Application of Machine Learning in Intensive Care Unit (ICU) Settings Using MIMIC Dataset: Systematic Review

Validation of a cancer population derived AKI machine learning algorithm in a general critical care scenario

PURPOSE

Acute Kidney Injury (AKI) is the sudden onset of kidney damage. This damage usually comes without warning and can lead to increased mortality and inpatient costs and is of particular significance to patients undergoing cancer treatment. In previous work, we developed a machine learning algorithm to predict AKI up to 30 days prior to the event, trained on cancer patient data. Here, we validate this model on non-cancer data.

METHODS/PATIENTS

Medical Information Mart for Intensive Care (MIMIC) is a large, freely available database containing de-identified data from patients who were admitted to the critical care units of the Beth Israel Deaconess Medical Center. Data from 28,498 MIMIC patients were used to validate our algorithm, non-availability of Total Protein measure being the largest removal criterion.

RESULTS AND CONCLUSIONS

Applying our algorithm to MIMIC data generated an AUROC of 0.821 (95% CI 0.820-0.821) per blood test. Our cancer derived algorithm compares positively with other AKI models derived and/or tested on MIMIC, with our model predicting AKI at the longest time frame of up to 30 days. This suggests that our model can achieve a good performance on patient cohorts very different to those from which it was derived, demonstrating the transferability and applicability for implementation in a clinical setting.

The evolving role of nursing informatics in the era of artificial intelligence

AIM

This narrative review explores the integration of artificial intelligence (AI) into nursing informatics and examines its impact on nursing practice, healthcare delivery, education, and policy.

BACKGROUND

Nursing informatics, which merges nursing science with information management and communication technologies, is crucial in modern healthcare. The emergence of AI presents opportunities to improve diagnostics, treatment, and healthcare resource management. However, integrating AI into nursing practice also brings challenges, including ethical concerns and the need for specialized training.

SOURCES OF EVIDENCE

A comprehensive literature search was conducted from January 2013 to December 2023 using databases like PubMed, Google Scholar, and Scopus. Articles were selected based on their relevance to AI's role in nursing informatics, particularly in enhancing patient care and healthcare efficiency.

DISCUSSION

AI significantly enhances nursing practice by improving diagnostic accuracy, optimizing care plans, and supporting resource allocation. However, its adoption raises ethical issues, such as data privacy concerns and biases within AI algorithms. Ensuring that nurses are adequately trained in AI technologies is essential for safe and effective integration.

IMPLICATIONS FOR NURSING PRACTICE AND POLICY

Policymakers should promote AI literacy programs for healthcare professionals and develop ethical guidelines to govern the use of AI in healthcare. This will ensure that AI tools are implemented responsibly, protecting patient rights and enhancing healthcare outcomes.

CONCLUSION

AI offers promising advancements in nursing informatics, leading to more efficient patient care and improved decision-making. Nonetheless, overcoming ethical challenges and ensuring AI literacy among nurses are critical steps for successful implementation.

Predicting Agitation-Sedation Levels in Intensive Care Unit Patients: Development of an Ensemble Model

Background

Agitation and sedation management is critical in intensive care as it affects patient safety. Traditional nursing assessments suffer from low frequency and subjectivity. Automating these assessments can boost intensive care unit (ICU) efficiency, treatment capacity, and patient safety.

objectives

The aim of this study was to develop a machine-learning based assessment of agitation and sedation.

Methods

Using data from the Taichung Veterans General Hospital ICU database (2020), an ensemble learning model was developed for classifying the levels of agitation and sedation. Different ensemble learning model sequences were compared. In addition, an interpretable artificial intelligence approach, SHAP (Shapley additive explanations), was employed for explanatory analysis.

Results

With 20 features and 121,303 data points, the random forest model achieved high area under the curve values across all models (sedation classification: 0.97; agitation classification: 0.88). The ensemble learning model enhanced agitation sensitivity (0.82) while maintaining high AUC values across all categories (all >0.82). The model explanations aligned with clinical experience.

