Advanced Clinical Decision Support Systems Research Articles

CDSS Adoption and the Role of Artificial Intelligence in Saudi Arabian Primary Healthcare

The integration of advanced Clinical Decision Support Systems (CDSS) is crucial in healthcare, especially in developing countries. However, the challenges to CDSS implementation in these settings are often overlooked, leading to implementation gaps. This study investigates the adoption of CDSS in primary healthcare facilities in Saudi Arabia through in-depth interviews with general practitioners. The advancement of primary healthcare systems in developing nations necessitates appropriate technological solutions to address issues such as equitable access, service quality, and patient satisfaction. The adoption of CDSS can potentially improve patient satisfaction and enhance medical consultations. Artificial intelligence -driven technologies like the Velys robot are transforming the medical field by assisting doctors in surgical procedures. In healthcare, Artificial intelligence (AI) can address issues related to complex diseases, aging populations, and workforce shortages. AI solutions can streamline patient flow, improve care coordination, and support personalized healthcare management. By assessing the impact of AI on patient care, healthcare efficiency, and effectiveness can be enhanced. This research highlights the challenges and opportunities in CDSS adoption in developing countries, with implications for healthcare improvement in Saudi Arabia.

Clinical decision support system based on artificial intelligence for adjusting insulin pump parameters in children with type 1 diabetes mellitus

BACKGROUND: Widely available diabetes devices (continuous glucose monitoring, insulin pump etc.) generate large amount of data and development of an advanced clinical decision support system (CDSS), able to automatically evaluate and optimize insulin therapy, is relevant.AIM: Development of a mathematical model and an CDSS based on it to optimize insulin therapy in children with type 1 diabetes (T1D) and assessment of the agreement between the recommendations of the CDSS and the physician on insulin pump (IP) parameters: basal profile (BP), carbohydrate ratio (CR), correction factor (СF).MATERIALS AND METHODS: Data from 504 children with T1DM were analyzed over the period of 7875 days. The data included glucose, insulin, food, sex, age, height, weight, diabetes duration and HbA1c. We constructed recurrent neural network (RNN) to predict glucose concentration for 30-120 minutes, an algorithm for optimizing IP settings using prediction results. Next, a software product was developed — a CDSS. To assess the agreement of the recommendations of the CDSS and physicians, retrospective data from 40 remote telemedicine consultations of 40 patients with T1D (median age 11.6 years [7; 15]) were used and 960 points of possible adjustments were analyzed. Three degrees of agreement have been introduced: complete agreement, partial agreement, and complete disagreement. The magnitude of the adjustments was also analyzed.RESULTS: The accuracy of glycemic predictions was better or comparable with other similar models. The assessment of agreement for BP, CR and CF, according to the Kappa index, showed slight and weak agreement. The frequency of complete agreement between recommendations for adjusting the ongoing IP therapy between the CDSS and physicians is 37.5–53.8%, and complete inconsistency is 4.5–17.4%. From a clinical point of view, consistency in the frequency of occurrence of the indicator is more important. There were no differences in median IP settings between the CDSS and physicians.CONCLUSION: The CDSS has an acceptable accuracy of glycemic predictions. The CDSS and physicians provide comparable recommendations regarding CSII parameters.

Gut microbiota functional profiling in autism spectrum disorders: bacterial VOCs and related metabolic pathways acting as disease biomarkers and predictors.

