|
1
|
Abdalwahab Abdallah ABA, Hafez Sadaka SI, Ali EI, Mustafa Bilal SA, Abdelrahman MO, Fakiali Mohammed FB, Nimir Ahmed SD, Abdelrahim Saeed NE. The Role of Artificial Intelligence in Pediatric Intensive Care: A Systematic Review. Cureus 2025; 17:e80142. [PMID: 40190909 PMCID: PMC11971983 DOI: 10.7759/cureus.80142] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Accepted: 03/06/2025] [Indexed: 04/09/2025] Open
Abstract
Pediatric intensive care units (PICUs) could transform due to artificial intelligence (AI), which could improve patient outcomes, increase diagnostic accuracy, and streamline repetitive procedures. The goal of this systematic review was to outline how AI can be used to enhance any health outcomes in pediatric intensive care. We searched four databases (PubMed, Scopus, Web of Science, and IEEE Xplore) for relevant studies using a predefined systematic search. We found 267 studies in these four databases. The studies were first screened to remove the duplicates and then screened by titles to remove irrelevant studies. The studies were further screened based on inclusion and exclusion criteria, in which 32 studies were found suitable for inclusion in this study. The studies were assessed for risk of bias using the Prediction Model Risk Of Bias Assessment Tool (PROBAST) tool. After AI was implemented, almost 22% (n = 7) of studies showed an immediate effect and enhanced health outcomes. A small number of studies involved AI implementation in actual PICUs, while the majority focused on experimental testing. AI models outperformed conventional clinical modalities among the remaining 78% (n = 25) and might have indirectly impacted patient outcomes. Significant variation in metrics and standardization was found when health outcomes were quantitatively assessed using statistical measures, including specificity (38%; n = 12) and area under the receiver operating characteristic curve (AUROC) (56%; n = 18). There are not sufficient studies showing that AI has significantly enhanced pediatric critical care patients' health outcomes. To evaluate AI's impact, more prospective, experimental research is required, utilizing verified outcome measures, defined metrics, and established application frameworks.
Collapse
Affiliation(s)
| | | | - Elryah I Ali
- Department of Medical Laboratory Technology, College of Applied Medical Sciences, Northern Border University, Arar, SAU
| | | | | | | | | | - Nuha Elrayah Abdelrahim Saeed
- Department of Biochemistry, University of Khartoum, Khartoum, SDN
- Department of Pediatrics, Al Enjaz Medical Center, Riyadh, SAU
| |
Collapse
|
|
2
|
Govindan RB, Loparo KA. Bedside monitoring tools and advanced signal processing approaches to monitor critically-ill infants. Semin Fetal Neonatal Med 2024; 29:101544. [PMID: 39467727 DOI: 10.1016/j.siny.2024.101544] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 10/30/2024]
Abstract
There is a substantial body of literature that supports neonatal monitoring and signal analysis of the collected data to provide valuable insights for improving patient clinical care and to inform new research studies. This comprehensive monitoring approach extends beyond the collection of conventional vital signs to include the acquisition of continuous waveform data from patient monitors and other bedside medical devices. This paper discusses the necessary infrastructure for waveform retrieval from bedside monitors, and explores options provided by leading healthcare companies, third-party vendors or academic research teams to implement scalable monitoring systems across entire critical care units. Additionally, we discuss the application of advanced signal processing that transcend traditional statistics, including heart rate variability in both the time- and frequency-domains, spectral analysis of EEG, and cerebral pressure autoregulation. The infrastructures and signal processing techniques outlined here are indispensable tools for intensivists, empowering them to enhance care for critically ill infants. In addition, we briefly address the emergence of advanced tools for fetal monitoring.
