Development and validation of a clinical decision support system to prevent anticoagulant duplications.

BACKGROUND AND OBJECTIVE: Unintended duplicate prescriptions of anticoagulants increase the risk of serious adverse events. Clinical Decision Support Systems (CDSSs) can help prevent such medication errors; however, sophisticated algorithms are needed to avoid alert fatigue. This article describes the steps taken in our hospital to develop a CDSS to prevent anticoagulant duplication (AD). METHODS: The project was composed of three phases. In phase I, the status quo was established. In phase II, a clinical pharmacist developed an algorithm to detect ADs using daily data exports. In phase III, the algorithm was integrated into the hospital's electronic health record system. Alerts were reviewed by clinical pharmacists before being sent to the prescribing physician. We conducted a retrospective analysis of all three phases to assess the impact of the interventions on the occurrence and duration of ADs. Phase III was analyzed in more detail regarding the acceptance rate, sensitivity, and specificity of the alerts. RESULTS: We identified 91 ADs in 1581 patients receiving two or more anticoagulants during phase I, 70 ADs in 1692 patients in phase II, and 57 ADs in 1575 patients in phase III. Mean durations of ADs were 1.8, 1.4, and 1.1 calendar days during phases I, II, and III, respectively. In comparison to the baseline in phase I, the relative risk reduction of AD in patients treated with at least two different anticoagulants during phase III was 42% (RR: 0.58, CI: 0.42-0.81). A total of 429 alerts were generated during phase III, many of which were self-limiting, and 186 alerts were sent to the respective prescribing physician. The acceptance rate was high at 97%. We calculated a sensitivity of 87.4% and a specificity of 87.9%. CONCLUSION: The stepwise development of a CDSS for the detection of AD markedly reduced the frequency and duration of medication errors in our hospital, thereby improving patient safety.

Clinical decision support system supported interventions in hospitalized older patients: a matter of natural course and adequate timing.

BACKGROUND: Drug-related problems (DRPs) and potentially inappropriate prescribing (PIP) are associated with adverse patient and health care outcomes. In the setting of hospitalized older patients, Clinical Decision Support Systems (CDSSs) could reduce PIP and therefore improve clinical outcomes. However, prior research showed a low proportion of adherence to CDSS recommendations by clinicians with possible explanatory factors such as little clinical relevance and alert fatigue. OBJECTIVE: To investigate the use of a CDSS in a real-life setting of hospitalized older patients. We aim to (I) report the natural course and interventions based on the top 20 rule alerts (the 20 most frequently generated alerts per clinical rule) of generated red CDSS alerts (those requiring action) over time from day 1 to 7 of hospitalization; and (II) to explore whether an optimal timing can be defined (in terms of day per rule). METHODS: All hospitalized patients aged ≥ 60 years, admitted to Zuyderland Medical Centre (the Netherlands) were included. The evaluation of the CDSS was investigated using a database used for standard care. Our CDSS was run daily and was evaluated on day 1 to 7 of hospitalization. We collected demographic and clinical data, and moreover the total number of CDSS alerts; the total number of top 20 rule alerts; those that resulted in an action by the pharmacist and the course of outcome of the alerts on days 1 to 7 of hospitalization. RESULTS: In total 3574 unique hospitalized patients, mean age 76.7 (SD 8.3) years and 53% female, were included. From these patients, in total 8073 alerts were generated; with the top 20 of rule alerts we covered roughly 90% of the total. For most rules in the top 20 the highest percentage of resolved alerts lies somewhere between day 4 and 5 of hospitalization, after which there is equalization or a decrease. Although for some rules, there is a gradual increase in resolved alerts until day 7. The level of resolved rule alerts varied between the different clinical rules; varying from > 50-70% (potassium levels, anticoagulation, renal function) to less than 25%. CONCLUSION: This study reports the course of the 20 most frequently generated alerts of a CDSS in a setting of hospitalized older patients. We have shown that for most rules, irrespective of an intervention by the pharmacist, the highest percentage of resolved rules is between day 4 and 5 of hospitalization. The difference in level of resolved alerts between the different rules, could point to more or less clinical relevance and advocates further research to explore ways of optimizing CDSSs by adjustment in timing and number of alerts to prevent alert fatigue.

Early Prediction of Mortality for Septic Patients Visiting Emergency Room Based on Explainable Machine Learning: A Real-World Multicenter Study.