Conclusions

This study proposes an ICU agitation-sedation assessment automation using machine learning, enhancing efficiency and safety. Ensemble learning improves agitation sensitivity while maintaining accuracy. Real-time monitoring and future digital integration have the potential for advancements in intensive care.

Contribution of Open Access Databases to Intensive Care Medicine Research: Scoping Review

BACKGROUND

Intensive care units (ICUs) handle the most critical patients with a high risk of mortality. Due to those conditions, close monitoring is necessary and therefore, a large volume of data is collected. Collaborative ventures have enabled the emergence of large open access databases, leading to numerous publications in the field.

OBJECTIVE

The aim of this scoping review is to identify the characteristics of studies using open access intensive care databases and to describe the contribution of these studies to intensive care research.

METHODS

The research was conducted using 3 databases (PubMed-MEDLINE, Embase, and Web of Science) from the inception of each database to August 1, 2022. We included original articles based on 4 open databases of patients admitted to ICUs: Amsterdam University Medical Centers Database, eICU Collaborative Research Database, High time resolution ICU dataset, Medical Information Mart for Intensive Care (II to IV). A double-blinded screening for eligibility was performed, first on the title and abstract and subsequently on the full-text articles. Characteristics relating to publication journals, study design, and statistical analyses were extracted and analyzed.

RESULTS

We observed a consistent increase in the number of publications from these databases since 2016. The Medical Information Mart for Intensive Care databases were the most frequently used. The highest contributions came from China and the United States, with 689 (52.7%) and 370 (28.3%) publications respectively. The median impact factor of publications was 3.8 (IQR 2.8-5.8). Topics related to cardiovascular and infectious diseases were predominant, accounting for 333 (25.5%) and 324 (24.8%) articles, respectively. Logistic regression emerged as the most commonly used statistical model for both inference and prediction questions, featuring in 396 (55.5%) and 281 (47.5%) studies, respectively. A majority of the inference studies yielded statistically significant results (84.0%). In prediction studies, area under the curve was the most frequent performance measure, with a median value of 0.840 (IQR 0.780-0.890).

CONCLUSIONS

The abundance of scientific outputs resulting from these databases, coupled with the diversity of topics addressed, highlight the importance of these databases as valuable resources for clinical research. This suggests their potential impact on clinical practice within intensive care settings. However, the quality and clinical relevance of these studies remains highly heterogeneous, with a majority of articles being published in low-impact factor journals.

Development and validation of a deep learning algorithm for the prediction of serum creatinine in critically ill patients

Objectives

Serum creatinine (SCr) is the primary biomarker for assessing kidney function; however, it may lag behind true kidney function, especially in instances of acute kidney injury (AKI). The objective of the work is to develop Nephrocast, a deep-learning model to predict next-day SCr in adult patients treated in the intensive care unit (ICU).

Materials and Methods

Nephrocast was trained and validated, temporally and prospectively, using electronic health record data of adult patients admitted to the ICU in the University of California San Diego Health (UCSDH) between January 1, 2016 and June 22, 2024. The model features consisted of demographics, comorbidities, vital signs and laboratory measurements, and medications. Model performance was evaluated by mean absolute error (MAE) and root-mean-square error (RMSE) and compared against the prediction day's SCr as a reference.

Results

A total of 28 191 encounters met the eligibility criteria, corresponding to 105 718 patient-days. The median (interquartile range [IQR]) MAE and RMSE in the internal test set were 0.09 (0.085-0.09) mg/dL and 0.15 (0.146-0.152) mg/dL, respectively. In the prospective validation, the MAE and RMSE were 0.09 mg/dL and 0.14 mg/dL, respectively. The model's performance was superior to the reference SCr.

Discussion and Conclusion

Our model demonstrated good performance in predicting next-day SCr by leveraging clinical data routinely collected in the ICU. The model could aid clinicians in in identifying high-risk patients for AKI, predicting AKI trajectory, and informing the dosing of renally eliminated drugs.

