A Comparative Study of the Arden Syntax and GDL Clinical Knowledge Representation Languages.

The expressiveness of a medical knowledge representation language has significant impact on the effectiveness of a knowledge-based clinical decision support system. We assess the expressiveness of two such languages, Arden Syntax and the Guideline Definition Language. Using data extracted from both languages' specifications, we quantify expressiveness by means of language syntax and the number of supported operators. Preliminary results show that Arden Syntax is a more dynamic standard, having better readability and a higher number and more diverse operators than GDL. In contrast, GDL is a more rigid language that utilizes an underlying data model specification in the openEHR framework.

A Systematic Approach to Quality Requirement Management in Medical Software.

BACKGROUND: The quality requirements for medical software have become increasingly demanding. Several quality standards and models are already in place, but there is a debate on whether these are specific enough for medical software. Moreover, mapping requirements to quality criteria can be challenging but is required throughout the software development process. OBJECTIVES: We propose a workflow in which we apply proven methods and tools for systematic collection, analysis, and evaluation of software quality criteria based on the ISO/IEC 25010:2011. METHODS: We employ affinity diagrams, Kano analysis and quality function deployment for the systematic requirement development, analysis, and management. RESULTS: We outline a systematic approach on how to use the recommended process when developing medical software. CONCLUSION: The paper proposes a systematic approach for requirements management that could be used for mapping medical software quality criteria and stakeholder requirements, independent from the quality criteria (and the underlying model) itself.

Exploring Methods of Implementing Arden Syntax for CDS Hooks.

BACKGROUND: Specifications for Arden Syntax lack provisions for the standardized access of clinical decision support (CDS) services. The CDS Hooks standard provides such access. OBJECTIVES: To extend an ArdenSuite reference implementation of the Arden Syntax by providing a CDS-Hooks-compatible interface. METHODS: With the use case Hepaxpert, an Arden-Syntax-based expert system for the interpretation of hepatitis serology test results, a needs analysis was performed to identify changes required in the ArdenSuite reference implementation to support the CDS Hooks API. Arden Syntax language support for CDS Hooks was also assessed. RESULTS: The needs assessment was performed in three phases: hook assessment, hook context definition, and Card definition. For the use case, the ArdenSuite was modified to include a new hook and hook context, which defines the type of CDS service as well its input parameters. Card definitions were created in the ArdenSuite. Examples of Arden Syntax support for the use case are presented for all three phases. CONCLUSION: Minor changes in the ArdenSuite made it compatible with the CDS Hooks specification.

Separating Business Logic from Medical Knowledge in Digital Clinical Workflows Using Business Process Model and Notation and Arden Syntax.

BACKGROUND: Evidence-based clinical guidelines have a major positive effect on the physician's decision-making process. Computer-executable clinical guidelines allow for automated guideline marshalling during a clinical diagnostic process, thus improving the decision-making process. OBJECTIVES: Implementation of a digital clinical guideline for the prevention of mother-to-child transmission of hepatitis B as a computerized workflow, thereby separating business logic from medical knowledge and decision-making. METHODS: We used the Business Process Model and Notation language system Activiti for business logic and workflow modeling. Medical decision-making was performed by an Arden-Syntax-based medical rule engine, which is part of the ARDENSUITE software. RESULTS: We succeeded in creating an electronic clinical workflow for the prevention of mother-to-child transmission of hepatitis B, where institution-specific medical decision-making processes could be adapted without modifying the workflow business logic. CONCLUSION: Separation of business logic and medical decision-making results in more easily reusable electronic clinical workflows.

Assessing the feasibility of a mobile health-supported clinical decision support system for nutritional triage in oncology outpatients using Arden Syntax.