BACKGROUND: Worldwide, sepsis is the leading cause of death in hospitals. If mortality rates in patients with sepsis can be predicted early, medical resources can be allocated efficiently. We constructed machine learning (ML) models to predict the mortality of patients with sepsis in a hospital emergency department. METHODS: This study prospectively collected nationwide data from an ongoing multicenter cohort of patients with sepsis identified in the emergency department. Patients were enrolled from 19 hospitals between September 2019 and December 2020. For acquired data from 3,657 survivors and 1,455 deaths, six ML models (logistic regression, support vector machine, random forest, extreme gradient boosting [XGBoost], light gradient boosting machine, and categorical boosting [CatBoost]) were constructed using fivefold cross-validation to predict mortality. Through these models, 44 clinical variables measured on the day of admission were compared with six sequential organ failure assessment (SOFA) components (PaO2/FIO2 [PF], platelets (PLT), bilirubin, cardiovascular, Glasgow Coma Scale score, and creatinine). The confidence interval (CI) was obtained by performing 10,000 repeated measurements via random sampling of the test dataset. All results were explained and interpreted using Shapley's additive explanations (SHAP). RESULTS: Of the 5,112 participants, CatBoost exhibited the highest area under the curve (AUC) of 0.800 (95% CI, 0.756-0.840) using clinical variables. Using the SOFA components for the same patient, XGBoost exhibited the highest AUC of 0.678 (95% CI, 0.626-0.730). As interpreted by SHAP, albumin, lactate, blood urea nitrogen, and international normalization ratio were determined to significantly affect the results. Additionally, PF and PLTs in the SOFA component significantly influenced the prediction results. CONCLUSION: Newly established ML-based models achieved good prediction of mortality in patients with sepsis. Using several clinical variables acquired at the baseline can provide more accurate results for early predictions than using SOFA components. Additionally, the impact of each variable was identified.

A quality improvement study of the implementation and initial results of a pragmatic clinical decision support system in the community pharmacy setting.

BACKGROUND: A six year collaboration between academics, community pharmacists and informaticians, led to the development of nine guidelines for a clinical decision support system, enhancing community pharmacists' ability to address drug-related problems and improve care. AIM: The objective of this study was to assess the effectiveness of clinical decision support system rules in enhancing medication management within the community pharmacy setting. This was achieved through retrospective monitoring of real-world usage and measuring the pharmacotherapeutic impact of the rules. METHOD: In 2019, a retrospective observational evaluation appraised the acceptance rate of the clinical decision support system components in 490 Belgian pharmacies. Among these, 51 pharmacies underwent a longitudinal analysis involving (i) co-prescription of methotrexate and folic acid, (ii) gastroprotection with non-steroidal anti-inflammatory drugs, and (iii) drug combinations causing QT prolongation. The study period spanned one year pre-launch, one year post-launch, and two years post-launch. RESULTS: Of the targeted pharmacies, 80% used 7 of the 9 rules. After four years, methotrexate-folic acid co-prescription increased 4%, reaching 79.8%. Gastroprotection improved by 3% among older patients and 7.47% in younger individuals (< 70 year) with multiple risk factors. The QT prolongation rules faced implementation difficulties. CONCLUSION: Pharmacists' acceptance of the developed rules was high and coincided with a decline in drug-related problems, holding potential public health impact. This real-world data can inform the future implementation of such systems, as it demonstrated the need for more detailed data-gathering and more intensive training of pharmacists in the handling of more complex problems such as QT prolongation.

Design and implementation of an automatic nursing assessment system based on CDSS technology.

BACKGROUND: Various quantitative and quality assessment tools are currently used in nursing to evaluate a patient's physiological, psychological, and socioeconomic status. The results play important roles in evaluating the efficiency of healthcare, improving the treatment plans, and lowing relevant clinical risks. However, the manual process of the assessment imposes a substantial burden and can lead to errors in digitalization. To fill these gaps, we proposed an automatic nursing assessment system based on clinical decision support system (CDSS). The framework underlying the CDSS included experts, evaluation criteria, and voting roles for selecting electronic assessment sheets over paper ones. METHODS: We developed the framework based on an expert voting flow to choose electronic assessment sheets. The CDSS was constructed based on a nursing process workflow model. A multilayer architecture with independent modules was used. The performance of the proposed system was evaluated by comparing the adverse events' incidence and the average time for regular daily assessment before and after the implementation. RESULTS: After implementation of the system, the adverse nursing events' incidence decreased significantly from 0.43 % to 0.37 % in the first year and further to 0.27 % in the second year (p-value: 0.04). Meanwhile, the median time for regular daily assessments further decreased from 63 s to 51 s. CONCLUSIONS: The automatic assessment system helps to reduce nurses' workload and the incidence of adverse nursing events.