The Use of Deep Learning and Machine Learning on Longitudinal Electronic Health Records for the Early Detection and Prevention of Diseases: Scoping Review

BACKGROUND

Electronic health records (EHRs) contain patients' health information over time, including possible early indicators of disease. However, the increasing amount of data hinders clinicians from using them. There is accumulating evidence suggesting that machine learning (ML) and deep learning (DL) can assist clinicians in analyzing these large-scale EHRs, as algorithms thrive on high volumes of data. Although ML has become well developed, studies mainly focus on engineering but lack medical outcomes.

OBJECTIVE

This study aims for a scoping review of the evidence on how the use of ML on longitudinal EHRs can support the early detection and prevention of disease. The medical insights and clinical benefits that have been generated were investigated by reviewing applications in a variety of diseases.

METHODS

This study was conducted according to the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines. A literature search was performed in 2022 in collaboration with a medical information specialist in the following databases: PubMed, Embase, Web of Science Core Collection (Clarivate Analytics), and IEEE Xplore Digital Library and computer science bibliography. Studies were eligible when longitudinal EHRs were used that aimed for the early detection of disease via ML in a prevention context. Studies with a technical focus or using imaging or hospital admission data were beyond the scope of this review. Study screening and selection and data extraction were performed independently by 2 researchers.

RESULTS

In total, 20 studies were included, mainly published between 2018 and 2022. They showed that a variety of diseases could be detected or predicted, particularly diabetes; kidney diseases; diseases of the circulatory system; and mental, behavioral, and neurodevelopmental disorders. Demographics, symptoms, procedures, laboratory test results, diagnoses, medications, and BMI were frequently used EHR data in basic recurrent neural network or long short-term memory techniques. By developing and comparing ML and DL models, medical insights such as a high diagnostic performance, an earlier detection, the most important predictors, and additional health indicators were obtained. A clinical benefit that has been evaluated positively was preliminary screening. If these models are applied in practice, patients might also benefit from personalized health care and prevention, with practical benefits such as workload reduction and policy insights.

CONCLUSIONS

Longitudinal EHRs proved to be helpful for support in health care. Current ML models on EHRs can support the detection of diseases in terms of accuracy and offer preliminary screening benefits. Regarding the prevention of diseases, ML and specifically DL models can accurately predict or detect diseases earlier than current clinical diagnoses. Adding personally responsible factors allows targeted prevention interventions. While ML models based on textual EHRs are still in the developmental stage, they have high potential to support clinicians and the health care system and improve patient outcomes.

Use of Artificial Intelligence in Improving Outcomes in Heart Disease: A Scientific Statement From the American Heart Association

A major focus of academia, industry, and global governmental agencies is to develop and apply artificial intelligence and other advanced analytical tools to transform health care delivery. The American Heart Association supports the creation of tools and services that would further the science and practice of precision medicine by enabling more precise approaches to cardiovascular and stroke research, prevention, and care of individuals and populations. Nevertheless, several challenges exist, and few artificial intelligence tools have been shown to improve cardiovascular and stroke care sufficiently to be widely adopted. This scientific statement outlines the current state of the art on the use of artificial intelligence algorithms and data science in the diagnosis, classification, and treatment of cardiovascular disease. It also sets out to advance this mission, focusing on how digital tools and, in particular, artificial intelligence may provide clinical and mechanistic insights, address bias in clinical studies, and facilitate education and implementation science to improve cardiovascular and stroke outcomes. Last, a key objective of this scientific statement is to further the field by identifying best practices, gaps, and challenges for interested stakeholders.