Collapse
Affiliation(s)
- R B Govindan
- The Zickler Family Prenatal Pediatrics Institute, Children's National Hospital, Washington, DC, USA; Department of Pediatrics, The George Washington University School of Medicine, Washington, DC, USA; The Developing Brain Institute, Children's National Hospital, Washington, DC, USA.
| | - Kenneth A Loparo
- Institute for Smart, Secure and Connected Systems: ISSACS, Case Western Reserve University, Cleveland, OH, USA.
| |
Collapse
|
|
3
|
Sehgal A, Yeomans EJ, Nixon GM. Kangaroo mother care improves cardiorespiratory physiology in preterm infants: an observational study. Arch Dis Child Fetal Neonatal Ed 2024; 109:628-633. [PMID: 38538151 DOI: 10.1136/archdischild-2023-326748] [Citation(s) in RCA: 1] [Impact Index Per Article: 1.0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Journal Information] [Submit a Manuscript] [Subscribe] [Scholar Register] [Received: 12/11/2023] [Accepted: 03/12/2024] [Indexed: 10/20/2024]
Abstract
OBJECTIVES To evaluate whether kangaroo mother care (KMC) in preterm infants on non-invasive respiratory support improves indices of cardiorespiratory wellbeing. STUDY DESIGN Prospective quasi-experimental observational study. SETTING Tertiary perinatal neonatal unit. PATIENTS 50 very preterm infants being managed with nasal continuous positive airway pressure. INTERVENTIONS Continuous high-resolution preductal pulse-oximetry recordings using Masimo Radical-7 oximeter for 1 hour (incubator care) followed by 1 hour during KMC performed on the same day. MAIN OUTCOME MEASURES Measures of cardiorespiratory stability (dips in oxygen saturations (SpO2)) of ≥5% less than baseline, % time spent with oxygen saturations <90%, SpO2 variability and heart rate fluctuation and incidence of bradycardias. RESULTS The gestational age and birth weight of the cohort were 28.4±2.1 weeks and 1137±301 g, respectively. Dips in SpO2 of ≥5% less than baseline were significantly fewer with KMC, median (IQR) 24 (12 to 42) vs 13 (3 to 25), p=0.001. SpO2 variability (Delta 12 s and 2 s), (1.24±0.6 vs 0.9±0.4, p=0.005 and 4.1±1.7 vs 2.8±1.2, p<0.0001) and rapid resaturation and desaturation indices were significantly lower during KMC, compared with incubator care. Percentage time spent in oxygen saturations <90% was less with KMC (7.5% vs 2.7%, p=0.04). Mean heart rate was comparable although fluctuations in heart rate (rise by >8 bpm) were lower with KMC (43±22 vs 33±20, p=0.03). Seven (14%) infants had bradycardias during incubator care and none during KMC, p=0.012. CONCLUSIONS KMC improves cardiorespiratory stability in ventilated preterm infants. Regular KMC has the potential to improve clinical outcomes in this vulnerable cohort.
Collapse
Affiliation(s)
- Arvind Sehgal
- Monash Children's Hospital, Melbourne, Victoria, Australia
- Department of Paediatrics, Monash University, Clayton, Victoria, Australia
| | - Emma J Yeomans
- Monash Children's Hospital, Melbourne, Victoria, Australia
- Department of Paediatrics, Monash University, Clayton, Victoria, Australia
| | - Gillian M Nixon
- Department of Paediatrics, Monash University, Clayton, Victoria, Australia
- Melbourne Children's Sleep Centre, Monash Children's Hospital, Melbourne, VIC, Australia
| |
Collapse
|
|
4
|
Kainth D, Prakash S, Sankar MJ. Diagnostic Performance of Machine Learning-based Models in Neonatal Sepsis: A Systematic Review. Pediatr Infect Dis J 2024; 43:889-901. [PMID: 39079037 DOI: 10.1097/inf.0000000000004409] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [MESH Headings] [Track Full Text] [Journal Information] [Submit a Manuscript] [Subscribe] [Scholar Register] [Indexed: 08/21/2024]
Abstract
BACKGROUND Timely diagnosis of neonatal sepsis is challenging. We aimed to systematically evaluate the diagnostic performance of sophisticated machine learning (ML) techniques for the prediction of neonatal sepsis. METHODS We searched MEDLINE, Embase, Web of Science and Cochrane CENTRAL databases using "neonate," "sepsis" and "machine learning" as search terms. We included studies that developed or validated an ML algorithm to predict neonatal sepsis. Those incorporating automated vital-sign data were excluded. Among 5008 records, 74 full-text articles were screened. Two reviewers extracted information as per the CHARMS (CHecklist for critical Appraisal and data extraction for systematic Reviews of prediction Modelling Studies) checklist. We followed the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-analyses) guideline extension for diagnostic test accuracy reviews and used the PROBAST tool for risk of bias assessment. Primary outcome was a predictive performance of ML models in terms of sensitivity, specificity and positive and negative predictive values. We generated a hierarchical summary receiver operating characteristics curve for pooled analysis. RESULTS Of 19 studies (15,984 participants) with 76 ML models, the random forest algorithm was the most employed. The candidate predictors per model ranged from 5 to 93; most included birth weight and gestation. None performed external validation. The risk of bias was high (18 studies). For the prediction of any sepsis (14 studies), pooled sensitivity was 0.87 (95% credible interval: 0.75-0.94) and specificity was 0.89 (95% credible interval: 0.77-0.95). Pooled area under the receiver operating characteristics curve was 0.94 (95% credible interval: 0.92-0.96). All studies, except one, used data from high- or upper-middle-income countries. With unavailable probability thresholds, the performance could not be assessed with sufficient precision. CONCLUSIONS ML techniques have good diagnostic accuracy for neonatal sepsis. The need for the development of context-specific models from high-burden countries is highlighted.