BACKGROUND: Nutritional screening procedures followed by regular nutrition monitoring for oncological outpatients are no standard practice in many European hospital wards and outpatient settings. As a result, early signs of malnutrition are missed and nutritional treatment is initiated when patients have already experienced severe weight loss. OBJECTIVE: We report on a novel clinical decision support system (CDSS) for the global assessment and nutritional triage of the nutritional condition of oncology outpatients. The system combines clinical and laboratory data collected in the clinical setting with patient-generated data from a smartphone application for monitoring the patients' nutritional status. Our objective is to assess the feasibility of a CDSS that combines the aforementioned data sources and describe its integration into a hospital information system. Furthermore, we collected patients' opinions on the value of the system, and whether they would regard the system as a useful aid in coping with their condition. MATERIALS AND METHODS: The system implements the Patient-Generated Subjective Global Assessment (PG-SGA) to monitor nutritional status in the outpatient setting. A smartphone application is used to collect patient-generated data by performing weekly mini-surveys on patients concerning their eating habits, weight, and overall well-being. Data are uploaded on completion of each mini-survey and stored on a secure server at the Medical University of Vienna (MUV). The data are then combined with relevant clinical information from the Vienna General Hospital (VGH) information system. The knowledge base for the CDSS is implemented in medical logic modules (MLMs) using Arden Syntax. A three-month pilot clinical trial was performed to test the feasibility of the system. Qualitative questionnaires were used to obtain the patients' opinions on the usability and personal value of the system during the four-week test period. RESULTS: We used the existing separation between the scientific and clinical data domains in the secured network environment (SNE) at the MUV and VGH to our advantage by importing, storing, and processing both patient-generated and routine data in the scientific data domain. To limit exposure to the SNE, patient-generated data stored outside the SNE were imported to the scientific domain once a day. The CDSS created for nutritional assessment and triage comprised ten MLMs, each including either a sub-assessment or the final results of the PG-SGA. Finally, an interface created for the hospital information system showed the results directly in clinical routine. In all 22 patients completed the clinical study. The results of the questionnaires showed that 91% of the patients were generally happy with the usability of the system, 91% believed that the application was of additional value in detecting cancer-related malnutrition, and 82% found it helpful as a long-term monitoring tool. DISCUSSION AND CONCLUSION: Despite strict protection of the clinical data domain, a CDSS employing patient-generated data can be integrated into clinical routine. The CDSS discussed in this report combined the information entered into a smartphone application with clinical data in order to inform the physician of a patient's nutritional status and thus permit suitable and timely intervention. The initial results show that the smartphone application was well accepted by patients, who considered it useful, but not many oncological outpatients were willing to participate in the clinical study because they did not possess an Android phone or lacked smartphone expertise. Furthermore, the results indicate that patient-generated data could be employed to augment clinical data and calculate metrics such as the PG-SGA without excessive effort by using a secure intermediate location as the locus of data storage and processing.

Clinical decision support systems at the Vienna General Hospital using Arden Syntax: Design, implementation, and integration.

INTRODUCTION: The Allgemeines Krankenhaus Informations Management (AKIM) project was started at the Vienna General Hospital (VGH) several years ago. This led to the introduction of a new hospital information system (HIS), and the installation of the expert system platform (EXP) for the integration of Arden-Syntax-based clinical decision support systems (CDSSs). In this report we take a look at the milestones achieved and the challenges faced in the creation and modification of CDSSs, and their integration into the HIS over the last three years. MATERIALS AND METHODS: We introduce a three-stage development method, which is followed in nearly all CDSS projects at the Medical University of Vienna and the VGH. Stage one comprises requirements engineering and system conception. Stage two focuses on the implementation and testing of the system. Finally, stage three describes the deployment and integration of the system in the VGH HIS. The HIS provides a clinical work environment for healthcare specialists using customizable graphical interfaces known as parametric medical documents. Multiple Arden Syntax servers are employed to host and execute the CDSS knowledge bases: two embedded in the EXP for production and development, and a further three in clinical routine for production, development, and quality assurance. RESULTS: Three systems are discussed; the systems serve different purposes in different clinical areas, but are all implemented with Arden Syntax. MONI-ICU is an automated surveillance system for monitoring healthcare-associated infections in the intensive care setting. TSM-CDS is a CDSS used for risk prediction in the formation of cutaneous melanoma metastases. Finally, TacroDS is a CDSS for the manipulation of dosages for tacrolimus, an immunosuppressive agent used after kidney transplantation. Problems in development and integration were related to data quality or availability, although organizational difficulties also caused delays in development and integration. DISCUSSION AND CONCLUSION: Since the inception of the AKIM project at the VGH and its ability to support standards such as Arden Syntax and integrate CDSSs into clinical routine, the clinicians' interest in, and demand for, decision support has increased substantially. The use of Arden Syntax as a standard for CDSSs played a substantial role in the ability to rapidly create high-quality CDSS systems, whereas the ability to integrate these systems into the HIS made CDSSs more popular among physicians. Despite these successes, challenges such as lack of (consistent and high-quality) electronic data, social acceptance among healthcare personnel, and legislative issues remain. These have to be addressed effectively before CDSSs can be more widely accepted and adopted.

Arden Syntax MLM Building Blocks for Microbiological Concepts and Their Application in Infection Surveillance.

