scholarly journals Reinforcement Learning for Clinical Decision Support in Critical Care: Comprehensive Review (Preprint)

2020 ◽  
Author(s):  
Siqi Liu ◽  
Kay Choong See ◽  
Kee Yuan Ngiam ◽  
Leo Anthony Celi ◽  
Xingzhi Sun ◽  
...  
Keyword(s):  
Critical Care ◽  
Reinforcement Learning ◽  
Decision Support ◽  
Clinical Decision Support ◽  
Clinical Decision ◽  
Evaluation Metrics ◽  
Comprehensive Review ◽  
Drug Dosing ◽  
Personalized Care ◽  
Clinical Environments

BACKGROUND Decision support systems based on reinforcement learning (RL) have been implemented to facilitate the delivery of personalized care. This paper aimed to provide a comprehensive review of RL applications in the critical care setting. OBJECTIVE This review aimed to survey the literature on RL applications for clinical decision support in critical care and to provide insight into the challenges of applying various RL models. METHODS We performed an extensive search of the following databases: PubMed, Google Scholar, Institute of Electrical and Electronics Engineers (IEEE), ScienceDirect, Web of Science, Medical Literature Analysis and Retrieval System Online (MEDLINE), and Excerpta Medica Database (EMBASE). Studies published over the past 10 years (2010-2019) that have applied RL for critical care were included. RESULTS We included 21 papers and found that RL has been used to optimize the choice of medications, drug dosing, and timing of interventions and to target personalized laboratory values. We further compared and contrasted the design of the RL models and the evaluation metrics for each application. CONCLUSIONS RL has great potential for enhancing decision making in critical care. Challenges regarding RL system design, evaluation metrics, and model choice exist. More importantly, further work is required to validate RL in authentic clinical environments.

scholarly journals Reinforcement Learning for Clinical Decision Support in Critical Care: Comprehensive Review

10.2196/18477 ◽  
2020 ◽  
Vol 22 (7) ◽  
pp. e18477 ◽  
Author(s):  
Siqi Liu ◽  
Kay Choong See ◽  
Kee Yuan Ngiam ◽  
Leo Anthony Celi ◽  
Xingzhi Sun ◽  
...  
Keyword(s):  
Critical Care ◽  
Reinforcement Learning ◽  
Decision Support ◽  
Clinical Decision Support ◽  
Clinical Decision ◽  
Evaluation Metrics ◽  
Comprehensive Review ◽  
Drug Dosing ◽  
Personalized Care ◽  
Clinical Environments

Background Decision support systems based on reinforcement learning (RL) have been implemented to facilitate the delivery of personalized care. This paper aimed to provide a comprehensive review of RL applications in the critical care setting. Objective This review aimed to survey the literature on RL applications for clinical decision support in critical care and to provide insight into the challenges of applying various RL models. Methods We performed an extensive search of the following databases: PubMed, Google Scholar, Institute of Electrical and Electronics Engineers (IEEE), ScienceDirect, Web of Science, Medical Literature Analysis and Retrieval System Online (MEDLINE), and Excerpta Medica Database (EMBASE). Studies published over the past 10 years (2010-2019) that have applied RL for critical care were included. Results We included 21 papers and found that RL has been used to optimize the choice of medications, drug dosing, and timing of interventions and to target personalized laboratory values. We further compared and contrasted the design of the RL models and the evaluation metrics for each application. Conclusions RL has great potential for enhancing decision making in critical care. Challenges regarding RL system design, evaluation metrics, and model choice exist. More importantly, further work is required to validate RL in authentic clinical environments.

scholarly journals Evidence-based Clinical Decision Support Systems for the prediction and detection of three disease states in critical care: A systematic literature review

F1000Research ◽  
2019 ◽  
Vol 8 ◽  
pp. 1728
Author(s):  
Goran Medic ◽  
Melodi Kosaner Kließ ◽  
Louis Atallah ◽  
Jochen Weichert ◽  
Saswat Panda ◽  
...  
Keyword(s):  
Machine Learning ◽  
Critical Care ◽  
Decision Support ◽  
Respiratory Distress ◽  
Decision Support Systems ◽  
Clinical Decision Support ◽  
Clinical Decision Support Systems ◽  
Clinical Decision ◽  
Hemodynamic Instability ◽  
Evidence Based