Early prediction of mortality at sepsis diagnosis time in critically ill patients by using interpretable machine learning

This study applied machine learning for the early prediction of 30-day mortality at sepsis diagnosis time in critically ill patients. Retrospective study using data collected from the Medical Information Mart for Intensive Care IV database. The data of the patient cohort was divided on the basis of the year of hospitalization, into training (2008-2013), validation (2014-2016), and testing (2017-2019) datasets. 24,377 patients with the sepsis diagnosis time < 24 h after intensive care unit (ICU) admission were included. A gradient boosting tree-based algorithm (XGBoost) was used for training the machine learning model to predict 30-day mortality at sepsis diagnosis time in critically ill patients. Model performance was measured in both discrimination and calibration aspects. The model was interpreted using the SHapley Additive exPlanations (SHAP) module. The 30-day mortality rate of the testing dataset was 17.9%, and 39 features were selected for the machine learning model. Model performance on the testing dataset achieved an area under the receiver operating characteristic curve (AUROC) of 0.853 (95% CI 0.837-0.868) and an area under the precision-recall curves of 0.581 (95% CI 0.541-0.619). The calibration plot for the model revealed a slope of 1.03 (95% CI 0.94-1.12) and intercept of 0.14 (95% CI 0.04-0.25). The SHAP revealed the top three most significant features, namely age, increased red blood cell distribution width, and respiratory rate. Our study demonstrated the feasibility of using the interpretable machine learning model to predict mortality at sepsis diagnosis time.

Systems of Care Delivery and Optimization in the Intensive Care Unit

As the volume and complexity of patients requiring intensive care grows, so do the barriers and challenges to the delivery of that care. This article summarizes these challenges, outlines strategies used to overcome them, and presents new developments and concepts within the care of the ICU patient.

Machine learning prediction will be part of future treatment of depression

Machine learning (ML) is changing the way that medicine is practiced. While already clinically utilised in diagnostic radiology and outcome prediction in intensive care unit, ML approaches in psychiatry remain nascent. Implementing ML algorithms in psychiatry, particularly in the treatment of depression, is significantly more challenging than other areas of medicine in part because of the less demarcated disease nosology and greater variability in practice. Given the current exiguous capacity of clinicians to predict patient and treatment outcomes in depression, there is a significantly greater need for better predictive capability. Early studies have shown promising results. ML predictions were significantly better than chance within the sequenced treatment alternatives to relieve depression (STAR*D) trial (accuracy 64.6%, p < 0.0001) and combining medications to enhance depression outcomes (COMED) randomised Controlled Trial (RCT) (accuracy 59.6%, p = 0.043), with similar results found in larger scale, retrospective studies. The greater flexibility and dimensionality of ML approaches has been demonstrated in studies incorporating diverse input variables including electroencephalography scans, achieving 88% accuracy for treatment response, and cognitive test scores, achieving up to 72% accuracy for treatment response. The predicting response to depression treatment (PReDicT) trial tested ML informed prescribing of antidepressants against standard therapy and found there was both better outcomes for anxiety and functional endpoints despite the algorithm only having a balanced accuracy of 57.5%. Impeding the progress of ML algorithms in psychiatry are pragmatic hurdles, including accuracy, expense, acceptability and comprehensibility, and ethical hurdles, including medicolegal liability, clinical autonomy and data privacy. Notwithstanding impediments, it is clear that ML prediction algorithms will be part of depression treatment in the future and clinicians should be prepared for their arrival.

Medical transformer for multimodal survival prediction in intensive care: integration of imaging and non-imaging data

When clinicians assess the prognosis of patients in intensive care, they take imaging and non-imaging data into account. In contrast, many traditional machine learning models rely on only one of these modalities, limiting their potential in medical applications. This work proposes and evaluates a transformer-based neural network as a novel AI architecture that integrates multimodal patient data, i.e., imaging data (chest radiographs) and non-imaging data (clinical data). We evaluate the performance of our model in a retrospective study with 6,125 patients in intensive care. We show that the combined model (area under the receiver operating characteristic curve [AUROC] of 0.863) is superior to the radiographs-only model (AUROC = 0.811, p < 0.001) and the clinical data-only model (AUROC = 0.785, p < 0.001) when tasked with predicting in-hospital survival per patient. Furthermore, we demonstrate that our proposed model is robust in cases where not all (clinical) data points are available.