BACKGROUND: The diagnosis - and hence definitions - of healthcare-associated infections (HAIs) rely on microbiological laboratory test results in specific constellations. OBJECTIVES: To construct a library that provides interoperable building blocks for the analysis of microbiological laboratory test results. METHODS: We used Java for preprocessing raw microbiological laboratory test results and Arden Syntax for knowledge-based querying of data based on microbiology information elements used in European surveillance criteria for HAIs. To test the library and quantify how often these information elements occur in the data, we performed a retrospective cohort study on adult patients admitted for at least 24 hours to an intensive care unit at the Vienna General Hospital in 2013. RESULTS: We identified eleven information elements for which information was electronically available. These elements were identified positively 1,239 times in 1,184 positive microbiology tests from 563 patients. DISCUSSION: The availability of a library for the analysis of microbiology laboratory test results in HAI terms facilitates electronic HAI surveillance.

Creating Clinical Fuzzy Automata with Fuzzy Arden Syntax.

Formal constructs for fuzzy sets and fuzzy logic are incorporated into Arden Syntax version 2.9 (Fuzzy Arden Syntax). With fuzzy sets, the relationships between measured or observed data and linguistic terms are expressed as degrees of compatibility that model the unsharpness of the boundaries of linguistic terms. Propositional uncertainty due to incomplete knowledge of relationships between clinical linguistic concepts is modeled with fuzzy logic. Fuzzy Arden Syntax also supports the construction of fuzzy state monitors. The latter are defined as monitors that employ fuzzy automata to observe gradual transitions between different stages of disease. As a use case, we re-implemented FuzzyARDS, a previously published clinical monitoring system for patients suffering from acute respiratory distress syndrome (ARDS). Using the re-implementation as an example, we show how key concepts of fuzzy automata, i.e., fuzzy states and parallel fuzzy state transitions, can be implemented in Fuzzy Arden Syntax. The results showed that fuzzy state monitors can be implemented in a straightforward manner.

Arden Syntax Clinical Foundation Framework for Event Monitoring in Intensive Care Units: Report on a Pilot Study.

The creation of clinical decision support systems has received a strong impulse over the last years, but their integration into a clinical routine has lagged behind, partly due to a lack of interoperability and trust by physicians. We report on the implementation of a clinical foundation framework in Arden Syntax, comprising knowledge units for (a) preprocessing raw clinical data, (b) the determination of single clinical concepts, and (c) more complex medical knowledge, which can be modeled through the composition and configuration of knowledge units in this framework. Thus, it can be tailored to clinical institutions or patients' caregivers. In the present version, we integrated knowledge units for several infection-related clinical concepts into the framework and developed a clinical event monitoring system over the framework that employs three different scenarios for monitoring clinical signs of bloodstream infection. The clinical event monitoring system was tested using data from intensive care units at Vienna General Hospital, Austria.

Adverse Drug Event Monitoring with Clinical and Laboratory Data Using Arden Syntax.

In times of steadily increasing numbers of administered drugs, the detection of adverse drug events (ADEs) is an important aspect of improving patient safety. At present only about 1-13% of detected ADEs are reported. Raising the number of reported ADEs will result in greater and more efficient support of pharmacovigilance. Potential ADE's must be identified early. In the iMedication system, which is a rule-based application, triggers are used for computerized detection of possible ADEs. Creating a pilot system, we defined the relevant use cases hyperkalemia, hyponatremia, renal failure, and over-anticoagulation; knowledge bases were implemented in Arden Syntax for each use case. The objective of these knowledge bases is to interpret patient-specific clinical data and generate notifications based on a calculated ADE risk score, which may indicate possible ADEs. This will permit appropriate monitoring of potential ADE situations over time in the interest of patient care, quality assurance, and pharmacovigilance.

Context-Sensitive Clinical Alert Packages Written in Arden Syntax.

An increasing body of raw patient data is generated on each day of a patient's stay at a hospital. It is of paramount importance that critical patient information be extracted from these large data volumes and presented to the patient's clinical caregivers as early as possible. Contemporary clinical alert systems attempt to provide this service with moderate success. The efficacy of the systems is limited by the fact that they are too general to fit specific patient populations or healthcare institutions. In this study we present an extendable alerting framework implemented in Arden Syntax, which can be configured to the needs and preferences of healthcare institutions and individual patient caregivers. We illustrate the potential of this alerting framework via an alert package that analyzes hematological laboratory results with data from intensive care units at the Vienna General Hospital, Austria. The results show the effectiveness of this alert package and its ability to generate key alerts while avoiding over-alerting.

Clinical Workflow Modeling in Obstetrics: Hepatitis B in Pregnancy.

Evidence-based clinical guidelines positively effect physician decision-making. Actionable clinical guidelines that actively trigger alerts, reminders, and instructive texts will increase effectiveness. We applied Activiti, a Business Process Model and Notation language system to model a clinical guideline for the prevention of mother-to-child transmission of hepatitis B as a computerized clinical workflow. Furthermore, we implemented an interconnected Arden-Syntax-based medical rule engine, which is part of the ARDENSUITE software.