Collapse
Affiliation(s)
- Deepika Kainth
- From the Department of Pediatrics, All India Institute of Medical Sciences
| | - Satya Prakash
- Department of Pediatrics, All India Institute of Medical Sciences, New Delhi, India
| | - M Jeeva Sankar
- From the Department of Pediatrics, All India Institute of Medical Sciences
| |
Collapse
|
|
5
|
Song Y, Zhai W, Ma S, Wu Y, Ren M, Van den Eynde J, Nardi P, Pang PYK, Ali JM, Han J, Guo Z. Machine learning-based prediction of off-pump coronary artery bypass grafting-associated acute kidney injury. J Thorac Dis 2024; 16:4535-4542. [PMID: 39144311 PMCID: PMC11320255 DOI: 10.21037/jtd-24-711] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 05/01/2024] [Accepted: 06/27/2024] [Indexed: 08/16/2024]
Abstract
Background The cardiac surgery-associated acute kidney injury (CSA-AKI) occurs in up to 1 out of 3 patients. Off-pump coronary artery bypass grafting (OPCABG) is one of the major cardiac surgeries leading to CSA-AKI. Early identification and timely intervention are of clinical significance for CSA-AKI. In this study, we aimed to establish a prediction model of off-pump coronary artery bypass grafting-associated acute kidney injury (OPCABG-AKI) after surgery based on machine learning methods. Methods The preoperative and intraoperative data of 1,041 patients who underwent OPCABG in Chest Hospital, Tianjin University from June 1, 2021 to April 30, 2023 were retrospectively collected. The definition of OPCABG-AKI was based on the 2012 Kidney Disease Improving Global Outcomes (KDIGO) criteria. The baseline data and intraoperative time series data were included in the dataset, which were preprocessed separately. A total of eight machine learning models were constructed based on the baseline data: logistic regression (LR), gradient-boosting decision tree (GBDT), eXtreme gradient boosting (XGBoost), adaptive boosting (AdaBoost), random forest (RF), support vector machine (SVM), k-nearest neighbor (KNN), and decision tree (DT). The intraoperative time series data were extracted using a long short-term memory (LSTM) deep learning model. The baseline data and intraoperative features were then integrated through transfer learning and fused into each of the eight machine learning models for training. Based on the calculation of accuracy and area under the curve (AUC) of the prediction model, the best model was selected to establish the final OPCABG-AKI risk prediction model. The importance of features was calculated and ranked by DT model, to identify the main risk factors. Results Among 701 patients included in the study, 73 patients (10.4%) developed OPCABG-AKI. The GBDT model was shown to have the best predictions, both based on baseline data only (AUC =0.739, accuracy: 0.943) as well as based on baseline and intraoperative datasets (AUC =0.861, accuracy: 0.936). The ranking of importance of features of the GBDT model showed that use of insulin aspart was the most important predictor of OPCABG-AKI, followed by use of acarbose, spironolactone, alfentanil, dezocine, levosimendan, clindamycin, history of myocardial infarction, and gender. Conclusions A GBDT-based model showed excellent performance for the prediction of OPCABG-AKI. The fusion of preoperative and intraoperative data can improve the accuracy of predicting OPCABG-AKI.