Evaluation of machine learning-based models for prediction of clinical deterioration: A systematic literature review

BACKGROUND AND OBJECTIVE

Early identification of patients at risk of deterioration can prevent life-threatening adverse events and shorten length of stay. Although there are numerous models applied to predict patient clinical deterioration, most are based on vital signs and have methodological shortcomings that are not able to provide accurate estimates of deterioration risk. The aim of this systematic review is to examine the effectiveness, challenges, and limitations of using machine learning (ML) techniques to predict patient clinical deterioration in hospital settings.

METHODS

A systematic review was performed in accordance with the Preferred Reporting Items for Systematic Reviews and meta-Analyses (PRISMA) guidelines using EMBASE, MEDLINE Complete, CINAHL Complete, and IEEExplore databases. Citation searching was carried out for studies that met inclusion criteria. Two reviewers used the inclusion/exclusion criteria to independently screen studies and extract data. To address any discrepancies in the screening process, the two reviewers discussed their findings and a third reviewer was consulted as needed to reach a consensus. Studies focusing on use of ML in predicting patient clinical deterioration that were published from inception to July 2022 were included.

RESULTS

A total of 29 primary studies that evaluated ML models to predict patient clinical deterioration were identified. After reviewing these studies, we found that 15 types of ML techniques have been employed to predict patient clinical deterioration. While six studies used a single technique exclusively, several others utilised a combination of classical techniques, unsupervised and supervised learning, as well as other novel techniques. Depending on which ML model was applied and the type of input features, ML models predicted outcomes with an area under the curve from 0.55 to 0.99.

CONCLUSIONS

Numerous ML methods have been employed to automate the identification of patient deterioration. Despite these advancements, there is still a need for further investigation to examine the application and effectiveness of these methods in real-world situations.

Enhanced Cyber Attack Detection Process for Internet of Health Things (IoHT) Devices Using Deep Neural Network

Internet of Health Things plays a vital role in day-to-day life by providing electronic healthcare services and has the capacity to increase the quality of patient care. Internet of Health Things (IoHT) devices and applications have been growing rapidly in recent years, becoming extensively vulnerable to cyber-attacks since the devices are small and heterogeneous. In addition, it is doubly significant when IoHT involves devices used in healthcare domain. Consequently, it is essential to develop a resilient cyber-attack detection system in the Internet of Health Things environment for mitigating the security risks and preventing Internet of Health Things devices from becoming exposed to cyber-attacks. Artificial intelligence plays a primary role in anomaly detection. In this paper, a deep neural network-based cyber-attack detection system is built by employing artificial intelligence on latest ECU-IoHT dataset to uncover cyber-attacks in Internet of Health Things environment. The proposed deep neural network system achieves average higher performance accuracy of 99.85%, an average area under receiver operator characteristic curve 0.99 and the false positive rate is 0.01. It is evident from the experimental result that the proposed system attains higher detection rate than the existing methods.

Artificial intelligence for clinical decision support for monitoring patients in cardiovascular ICUs: A systematic review

Background

Artificial intelligence (AI) and machine learning (ML) models continue to evolve the clinical decision support systems (CDSS). However, challenges arise when it comes to the integration of AI/ML into clinical scenarios. In this systematic review, we followed the Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA), the population, intervention, comparator, outcome, and study design (PICOS), and the medical AI life cycle guidelines to investigate studies and tools which address AI/ML-based approaches towards clinical decision support (CDS) for monitoring cardiovascular patients in intensive care units (ICUs). We further discuss recent advances, pitfalls, and future perspectives towards effective integration of AI into routine practices as were identified and elaborated over an extensive selection process for state-of-the-art manuscripts.

Methods

Studies with available English full text from PubMed and Google Scholar in the period from January 2018 to August 2022 were considered. The manuscripts were fetched through a combination of the search keywords including AI, ML, reinforcement learning (RL), deep learning, clinical decision support, and cardiovascular critical care and patients monitoring. The manuscripts were analyzed and filtered based on qualitative and quantitative criteria such as target population, proper study design, cross-validation, and risk of bias.

