the residence-property addresses they contain and the geographical coordinate setting of these, to map the population as a point coverage. Using standard GIS methods this coverage could be linked, merged or intersected with any other map to create new subsets of population. Representation of populations down to the individual level by automatised spatialisation of available census data is in its simplicity a new informatics method which in the designated GIS medium adds a new power of resolution. RESULTS: We demonstrate this using the radon maps provided by the local communes. The Swedish annual population registration records of 1991 for the county of Ostergötland and the property register available at the Central Statistical Bureau of Sweden formed the main data sources. By coupling the address in the population register to the property register each individual was mapped to the centroid of a property. By intersecting the population coverage with the radon maps, the population living in high, normal or low risk areas was identified and then analysed and stratified by commune, sex and age. The resulting tables can be linked to other database registers, to visualise and analyse geographical and related patterns. The methodology can be adapted for use with any other environmental map or small area. It can also be expanded to the fourth dimension by linking likewise available migration information to generate immediately coordinate-set, accumulated exposition and similar data.

Bhorugram (India): revisited. A 4 year follow-up of a computer-based information system for distributed MCH services.

Impact of a health informatics research project on the management and delivery of primary health care, focusing on maternal and child health care (MCH), was reported earlier in 1992, with encouraging results in terms of increased antenatal care coverage and decrease in Immunisation drop-outs. The software (MCHS) was based on essential data sets (EDS) to assist the care providers in information management for MCH care. The site was revisited after four years to evaluate the status of the project and to ascertain whether computers were still in use and was the dispensary staff using the MCHS package or not? The results are presented in this paper to demonstrate that 'information is a difference that makes a difference'. Information generated by the MCHS acted as a catalyst for behavioural change in the community from indifferent users to active users of the health care services. This is reflected by the increased acceptance of antenatal care, 80% fully immunised child (FIC) and participation in other primary health care activities.

Extended telemedical consultation using Arden Syntax based decision support, hypertext and WWW technique.

There is an obvious need for geographic distribution of expert knowledge among several health care units without increasing the cost of on-site expertise in locations where health care is provided. This paper describes the design of a knowledge-based decision-support system for extended consultation in clinical medicine. The system is based on Arden Syntax for Medical Logic Modules and hypertext using World Wide Web technology. It provides advice and explanations regarding the given advice. The explanations are presented in a hypertext format allowing the user to browse related information and to verify the relevance of the given advice. The system is intended to be used in a closed local network. With special precautions regarding issues of safety and patient security, the system can be used over wider areas such as in rural medicine. A prototype has been developed in the field of clinical microbiology and infectious diseases regarding infective endocarditis.

A qualitative study of clinicians ways of using a decision-support system.

We have studied how clinicians approached a decision-support system to manage patient cases. The design of the system under study was based on an integration of hypertext and rule-based systems. World-Wide Web technology was used for the implementation of the system. By using grounded theory and stimulated recall, we found that getting patient-specific support and continuing medical education were the two major usages of the system and that the three parameters relevance, validity, and work were important in describing how the system was experienced by the users.

Machine learning for the quality of life in inflammatory bowel disease.

Presence of a chronic disease influences patients' lives and reinforces demands to accept and then cope with the illness. In the case of inflammatory bowel disease, quality of life greatly differs through phases of remissions and relapses. Could the quality of life questionnaire tell the difference? In this study we are disclosing possibilities of assessing patients' perspectives by analysing analogue scale statements regarding concerns and worries related to ulcerative colitis. Some two hundred Swedish patients, 3/4 in remission and 1/4 in relapse, filled out a booklet containing 36 statements. To characterise the disease activity, we have used multivariate discrimination. To structure and describe in details paths distinguishing the remission from relapse, we have used an artificial intelligence procedure. Applications of the CART (Classification And Regression Trees) algorithm resulted in a set of classifiers which are, based on the similar subsets of significant variables, i.e. statements. Best reached classification accuracy did not exceed 80% in any case. Other classifiers namely, K-nearest-neighbour (KNN), Learning Vector Quantization (LVQ) and Back Propagation Neural Network (BPNN) confirmed that outcome. An expectation that the disease activity should clearly speak throughout the questionnaire held for a certain number of the observations such as pain and suffering, loss of bowel control, dying early, feeling alone, ability to have children, being treated as different and concerns regarding the medication. To highlight the difference of incorrect 20%, K-means clustering was performed. The results settled a basis for a hypothesis that the studied quality of life instrument captures more than the disease activity.

Primary care informatics: Bhorugram, India: revisited.

A Primary Health Care Informatics project was initiated in 1990 at Bhorugram, Rajasthan, India. It was reported in MIE91 with initial, encouraging results. The evaluation of the information system has been successful in all operational terms, and the Fully Immunised Child (FIC) could be used as a local cardinal index to monitor and evaluate the quality of care, cost-effectiveness and community participation. Health Systems and Health Economy Research alike have long searched for apt quality and performance markers that in themselves could embrace and express a number of aspects and factors. The FIC index is shown to well meet the criteria of such a cardinal measure in the MCHC services. We, also, demonstrate an implementation strategy for establishing Information Systems in Primary Health Care Centres in Developing countries.

Database and knowledge base integration--a data mapping method for Arden Syntax knowledge modules.

One of the most important categories of decision-support systems in medicine are data driven systems where the inference engine is linked to a database. It is, therefore, important to find methods that facilitate the implementation of database queries referred to in the knowledge modules. A method is described for linking clinical databases to a knowledge base with Arden Syntax modules. The method is based on a query meta-database including templates for SQL queries which is maintained by a database administrator. During knowledge module authoring the medical expert refers only to a code in the query meta-database; no knowledge is needed about the database model or the naming of attributes and relations. The method uses standard tools, such as C+2 and ODBC, which makes it possible to implement the method at many platforms and to link to different clinical databases in a standardized way.

