Clinical Decision Support Systems Systems Research Articles

Advancements in AI-driven diagnostic radiology: Enhancing accuracy and efficiency

Background: Healthcare delivery has transformed significantly with the integration of clinical decision support systems (CDS) and medical imaging. Convolutional neural networks (CNNs), a type of artificial intelligence (AI) algorithm, have exhibited remarkable accuracy in discerning intricate patterns and anomalies within medical images, surpassing human capability. Aim: This study aims to explore the impact of AI augmentation on diagnostic tasks, focusing on enhancing sensitivity, accuracy, and interrater agreement across various medical conditions. Additionally, it seeks to investigate how AI simplifies complex processes and integrates with existing technologies, extending its role in CDS systems beyond diagnostic accuracy. Methods: The research examines the effectiveness of AI in interpreting CT imaging and diagnosis. Furthermore, it assesses the integration of AI with radiology to enhance the detection of cerebral hemorrhages on head CT scans in time-pressed clinical settings. The research was performed using search engines such as google scholar and Pubmed. Results: The findings indicate that AI augmentation significantly enhances diagnostic capabilities, improves physician confidence, reduces interpretation time, and optimizes workflow efficiency. AI not only improves accuracy but also simplifies processes, thereby revolutionizing healthcare delivery. Conclusion: As artificial intelligence continues to evolve, its revolutionary potential in healthcare becomes increasingly evident.

How AI-Based Systems Can Induce Reflections: The Case of AI-Augmented Diagnostic Work

This paper addresses a thus-far neglected dimension in human-artificial intelligence (AI) augmentation: machine-induced reflections. By establishing a grounded theoretical-informed model of machine-induced reflection, we contribute to the ongoing discussion in information systems (IS) regarding AI and research on reflection theories. In our multistage study, physicians used a machine learning-based (ML) clinical decision support system (CDSS) to see if and how this interaction can stimulate reflective practice in the context of an X-ray diagnosis task. By analyzing verbal protocols, performance metrics, and survey data, we developed an integrative theoretical foundation to explain how ML-based systems can help stimulate reflective practice. Individuals engage in more critical or shallower modes depending on whether they perceive a conflict or agreement with these CDSS systems, which in turn leads to different levels of reflection depth. By uncovering the process of machine-induced reflections, we offer IS research a different perspective on how such AI-based systems can help individuals become more reflective, and consequently more effective, professionals. This perspective stands in stark contrast to the traditional, efficiency-focused view of ML-based decision support systems and also enriches theories on human-AI augmentation.

Improving computerized decision support system interventions: a qualitative study combining the theoretical domains framework with the GUIDES Checklist

BackgroundComputerized clinical decision support systems (CDSSs) can improve care by bridging knowledge to practice gaps. However, the real-world uptake of such systems in health care settings has been suboptimal. We sought to: (1) use the Theoretical Domains Framework (TDF) to identify determinants (barriers/enablers) of uptake of the Electronic Asthma Management System (eAMS) CDSS; (2) match identified TDF belief statements to elements in the Guideline Implementation with Decision Support (GUIDES) Checklist; and (3) explore the relationship between the TDF and GUIDES frameworks and the usefulness of this sequential approach for identifying opportunities to improve CDSS uptake.MethodsIn Phase 1, we conducted semistructured interviews with primary care physicians in Toronto, Canada regarding the uptake of the eAMS CDSS. Using content analysis, two coders independently analyzed interview transcripts guided by the TDF to generate themes representing barriers and enablers to CDSS uptake. In Phase 2, the same reviewers independently mapped each belief statement to a GUIDES domain and factor. We calculated the proportion of TDF belief statements that linked to each GUIDES domain and the proportion of TDF domains that linked to GUIDES factors (and vice-versa) and domains.ResultsWe interviewed 10 participants before data saturation. In Phase 1, we identified 53 belief statements covering 12 TDF domains; 18 (34.0%) were barriers, and 35 (66.0%) were enablers. In Phase 2, 41 statements (77.4%) linked to at least one GUIDES factor, while 12 (22.6%) did not link to any specific factor. The GUIDES Context Domain was linked to the largest number of belief statements (19/53; 35.8%). Each TDF domain linked to one or more GUIDES factor, with 6 TDF domains linking to more than 1 factor and 8 TDF domains linking to more than 1 GUIDES domain.ConclusionsThe TDF provides unique insights into barriers and enablers to CDSS uptake, which can then be mapped to GUIDES domains and factors to identify required changes to CDSS context, content, and system. This can be followed by conventional mapping of TDF domains to behaviour change techniques to optimize CDSS implementation. This novel step-wise approach combines two established frameworks to optimize CDSS interventions, and requires prospective validation.