Results

More than 100 queries over two medical search engines and subjective literature research were developed which identified 89 studies. After extensive assessments of the studies both technically and medically, 21 studies were selected for the final qualitative assessment.

Discussion

Clinical time series and electronic health records (EHR) data were the most common input modalities, while methods such as gradient boosting, recurrent neural networks (RNNs) and RL were mostly used for the analysis. Seventy-five percent of the selected papers lacked validation against external datasets highlighting the generalizability issue. Also, interpretability of the AI decisions was identified as a central issue towards effective integration of AI in healthcare.

MIMIC-IV on FHIR: converting a decade of in-patient data into an exchangeable, interoperable format

OBJECTIVE

Convert the Medical Information Mart for Intensive Care (MIMIC)-IV database into Health Level 7 Fast Healthcare Interoperability Resources (FHIR). Additionally, generate and publish an openly available demo of the resources, and create a FHIR Implementation Guide to support and clarify the usage of MIMIC-IV on FHIR.

MATERIALS AND METHODS

FHIR profiles and terminology system of MIMIC-IV were modeled from the base FHIR R4 resources. Data and terminology were reorganized from the relational structure into FHIR according to the profiles. Resources generated were validated for conformance with the FHIR profiles. Finally, FHIR resources were published as newline delimited JSON files and the profiles were packaged into an implementation guide.

RESULTS

The modeling of MIMIC-IV in FHIR resulted in 25 profiles, 2 extensions, 35 ValueSets, and 34 CodeSystems. An implementation guide encompassing the FHIR modeling can be accessed at mimic.mit.edu/fhir/mimic. The generated demo dataset contained 100 patients and over 915 000 resources. The full dataset contained 315 000 patients covering approximately 5 840 000 resources. The final datasets in NDJSON format are accessible on PhysioNet.

DISCUSSION

Our work highlights the challenges and benefits of generating a real-world FHIR store. The challenges arise from terminology mapping and profiling modeling decisions. The benefits come from the extensively validated openly accessible data created as a result of the modeling work.

CONCLUSION

The newly created MIMIC-IV on FHIR provides one of the first accessible deidentified critical care FHIR datasets. The extensive real-world data found in MIMIC-IV on FHIR will be invaluable for research and the development of healthcare applications.

Applications of Artificial Intelligence in Thrombocytopenia

Thrombocytopenia is a medical condition where blood platelet count drops very low. This drop in platelet count can be attributed to many causes including medication, sepsis, viral infections, and autoimmunity. Clinically, the presence of thrombocytopenia might be very dangerous and is associated with poor outcomes of patients due to excessive bleeding if not addressed quickly enough. Hence, early detection and evaluation of thrombocytopenia is essential for rapid and appropriate intervention for these patients. Since artificial intelligence is able to combine and evaluate many linear and nonlinear variables simultaneously, it has shown great potential in its application in the early diagnosis, assessing the prognosis and predicting the distribution of patients with thrombocytopenia. In this review, we conducted a search across four databases and identified a total of 13 original articles that looked at the use of many machine learning algorithms in the diagnosis, prognosis, and distribution of various types of thrombocytopenia. We summarized the methods and findings of each article in this review. The included studies showed that artificial intelligence can potentially enhance the clinical approaches used in the diagnosis, prognosis, and treatment of thrombocytopenia.

Strategies of Predictive Schemes and Clinical Diagnosis for Prognosis Using MIMIC-III: A Systematic Review

The prime purpose of the proposed study is to construct a novel predictive scheme for assisting in the prognosis of criticality using the MIMIC-III dataset. With the adoption of various analytics and advanced computing in the healthcare system, there is an increasing trend toward developing an effective prognostication mechanism. Predictive-based modeling is the best alternative to work in this direction. This paper discusses various scientific contributions using desk research methodology towards the Medical Information Mart for Intensive Care (MIMIC-III). This open-access dataset is meant to help predict patient trajectories for various purposes ranging from mortality forecasting to treatment planning. With a dominant machine learning approach in this perspective, there is a need to discover the effectiveness of existing predictive methods. The resultant outcome of this paper offers an inclusive discussion about various available predictive schemes and clinical diagnoses using MIMIC-III in order to contribute toward better information associated with its strengths and weaknesses. Therefore, the paper provides a clear visualization of existing schemes for clinical diagnosis using a systematic review approach.