Background: Clinical decision support (CDS) systems have emerged as tools providing intelligent decision making to address challenges of critical care. CDS systems can be based on existing guidelines or best practices; and can also utilize machine learning to provide a diagnosis, recommendation, or therapy course. Methods: This research aimed to identify evidence-based study designs and outcome measures to determine the clinical effectiveness of clinical decision support systems in the detection and prediction of hemodynamic instability, respiratory distress, and infection within critical care settings. PubMed, ClinicalTrials.gov and Cochrane Database of Systematic Reviews were systematically searched to identify primary research published in English between 2013 and 2018. Studies conducted in the USA, Canada, UK, Germany and France with more than 10 participants per arm were included. Results: In studies on hemodynamic instability, the prediction and management of septic shock were the most researched topics followed by the early prediction of heart failure. For respiratory distress, the most popular topics were pneumonia detection and prediction followed by pulmonary embolisms. Given the importance of imaging and clinical notes, this area combined Machine Learning with image analysis and natural language processing. In studies on infection, the most researched areas were the detection, prediction, and management of sepsis, surgical site infections, as well as acute kidney injury. Overall, a variety of Machine Learning algorithms were utilized frequently, particularly support vector machines, boosting techniques, random forest classifiers and neural networks. Sensitivity, specificity, and ROC AUC were the most frequently reported performance measures. Conclusion: This review showed an increasing use of Machine Learning for CDS in all three areas. Large datasets are required for training these algorithms; making it imperative to appropriately address, challenges such as class imbalance, correct labelling of data and missing data. Recommendations are formulated for the development and successful adoption of CDS systems.

scholarly journals Renal Drug Dosing

2016 ◽  
Vol 07 (03) ◽  
pp. 731-744 ◽  
Author(s):  
Erin Vogel ◽  
Sarah Billups ◽  
Sheryl Herner ◽  
Thomas Delate
Keyword(s):  
Renal Function ◽  
Decision Support ◽  
Renal Impairment ◽  
Clinical Decision Support ◽  
Error Rate ◽  
Dose Adjustment ◽  
Clinical Decision Support Systems ◽  
Clinical Decision ◽  
Error Rates ◽  
Drug Dosing

SummaryThe purpose of this study was to compare the effectiveness of an outpatient renal dose adjustment alert via a computerized provider order entry (CPOE) clinical decision support system (CDSS) versus a CDSS with alerts made to dispensing pharmacists.This was a retrospective analysis of patients with renal impairment and 30 medications that are contraindicated or require dose-adjustment in such patients. The primary outcome was the rate of renal dosing errors for study medications that were dispensed between August and December 2013, when a pharmacist-based CDSS was in place, versus August through December 2014, when a prescriber-based CDSS was in place. A dosing error was defined as a prescription for one of the study medications dispensed to a patient where the medication was contraindicated or improperly dosed based on the patient’s renal function. The denominator was all prescriptions for the study medications dispensed during each respective study period.During the pharmacist-and prescriber-based CDSS study periods, 49,054 and 50,678 prescriptions, respectively, were dispensed for one of the included medications. Of these, 878 (1.8%) and 758 (1.5%) prescriptions were dispensed to patients with renal impairment in the respective study periods. Patients in each group were similar with respect to age, sex, and renal function stage. Overall, the five-month error rate was 0.38%. Error rates were similar between the two groups: 0.36% and 0.40% in the pharmacist-and prescriber-based CDSS, respectively (p=0.523). The medication with the highest error rate was dofetilide (0.51% overall) while the medications with the lowest error rate were dabigatran, fondaparinux, and spironolactone (0.00% overall).Prescriber-and pharmacist-based CDSS provided comparable, low rates of potential medication errors. Future studies should be undertaken to examine patient benefits of the prescriber-based CDSS. Citation: Vogel EA, Billups SJ, Herner SJ, Delate T. Renal drug dosing: Effectiveness of outpatient pharmacist-based vs. prescriber-based clinical decision support systems.

scholarly journals A New Real Time Clinical Decision Support System Using Machine Learning for Critical Care Units

IEEE Access ◽  
2020 ◽  
Vol 8 ◽  
pp. 185676-185687
Author(s):  
Noha Ossama El-Ganainy ◽  
Ilangko Balasingham ◽  
Per Steinar Halvorsen ◽  
Leiv Arne Rosseland
Keyword(s):  
Machine Learning ◽  
Critical Care ◽  
Decision Support ◽  
Decision Support System ◽  
Real Time ◽  
Clinical Decision Support ◽  
Support System ◽  
Clinical Decision Support System ◽  
Clinical Decision ◽  
Critical Care Units

scholarly journals Description of a clinical decision support tool with integrated dose calculator for paediatrics

2021 ◽  
Author(s):  
Lukas Higi ◽  
Karin Käser ◽  
Monika Wälti ◽  
Michael Grotzer ◽  
Priska Vonbach
Keyword(s):  
Decision Support ◽  
Medical Device ◽  
Medication Errors ◽  
Clinical Decision Support ◽  
Clinical Information ◽  
Decision Support Tool ◽  
Clinical Decision ◽  
Support Tool ◽  
Drug Dosing ◽  
Clinical Decision Support Tool