Realizing the promise of big data: how Taiwan can help the world reduce medical errors and advance precision medicine

PurposeThe study explores how Taiwan’s electronic health data systems can be used to build algorithms that reduce or eliminate medical errors and to advance precision medicine.Design/methodology/approachThis study is a narrative review of the literature.FindingsThe body of medical knowledge has grown far too large for human clinicians to parse. In theory, electronic health records could augment clinical decision-making with electronic clinical decision support systems (CDSSs). However, computer scientists and clinicians have made remarkably little progress in building CDSSs, because health data tend to be siloed across many different systems that are not interoperable and cannot be linked using common identifiers. As a result, medicine in the USA is often practiced inconsistently with poor adherence to the best preventive and clinical practices. Poor information technology infrastructure contributes to medical errors and waste, resulting in suboptimal care and tens of thousands of premature deaths every year. Taiwan’s national health system, in contrast, is underpinned by a coordinated system of electronic data systems but remains underutilized. In this paper, the authors present a theoretical path toward developing artificial intelligence (AI)-driven CDSS systems using Taiwan’s National Health Insurance Research Database. Such a system could in theory not only optimize care and prevent clinical errors but also empower patients to track their progress in achieving their personal health goals.Originality/valueWhile research teams have previously built AI systems with limited applications, this study provides a framework for building global AI-based CDSS systems using one of the world’s few unified electronic health data systems.

A randomized controlled trial for evaluating the impact of integrating a computerized clinical decision support system and a socially assistive humanoid robot into grand rounds during pre/post-operative care.

Although clinical decision support systems (CDSSs) are increasingly emphasized as one of the possible levers for improving care, they are still not widely used due to different barriers, such as doubts about systems’ performance, their complexity and poor design, practitioners’ lack of time to use them, poor computer skills, reluctance to use them in front of patients, and deficient integration into existing workflows. While several studies on CDSS exist, there is a need for additional high-quality studies using large samples and examining the differences between outcomes following a decision based on CDSS support and those following decisions without this kind of information. Even less is known about the effectiveness of a CDSS that is delivered during a grand round routine and with the help of socially assistive humanoid robots (SAHRs). In this study, 200 patients will be randomized into a Control Group (i.e. standard care) and an Intervention Group (i.e. standard care and novel CDSS delivered via a SAHR). Health care quality and Quality of Life measures will be compared between the two groups. Additionally, approximately 22 clinicians, who are also active researchers at the University Clinical Center Maribor, will evaluate the acceptability and clinical usability of the system. The results of the proposed study will provide high-quality evidence on the effectiveness of CDSS systems and SAHR in the grand round routine.

57 EMBED: A Pragmatic Trial of User-Centered Clinical Decision Support to Implement EMergency Department-Initiated BuprenorphinE for Opioid Use Disorder