Autism spectrum disorder (ASD) is a multifactorial neurodevelopmental disorder. Major interplays between the gastrointestinal (GI) tract and the central nervous system (CNS) seem to be driven by gut microbiota (GM). Herein, we provide a GM functional characterization, based on GM metabolomics, mapping of bacterial biochemical pathways, and anamnestic, clinical, and nutritional patient metadata. Fecal samples collected from children with ASD and neurotypical children were analyzed by gas-chromatography mass spectrometry coupled with solid phase microextraction (GC-MS/SPME) to determine volatile organic compounds (VOCs) associated with the metataxonomic approach by 16S rRNA gene sequencing. Multivariate and univariate statistical analyses assessed differential VOC profiles and relationships with ASD anamnestic and clinical features for biomarker discovery. Multiple web-based and machine learning (ML) models identified metabolic predictors of disease and network analyses correlated GM ecological and metabolic patterns. The GM core volatilome for all ASD patients was characterized by a high concentration of 1-pentanol, 1-butanol, phenyl ethyl alcohol; benzeneacetaldehyde, octadecanal, tetradecanal; methyl isobutyl ketone, 2-hexanone, acetone; acetic, propanoic, 3-methyl-butanoic and 2-methyl-propanoic acids; indole and skatole; and o-cymene. Patients were stratified based on age, GI symptoms, and ASD severity symptoms. Disease risk prediction allowed us to associate butanoic acid with subjects older than 5 years, indole with the absence of GI symptoms and low disease severity, propanoic acid with the ASD risk group, and p-cymene with ASD symptoms, all based on the predictive CBCL-EXT scale. The HistGradientBoostingClassifier model classified ASD patients vs. CTRLs by an accuracy of 89%, based on methyl isobutyl ketone, benzeneacetaldehyde, phenyl ethyl alcohol, ethanol, butanoic acid, octadecane, acetic acid, skatole, and tetradecanal features. LogisticRegression models corroborated methyl isobutyl ketone, benzeneacetaldehyde, phenyl ethyl alcohol, skatole, and acetic acid as ASD predictors. Our results will aid the development of advanced clinical decision support systems (CDSSs), assisted by ML models, for advanced ASD-personalized medicine, based on omics data integrated into electronic health/medical records. Furthermore, new ASD screening strategies based on GM-related predictors could be used to improve ASD risk assessment by uncovering novel ASD onset and risk predictors.

Deep Learning Based Severity Grading of Knee Osteoarthritis Based on Radiological Images

Abstract: Knee OA is a prevalent degenerative joint condition affecting a significant portion of the global population, and timely identification and accurate grading are crucial for optimal patient care and treatment planning. This system aimed at developing an advanced clinical decision support system for the early detection and precise grading of knee osteoarthritis (OA) using cutting-edge deep learning techniques. Traditionally, the diagnosis of knee osteoarthritis has relied on manual assessments of radiological images, which can be resource-intensive and subject to variations among evaluators. Deep learning, a subset of artificial intelligence, offers a promising solution by automating this process with exceptional precision. The core objectives of this project include the development of a CNN model based on the EfficientNet-B5 architecture, designed to autonomously analyze radiological images of knee joints. Additionally, a user-friendly web-based interface is being constructed to facilitate image uploads and present severity grades. The project involves rigorous training and fine-tuning of the EfficientNet-B5 using a dataset of knee OA images to achieve a high degree of accuracy in severity grading. The system will ultimately be deployed on a scalable and secure cloud platform for practical utility.

Multidrug Interactions: Why Do They Occur and How to Handle?

A nonoptimized medication therapy (NOMT) event is an iatrogenic hazard or incident associated with medications and is a leading cause of death, serious injury, and illness. NOMT events are often related to multidrug interactions in patients with polypharmacy. In these patients, NOMT events can be avoided by using advanced clinical decision support systems and clinical interventions such as separating the time of administration of certain drugs during the day. At the individual level, medication reconciliation is a first logical step for reducing adverse side effects. Then, intersubject variability in drug response should be considered to optimize patient drug regimens. Furthermore, patient pharmacogenomic status information can help ensure appropriateness of drug therapy. However, in patients with polypharmacy, such information is most valuable when combined with phenoconversion probability. At a population level, the virtual addition of drugs to various drug regimens and the use of a medication risk score can help predict the risk of NOMT events. This review outlines some of the mechanisms behind multidrug interactions and their association with drug safety and NOMTs, polypharmacy and its impact on patient outcomes, the value of pharmacogenomics, and an assessment of simulation studies and the virtual addition of drugs to a drug regimen using real-world data.

Development of knowledge-based clinical decision support system for the management of cardiovascular diseases