On Scene Injury Severity Prediction (OSISP) model for trauma developed using the Swedish Trauma Registry.

BACKGROUND: Providing optimal care for trauma, the leading cause of death for young adults, remains a challenge e.g., due to field triage limitations in assessing a patient's condition and deciding on transport destination. Data-driven On Scene Injury Severity Prediction (OSISP) models for motor vehicle crashes have shown potential for providing real-time decision support. The objective of this study is therefore to evaluate if an Artificial Intelligence (AI) based clinical decision support system can identify severely injured trauma patients in the prehospital setting. METHODS: The Swedish Trauma Registry was used to train and validate five models - Logistic Regression, Random Forest, XGBoost, Support Vector Machine and Artificial Neural Network - in a stratified 10-fold cross validation setting and hold-out analysis. The models performed binary classification of the New Injury Severity Score and were evaluated using accuracy metrics, area under the receiver operating characteristic curve (AUC) and Precision-Recall curve (AUCPR), and under- and overtriage rates. RESULTS: There were 75,602 registrations between 2013-2020 and 47,357 (62.6%) remained after eligibility criteria were applied. Models were based on 21 predictors, including injury location. From the clinical outcome, about 40% of patients were undertriaged and 46% were overtriaged. Models demonstrated potential for improved triaging and yielded AUC between 0.80-0.89 and AUCPR between 0.43-0.62. CONCLUSIONS: AI based OSISP models have potential to provide support during assessment of injury severity. The findings may be used for developing tools to complement field triage protocols, with potential to improve prehospital trauma care and thereby reduce morbidity and mortality for a large patient population.

Predicting Successful Placements for Youth in Child Welfare with Machine Learning.

Out-of-home placement decisions have extremely high stakes for the present and future well-being of children in care because some placement types, and multiple placements, are associated with poor outcomes. We propose that a clinical decision support system (CDSS) using existing data about children and their previous placement success could inform future placement decision-making for their peers. The objective of this study was to test the feasibility of developing machine learning models to predict the best level of care placement (i.e., the placement with the highest likelihood of doing well in treatment) based on each youth's behavioral health needs and characteristics. We developed machine learning models to predict the probability of each youth's treatment success in psychiatric residential care (i.e., Psychiatric Residential Treatment Facility [PRTF]) versus any other placement (AUROCs > 0.70) using data collected in standard care at a behavioral health organization. Placement recommendations based on these machine learning models distinguished between youth who did well in residential care versus non-residential care (e.g., 80% of those who received care in the recommended setting with the highest predicted likelihood of success had above average risk-adjusted outcomes). Then we developed and validated machine learning models to predict the probability of each youth's treatment success across specific placement types in a state-wide system, achieving an average AUROC score of greater than 0.75. Machine learning models based on risk-adjusted behavioral health and functional data show promise in predicting positive placement outcomes and informing future placement decisions for youth in care. Related ethical considerations are discussed.

Mejora en el manejo de la sospecha del síndrome aórtico agudo en urgencias mediante un algoritmo clínico y el estudio de factores predictivos / Improvement in the management of suspected acute aortic syndrome in the emergency room through a clinical algorithm and study of predictive factors

Antecedentes y objetivo: El síndrome aórtico agudo (SAA) es poco frecuente y difícil de diagnosticar, con una gran variabilidad en su cuadro clínico inicial. Los objetivos son: 1) desarrollar un algoritmo informático, o un sistema de apoyo a las decisiones clínicas (SADC), para el manejo y la solicitud de estudios de diagnóstico por imagen en el servicio de Urgencias, en concreto de una tomografía computarizada (TC) de la aorta, ante la sospecha de SAA, 2) determinar el efecto de la implantación de este sistema, y 3) determinar los factores asociados a un diagnóstico radiológico positivo que mejoren la capacidad predictiva de los hallazgos de la TC de aorta. Material y métodos: Tras desarrollar e implementar un algoritmo basado en la evidencia, se estudiaron casos de sospecha de SAA. Se utilizó el test de la χ2 para analizar la asociación entre las variables incluidas en el algoritmo y el diagnóstico radiológico, con 3 categorías: sin hallazgos relevantes, positivo para SAA y diagnósticos alternativos. Resultados: Se identificaron 130 solicitudes; 19 (14,6%) tenían SAA y 34 (26,2%) tenían otra patología aguda. De las 19 con SAA, 15 habían sido estratificadas como de alto riesgo y 4 como de riesgo intermedio. La probabilidad de SAA era 3,4 veces mayor en los pacientes con aneurisma aórtico conocido (p=0,021, IC del 95%: 1,2-9,6) y 5,1 veces mayor en los pacientes con un soplo de insuficiencia vascular aórtica de novo(p=0,019, IC del 95 %: 1,3-20,1). La probabilidad de tener una enfermedad aguda grave alternativa fue 3,2 veces mayor en los pacientes con hipotensión o choque (p=0,02, IC del 95 %: 1,2-8,5). Conclusión: El uso de un SADC en el servicio de Urgencias puede ayudar a optimizar el diagnóstico del SAA. Se demostró que la presencia de un aneurisma aórtico conocido y de insuficiencia valvular aórtica de nueva aparición aumentan significativamente la probabilidad de SAA. Se necesitan más estudios para establecer una regla de predicción clínica.(AU)