Emergency departments (ED) are a critical access point for individuals with opioid use disorder (OUD). Medications for OUD decrease: (1) mortality after nonfatal overdose, (2) withdrawal symptoms, (3) craving, and (4) illicit opioid use. Buprenorphine (BUP) can be safely initiated in the ED and increases retention in outpatient treatment programs. Adoption of ED-initiated BUP has been slow due to lack of physician knowledge, experience, and time. The study objective was to determine the effect of a user-friendly clinical decision support system (CDS) on rates of ED-initiated BUP in the routine emergency care of individuals with OUD. A parallel group randomized pragmatic trial was conducted in 18 EDs across 5 health care systems in the United States. The EMBED intervention is a user-centered, physician-facing CDS system integrated into the electronic health record (EHR) to facilitate ED-initiation of BUP by: (1) determining a diagnosis of OUD, (2) assessing withdrawal severity, (3) motivating patients to accept treatment, and (4) automating EHR workflow including clinical and after visit documentation, order entry, prescribing, and referral. EDs were randomly allocated to the intervention versus usual care using stratified covariate constrained randomization of clusters to arms with a ratio of 1:1. Constraining covariates included EHR vendor, patient volume, practice setting, ongoing resources for OUD patients, and proportion of attending physicians waivered to prescribe BUP. Participants included adult ED OUD patients meeting a predetermined EHR phenotype for OUD and the attending emergency physicians caring for them. Control site visits associated with attending physicians who practiced at both intervention and control sites were excluded. The primary outcome was the rate of ED-initiated BUP (administration or prescription of BUP). Implementation outcomes as compared to usual care were measured using the RE-AIM (reach, effectiveness, adoption, implementation, and maintenance) framework. 354, 800 ED visits (166, 184 intervention and 188, 616 control) from November 2019-March 2021 were assessed for eligibility. Overall, 4, 561 OUD patients (2, 457 intervention, 2, 104 controls) under the care of 607 attending physicians (335 intervention, 272 controls) were included in the analysis. Among these patients, 33% and 36% were female and the mean age (SD) was 39.4 y (13.3) and 38.5 y (13.0) in the intervention and control groups, respectively. Though the analysis is not complete at this time, full results are forthcoming and will be available in advance of the ACEP21 Research Forum. This pragmatic trial involving patients with OUD will report the effect of a user-centered CDS system that automates EHR workflow to initiate BUP in the ED.

Decision support systems, assistance systems and telemedicine in epileptology

HintergrundDie wissenschaftlichen Erkenntnisse über Epilepsien und deren klinische Implikationen nehmen rasant zu. Für Nichtexperten stellt sich die zunehmende Herausforderung, den Überblick hierüber zu bewahren. Hier setzen Clinical-decision-support-Systeme (CDSS) an, indem sie standard- und expertengetriggertes Wissen zur Diagnostik und Therapie individualisiert und automatisiert liefern. Zudem sind Medizin-Apps und telemedizinische Verfahren zur Diagnostik und Therapie sowie Assistenzsysteme zur Anfallsdetektion bei Epilepsien verfügbar.Ziel der ArbeitEs soll ein Überblick über die aktuellen Entwicklungen und Anwendungsmöglichkeiten verfügbarer tele-epileptologischer Methoden gegeben werden.Material und MethodenAuf der Basis persönlicher Kenntnis und eines Literaturreviews werden epilepsiespezifische CDSS, Medizin-Apps, Assistenzsysteme sowie telemedizinische Anwendungen charakterisiert und deren klinische Einsatzmöglichkeiten dargestellt.Ergebnisse und DiskussionPersonen mit Epilepsie könnten aufgrund des chronischen Verlaufs und der Komplexität der Erkrankung und ihrer Folgen von CDSS profitieren. Es erscheint wünschenswert, dass epilepsiespezifische CDSS sowohl für die Behandelnden als auch für Patienten nutzbar werden. Apps für Menschen mit Epilepsie dienen derzeit meist der Verlaufsdokumentation von Anfallsfrequenz, Medikamentencompliance und Nebenwirkungen. Gegenwärtige Anfallsdetektionssysteme erkennen vor allem generalisiert tonisch-klonische Anfälle (GTKA). Ein klinischer Nutzen ist noch nicht hinreichend belegt, erscheint aber wahrscheinlich, insbesondere da GTKA mit dem Risiko eines plötzlichen Todes von Epilepsiepatienten assoziiert sind und Interventionen als wirksam gelten.

Performance Evaluation of Clinical Decision Support Systems (CDSS): Developing a Business Intelligence (BI) Dashboard.

This document describes the development of a Business Intelligence (BI) dashboard for tracking the drug-drug interaction (DDI) alerts implemented as Clinical Decision Support Systems (CDSS) in Electronic Health Records (EHR). CDSS are known for their potential to reduce medical error. The use of requirements in the development of BI dashboards is crucial to obtain successful software. In this work, the requirements were analysed using a score methodology, considering the relevance of the indicators and visualization methods. CDSS effectiveness and acceptance have been questioned, so it is fundamental to monitor their behaviour and performance. The dashboard was designed in order to satisfy the needed indicators. Using BI as a tool for monitoring the CDSS performance made it possible to operationalize the EHR content repository, maximizing the understanding in relation to the override and, by inference, to optimize the CDSS system by opening new lines of work.