Abstract Background Knowledge-based clinical decision support systems (CDSS) are technological tools that analyze patient information and present justified diagnostic and therapeutic suggestions to the user. They improve adherence to evidence-based medicine and reduce medical errors. Material and methods We have developed an advanced CDSS with the following components: 1. Knowledge Management System, which allows the introduction of interoperable clinical variables organized in taxonomies, and the generation, validation and direct maintenance of decision rules by experts doctors. 2. Smart Assistant for Medical Reports Generation, which facilitates user interaction for the introduction of clinical data (natural language and/or data entry forms) and decission making supported by real-time suggestions that are automatically generated by 3. Inference Engine. Rules and variables derived from the most relevant cardiovascular Clinical Practice Guidelines (CPGs) are introduced by cardiologists into the system. The rules undergo a process of peer review and internal and external validation with anonymized clinical cases. Results More than 2,000 clinical variables and more than 5,000 rules corresponding to more than 80 CPGs and consensuns papers have been created and validated. An internal and external pilot validation of the system has been successfully carried out for the main cardiovascular problems, such as heart failure, atrial fibrillation, valvulopathies, coronary syndromes, cardiomyopathies, hypertension, diabetes and dyslipidaemias, using more than 400 cases. Conclusion Our CDSS is capable of providing real-time diagnostic and therapeutic recommendations for complex scenarios in cardiovascular disease, combining thousands of variables and decision rules in a single system. Prospective evaluation of the system in real clinical environments, certification and integration in Electronic Medical Records for clinical use is required. Funding Acknowledgement Type of funding sources: Private company. Main funding source(s): DILEMMA solutions, SL

Development of knowledge-based clinical decision support system for the management of cardiovascular diseases

Abstract Background Knowledge-based clinical decision support systems (CDSS) are technological tools that analyze patient information and present justified diagnostic and therapeutic suggestions to the user. They improve adherence to evidence-based medicine and reduce medical errors. Material and methods We have developed an advanced CDSS with the following components: 1. Knowledge Management System, which allows the introduction of interoperable clinical variables organized in taxonomies, and the generation, validation and direct maintenance of decision rules by experts doctors. 2. Smart Assistant for Medical Reports Generation, which facilitates user interaction for the introduction of clinical data (natural language and/or data entry forms) and decission making supported by real-time suggestions that are automatically generated by 3. Inference Engine. Rules and variables derived from the most relevant cardiovascular Clinical Practice Guidelines (CPGs) are introduced by cardiologists into the system. The rules undergo a process of peer review and internal and external validation with anonymized clinical cases. Results More than 2,000 clinical variables and more than 5,000 rules corresponding to more than 80 CPGs and consensuns papers have been created and validated. An internal and external pilot validation of the system has been successfully carried out for the main cardiovascular problems, such as heart failure, atrial fibrillation, valvulopathies, coronary syndromes, cardiomyopathies, hypertension, diabetes and dyslipidaemias, using more than 400 cases. Conclusion Our CDSS is capable of providing real-time diagnostic and therapeutic recommendations for complex scenarios in cardiovascular disease, combining thousands of variables and decision rules in a single system. Prospective evaluation of the system in real clinical environments, certification and integration in Electronic Medical Records for clinical use is required. Funding Acknowledgement Type of funding sources: Private company. Main funding source(s): DILEMMA solutions, SL

Artificial Intelligence in Oncology: Current Capabilities, Future Opportunities, and Ethical Considerations.

The promise of highly personalized oncology care using artificial intelligence (AI) technologies has been forecasted since the emergence of the field. Cumulative advances across the science are bringing this promise to realization, including refinement of machine learning- and deep learning algorithms; expansion in the depth and variety of databases, including multiomics; and the decreased cost of massively parallelized computational power. Examples of successful clinical applications of AI can be found throughout the cancer continuum and in multidisciplinary practice, with computer vision-assisted image analysis in particular having several U.S. Food and Drug Administration-approved uses. Techniques with emerging clinical utility include whole blood multicancer detection from deep sequencing, virtual biopsies, natural language processing to infer health trajectories from medical notes, and advanced clinical decision support systems that combine genomics and clinomics. Substantial issues have delayed broad adoption, with data transparency and interpretability suffering from AI's "black box" mechanism, and intrinsic bias against underrepresented persons limiting the reproducibility of AI models and perpetuating health care disparities. Midfuture projections of AI maturation involve increasing a model's complexity by using multimodal data elements to better approximate an organic system. Far-future positing includes living databases that accumulate all aspects of a person's health into discrete data elements; this will fuel highly convoluted modeling that can tailor treatment selection, dose determination, surveillance modality and schedule, and more. The field of AI has had a historical dichotomy between its proponents and detractors. The successful development of recent applications, and continued investment in prospective validation that defines their impact on multilevel outcomes, has established a momentum of accelerated progress.

Assessing the Impact of an Advanced Clinical Decision Support System on Medication Safety and Hospital Readmissions in an Innovative Transitional Care Model: A Pilot Study.