Collapse
Affiliation(s)
- Yuezi Song
- Department of Anesthesiology, Chest Hospital, Tianjin University, Tianjin, China
| | - Wenqian Zhai
- Department of Anesthesiology, Chest Hospital, Tianjin University, Tianjin, China
- Tianjin Key Laboratory of Cardiovascular Emergencies and Critical Care, Tianjin, China
| | - Songnan Ma
- Academy of Medical Engineering and Translational Medicine, Tianjin University, Tianjin, China
| | - Yubo Wu
- Academy of Medical Engineering and Translational Medicine, Tianjin University, Tianjin, China
| | - Min Ren
- Tianjin Institute of Cardiovascular Disease, Tianjin, China
| | | | - Paolo Nardi
- Department of Cardiac Surgery, Tor Vergata University Hospital of Rome, Rome, Italy
| | - Philip Y. K. Pang
- Department of Cardiothoracic Surgery, National Heart Centre Singapore, Singapore, Singapore
| | - Jason M. Ali
- Department of Cardiothoracic Surgery, Royal Papworth Hospital, Cambridge, UK
| | - Jiange Han
- Department of Anesthesiology, Chest Hospital, Tianjin University, Tianjin, China
- Tianjin Key Laboratory of Cardiovascular Emergencies and Critical Care, Tianjin, China
| | - Zhigang Guo
- Tianjin Key Laboratory of Cardiovascular Emergencies and Critical Care, Tianjin, China
- Department of Cardiovascular Surgery, Chest Hospital, Tianjin University, Tianjin, China
| |
Collapse
|
|
6
|
Li A, Mullin S, Elkin PL. Improving Prediction of Survival for Extremely Premature Infants Born at 23 to 29 Weeks Gestational Age in the Neonatal Intensive Care Unit: Development and Evaluation of Machine Learning Models. JMIR Med Inform 2024; 12:e42271. [PMID: 38354033 PMCID: PMC10902770 DOI: 10.2196/42271] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 08/30/2022] [Revised: 02/02/2023] [Accepted: 12/28/2023] [Indexed: 03/02/2024] Open
Abstract
BACKGROUND Infants born at extremely preterm gestational ages are typically admitted to the neonatal intensive care unit (NICU) after initial resuscitation. The subsequent hospital course can be highly variable, and despite counseling aided by available risk calculators, there are significant challenges with shared decision-making regarding life support and transition to end-of-life care. Improving predictive models can help providers and families navigate these unique challenges. OBJECTIVE Machine learning methods have previously demonstrated added predictive value for determining intensive care unit outcomes, and their use allows consideration of a greater number of factors that potentially influence newborn outcomes, such as maternal characteristics. Machine learning-based models were analyzed for their ability to predict the survival of extremely preterm neonates at initial admission. METHODS Maternal and newborn information was extracted from the health records of infants born between 23 and 29 weeks of gestation in the Medical Information Mart for Intensive Care III (MIMIC-III) critical care database. Applicable machine learning models predicting survival during the initial NICU admission were developed and compared. The same type of model was also examined using only features that would be available prepartum for the purpose of survival prediction prior to an anticipated preterm birth. Features most correlated with the predicted outcome were determined when possible for each model. RESULTS Of included patients, 37 of 459 (8.1%) expired. The resulting random forest model showed higher predictive performance than the frequently used Score for Neonatal Acute Physiology With Perinatal Extension II (SNAPPE-II) NICU model when considering extremely preterm infants of very low birth weight. Several other machine learning models were found to have good performance but did not show a statistically significant difference from previously available models in this study. Feature importance varied by model, and those of greater importance included gestational age; birth weight; initial oxygenation level; elements of the APGAR (appearance, pulse, grimace, activity, and respiration) score; and amount of blood pressure support. Important prepartum features also included maternal age, steroid administration, and the presence of pregnancy complications. CONCLUSIONS Machine learning methods have the potential to provide robust prediction of survival in the context of extremely preterm births and allow for consideration of additional factors such as maternal clinical and socioeconomic information. Evaluation of larger, more diverse data sets may provide additional clarity on comparative performance.