A mobile health monitoring-and-treatment system based on integration of the SSN sensor ontology and the HL7 FHIR standard

BackgroundMobile health (MH) technologies including clinical decision support systems (CDSS) provide an efficient method for patient monitoring and treatment. A mobile CDSS is based on real-time sensor data and historical electronic health record (EHR) data. Raw sensor data have no semantics of their own; therefore, a computer system cannot interpret these data automatically. In addition, the interoperability of sensor data and EHR medical data is a challenge. EHR data collected from distributed systems have different structures, semantics, and coding mechanisms. As a result, building a transparent CDSS that can work as a portable plug-and-play component in any existing EHR ecosystem requires a careful design process. Ontology and medical standards support the construction of semantically intelligent CDSSs.MethodsThis paper proposes a comprehensive MH framework with an integrated CDSS capability. This cloud-based system monitors and manages type 1 diabetes mellitus. The efficiency of any CDSS depends mainly on the quality of its knowledge and its semantic interoperability with different data sources. To this end, this paper concentrates on constructing a semantic CDSS based on proposed FASTO ontology.ResultsThis realistic ontology is able to collect, formalize, integrate, analyze, and manipulate all types of patient data. It provides patients with complete, personalized, and medically intuitive care plans, including insulin regimens, diets, exercises, and education sub-plans. These plans are based on the complete patient profile. In addition, the proposed CDSS provides real-time patient monitoring based on vital signs collected from patients’ wireless body area networks. These monitoring include real-time insulin adjustments, mealtime carbohydrate calculations, and exercise recommendations. FASTO integrates the well-known standards of HL7 fast healthcare interoperability resources (FHIR), semantic sensor network (SSN) ontology, basic formal ontology (BFO) 2.0, and clinical practice guidelines. The current version of FASTO includes 9577 classes, 658 object properties, 164 data properties, 460 individuals, and 140 SWRL rules. FASTO is publicly available through the National Center for Biomedical Ontology BioPortal at https://bioportal.bioontology.org/ontologies/FASTO.ConclusionsThe resulting CDSS system can help physicians to monitor more patients efficiently and accurately. In addition, patients in rural areas can depend on the system to manage their diabetes and emergencies.

CDSS-RM: a clinical decision support system reference model

Clinical Decision Support Systems (CDSS) provide aid in clinical decision making and therefore need to take into consideration human, data interactions, and cognitive functions of clinical decision makers. The objective of this paper is to introduce a high level reference model that is intended to be used as a foundation to design successful and contextually relevant CDSS systems. The paper begins by introducing the information flow, use, and sharing characteristics in a hospital setting, and then it outlines the referential context for the model, which are clinical decisions in a hospital setting. Important characteristics of the Clinical decision making process include: (i) Temporally ordered steps, each leading to new data, which in turn becomes useful for a new decision, (ii) Feedback loops where acquisition of new data improves certainty and generates new questions to examine, (iii) Combining different kinds of clinical data for decision making, (iv) Reusing the same data in two or more different decisions, and (v) Clinical decisions requiring human cognitive skills and knowledge, to process the available information. These characteristics form the foundation to delineate important considerations of Clinical Decision Support Systems design. The model includes six interacting and interconnected elements, which formulate the high-level reference model (CDSS-RM). These elements are introduced in the form of questions, as considerations, and are examined with the use of illustrated scenario-based and data-driven examples. The six elements /considerations of the reference model are: (i) Do CDSS mimic the cognitive process of clinical decision makers? (ii) Do CDSS provide recommendations with longitudinal insight? (iii) Is the model performance contextually realistic? (iv) Is the ‘Historical Decision’ bias taken into consideration in CDSS design? (v) Do CDSS integrate established clinical standards and protocols? (vi) Do CDSS utilize unstructured data? The CDSS-RM reference model can contribute to optimized design of modeling methodologies, in order to improve response of health systems to clinical decision-making challenges.