(1) Background: Adverse drug events and inappropriate use of medications lead to hospitalizations, medication-related morbidity, and mortality. This study examined whether a novel medication risk prediction tool, the MedWise Risk Score™, was associated with medication safety-related problem (MRP) identification and whether integration into an existing innovative transitions of care (TOC) service could decrease readmissions. (2) Methods: This retrospective comparator group study assessed patients discharged from a hospital in southern Arizona between January and December 2020. Participants were included in the study if they were 18 years of age or older, referred to the pharmacist for TOC services, and received a pharmacist consultation within one-week post discharge. Patients were categorized into two groups: (1) medication safety review (MSR)-TOC service (intervention) or (2) existing innovative TOC service (control). (3) Results: Of 164 participants, most were male (57%) and were between 70–79 years of age. Overall, there were significantly more drug-drug interactions (DDI) MRPs identified per patient in the intervention vs. control group for those who were readmitted (3.7 ± 1.5 vs. 0.9 ± 0.6, p < 0.001) and those who were not readmitted (2 ± 1.3 vs. 1.3 ± 1.2, p = 0.0120). Furthermore, of those who were readmitted, the average number of identified MRPs per patient was greater in the intervention group compared to the control (6.3 vs. 2.5, respectively, p > 0.05). Relative to the control, the readmission frequency was 30% lower in the treatment group; however, there was insufficient power to detect significant differences between groups. (4) Conclusions: The integration of a medication risk prediction tool into this existing TOC service identified more DDI MRPs compared to the previous innovative TOC service, which lends evidence that supports its ability to prevent readmissions. Future work is warranted to demonstrate the longitudinal impact of this intervention in a larger sample size.

Medication-related problems identified by pharmacists in an enhanced medication therapy management model.

To describe the types of clinically actionable medication-related problems (MRPs) identified and the types of resolving recommendations issued by pharmacists using an advanced clinical decision support system (CDSS) for Medicare Part D beneficiaries. Retrospective cross-sectional study. We examined frequencies of MRPs and recommendations for beneficiaries who received a first-ever medication safety review (MSR) during plan years 2018-2019. MRPs were considered clinically actionable if implementation of a recommendation would alter the medication regimen. Pharmacists identified ≥ 1 clinically actionable MRP for 82.4% (18,703/22,696) beneficiaries receiving an MSR. Among these beneficiaries, 36,455 MRPs were identified (mean [SD] number of MRPs: 1.9 [1.0]). "Adverse drug reaction" (n = 14,788; 40.6%), "drug interaction" (n = 9716; 26.7%), and "medication use without indication" (n = 6496; 17.8%) represented 85.0% of all MRPs. "Start alternative therapy" was most frequently recommended to resolve "adverse drug reactions" (6724/14,788; 45.5%), followed by "change time of administration" to resolve "drug interactions" (5189/9716; 53.4%) and "discontinue medication" to resolve "drug use without indication" (5718/6496; 88.0%). Overall, "start alternative therapy" (n = 12,219) and "discontinue medication" (n = 12,059) made up 66.6% of all recommendations. In Medicare Part D beneficiaries, pharmacists using an advanced CDSS identified a substantial number of MRPs pertaining to medication safety and issued recommendations to decrease the risk of adverse drug events.

Artificial Intelligence in Medical Sensors for Clinical Decisions.

Due to the limited ability of conventional methods and the limited perspective of human diagnostics, patients are often diagnosed incorrectly or at a late stage as their disease condition progresses. They may then undergo unnecessary treatments due to inaccurate diagnoses. In this Perspective, we offer a brief overview on the integration of nanotechnology-based medical sensors and artificial intelligence (AI) for advanced clinical decision support systems to help decision-makers and healthcare systems improve how they approach information, insights, and the surrounding contexts, as well as to promote the uptake of personalized medicine on an individualized basis. Relying on these milestones, wearable sensing devices could enable interactive and evolving clinical decisions that could be used for evidence-based analysis and recommendations as well as for personalized monitoring of disease progress and treatment. We present and discuss the ongoing challenges and future opportunities associated with AI-enabled medical sensors in clinical decisions.

Effectiveness of an Advanced Clinical Decision Support System on Clinical Decision-Making Skills in a Call Center Medication Therapy Management Pharmacy Setting: A Pilot Study.