Background and objective: Acute aortic syndrome (AAS) is uncommon and difficult to diagnose, with great variability in clinical presentation. To develop a computerized algorithm, or clinical decision support system (CDSS), for managing and requesting imaging in the emergency department, specifically computerized tomography of the aorta (CTA), when there is suspicion of AAS, and to determine the effect of implementing this system. To determine the factors associated with a positive radiological diagnosis that improve the predictive capacity of CTA findings. Materials and methods: After developing and implementing an evidence-based algorithm, we studied suspected cases of AAS. Chi-squared test was used to analyze the association between the variables included in the algorithm and radiological diagnosis, with 3 categories: no relevant findings, positive for AAS, and alternative diagnoses. Results: 130 requests were identified; 19 (14.6%) had AAS and 34 (26.2%) had a different acute pathology. Of the 19 with AAS, 15 had been stratified as high risk and 4 as intermediate risk. The probability of AAS was 3.4 times higher in patients with known aortic aneurysm (P=.021, 95% CI 1.2–9.6) and 5.1 times higher in patients with a new aortic regurgitation murmur (P=.019, 95% CI 1.3–20.1). The probability of having an alternative severe acute pathology was 3.2 times higher in patients with hypotension or shock (P=.02, 95% CI 1.2–8.5). Conclusion: The use of a CDSS in the emergency department can help optimize AAS diagnosis. The presence of a known aortic aneurysm and new-onset aortic regurgitation were shown to significantly increase the probability of AAS. Further studies are needed to establish a clinical prediction rule.(AU)

Improvement in the management of suspected acute aortic syndrome in the emergency room through a clinical algorithm and study of predictive factors.

BACKGROUND AND OBJECTIVE: Acute aortic syndrome (AAS) is uncommon and difficult to diagnose, with great variability in clinical presentation. To develop a computerized algorithm, or clinical decision support system (CDSS), for managing and requesting imaging in the emergency department, specifically computerized tomography of the aorta (CTA), when there is suspicion of AAS, and to determine the effect of implementing this system. To determine the factors associated with a positive radiological diagnosis that improve the predictive capacity of CTA findings. MATERIALS AND METHODS: After developing and implementing an evidence-based algorithm, we studied suspected cases of AAS. Chi-squared test was used to analyze the association between the variables included in the algorithm and radiological diagnosis, with 3 categories: no relevant findings, positive for AAS, and alternative diagnoses. RESULTS: 130 requests were identified; 19 (14.6%) had AAS and 34 (26.2%) had a different acute pathology. Of the 19 with AAS, 15 had been stratified as high risk and 4 as intermediate risk. The probability of AAS was 3.4 times higher in patients with known aortic aneurysm (P = .021, 95% CI 1.2-9.6) and 5.1 times higher in patients with a new aortic regurgitation murmur (P = .019, 95% CI 1.3-20.1). The probability of having an alternative severe acute pathology was 3.2 times higher in patients with hypotension or shock (P = .02, 95% CI 1.2-8.5). CONCLUSION: The use of a CDSS in the emergency department can help optimize AAS diagnosis. The presence of a known aortic aneurysm and new-onset aortic regurgitation were shown to significantly increase the probability of AAS. Further studies are needed to establish a clinical prediction rule.

Acceptance of a digital therapy recommender system for psoriasis.