Collapse
Affiliation(s)
- Angie Li
- Department of Biomedical Informatics, Jacobs School of Medicine and Biomedical Sciences, University at Buffalo, Buffalo, NY, United States
| | - Sarah Mullin
- Department of Biomedical Informatics, Jacobs School of Medicine and Biomedical Sciences, University at Buffalo, Buffalo, NY, United States
| | - Peter L Elkin
- Department of Biomedical Informatics, Jacobs School of Medicine and Biomedical Sciences, University at Buffalo, Buffalo, NY, United States
| |
Collapse
|
|
7
|
Keles E, Bagci U. The past, current, and future of neonatal intensive care units with artificial intelligence: a systematic review. NPJ Digit Med 2023; 6:220. [PMID: 38012349 PMCID: PMC10682088 DOI: 10.1038/s41746-023-00941-5] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 01/29/2023] [Accepted: 10/05/2023] [Indexed: 11/29/2023] Open
Abstract
Machine learning and deep learning are two subsets of artificial intelligence that involve teaching computers to learn and make decisions from any sort of data. Most recent developments in artificial intelligence are coming from deep learning, which has proven revolutionary in almost all fields, from computer vision to health sciences. The effects of deep learning in medicine have changed the conventional ways of clinical application significantly. Although some sub-fields of medicine, such as pediatrics, have been relatively slow in receiving the critical benefits of deep learning, related research in pediatrics has started to accumulate to a significant level, too. Hence, in this paper, we review recently developed machine learning and deep learning-based solutions for neonatology applications. We systematically evaluate the roles of both classical machine learning and deep learning in neonatology applications, define the methodologies, including algorithmic developments, and describe the remaining challenges in the assessment of neonatal diseases by using PRISMA 2020 guidelines. To date, the primary areas of focus in neonatology regarding AI applications have included survival analysis, neuroimaging, analysis of vital parameters and biosignals, and retinopathy of prematurity diagnosis. We have categorically summarized 106 research articles from 1996 to 2022 and discussed their pros and cons, respectively. In this systematic review, we aimed to further enhance the comprehensiveness of the study. We also discuss possible directions for new AI models and the future of neonatology with the rising power of AI, suggesting roadmaps for the integration of AI into neonatal intensive care units.
Collapse
Affiliation(s)
- Elif Keles
- Northwestern University, Feinberg School of Medicine, Department of Radiology, Chicago, IL, USA.
| | - Ulas Bagci
- Northwestern University, Feinberg School of Medicine, Department of Radiology, Chicago, IL, USA
- Northwestern University, Department of Biomedical Engineering, Chicago, IL, USA
- Department of Electrical and Computer Engineering, Chicago, IL, USA
| |
Collapse
|
|
8
|
Cheng TY, Yu-Chieh Ho S, Chien TW, Chou W. Global research trends in artificial intelligence for critical care with a focus on chord network charts: Bibliometric analysis. Medicine (Baltimore) 2023; 102:e35082. [PMID: 37746962 PMCID: PMC10519472 DOI: 10.1097/md.0000000000035082] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Journal Information] [Submit a Manuscript] [Subscribe] [Scholar Register] [Received: 03/05/2023] [Revised: 08/14/2023] [Accepted: 08/15/2023] [Indexed: 09/26/2023] Open
Abstract
BACKGROUND The field of critical care-related artificial intelligence (AI) research is rapidly gaining interest. However, there is still a lack of comprehensive bibliometric studies that measure and analyze scientific publications on a global scale. Network charts have traditionally been used to highlight author collaborations and coword phenomena (ACCP). It is necessary to determine whether chord network charts (CNCs) can provide a better understanding of ACCP, thus requiring clarification. This study aimed to achieve 2 objectives: evaluate global research trends in AI in intensive care medicine on publication outputs, coauthorships between nations, citations, and co-occurrences of keywords; and demonstrate the use of CNCs for ACCP in bibliometric analysis. METHODS The web of science database was searched for a total of 1992 documents published between 2013 and 2022. The document type was limited to articles and article reviews, and titles and abstracts were screened for eligibility. The characteristics of the publications, including preferred journals, leading research countries, international collaborations, top institutions, and major keywords, were analyzed using the category-journal rank-authorship-L-index score and trend analysis. The 100 most highly cited articles are also listed in detail. RESULTS Between 2018 and 2022, there was a sharp increase in publications, which accounted for 92.8% (1849/1992) of all papers included in the study. The United States and China were responsible for nearly 50% (936/1992) of the total publications. The leading countries, institutes, departments, authors, and journals in terms of publications were the US, Massachusetts Gen Hosp (US), Medical School, Zhongheng Zhang (China), and Science Reports. The top 3 primary keywords denoting research hotspots for AI in critically ill patients were mortality, model, and intensive care unit, with mortality having the highest burst strength (4.49). The keywords risk and system showed the highest growth trend (0.98) in counts over the past 4 years. CONCLUSIONS This study provides valuable insights into the potential for ACCP and future research opportunities. For AI-based clinical research to become widely accepted in critical care practice, collaborative research efforts are necessary to strengthen the maturity and robustness of AI-driven models using CNCs for display.