DMTO: a realistic ontology for standard diabetes mellitus treatment

BackgroundTreatment of type 2 diabetes mellitus (T2DM) is a complex problem. A clinical decision support system (CDSS) based on massive and distributed electronic health record data can facilitate the automation of this process and enhance its accuracy. The most important component of any CDSS is its knowledge base. This knowledge base can be formulated using ontologies. The formal description logic of ontology supports the inference of hidden knowledge. Building a complete, coherent, consistent, interoperable, and sharable ontology is a challenge.ResultsThis paper introduces the first version of the newly constructed Diabetes Mellitus Treatment Ontology (DMTO) as a basis for shared-semantics, domain-specific, standard, machine-readable, and interoperable knowledge relevant to T2DM treatment. It is a comprehensive ontology and provides the highest coverage and the most complete picture of coded knowledge about T2DM patients’ current conditions, previous profiles, and T2DM-related aspects, including complications, symptoms, lab tests, interactions, treatment plan (TP) frameworks, and glucose-related diseases and medications. It adheres to the design principles recommended by the Open Biomedical Ontologies Foundry and is based on ontological realism that follows the principles of the Basic Formal Ontology and the Ontology for General Medical Science. DMTO is implemented under Protégé 5.0 in Web Ontology Language (OWL) 2 format and is publicly available through the National Center for Biomedical Ontology’s BioPortal at http://bioportal.bioontology.org/ontologies/DMTO. The current version of DMTO includes more than 10,700 classes, 277 relations, 39,425 annotations, 214 semantic rules, and 62,974 axioms. We provide proof of concept for this approach to modeling TPs.ConclusionThe ontology is able to collect and analyze most features of T2DM as well as customize chronic TPs with the most appropriate drugs, foods, and physical exercises. DMTO is ready to be used as a knowledge base for semantically intelligent and distributed CDSS systems.

Clinical Decision Support System for Liver Fibrosis Prediction in Hepatitis Patients: A Case Comparison of Two Soft Computing Techniques

Diagnosis of deadly diseases, such as liver fibrosis, is very important. Clinical decision support systems (CDSSs) based on patient’s historical medical data and accurate AI techniques can aid physicians in their decision-making process. The task of arriving at an accurate and timely diagnosis decision is always complex because of the dynamic, vagueness, and uncertainty associated with this disease. Fuzzy logic can perfectly handle these issues. In recent years, two of the most interesting techniques are a fuzzy analytical hierarchy process (FAHP) and an adaptive neuro-fuzzy inference system (ANFIS). The FAHP is popular for dealing with uncertainty in multi-criteria decision-making, and the ANFIS is popular in learning fuzzy inference system from data based on artificial neural networks. To the best of our knowledge, these two methods have not been used to model CDSSs in fibrosis stage detection domain. In this paper, we develop a CDSS based on a case comparison of the effectiveness of the FAHP and the ANFIS in the medical diagnosis of the fibrosis disease. We carefully design and implement two frameworks based on these two techniques. Diagnostic real data of 119 cases infected by chronic viral hepatitis C from the Liver Institute at Mansoura University in Egypt are used to train and test both the FAHP and ANFIS. Criteria and subcriteria weights are based on opinions of two domain experts. The ANFIS model is designed using trial and error based on the analysis of various experiments. Results are later compared with the diagnostic conclusions of medical expert and other three medical and fuzzy techniques. The comparison results show that these two techniques can successfully be employed in designing a diagnostic CDSS system for fibrosis diagnosis. The two techniques achieve a classification accuracy of 93.3%. The results confirm the efficiency and effectiveness of both methods. Therefore, both the FAHP and ANFIS are viable approaches in modeling CDSS for diagnosis of a liver fibrosis stage.

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

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. 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. 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. 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.

Transforming User Needs into Functional Requirements for an Antibiotic Clinical Decision Support System

Many informatics studies use content analysis to generate functional requirements for system development. Explication of this translational process from qualitative data to functional requirements can strengthen the understanding and scientific rigor when applying content analysis in informatics studies. To describe a user-centered approach transforming emergent themes derived from focus group data into functional requirements for informatics solutions and to illustrate these methods to the development of an antibiotic clinical decision support system (CDS). THE APPROACH CONSISTED OF FIVE STEPS: 1) identify unmet therapeutic planning information needs via Focus Group Study-I, 2) develop a coding framework of therapeutic planning themes to refine the domain scope to antibiotic therapeutic planning, 3) identify functional requirements of an antibiotic CDS system via Focus Group Study-II, 4) discover informatics solutions and functional requirements from coded data, and 5) determine the types of information needed to support the antibiotic CDS system and link with the identified informatics solutions and functional requirements. The coding framework for Focus Group Study-I revealed unmet therapeutic planning needs. Twelve subthemes emerged and were clustered into four themes; analysis indicated a need for an antibiotic CDS intervention. Focus Group Study-II included five types of information needs. Comments from the Barrier/Challenge to information access and Function/Feature themes produced three informatics solutions and 13 functional requirements of an antibiotic CDS system. Comments from the Patient, Institution, and Domain themes generated required data elements for each informatics solution. This study presents one example explicating content analysis of focus group data and the analysis process to functional requirements from narrative data. Illustration of this 5-step method was used to develop an antibiotic CDS system, resolving unmet antibiotic prescribing needs. As a reusable approach, these techniques can be refined and applied to resolve unmet information needs with informatics interventions in additional domains.