Predicting Readmission or Death After Discharge From the ICU: External Validation and Retraining of a Machine Learning Model

OBJECTIVES

Many machine learning (ML) models have been developed for application in the ICU, but few models have been subjected to external validation. The performance of these models in new settings therefore remains unknown. The objective of this study was to assess the performance of an existing decision support tool based on a ML model predicting readmission or death within 7 days after ICU discharge before, during, and after retraining and recalibration.

DESIGN

A gradient boosted ML model was developed and validated on electronic health record data from 2004 to 2021. We performed an independent validation of this model on electronic health record data from 2011 to 2019 from a different tertiary care center.

SETTING

Two ICUs in tertiary care centers in The Netherlands.

PATIENTS

Adult patients who were admitted to the ICU and stayed for longer than 12 hours.

INTERVENTIONS

None.

MEASUREMENTS AND MAIN RESULTS

We assessed discrimination by area under the receiver operating characteristic curve (AUC) and calibration (slope and intercept). We retrained and recalibrated the original model and assessed performance via a temporal validation design. The final retrained model was cross-validated on all data from the new site. Readmission or death within 7 days after ICU discharge occurred in 577 of 10,052 ICU admissions (5.7%) at the new site. External validation revealed moderate discrimination with an AUC of 0.72 (95% CI 0.67-0.76). Retrained models showed improved discrimination with AUC 0.79 (95% CI 0.75-0.82) for the final validation model. Calibration was poor initially and good after recalibration via isotonic regression.

CONCLUSIONS

In this era of expanding availability of ML models, external validation and retraining are key steps to consider before applying ML models to new settings. Clinicians and decision-makers should take this into account when considering applying new ML models to their local settings.

Enabling Timely Medical Intervention by Exploring Health-Related Multivariate Time Series with a Hybrid Attentive Model

Modern healthcare practice, especially in intensive care units, produces a vast amount of multivariate time series of health-related data, e.g., multi-lead electrocardiogram (ECG), pulse waveform, blood pressure waveform and so on. As a result, timely and accurate prediction of medical intervention (e.g., intravenous injection) becomes possible, by exploring such semantic-rich time series. Existing works mainly focused on onset prediction at the granularity of hours that was not suitable for medication intervention in emergency medicine. This research proposes a Multi-Variable Hybrid Attentive Model (MVHA) to predict the impending need of medical intervention, by jointly mining multiple time series. Specifically, a two-level attention mechanism is designed to capture the pattern of fluctuations and trends of different time series. This work applied MVHA to the prediction of the impending intravenous injection need of critical patients at the intensive care units. Experiments on the MIMIC Waveform Database demonstrated that the proposed model achieves a prediction accuracy of 0.8475 and an ROC-AUC of 0.8318, which significantly outperforms baseline models.

Machine Learning-Based Pressure Ulcer Prediction in Modular Critical Care Data

Increasingly available open medical and health datasets encourage data-driven research with a promise of improving patient care through knowledge discovery and algorithm development. Among efficient approaches to such high-dimensional problems are a number of machine learning methods, which are applied in this paper to pressure ulcer prediction in modular critical care data. An inherent property of many health-related datasets is a high number of irregularly sampled time-variant and scarcely populated features, often exceeding the number of observations. Although machine learning methods are known to work well under such circumstances, many choices regarding model and data processing exist. In particular, this paper address both theoretical and practical aspects related to the application of six classification models to pressure ulcers, while utilizing one of the largest available Medical Information Mart for Intensive Care (MIMIC-IV) databases. Random forest, with an accuracy of 96%, is the best-performing approach among the considered machine learning algorithms.