(1) Background: There is limited evidence related to the efficacy of advanced clinical decision support systems (CDSS) on the quantity of high-quality clinical recommendations in a pharmacy-related medication therapy management (MTM) setting. The study aimed to assess the effect of an advanced CDSS on the quantity of relevant clinical pharmacist recommendations in a call center MTM setting. (2) Methods: This pre-test/post-test with comparator group study compared clinical skills assessment scores between certified MTM pharmacists in March 2020. A Wilcoxon Signed Rank test assessed the difference between pre- and post-test scores in both groups. (3) Results: Of 20 participants, the majority were less than 40 years old (85%) with a Doctor of Pharmacy degree (90%). Nine were female. Intervention group participants had less than three years of experience as a pharmacist. The control group had less than three years (40%) or seven to ten years (40%) of experience. There was a significant increase in intervention group scores between pre- (median = 3.0, IQR = 3.0) and post-test segments (median = 6.5, IQR = 4.0, p = 0.02). There was no significant change between control group pre- and post-test segments (p = 0.48). (4) Conclusion: Pharmacist exposure to an advanced CDSS was associated with significantly increased quantity of relevant clinical recommendations in an MTM pharmacy setting.

The effect of medication related clinical decision support at the time of physician order entry

Background In advanced clinical decision support systems, patient characteristics and laboratory values are included in the algorithms that generate alerts. These alerts have a higher specificity than basic medication surveillance alerts. The alerts of advanced clinical decision support systems can be shown directly to the prescriber during order entry, without the risk of generating an overload of irrelevant alerts. We implemented five advanced algorithms that are shown directly to the prescriber. These algorithms are for gastrointestinal prophylaxis, folic or folinic acid prescribed with orally or subcutaneously administered methotrexate, vitamin D prescribed with bisphosphonates, hyponatremia and measuring plasma levels for vancomycin and gentamicin. Objective We evaluated the effect of the implementation of the algorithms. Setting We performed prospective intervention studies with a historical group for comparison in both inpatients and outpatients at a teaching hospital in the Netherlands. Methods We compared the time period after implementation of the algorithm with the time period before implementation, using data from the hospital information system Epic. Difference in guideline adherence were analyzed using Chi square tests. Main outcome measure The outcome measures were the number of alerts, the acceptance rate of the advice in the alert, and for the algorithm measuring plasma levels for vancomycin and gentamicin the time to the correct dose. Results For all algorithms, the implementation resulted in a significant increase in guideline adherence, varying from 11 to 36%. The acceptance rate varied from 14% for hyponatremia to 90% for methotrexate. For gastrointestinal prophylaxis the acceptance rate was 4.4% for basic drug-drug interaction alerts when no gastrointestinal prophylaxis was prescribed and increased to 44.7% after implementation of the advanced algorithm. This algorithm substantially decreased the number of alerts from 812 before implementation to 217 after implementation. After implementation of the algorithm for measuring plasma levels for vancomycin and gentamicin, the proportion of patients receiving the correct dose after 48h increased from 73 to 84% (p = 0.03). Conclusion Implementation of advanced algorithms that take patient characteristics into account and are shown directly to the physician during order entry, result in an increased guideline adherence.

Proteomics and Metabolomics Approaches towards a Functional Insight onto AUTISM Spectrum Disorders: Phenotype Stratification and Biomarker Discovery.

Autism spectrum disorders (ASDs) are neurodevelopmental disorders characterized by behavioral alterations and currently affect about 1% of children. Significant genetic factors and mechanisms underline the causation of ASD. Indeed, many affected individuals are diagnosed with chromosomal abnormalities, submicroscopic deletions or duplications, single-gene disorders or variants. However, a range of metabolic abnormalities has been highlighted in many patients, by identifying biofluid metabolome and proteome profiles potentially usable as ASD biomarkers. Indeed, next-generation sequencing and other omics platforms, including proteomics and metabolomics, have uncovered early age disease biomarkers which may lead to novel diagnostic tools and treatment targets that may vary from patient to patient depending on the specific genomic and other omics findings. The progressive identification of new proteins and metabolites acting as biomarker candidates, combined with patient genetic and clinical data and environmental factors, including microbiota, would bring us towards advanced clinical decision support systems (CDSSs) assisted by machine learning models for advanced ASD-personalized medicine. Herein, we will discuss novel computational solutions to evaluate new proteome and metabolome ASD biomarker candidates, in terms of their recurrence in the reviewed literature and laboratory medicine feasibility. Moreover, the way to exploit CDSS, performed by artificial intelligence, is presented as an effective tool to integrate omics data to electronic health/medical records (EHR/EMR), hopefully acting as added value in the near future for the clinical management of ASD.