Collapse
Affiliation(s)
- Teng-Yun Cheng
- Department of Emergency Medicine, Chi-Mei Medical Center, Liouying, Tainan, Taiwan
| | - Sam Yu-Chieh Ho
- Department of Emergency Medicine, Chi-Mei Medical Center, Tainan, Taiwan
| | - Tsair-Wei Chien
- Department of Medical Research, Chi-Mei Medical Center, Tainan, Taiwan
| | - Willy Chou
- Department of Physical Medicine and Rehabilitation, Chi Mei Medical Center, Jiali, Tainan, Taiwan
- Department of Physical Medicine and Rehabilitation, Chung San Medical University Hospital, Taichung, Taiwan
| |
Collapse
|
|
9
|
Balch JA, Ruppert MM, Shickel B, Ozrazgat-Baslanti T, Tighe PJ, Efron PA, Upchurch GR, Rashidi P, Bihorac A, Loftus TJ. Building an automated, machine learning-enabled platform for predicting post-operative complications. Physiol Meas 2023; 44:024001. [PMID: 36657179 PMCID: PMC9910093 DOI: 10.1088/1361-6579/acb4db] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Grants] [Track Full Text] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 09/04/2022] [Revised: 12/29/2022] [Accepted: 01/19/2023] [Indexed: 01/21/2023]
Abstract
Objective. In 2019, the University of Florida College of Medicine launched theMySurgeryRiskalgorithm to predict eight major post-operative complications using automatically extracted data from the electronic health record.Approach. This project was developed in parallel with our Intelligent Critical Care Center and represents a culmination of efforts to build an efficient and accurate model for data processing and predictive analytics.Main Results and Significance. This paper discusses how our model was constructed and improved upon. We highlight the consolidation of the database, processing of fixed and time-series physiologic measurements, development and training of predictive models, and expansion of those models into different aspects of patient assessment and treatment. We end by discussing future directions of the model.