Identification of data elements for blood gas analysis dataset: a base for developing registries and artificial intelligence-based systems

Background

One of the challenging decision-making tasks in healthcare centers is the interpretation of blood gas tests. One of the most effective assisting approaches for the interpretation of blood gas analysis (BGA) can be artificial intelligence (AI)-based decision support systems. A primary step to develop intelligent systems is to determine information requirements and automated data input for the secondary analyses. Datasets can help the automated data input from dispersed information systems. Therefore, the current study aimed to identify the data elements required for supporting BGA as a dataset.

Materials and methods

This cross-sectional descriptive study was conducted in Nemazee Hospital, Shiraz, Iran. A combination of literature review, experts’ consensus, and the Delphi technique was used to develop the dataset. A review of the literature was performed on electronic databases to find the dataset for BGA. An expert panel was formed to discuss on, add, or remove the data elements extracted through searching the literature. Delphi technique was used to reach consensus and validate the draft dataset.

Results

The data elements of the BGA dataset were categorized into ten categories, namely personal information, admission details, present illnesses, past medical history, social status, physical examination, paraclinical investigation, blood gas parameter, sequential organ failure assessment (SOFA) score, and sampling technique errors. Overall, 313 data elements, including 172 mandatory and 141 optional data elements were confirmed by the experts for being included in the dataset.

Conclusions

We proposed a dataset as a base for registries and AI-based systems to assist BGA. It helps the storage of accurate and comprehensive data, as well as integrating them with other information systems. As a result, high-quality care is provided and clinical decision-making is improved.

DeIDNER Model: A Neural Network Named Entity Recognition Model for Use in the De-identification of Clinical Notes

Clinical named entity recognition (NER) is an essential building block for many downstream natural language processing (NLP) applications such as information extraction and de-identification. Recently, deep learning (DL) methods that utilize word embeddings have become popular in clinical NLP tasks. However, there has been little work on evaluating and combining the word embeddings trained from different domains. The goal of this study is to improve the performance of NER in clinical discharge summaries by developing a DL model that combines different embeddings and investigate the combination of standard and contextual embeddings from the general and clinical domains. We developed: 1) A human-annotated high-quality internal corpus with discharge summaries and 2) A NER model with an input embedding layer that combines different embeddings: standard word embeddings, context-based word embeddings, a character-level word embedding using a convolutional neural network (CNN), and an external knowledge sources along with word features as one-hot vectors. Embedding was followed by bidirectional long short-term memory (Bi-LSTM) and conditional random field (CRF) layers. The proposed model reaches or overcomes state-of-the-art performance on two publicly available data sets and an F1 score of 94.31% on an internal corpus. After incorporating mixed-domain clinically pre-trained contextual embeddings, the F1 score further improved to 95.36% on the internal corpus. This study demonstrated an efficient way of combining different embeddings that will improve the recognition performance aiding the downstream de-identification of clinical notes.

Machine Learning Approach to Optimize Sedation Use in Endoscopic Procedures

Endoscopy procedures are often performed with either moderate or deep sedation. While deep sedation is costly, procedures with moderate sedation are not always well tolerated resulting in patient discomfort, and are often aborted. Due to lack of clear guidelines, the decision to utilize moderate sedation or anesthesia for a procedure is made by the providers, leading to high variability in clinical practice. The objective of this study was to build a Machine Learning (ML) model that predicts if a colonoscopy can be successfully completed with moderate sedation based on patients' demographics, comorbidities, and prescribed medications. XGBoost model was trained and tested on 10,025 colonoscopies (70% - 30%) performed at University of Arkansas for Medical Sciences (UAMS). XGBoost achieved average area under receiver operating characteristic curve (AUC) of 0.762, F1-score to predict procedures that need moderate sedation was 0.85, and precision and recall were 0.81 and 0.89 respectively. The proposed model can be employed as a decision support tool for physicians to bolster their confidence while choosing between moderate sedation and anesthesia for a colonoscopy procedure.