A Duel between Clinical Decision Support System and Healthcare Professionals: A Study in Malaysia

As clinical IT can improve the productivity and performance (at different levels) of hospitals, researchers have tried to find out factors affecting the successful adoption of technological advancement (especially, progression in information technology) among healthcare professionals. One of the most important determinants influencing the success of clinical decision support system (CDS) adoption is that to what degree the system is accepted by healthcare professionals. Hospitals invest in CDS systems with the hope of cutting medication errors and increasing the quality of products or services. But if users are not likely to accept the systems, hospitals can not benefit significantly from the systems. On the other hand, if users accept CDS system they become more willing to make use of the systems in their practice patterns. The usage of CDS can be a sign of the system success in hospitals. Therefore, the purpose of this study is to investigate the reasons that motivate healthcare professionals to use CDS in hospitals. In this study, a series of interviews has been conducted among 21 healthcare professionals in Malaysia. The content analysis has been used to analyze the interviews in this context. The results reflect the importance of perceived threat to professional autonomy, level of interactivity with CDS, level of involvement in decision making, computer self efficacy and subjective norms in predicting healthcare professionals’ intention to use CDS system in Malaysia. Based on the qualitative study, a conceptual mode has been proposed for an empirical study in future research.

Comparing Bayesian inference and case-based reasoning as support techniques in the diagnosis of Acute Bacterial Meningitis

The amount of information available for physicians has dramatically increased in the recent past. In contrast, the specialist’s ability to understand, synthesize and take into account such information is severely constrained by the short time available for the appointments. Therefore, systems reusing available knowledge and implementing reasoning processes become critical to support the tasks of the doctors. As a number of different techniques for building such systems are available, contrasting their effectiveness becomes a major concern. This is especially important in the case of infectious diseases that can be lethal within hours such as the Acute Bacterial Meningitis (ABM) for which implementing and contrasting different techniques allows for an increased reliability and speed in supporting the process of diagnosis. This work focuses on the construction of diagnosis support tools for ABM, reporting a comparative assessment of the quality of a Clinical Decision Support System (CDSS) resulting from the application of Case Based Reasoning (CBR), to that of an existing CDSS system developed using a Bayesian expert system. Although both approaches proved to be useful, the one based in CBR techniques show some interesting capabilities as higher precision, automatic learning or experience capturing, and also a better response to lack of input data. The three developed systems perform with high levels of accuracy– e.g. propose correct diagnostics based on a certain set of symptoms – but the one based on CBR present some additional capabilities that look very promising for implementing these kind of systems in a real world scenario.

Clinical decision support implemented with academic detailing improves prescribing of key renally cleared drugs in the hospital setting

Lack of dose adjustment for renally cleared drugs in the presence of poor renal function is a common problem in the hospital setting. The absence of a clinical decision support system (CDSS) from direct clinician workflows such as computerized provider order entry (CPOE) hinders the uptake of CDSS. This study implemented CDSS in an environment independent of CPOE, introduced to prescribers via academic detailing, to address the dosing of renally cleared drugs. GFR+ was designed to automatically calculate and update renal function, doses of key drugs adjusted for renal function, and highlight clinically significant decreases in renal function. Prescribers were made aware of GFR+, its navigation, and surrounding clinical issues, using academic detailing. The rate of dosing conformity and management for key renally cleared drugs in hospitalized patients, before and after GFR+ implementation. Improvements were seen in dosing conformity for enoxaparin (from 68% to 86%, p=0.03), gentamicin (63-87%, p=0.01), and vancomycin (47-77%, p=0.07), as well as the appropriate use of gentamicin therapeutic drug monitoring (70-90%, p=0.02). During episodes of acute renal impairment, renally cleared drugs were held on 38% of instances in the pre-intervention period compared with 62% post-intervention (p=0.01). Clinical decision support implemented with academic detailing improved dosing conformity and management of key renally cleared drugs in a hospitalized population. Academic detailing should be strongly considered to facilitate the introduction of CDSS systems that cannot be placed directly into the clinician workflow.