BACKGROUND: About 2% of the German population are affected by psoriasis. A growing number of cost-intensive systemic treatments are available. Surveys have shown high proportions of patients with moderate to severe psoriasis are not adequately treated despite a high disease burden. Digital therapy recommendation systems (TRS) may help implement guideline-based treatment. However, little is known about the acceptance of such clinical decision support systems (CDSSs). Therefore, the aim of the study was to access the acceptance of a prototypical TRS demonstrator. METHODS: Three scenarios (potential test patients with psoriasis but different sociodemographic and clinical characteristics, previous treatments, desire to have children, and multiple comorbidities) were designed in the demonstrator. The TRS demonstrator and test patients were presented to a random sample of 76 dermatologists attending a national dermatology conference in a cross-sectional face-to-face survey with case vignettes. The dermatologist were asked to rate the demonstrator by system usability scale (SUS), whether they would use it for certain patients populations and barriers of usage. Reasons for potential usage of the TRS demonstrator were tested via a Poisson regression with robust standard errors. RESULTS: Acceptance of the TRS was highest for patients eligible for systemic therapy (82%). 50% of participants accepted the system for patients with additional comorbidities and 43% for patients with special subtypes of psoriasis. Dermatologists in the outpatient sector or with many patients per week were less willing to use the TRS for patients with special psoriasis-subtypes. Dermatologists rated the demonstrator as acceptable with an mean SUS of 76.8. Participants whose SUS was 10 points above average were 27% more likely to use TRS for special psoriasis-subtypes. The main barrier in using the TRS was time demand (47.4%). Participants who perceived time as an obstacle were 22.3% less willing to use TRS with systemic therapy patients. 27.6% of physicians stated that they did not understand exactly how the recommendation was generated by the TRS, with no effect on the preparedness to use the system. CONCLUSION: The considerably high acceptance and the preparedness to use the psoriasis CDSS suggests that a TRS appears to be implementable in routine healthcare and may improve clinical care. Main barrier is the additional time demand posed on dermatologists in a busy clinical setting. Therefore, it will be a major challenge to identify a limited set of variables that still allows a valid recommendation with precise prediction of the patient-individual benefits and harms.

Multimodality Video Acquisition System for the Assessment of Vital Distress in Children.

In children, vital distress events, particularly respiratory, go unrecognized. To develop a standard model for automated assessment of vital distress in children, we aimed to construct a prospective high-quality video database for critically ill children in a pediatric intensive care unit (PICU) setting. The videos were acquired automatically through a secure web application with an application programming interface (API). The purpose of this article is to describe the data acquisition process from each PICU room to the research electronic database. Using an Azure Kinect DK and a Flir Lepton 3.5 LWIR attached to a Jetson Xavier NX board and the network architecture of our PICU, we have implemented an ongoing high-fidelity prospectively collected video database for research, monitoring, and diagnostic purposes. This infrastructure offers the opportunity to develop algorithms (including computational models) to quantify vital distress in order to evaluate vital distress events. More than 290 RGB, thermographic, and point cloud videos of each 30 s have been recorded in the database. Each recording is linked to the patient's numerical phenotype, i.e., the electronic medical health record and high-resolution medical database of our research center. The ultimate goal is to develop and validate algorithms to detect vital distress in real time, both for inpatient care and outpatient management.

Learning meaningful latent space representations for patient risk stratification: Model development and validation for dengue and other acute febrile illness.

Background: Increased data availability has prompted the creation of clinical decision support systems. These systems utilise clinical information to enhance health care provision, both to predict the likelihood of specific clinical outcomes or evaluate the risk of further complications. However, their adoption remains low due to concerns regarding the quality of recommendations, and a lack of clarity on how results are best obtained and presented. Methods: We used autoencoders capable of reducing the dimensionality of complex datasets in order to produce a 2D representation denoted as latent space to support understanding of complex clinical data. In this output, meaningful representations of individual patient profiles are spatially mapped in an unsupervised manner according to their input clinical parameters. This technique was then applied to a large real-world clinical dataset of over 12,000 patients with an illness compatible with dengue infection in Ho Chi Minh City, Vietnam between 1999 and 2021. Dengue is a systemic viral disease which exerts significant health and economic burden worldwide, and up to 5% of hospitalised patients develop life-threatening complications. Results: The latent space produced by the selected autoencoder aligns with established clinical characteristics exhibited by patients with dengue infection, as well as features of disease progression. Similar clinical phenotypes are represented close to each other in the latent space and clustered according to outcomes broadly described by the World Health Organisation dengue guidelines. Balancing distance metrics and density metrics produced results covering most of the latent space, and improved visualisation whilst preserving utility, with similar patients grouped closer together. In this case, this balance is achieved by using the sigmoid activation function and one hidden layer with three neurons, in addition to the latent dimension layer, which produces the output (Pearson, 0.840; Spearman, 0.830; Procrustes, 0.301; GMM 0.321). Conclusion: This study demonstrates that when adequately configured, autoencoders can produce two-dimensional representations of a complex dataset that conserve the distance relationship between points. The output visualisation groups patients with clinically relevant features closely together and inherently supports user interpretability. Work is underway to incorporate these findings into an electronic clinical decision support system to guide individual patient management.