AbstractMedication errors, especially dosing errors are a leading cause of preventable harm in paediatric patients. The paediatric patient population is particularly vulnerable to dosing errors due to immaturity of metabolising organs and developmental changes. Moreover, the lack of clinical trial data or suitable drug forms, and the need for weight-based dosing, does not simplify drug dosing in paediatric or neonatal patients. Consequently, paediatric pharmacotherapy often requires unlicensed and off-label use including manipulation of adult dosage forms. In practice, this results in the need to calculate individual dosages which in turn increases the likelihood of dosing errors. In the age of digitalisation, clinical decision support (CDS) tools can support healthcare professionals in their daily work. CDS tools are currently amongst the gold standards in reducing preventable errors. In this publication, we describe the development and core functionalities of the CDS tool PEDeDose, a Class IIa medical device software certified according to the European Medical Device Regulation. The CDS tool provides a drug dosing formulary with an integrated calculator to determine individual dosages for paediatric, neonatal, and preterm patients. Even a technical interface is part of the CDS tool to facilitate integration into primary systems. This enables the support of the paediatrician directly during the prescribing process without changing the user interface.Conclusion: PEDeDose is a state-of-the-art CDS tool for individualised paediatric drug dosing that includes a certified calculator. What is Known:• Dosing errors are the most common type of medication errors in paediatric patients.• Clinical decision support tools can reduce medication errors effectively. Integration into the practitioner’s workflow improves usability and user acceptance. What is New:• A clinical decision support tool with a certified integrated dosing calculator for paediatric drug dosing.• The tool was designed to facilitate integration into clinical information systems to directly support the prescribing process. Any clinical information system available can interoperate with the PEDeDose web service.

scholarly journals Implementing pharmacogenomics decision support across seven European countries: The Ubiquitous Pharmacogenomics (U-PGx) project

2018 ◽  
Vol 25 (7) ◽  
pp. 893-898 ◽  
Author(s):  
Kathrin Blagec ◽  
Rudolf Koopmann ◽  
Mandy Crommentuijn – van Rhenen ◽  
Inge Holsappel ◽  
Cathelijne H van der Wouden ◽  
...  
Keyword(s):  
Decision Support ◽  
Clinical Decision Support ◽  
Clinical Care ◽  
Clinical Decision ◽  
European Countries ◽  
Support Intervention ◽  
Drug Dosing ◽  
Decision Support Intervention ◽  
Support Tools ◽  
Clinical Decision Support Tools

Abstract Clinical pharmacogenomics (PGx) has the potential to make pharmacotherapy safer and more effective by utilizing genetic patient data for drug dosing and selection. However, widespread adoption of PGx depends on its successful integration into routine clinical care through clinical decision support tools, which is often hampered by insufficient or fragmented infrastructures. This paper describes the setup and implementation of a unique multimodal, multilingual clinical decision support intervention consisting of digital, paper-, and mobile-based tools that are deployed across implementation sites in seven European countries participating in the Ubiquitous PGx (U-PGx) project.

scholarly journals Quantum deep reinforcement learning for clinical decision support in oncology: application to adaptive radiotherapy

2021 ◽  
Vol 11 (1) ◽  
Author(s):  
Dipesh Niraula ◽  
Jamalina Jamaluddin ◽  
Martha M. Matuszak ◽  
Randall K. Ten Haken ◽  
Issam El Naqa
Keyword(s):  
Decision Making ◽  
Reinforcement Learning ◽  
Decision Support ◽  
Clinical Decision Support ◽  
Dose Response ◽  
Clinical Decision Making ◽  
Clinical Decision ◽  
Specific Information ◽  
Clinical Decisions ◽  
Nsclc Patients

AbstractSubtle differences in a patient’s genetics and physiology may alter radiotherapy (RT) treatment responses, motivating the need for a more personalized treatment plan. Accordingly, we have developed a novel quantum deep reinforcement learning (qDRL) framework for clinical decision support that can estimate an individual patient’s dose response mid-treatment and recommend an optimal dose adjustment. Our framework considers patients’ specific information including biological, physical, genetic, clinical, and dosimetric factors. Recognizing that physicians must make decisions amidst uncertainty in RT treatment outcomes, we employed indeterministic quantum states to represent human decision making in a real-life scenario. We paired quantum decision states with a model-based deep q-learning algorithm to optimize the clinical decision-making process in RT. We trained our proposed qDRL framework on an institutional dataset of 67 stage III non-small cell lung cancer (NSCLC) patients treated on prospective adaptive protocols and independently validated our framework in an external multi-institutional dataset of 174 NSCLC patients. For a comprehensive evaluation, we compared three frameworks: DRL, qDRL trained in a Qiskit quantum computing simulator, and qDRL trained in an IBM quantum computer. Two metrics were considered to evaluate our framework: (1) similarity score, defined as the root mean square error between retrospective clinical decisions and the AI recommendations, and (2) self-evaluation scheme that compares retrospective clinical decisions and AI recommendations based on the improvement in the observed clinical outcomes. Our analysis shows that our framework, which takes into consideration individual patient dose response in its decision-making, can potentially improve clinical RT decision-making by at least about 10% compared to unaided clinical practice. Further validation of our novel quantitative approach in a prospective study will provide a necessary framework for improving the standard of care in personalized RT.

Sign in / Sign up

Export Citation Format

Share Document