Detecting borderline infection in an automated monitoring system for healthcare-associated infection using fuzzy logic.

BACKGROUND: Many electronic infection detection systems employ dichotomous classification methods, classifying patient data as pathological or normal with respect to one or several types of infection. An electronic monitoring and surveillance system for healthcare-associated infections (HAIs) known as Moni-ICU is being operated at the intensive care units (ICUs) of the Vienna General Hospital (VGH) in Austria. Instead of classifying patient data as pathological or normal, Moni-ICU introduces a third borderline class. Patient data classified as borderline with respect to an infection-related clinical concept or HAI surveillance definition signify that the data nearly or partly fulfill the definition for the respective concept or HAI, and are therefore neither fully pathological nor fully normal. OBJECTIVE: Using fuzzy sets and propositional fuzzy rules, we calculated how frequently patient data are classified as normal, borderline, or pathological with respect to infection-related clinical concepts and HAI definitions. In dichotomous classification methods, borderline classification results would be confounded by normal. Therefore, we also assessed whether the constructed fuzzy sets and rules employed by Moni-ICU classified patient data too often or too infrequently as borderline instead of normal. PARTICIPANTS AND METHODS: Electronic surveillance data were collected from adult patients (aged 18 years or older) at ten ICUs of the VGH. All adult patients admitted to these ICUs over a two-year period were reviewed. In all 5099 patient stays (4120 patients) comprising 49,394 patient days were evaluated. For classification, a part of Moni-ICU's knowledge base comprising fuzzy sets and rules for ten infection-related clinical concepts and four top-level HAI definitions was employed. Fuzzy sets were used for the classification of concepts directly related to patient data; fuzzy rules were employed for the classification of more abstract clinical concepts, and for top-level HAI surveillance definitions. Data for each clinical concept and HAI definition were classified as either normal, borderline, or pathological. For the assessment of fuzzy sets and rules, we compared how often a borderline value for a fuzzy set or rule would result in a borderline value versus a normal value for its associated HAI definition(s). The statistical significance of these comparisons was expressed in p-values calculated with Fisher's exact test. RESULTS: The results showed that, for clinical concepts represented by fuzzy sets, 1-17% of the data were classified as borderline. The number was substantially higher (20-81%) for fuzzy rules representing more abstract clinical concepts. A small body of data were found to be in the borderline range for the four top-level HAI definitions (0.02-2.35%). Seven of ten fuzzy sets and rules were associated significantly more often with borderline values than with normal values for their respective HAI definition(s) (p<0.001). CONCLUSION: The study showed that Moni-ICU was effective in classifying patient data as borderline for infection-related concepts and top-level HAI surveillance definitions.

Knowledge-based immunosuppressive therapy for kidney transplant patients--from theoretical model to clinical integration.

Immunosuppressive therapy is a risky necessity after a patient received a kidney transplant. To reduce risks, a knowledge-based system was developed that determines the right dosage of the immunosuppresive agent Tacrolimus. A theoretical model, to classify medication blood levels as well as medication adaptions, was created using data from almost 500 patients, and over 13.000 examinations. This model was then translated into an Arden Syntax knowledge base, and integrated directly into the hospital information system of the Vienna General Hospital. In this paper we give an overview of the construction and integration of such a system.

Advances In Infection Surveillance and Clinical Decision Support With Fuzzy Sets and Fuzzy Logic.

By the use of extended intelligent information technology tools for fully automated healthcare-associated infection (HAI) surveillance, clinicians can be informed and alerted about the emergence of infection-related conditions in their patients. Moni--a system for monitoring nosocomial infections in intensive care units for adult and neonatal patients--employs knowledge bases that were written with extensive use of fuzzy sets and fuzzy logic, allowing the inherent un-sharpness of clinical terms and the inherent uncertainty of clinical conclusions to be a part of Moni's output. Thus, linguistic as well as propositional uncertainty became a part of Moni, which can now report retrospectively on HAIs according to traditional crisp HAI surveillance definitions, as well as support clinical bedside work by more complex crisp and fuzzy alerts and reminders. This improved approach can bridge the gap between classical retrospective surveillance of HAIs and ongoing prospective clinical-decision-oriented HAI support.

Data use and effectiveness in electronic surveillance of healthcare associated infections in the 21st century: a systematic review.