Impulse Classification Network for Video Head Impulse Test.

The vestibulo-ocular reflex (VOR) is a dynamic system of the human brain that helps to maintain balance and to stabilize vision during head movement. The video head impulse test (vHIT) is a clinical test that uses lightweight, high-speed video goggles to examine the VOR function by calculating the ratio of eye-movement to head-movement velocities. The main problem with a patient's vHIT is that data coming from the goggles may have artifacts and other noise. This paper proposes an impulse classification network (ICN) using a one-dimensional convolutional neural network that can detect noisy data and classify human VOR impulses. Our ICN found actual classes of a patient's impulses with 95% accuracy.Clinical Relevance-ICN is a high-performance classification method that works on a patient's vHIT impulse data by identifying abnormalities and artifacts. This method is an advanced clinical decision support system that can help doctors quickly make decisions.

A Clinical Decision-Support System Based on Three-Stage Integrated Image Analysis for Diagnosing Lung Disease

Thoracic computed tomography (CT) technology has been used for lung cancer screening in high-risk populations, and this technique is highly effective in the identification of early lung cancer. With the rapid development of intelligent image analysis in the field of medical science and technology, many researchers have proposed computer-aided automatic diagnosis methods for facilitating medical experts in detecting lung nodules. This paper proposes an advanced clinical decision-support system for analyzing chest CT images of lung disease. Three advanced methods are utilized in the proposed system: the three-stage automated segmentation method (TSASM), the discrete wavelet packets transform (DWPT) with singular value decomposition (SVD), and the algorithms of the rough set theory, which comprise a classification-based method. Two collected medical CT image datasets were prepared to evaluate the proposed system. The CT image datasets were labeled (nodule, non-nodule, or inflammation) by experienced radiologists from a regional teaching hospital. According to the results, the proposed system outperforms other classification methods (trees, naïve Bayes, multilayer perception, and sequential minimal optimization) in terms of classification accuracy and can be employed as a clinical decision-support system for diagnosing lung disease.

Enabling the integration of clinical event and physiological data for real-time and retrospective analysis

There is a growing trend of developing advanced clinical decision support systems that analyze physiological data streams for early detection of a variety of clinical diagnoses. This paper presents prototype architecture for both a real-time mobile clinical event data capture application and an Artemis-based replay system for retrospective analysis and validation of physiological data analytics. These two components provide important information for improving the ability of clinical decision support systems and patient monitoring algorithms to detect and adjust for artifacts caused by clinical events. A description of the prototypes, as well as results from initial prototype testing is provided. Although the sample size is small for the initial testing, significant information with respect to design principles and infrastructure needs were uncovered. Future research directions are identified to improve the mobile application through increased security, robustness, further integration into data mining analysis, and future clinical decision support algorithms.

Analysis of Potential Drug‐Drug Interactions in Medical Intensive Care Unit Patients

To describe the frequency and type of potential drug-drug interactions (pDDIs) in a general intensive care unit (ICU) and to make recommendations to improve the management of these pDDIs. Retrospective observational study. General ICU of a tertiary care hospital. All patients admitted for more than 24hours between May 2009 and December 2010 who were prescribed at least one medication. Based on the G-Standaard, the Dutch national drug database, pDDIs were identified and classified into categories of potential clinical outcome and management advice. In total, 35,784 medication episodes were identified, resulting in 2887 pDDIs (8.1%). These 2887 pDDIs occurred in 1659 patients for a mean frequency of 1.7 (95% confidence interval [CI] 1.6-1.9) pDDIs per patient. Overall, 54% of the patients experienced at least one pDDI with pDDIs present during 27% of all ICU admission days. All pDDIs could be reconstructed using 81 of the 358 (23%) relevant unique pDDI pairs described in the G-Standaard. The most frequently occurring potential clinical consequence was an increased risk of side effects or toxicity (91% of the pDDIs) such as electrolyte disturbances and masking of hypoglycemia. The most important advised management strategy was monitoring (81%), consisting of monitoring of laboratory values (52%), clinical monitoring of toxicity or effectiveness (48%), or monitoring of physical parameters such as electrocardiogram and blood pressure (11%). Potential drug-drug interactions occur in 54% of all ICU patients, which is two times more than the rate seen in patients on general wards. A limited set of 20 pDDI pairs is responsible for more than 90% of all pDDIs. Therefore, it is worthwhile to develop guidelines for the management of these specific pDDIs. As the vast majority of the interactions can be managed by monitoring, advanced clinical decision support systems linking laboratory data to prescription data may be an effective risk management strategy.