Terminology support for development of sharable knowledge modules.

Lack of an agreed infrastructure for terminology is identified as one of the major barriers to interchange of knowledge modules and integration of knowledge bases with other clinical information systems. The goal of the GALEN project is to bridge this gap between different terminology systems through the construction of a terminology server, which is based on a rich conceptual model with mapping facilities to natural language expressions and coding schemas. The long term goal is to support communication between medical information systems. Arden Syntax is a standard format for the creation of knowledge modules, with sharability as one of the main objectives. Since Arden Syntax is based on a data-driven approach, the data items used need to be adapted to locally available terminology. The GALEN approach appears to be complementary to Arden Syntax and to the development of sharable knowledge modules. The major theme of this paper is utilization of the GALEN terminology server for knowledge module authoring. Two systems are presented, a knowledge base manager and a client to the terminology server, allowing the user to navigate in the semantic network and to import concept definitions and terms into the knowledge modules. The benefit of the terminology server, allowing the user to navigate in the semantic network and to import concept definitions and terms into the knowledge modules. The benefit of the terminology services is discussed.

A design for a World Wide Web decision-support system using a controlled medical terminology.

We have developed a medical decision-support system based on Arden Syntax for Medical Logic Modules and hypertext using the World Wide Web. The two representations are integrated to provide a better platform for decision support. To manage the integration a controlled medical terminology has been used as a well-defined interface between the representations. The terminology is also used to facilitate communication between specialists and non-specialists.

Database and knowledge base integration in decision support systems.

Since decision support systems (DSS) in medicine often are linked to clinical databases it is important to find methods that facilitate the work for DSS developers to implement database queries in the knowledge base (KB). This paper presents a method for linking clinical databases to a KB with Arden Syntax modules. The method is based on a query meta database including templates for SQL queries. During knowledge module authoring the medical expert only refers to a code in the query meta database. Our method uses standard tools so it can be implemented on different platforms and linked to different clinical databases.

Distance from the Primary Health Center: a GIS method to study geographical access to health care.

Aerial distance from the dependent Primary Health Center is a crude but objective measure of geographical accessibility to Primary Health care facilities. This report describes a method for calculation of distances between the PHC and the population it serves using the data available from the local health authorities and the Swedish Central Statistical Bureau. The Swedish annual population registration records of 1991 and the property register available with the Central Statistical Bureau of Sweden formed the main data sources. By coupling the address in the population register to the property register each individual was mapped to the centroid of a property. The location of the PHCs as well as the areas covered by each were obtained from the local health authorities and mapped. By intersecting the population coverage with the PHC coverage the population for each PHC area was identified. Subsequently the distance to the PHC was calculated for each individual (property centroid). The population maps so generated can be linked to other databases to visualize and analyze the spatial dimension of health and disease. The methodology can be adapted for use with postal code districts, census enumeration tracts, or any other small area.

Evaluation of a knowledge-based decision-support system for ventilator therapy management.

Evaluation of knowledge-based systems differs from that of conventional systems in terms of verification and validation techniques. Furthermore, evaluating medical decision-support systems is difficult because the field is thus far comparatively unexplored. This paper presents an evaluation of a medical knowledge-based system called VentEx that supports decision-making in the management of ventilator therapy. Real patient data from 1300 hours of patient care involving 12 patients with 6 diagnoses are used to validate the knowledge base. The results range from 4.5% to 15.6% disagreement between the setting recommendations produced by VentEx and a gold standard, and 22.2% disagreement for recommendations for weaning. A comparison between the standard and two physicians showed that VentEx produced advice of the same quality as the physicians.

Arden syntax and GALEN terminology support: a powerful combination to represent medical knowledge.

Arden syntax is an ASTM standard defined a few years ago to represent rule-based knowledge in the medical domain in an application- and implementation-independent way. Knowledge is expressed as Medical Logic Modules (MLMs). So far, it has mainly been used for providing critique and reminders in data driven, medical decision support systems, where typically rules are evoked by certain database operations (e.g., "store systolic blood pressure"). The syntax is an important step towards a general medical knowledge representation, with version control, maintenance information, arithmetic operations, logic rule syntax, and temporal operations, but it is not yet complete. The standard does not define how medical concepts (e.g., "systolic blood pressure") are to be referenced. Existing coding schemes such as SNOMED and ICD10 may be used for this, but they have generally failed to provide the required level of detail. The use of Arden syntax to transfer knowledge between centers in the form of MLMs has been shown to be non-trivial [1,2]. GALEN (Generalized Architecture for Languages, Encyclopedias, and Nomenclatures in Medicine) is an EC AIM project aiming at a machine-readable system to represent medical concepts in a semantic network. The network allows for the combination and specialization of concepts to the required level of detail. The project has resulted in a notation (GRAIL: GALEN Representation and Integration Language) and several tools, such as browsers for manipulating medical concepts. In a GALEN subproject, we are exploring the potential utility of combining Arden syntax with semantics from the GALEN terminology, and we are developing tools to facilitate the creation of MLMs with terminology support. Our impression is that the combination is very useful: GALEN tools allow a domain expert to express her knowledge in purely medical terms, independent of an actual or intended patient database structure and data representation. They can also help a knowledge engineer or quite possibly an automatic interpreter to translate a well defined medical concept (expressed in GRAIL) to actual database queries (in SQL or other notation). This translation can be facilitated by a Data Dictionary describing the local patient database.