Collapse
Affiliation(s)
- Jeremy A Balch
- Intelligent Critical Care Center, University of Florida, Gainesville, FL, United States of America
- Department of Surgery, University of Florida, Gainesville, Florida, United States of America
| | - Matthew M Ruppert
- Intelligent Critical Care Center, University of Florida, Gainesville, FL, United States of America
- Department of Medicine, University of Florida, Gainesville, Florida, United States of America
| | - Benjamin Shickel
- Intelligent Critical Care Center, University of Florida, Gainesville, FL, United States of America
- Department of Medicine, University of Florida, Gainesville, Florida, United States of America
| | - Tezcan Ozrazgat-Baslanti
- Intelligent Critical Care Center, University of Florida, Gainesville, FL, United States of America
- Department of Medicine, University of Florida, Gainesville, Florida, United States of America
| | - Patrick J Tighe
- Department of Anesthesiology, University of Florida, Gainesville, Florida, United States of America
| | - Philip A Efron
- Department of Surgery, University of Florida, Gainesville, Florida, United States of America
| | - Gilbert R Upchurch
- Department of Surgery, University of Florida, Gainesville, Florida, United States of America
| | - Parisa Rashidi
- Intelligent Critical Care Center, University of Florida, Gainesville, FL, United States of America
- Department of Biomedical Engineering, University of Florida, Gainesville, Florida, United States of America
| | - Azra Bihorac
- Intelligent Critical Care Center, University of Florida, Gainesville, FL, United States of America
- Department of Medicine, University of Florida, Gainesville, Florida, United States of America
| | - Tyler J Loftus
- Intelligent Critical Care Center, University of Florida, Gainesville, FL, United States of America
- Department of Surgery, University of Florida, Gainesville, Florida, United States of America
| |
Collapse
|
|
10
|
Beam KS, Zupancic JAF. Machine learning: remember the fundamentals. Pediatr Res 2023; 93:291-292. [PMID: 36550355 DOI: 10.1038/s41390-022-02420-1] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Track Full Text] [Journal Information] [Submit a Manuscript] [Subscribe] [Scholar Register] [Received: 11/22/2022] [Accepted: 11/30/2022] [Indexed: 12/24/2022]
Affiliation(s)
- Kristyn S Beam
- Department of Neonatology, Beth Israel Deaconess Medical Center, Boston, MA, USA.,Department of Pediatrics, Harvard Medical School, Boston, MA, USA
| | - John A F Zupancic
- Department of Neonatology, Beth Israel Deaconess Medical Center, Boston, MA, USA. .,Department of Pediatrics, Harvard Medical School, Boston, MA, USA.
| |
Collapse
|
|
11
|
Sullivan BA, Kausch SL, Fairchild KD. Artificial and human intelligence for early identification of neonatal sepsis. Pediatr Res 2023; 93:350-356. [PMID: 36127407 PMCID: PMC11749885 DOI: 10.1038/s41390-022-02274-7] [Citation(s) in RCA: 3] [Impact Index Per Article: 1.5] [Reference Citation Analysis] [Abstract] [MESH Headings] [Grants] [Track Full Text] [Journal Information] [Submit a Manuscript] [Subscribe] [Scholar Register] [Received: 04/18/2022] [Revised: 07/29/2022] [Accepted: 08/05/2022] [Indexed: 11/09/2022]
Abstract
Artificial intelligence may have a role in the early detection of sepsis in neonates. Machine learning can identify patterns that predict high or increasing risk for clinical deterioration from a sepsis-like illness. In developing this potential addition to NICU care, careful consideration should be given to the data and methods used to develop, validate, and evaluate prediction models. When an AI system alerts clinicians to a change in a patient's condition that warrants a bedside evaluation, human intelligence and experience come into play to determine an appropriate course of action: evaluate and treat or wait and watch closely. With intelligently developed, validated, and implemented AI sepsis systems, both clinicians and patients stand to benefit. IMPACT: This narrative review highlights the application of AI in neonatal sepsis prediction. It describes issues in clinical prediction model development specific to this population. This article reviews the methods, considerations, and literature on neonatal sepsis model development and validation. Challenges of AI technology and potential barriers to using sepsis AI systems in the NICU are discussed.
Collapse
Affiliation(s)
- Brynne A Sullivan
- Department of Pediatrics, University of Virginia School of Medicine, Charlottesville, VA, USA.