Physicians’ responses to clinical decision support on an intensive care unit—Comparison of four different alerting methods

In intensive care environments, technology is omnipresent whereby ensuring constant monitoring and the administration of critical drugs to unstable patients. A clinical decision support system (CDSS), with its widespread possibilities, can be a valuable tool in supporting adequate patient care. However, it is still unclear how decision support alerts should be presented to physicians and other medical staff to ensure that they are used most effectively. To determine the effect of four different alert presentation methods on alert compliance after the implementation of an advanced CDSS on the intensive care unit (ICU) in our hospital. A randomized clinical trial was executed from August 2010 till December 2011, which included all patients admitted to the ICU of our hospital. The CDSS applied contained a set of thirteen locally developed clinical rules. The percentage of alert compliance was compared for four alert presentation methods: pharmacy intervention, physician alert list, electronic health record (EHR) section and pop-up alerts. Additionally, surveys were held to determine the method most preferred by users of the CDSS. In the study period, the CDSS generated 902 unique alerts, primarily due to drug dosing during decreased renal function and potassium disturbances. Alert compliance was highest for recommendations offered in pop-up alerts (41%, n=68/166), followed by pharmacy intervention (33%, n=80/244), the physician alert list (20%, n=40/199) and the EHR section (19%, n=55/293). The method most preferred by clinicians was pharmacy intervention, and pop-up alerts were found suitable as well if applied correctly. The physician alert list and EHR section were not considered suitable for CDSSs in the process of this study. The alert presentation method used for CDSSs is crucial for the compliance with alerts for the clinical rules and, consequently, for the efficacy of these systems. Active alerts such as pop-ups and pharmacy intervention were more effective than passive alerts, which do not automatically appear within the clinical workflow. In this pilot study, ICU clinicians also preferred pharmacy intervention and pop-up alerts. More research is required to expand these results to other departments and other hospitals, as well as to other types of CDSSs and different alert presentation methods.

A Computerized Adverse Drug Event Alerting System Using Clinical Rules

Adverse drug events (ADEs) are an important problem in hospital practice. Computerized physician order entry (CPOE) and clinical decision support systems (CDSS) are useful tools in the prevention of ADEs. In the Netherlands there are some basic CDSS within CPOE systems, but there is not much experience with sophisticated systems. We have recently developed a more advanced CDSS, a computerized adverse drug event alerting system (ADEAS). The aim of the study was to compare the newly developed ADEAS, which uses a set of clinical rules, with the conventional medication surveillance, a basic CDSS within a CPOE, to assess its additional value in detecting patients with a potential ADE. Leiden University Medical Center (LUMC), a university hospital in Leiden, the Netherlands. Two studies were carried out; one retrospective and one prospective. The retrospective comparison of ADEAS with conventional medication surveillance was conducted on all patients admitted to the hospital (except intensive care unit patients) during a 1-month period (15 November-15 December 2006). A prospective comparison of both systems was performed during a 6-month period (May-October 2007) on one general internal medicine ward. The endpoint was the total number of alerts and content of alerts generated by both methods. In the prospective study we also focused on the number of unique alerts and interventions by the hospital pharmacist following the alerts. In the retrospective study, ADEAS generated 2010 alerts compared with 2322 generated by the conventional method. In the prospective study, 248 and 177 alerts were generated by ADEAS and the conventional method, respectively. The number of unique alerts was 85 (of which 72 were considered true positive alerts) and 136, respectively. The hospital pharmacist made 14 (19.4%) interventions following a true positive alert with ADEAS and 5 (3.7%) with the conventional method. The contents of alerts generated by ADEAS were different to the safety alerts generated by conventional medication surveillance. The conventional medication surveillance generated safety alerts regarding drug-drug interactions and drug-overdosing. ADEAS generated alerts regarding declined renal function or other laboratory abnormalities and absence of essential concurrent medication. Compared with our conventional medication surveillance, the computerized alert system ADEAS selected different patients at risk for an ADE. This makes ADEAS in our hospital of additional value to the hospital pharmacist as a suitable tool in reducing the number of preventable ADEs.