Usability of the IDDEAS prototype in child and adolescent mental health services: A qualitative study for clinical decision support system development.

Introduction: Child and adolescent mental health services (CAMHS) clinical decision support system (CDSS) provides clinicians with real-time support as they assess and treat patients. CDSS can integrate diverse clinical data for identifying child and adolescent mental health needs earlier and more comprehensively. Individualized Digital Decision Assist System (IDDEAS) has the potential to improve quality of care with enhanced efficiency and effectiveness. Methods: We examined IDDEAS usability and functionality in a prototype for attention deficit hyperactivity disorder (ADHD), using a user-centered design process and qualitative methods with child and adolescent psychiatrists and clinical psychologists. Participants were recruited from Norwegian CAMHS and were randomly assigned patient case vignettes for clinical evaluation, with and without IDDEAS. Semi-structured interviews were conducted as one part of testing the usability of the prototype following a five-question interview guide. All interviews were recorded, transcribed, and analyzed following qualitative content analysis. Results: Participants were the first 20 individuals from the larger IDDEAS prototype usability study. Seven participants explicitly stated a need for integration with the patient electronic health record system. Three participants commended the step-by-step guidance as potentially helpful for novice clinicians. One participant did not like the aesthetics of the IDDEAS at this stage. All participants were pleased about the display of the patient information along with guidelines and suggested that wider guideline coverage will make IDDEAS much more useful. Overall, participants emphasized the importance of maintaining the clinician as the decision-maker in the clinical process, and the overall potential utility of IDDEAS within Norwegian CAMHS. Conclusion: Child and adolescent mental health services psychiatrists and psychologists expressed strong support for the IDDEAS clinical decision support system if better integrated in daily workflow. Further usability assessments and identification of additional IDDEAS requirements are necessary. A fully functioning, integrated version of IDDEAS has the potential to be an important support for clinicians in the early identification of risks for youth mental disorders and contribute to improved assessment and treatment of children and adolescents.

Detection of Drug-Related Problems through a Clinical Decision Support System Used by a Clinical Pharmacy Team.

Clinical decision support systems (CDSSs) are intended to detect drug-related problems in real time and might be of value in healthcare institutions with a clinical pharmacy team. The objective was to report the detection of drug-related problems through a CDSS used by an existing clinical pharmacy team over 22 months. It was a retrospective single-center study. A CDSS was integrated in the clinical pharmacy team in July 2019. The investigating clinical pharmacists evaluated the pharmaceutical relevance and physician acceptance rates for critical alerts (i.e., alerts for drug-related problems arising during on-call periods) and noncritical alerts (i.e., prevention alerts arising during the pharmacist's normal work day) from the CDSS. Of the 3612 alerts triggered, 1554 (43.0%) were critical, and 594 of these 1554 (38.2%) prompted a pharmacist intervention. Of the 2058 (57.0%) noncritical alerts, 475 of these 2058 (23.1%) prompted a pharmacist intervention. About two-thirds of the total pharmacist interventions (PI) were accepted by physicians; the proportion was 71.2% for critical alerts (i.e., 19 critical alerts per month vs. 12.5 noncritical alerts per month). Some alerts were pharmaceutically irrelevant-mainly due to poor performance by the CDSS. Our results suggest that a CDSS is a useful decision-support tool for a hospital pharmacist's clinical practice. It can help to prioritize drug-related problems by distinguishing critical and noncritical alerts. However, building an appropriate organizational structure around the CDSS is important for correct operation.

Challenge in hyponatremic patients - the potential of a laboratory-based decision support system for hyponatremia to improve patient's safety.