OBJECTIVE: As more electronic health records have become available during the last decade, we aimed to uncover recent trends in use of electronically available patient data by electronic surveillance systems for healthcare associated infections (HAIs) and identify consequences for system effectiveness. METHODS: A systematic review of published literature evaluating electronic HAI surveillance systems was performed. The PubMed service was used to retrieve publications between January 2001 and December 2011. Studies were included in the review if they accurately described what electronic data were used and if system effectiveness was evaluated using sensitivity, specificity, positive predictive value, or negative predictive value. Trends were identified by analyzing changes in the number and types of electronic data sources used. RESULTS: 26 publications comprising discussions on 27 electronic systems met the eligibility criteria. Trend analysis showed that systems use an increasing number of data sources which are either medico-administrative or clinical and laboratory-based data. Trends on the use of individual types of electronic data confirmed the paramount role of microbiology data in HAI detection, but also showed increased use of biochemistry and pharmacy data, and the limited adoption of clinical data and physician narratives. System effectiveness assessments indicate that the use of heterogeneous data sources results in higher system sensitivity at the expense of specificity. CONCLUSIONS: Driven by the increased availability of electronic patient data, electronic HAI surveillance systems use more data, making systems more sensitive yet less specific, but also allow systems to be tailored to the needs of healthcare institutes' surveillance programs.

Validation of fuzzy sets in an automated detection system for intensive-care-unit-acquired central-venous-catheter-related infections.

Central venous catheters play an important role in patient care in intensive care units (ICUs), but their use comes at the risk of catheter-related infections (CRIs). Electronic surveillance systems can detect CRIs more accurately than manual surveillance, but these systems often omit patients that do not exhibit all infection signs to their full degree, the so-called borderline group. By extending an electronic surveillance system with fuzzy constructs, the borderline group can be identified. In this study, we examined the size of the borderline group for systemic CRIs (CRI2) by calculating the frequency of fuzzy values for CRI2 and related infection parameters in patient data involving ten ICUs (75 beds) over one year. We also validated the expert-defined fuzzy constructs by comparing overall and CRI2-specific support. The study showed that more than 86% of the data contained fuzzy values, and that the borderline group for CRI2 consisted of 2% of the study group. It was also confirmed that most fuzzy constructs were good representatives of the borderline CRI2 patient group.

Delivering antibiotic resistance information specifically tailored to location and time.

Antibiotic resistance poses a significant threat to humanity. Hundred years since the beginning of the era of antibacterial drugs, we are facing increasing numbers of infections with multi-resistant pathogens. The current approach of distributing information on antibiotic resistance in printed form in the clinics has disadvantages with respect to the actuality of the data and the regional heterogeneity of resistance patterns. We developed an application named qRe using representational state transfer as a communication standard to deliver antibiotic resistance percentage information to the end user. The data is selected specifically for his/her geographic location. The user can display the information using either the application for Android smart phones or the web application. With the presented software we show the technical feasibility of delivering antibiotic resistance information specifically tailored to location and time. A short evaluation of the software showed an overall positive response from physicians. Based on recommendations of previous investigations, we expect a measurable clinical impact.

Effectiveness of an automated surveillance system for intensive care unit-acquired infections.

This study assessed the effectiveness of a fully automated surveillance system for the detection of healthcare-associated infections (HCAIs) in intensive care units. Manual ward surveillance (MS) and electronic surveillance (ES) were performed for two intensive care units of the Vienna General Hospital. All patients admitted for a period longer than 48 h between 13 November 2006 and 7 February 2007 were evaluated according to HELICS-defined rules for HCAI. The sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV) and personnel time spent per surveillance type were calculated. Ninety-three patient admissions were observed, whereby 30 HCAI episodes were taken as a reference standard. Results with MS versus ES were: sensitivity 40% versus 87%, specificity 94% versus 99%, PPV 71% versus 96%, NPV 80% versus 95%, and time spent per surveillance type 82.5 h versus 12.5 h. In conclusion, ES was found to be more effective than MS while consuming fewer personnel resources.

Assessing the clinical uses of fuzzy detection results in the automated detection of CVC-related infections: a preliminary report.

Central venous catheters (CVCs) play an essential role in the care of the critically ill, but their use comes at the risk of infection. By using fuzzy set theory and logic to model clinical linguistic CVC-related infection criteria, clinical detection systems can detect borderline infections where not all infection parameters have been (fully) met, also called fuzzy results. In this paper we analyzed the clinical use of these results. We used a fuzzy-logic-based computerized infection control system for the monitoring of healthcare-associated infections to uncover fuzzy results and periods, after which we classified them, and used these classifications together with knowledge of prior CVC-related infection episodes in temporal association rule mining. As a result, we uncovered several rules which can help with the early detection of re-occurring CVC-related infections.