Development strategy and potential impact on medication safety for clinical rules: the lithium case

Introduction: Prescribing, administration and monitoring of drugs are complex and error sensitive processes. Physicians have to take into account many drug- and patient specific characteristics continuously during therapy. Appropriate use of information technology, especially the introduction of clinical decision support systems (CDSS) can substantially reduce medication error rates and improve patient safety [1, 2]. Strategies for the development of clinical rules used in these CDSS systems are not available. Purpose: The aim of this study was: (1) to evaluate a strategy for the development of clinical rules and (2) to determine the potential impact on medication safety of these clinical rules. Lithium therapy was chosen as a test concept. Methods: For the development of clinical rules we followed a Plan – Do – Check – Act (PDCA) cycle with a panel of experts. Standard literature sources were screened for patient characteristics, monitoring parameters and other characteristics that have to be taken into account during lithium therapy. We translated this information into clinical rules for the CDSS system Gaston®# # Gaston®: P.A.de Clerq, MEDECS b.v. http://www.medecs.nl (plan). These rules were retrospectively tested on patients admitted to the Catharina-hospital Eindhoven in 2004 and 2005 using lithium (do). Signals were checked in the electronic patients record on: (1) appropriateness and (2) the action undertaken by the physician compared to the action warranted by the CDSS. For every rule the Positive Predictive Value (PPV) was determined: number of appropriate signals produced /total number of signals. When the action undertaken was not right according to the clinical rules we scored this as a Potential Adverse Drug Event (PADE) (check). Results were discussed with an expert panel and the clinical rules adjusted according their consensus (act). Results: The PDCA strategy with an expert panel resulted in a package of clinical rules for patients starting with, or using lithium therapy. To design the definite clinical rules the input of the expert team was essential, because literature left to much room for interpretation in clinical practice. The clinical rules were tested on 38 patients starting with and 93 patients using lithium in 2004 or 2005. The PPV’s of these clinical rules varied from 0.2 to 1.0 (mean 0.8). We identified 81 PADE’s in 71 patients, that could have been prevented when the CDSS would have been active. Most important PADE’s were: 1) omitted determination of lithium plasma levels after reduction of creatinin clearance < 50 ml/min (4 times) and starting or stopping of an interacting drug (17 times); 2) determination of lithium, Thyroid Stimulating Hormone (TSH), leucocyte and calcium plasma levels every 3 or 6 months (resp. 17, 6, 2 and 4 times). An important limitation for the creation of higher PPV’s and for the identification of PADE’s was the lack of outpatient data. Further research is needed to show that this system is clinically applicable. Conclusion: A strategy of PDCA combined with an expert panel is a prerequisite for the development of clinical rules for drugs in a CDSS. We demonstrated that a CDSS is able to identify patients with potential adverse drug events. Therefore we conclude that application of CDSS can have a relevant impact on medication safety in clinical practice.

Verbal prescribing in general practice consultations

This paper looks at aspects of doctor–patient communication and focuses on how prescribing decisions fit into the consultation within the context of the use (and non-use) of a technological clinical decision support system (CDSS) in the UK. Analysis of 6 simulated consultations filmed as part of the evaluation of a CDSS system indicated that the general practitioners (GPs) used their computers for a short time during consultations. The data showed that doctors’ utterances, occurring at an early stage of the consultations, signalled the prescribing decision and eventual outcome of the consultation. The concept of ‘verbal prescriptions’ is used to describe these utterances of the GPs, and facilitates an understanding of how prescribing decisions are routinely achieved. Prescribing decisions can occur in the relatively early stages of the consultation, and both prior to and independently of the CDSS. Consequently, we suggest that the pattern of GP decision-making needs to be taken into account in CDSS design. However, this is not just an issue for CDSS design and implementation, as the verbal prescription phenomenon may impact upon patient involvement in decision-making, and even the appropriate use of evidence based medicine.