OBJECTIVES: Hyponatremia is the most frequent electrolyte disorder in hospitalized patients with increased mortality and morbidity. In this study, we evaluated the follow-up diagnostic, the risk of inadequate fast correction and the outcome of patients with profound hyponatremia (pHN), defined as a blood sodium concentration below 120 mmol/L. The aim was to identify a promising approach for a laboratory-based clinical decision support system (CDSS). METHODS: This retrospective study included 378,980 blood sodium measurements of 83,315 cases at a German tertiary care hospital. Hospitalized cases with pHN (n=211) were categorized into two groups by the time needed for a follow-up measurement to be performed (time to control, TTC) as either <12 h (group 1: "TTC≤12 h", n=118 cases) or >12 h (group 2: "TTC>12 h", n=93 cases). Length of hospital stay, sodium level at discharge, ward transfers, correction of hyponatremia, and risk of osmotic demyelination syndrome (ODS) due to inadequate fast correction were evaluated with regard to the TTC of sodium blood concentration. RESULTS: pHN was detected in 1,050 measurements (0.3%) in 211 cases. Cases, in which follow-up diagnostics took longer (TTC>12 h), achieved a significantly lower sodium correction during their hospitalization (11.2 vs. 16.7 mmol/L, p<0.001), were discharged more frequently in hyponatremic states (<135 mmol/L; 58 (62.4%) vs. 43 (36.4%), p<0.001) and at lower sodium blood levels (131.2 vs. 135.0 mmol/L, p<0.001). Furthermore, for these patients there was a trend toward an increased length of hospital stay (13.1 vs. 8.5 days, p=0.089), as well as an increased risk of inadequate fast correction (p<0.001). CONCLUSIONS: Our study shows that less frequent follow-up sodium measurements in pHN are associated with worse outcomes. Patients with a prolonged TTC are at risk of insufficient correction of hyponatremia, reduced sodium values at discharge, and possible overcorrection. Our results suggest that a CDSS that alerts treating physicians when a control time of >12 h is exceeded could improve patient care in the long term. We are initiating a prospective study to investigate the benefits of our self-invented CDSS (www.ampel.care) for patients with pHN.

Patient views on asthma diagnosis and how a clinical decision support system could help: A qualitative study.

INTRODUCTION: Making a diagnosis of asthma can be challenging for clinicians and patients. A clinical decision support system (CDSS) for use in primary care including a patient-facing mode, could change how information is shared between patients and healthcare professionals and improve the diagnostic process. METHODS: Participants diagnosed with asthma within the last 5 years were recruited from general practices across four UK regions. In-depth interviews were used to explore patient experiences relating to their asthma diagnosis and to understand how a CDSS could be used to improve the diagnostic process for patients. Interviews were audio recorded, transcribed verbatim and analysed using a thematic approach. RESULTS: Seventeen participants (12 female) undertook interviews, including 14 individuals and 3 parents of children with asthma. Being diagnosed with asthma was generally considered an uncertain process. Participants felt a lack of consultation time and poor communication affected their understanding of asthma and what to expect. Had the nature of asthma and the steps required to make a diagnosis been explained more clearly, patients felt their understanding and engagement in asthma self-management could have been improved. Participants considered that a CDSS could provide resources to support the diagnostic process, prompt dialogue, aid understanding and support shared decision-making. CONCLUSION: Undergoing an asthma diagnosis was uncertain for patients if their ideas and concerns were not addressed by clinicians and were influenced by a lack of consultation time and limitations in communication. An asthma diagnosis CDSS could provide structure and an interface to prompt dialogue, provide visuals about asthma to aid understanding and encourage patient involvement. PATIENT AND PUBLIC CONTRIBUTION: Prespecified semistructured interview topic guides (young person and adult versions) were developed by the research team and piloted with members of the Asthma UK Centre for Applied Research Patient and Public Involvement (PPI) group. Findings were regularly discussed within the research group and with PPI colleagues to aid the interpretation of data.

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

Background: 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. Methods: 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. Results: 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. Conclusion: 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.

Preliminary evaluation of the practicability of a clinical decision support system for diagnosis and treatment of thyroid cancer (CDSS-TC) / 中华内分泌外科杂志

Objective:To study the practical efficacy of the clinical decision support system for diagnosis and treatment of thyroid cancer (CDSS-TC) in assisting doctors to complete several diagnosis and treatment tasks, and to make a preliminary evaluation of its clinical practicability according to the test results.Methods:From Jan. 2022 to Mar. 2022, 90 patients with thyroid cancer who were admitted to the Head and Neck Surgery Department of Shaw Hospital affiliated to Zhejiang University were prospectively analyzed, and the average time spent in reading the pre-operative B-ultrasound report, as well as the individual fitness of the dose adjustment of eugenol in 70 patients with thyroid cancer after surgery. A retrospective analysis was made of the compliance of the basis of the "recommended scheme" and the deviation of the basis of the doctor’s "final scheme" for the preoperative surgery of 120 patients with thyroid cancer who were treated for the first time in the head and neck surgery of Shaw Hospital affiliated to Zhejiang University from Mar. 2021 to May. 2021. All cases were treated by pure artificial (group A) and CDSS-TC assisted (group B) , and the differences in organization were compared.Results:The average time for disposal of a single B-ultrasound report in Group B was much shorter than that in Group A ( P=5.600E-04) ; The number of patients with excellent grade and the total number of patients with excellent grade and qualified grade recommended by the doctor in group B were significantly higher than those in group A ( P=7.819E-20 and P=1.335E-18) ; The conformity rate of the basis of CDSS-TC "Recommended Scheme" ≥ 98%; The deviation rate of the basis for "final protocol" of doctors in group B was lower than that in group A ( P=0.059 for total resection or not, P=0.075 for lateral neck dissection or not) . Conclusions:CDSS-TC can accurately extract the disease-related source information in all the original examination/laboratory reports, and provide accurate decision-making suggestions through efficient correlation analysis. In view of the accurate and objective conclusions of its analysis, it can provide high-quality and all-link decision support for doctors’ clinical diagnosis and treatment, and is an ideal information work platform.

A Clinical Decision Support System for Diabetes Patients with Deep Learning: Experience of a Taiwan Medical Center.

Background: Diabetes mellitus (DM) is a major public health problem worldwide. It involves dysfunction of blood sugar regulation resulting from insulin resistance, inadequate insulin secretion, or excessive glucagon secretion. Methods: This study collated 971,401 drug usage records of 51,009 DM patients. These data include patient identification code, age, gender, outpatient visiting dates, visiting code, medication features (included items, doses, and frequencies of drugs), HbA1c results, and testing time. We apply a random forest (RF) model for feature selection and implement a regression model with the bidirectional long short-term memory (Bi-LSTM) deep learning architecture. Finally, we use the root mean square error (RMSE) as the evaluation index for the prediction model. Results: After data cleaning, the data included 8,729 male and 9,115 female cases. Metformin was the most important feature suggested by the RF model, followed by glimepiride, acarbose, pioglitazone, glibenclamide, gliclazide, repaglinide, nateglinide, sitagliptin, and vildagliptin. The model performed better with the past two seasons in the training data than with additional seasons. Further, the Bi-LSTM architecture model performed better than support vector machines (SVMs). Discussion & Conclusion: This study found that Bi-LSTM models is a well kernel in a CDSS which help physicians' decision-making, and the increasing the number of seasons will negative impact the performance. In addition, this study found that the most important drug is metformin, which is recommended as first-line treatment OHA in various situations for DM patients.

Development and assessment of PharmaCheck: an electronic screening tool for the prevention of twenty major adverse drug events.

BACKGROUND: Adverse drug events (ADEs) can be prevented by deploying clinical decision support systems (CDSS) that directly assist physicians, via computerized order entry systems, and clinical pharmacists performing medication reviews as part of medical rounds. However, physicians using CDSS are known to be exposed to the alert-fatigue phenomenon. Our study aimed to assess the performance of PharmaCheck-a CDSS to help clinical pharmacists detect high-risk situations with the potential to lead to ADEs-and its impact on clinical pharmacists' activities. METHODS: Twenty clinical rules, divided into four risk classes, were set for the daily screening of high-risk situations in the electronic health records of patients admitted to our General Internal Medicine Department. Alerts to clinical pharmacists encouraged them to telephone prescribers and suggest any necessary treatment adjustments. PharmaCheck's performance was assessed using the intervention's positive predictive value (PPV), which characterizes the proportion of interventions for each alert triggered. PharmaCheck's impact was assessed by considering clinical pharmacists as a filter for ruling out futile alerts and by comparing the final clinical PPV with a pharmacist (the proportion of interventions that led to a change in the medical regimen) to the final clinical PPV without a pharmacist. RESULTS: Over 132 days, 447 alerts were triggered for 383 patients, leading to 90 interventions (overall intervention PPV = 20.1%). By risk class, intervention PPVs made up 26.9% (n = 65/242) of abnormal laboratory value alerts, 3.1% (4/127) of alerts for contraindicated medications or medications to be used with caution, 28.2% (20/71) of drug-drug interaction alerts, and 14.3% (1/7) of inadequate mode of administration alerts. Clinical PPVs reached 71.0% (64/90) when pharmacists filtered alerts and 14% (64/242) if they were not doing it. CONCLUSION: PharmaCheck enabled clinical pharmacists to improve their traditional processes and broaden their coverage by focusing on 20 high-risk situations. Alert management by pharmacists seemed to be a more effective way of preventing risky situations and alert-fatigue than a model addressing alerts to physicians exclusively. Some fine-tuning could enhance PharmaCheck's performance by considering the information quality of triggers, the variability of clinical settings, and the fact that some prescription processes are already highly secured.