| | - Sherry L Kausch
- Department of Pediatrics, University of Virginia School of Medicine, Charlottesville, VA, USA
| | - Karen D Fairchild
- Department of Pediatrics, University of Virginia School of Medicine, Charlottesville, VA, USA
| |
Collapse
|
|
12
|
Coleman J, Ginsburg AS, Macharia WM, Ochieng R, Chomba D, Zhou G, Dunsmuir D, Karlen W, Ansermino JM. Assessment of neonatal respiratory rate variability. J Clin Monit Comput 2022; 36:1869-1879. [PMID: 35332406 PMCID: PMC9637627 DOI: 10.1007/s10877-022-00840-2] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Grants] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 08/18/2021] [Accepted: 03/02/2022] [Indexed: 11/30/2022]
Abstract
Accurate measurement of respiratory rate (RR) in neonates is challenging due to high neonatal RR variability (RRV). There is growing evidence that RRV measurement could inform and guide neonatal care. We sought to quantify neonatal RRV during a clinical study in which we compared multiparameter continuous physiological monitoring (MCPM) devices. Measurements of capnography-recorded exhaled carbon dioxide across 60-s epochs were collected from neonates admitted to the neonatal unit at Aga Khan University-Nairobi hospital. Breaths were manually counted from capnograms and using an automated signal detection algorithm which also calculated mean and median RR for each epoch. Outcome measures were between- and within-neonate RRV, between- and within-epoch RRV, and 95% limits of agreement, bias, and root-mean-square deviation. Twenty-seven neonates were included, with 130 epochs analysed. Mean manual breath count (MBC) was 48 breaths per minute. Median RRV ranged from 11.5% (interquartile range (IQR) 6.8-18.9%) to 28.1% (IQR 23.5-36.7%). Bias and limits of agreement for MBC vs algorithm-derived breath count, MBC vs algorithm-derived median breath rate, MBC vs algorithm-derived mean breath rate were - 0.5 (- 2.7, 1.66), - 3.16 (- 12.12, 5.8), and - 3.99 (- 11.3, 3.32), respectively. The marked RRV highlights the challenge of performing accurate RR measurements in neonates. More research is required to optimize the use of RRV to improve care. When evaluating MCPM devices, accuracy thresholds should be less stringent in newborns due to increased RRV. Lastly, median RR, which discounts the impact of extreme outliers, may be more reflective of the underlying physiological control of breathing.
Collapse
Affiliation(s)
- Jesse Coleman
- Evaluation of Technologies for Neonates in Africa (ETNA), Nairobi, Kenya.
- Centre for International Child Health, 305 - 4088 Cambie Street, Vancouver, BC, V5Z 2X8, Canada.
| | | | | | | | - Dorothy Chomba
- Department of Pediatrics, Aga Khan University, Nairobi, Kenya
| | - Guohai Zhou
- Center for Clinical Investigation, Brigham and Women's Hospital, Boston, MA, USA
| | - Dustin Dunsmuir
- Anesthesiology, Pharmacology & Therapeutics, The University of British Columbia, Vancouver, BC, Canada
| | - Walter Karlen
- Mobile Health Systems Lab, Department of Health Sciences and Technology, ETH Zurich, Zurich, Switzerland
| | - J Mark Ansermino
- Anesthesiology, Pharmacology & Therapeutics, The University of British Columbia, Vancouver, BC, Canada
| |
Collapse
|
|
13
|
Suresh S, Newton DT, Everett TH, Lin G, Duerstock BS. Feature Selection Techniques for a Machine Learning Model to Detect Autonomic Dysreflexia. Front Neuroinform 2022; 16:901428. [PMID: 36033642 PMCID: PMC9416695 DOI: 10.3389/fninf.2022.901428] [Citation(s) in RCA: 8] [Impact Index Per Article: 2.7] [Reference Citation Analysis] [Abstract] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 03/21/2022] [Accepted: 05/23/2022] [Indexed: 11/15/2022] Open
Abstract
Feature selection plays a crucial role in the development of machine learning algorithms. Understanding the impact of the features on a model, and their physiological relevance can improve the performance. This is particularly helpful in the healthcare domain wherein disease states need to be identified with relatively small quantities of data. Autonomic Dysreflexia (AD) is one such example, wherein mismanagement of this neurological condition could lead to severe consequences for individuals with spinal cord injuries. We explore different methods of feature selection needed to improve the performance of a machine learning model in the detection of the onset of AD. We present different techniques used as well as the ideal metrics using a dataset of thirty-six features extracted from electrocardiograms, skin nerve activity, blood pressure and temperature. The best performing algorithm was a 5-layer neural network with five relevant features, which resulted in 93.4% accuracy in the detection of AD. The techniques in this paper can be applied to a myriad of healthcare datasets allowing forays into deeper exploration and improved machine learning model development. Through critical feature selection, it is possible to design better machine learning algorithms for detection of niche disease states using smaller datasets.
Collapse
Affiliation(s)
- Shruthi Suresh
- Weldon School of Biomedical Engineering, Purdue University, West Lafayette, IN, United States
| | - David T. Newton
- Department of Statistics, Purdue University, West Lafayette, IN, United States
| | - Thomas H. Everett
- Krannert Cardiovascular Research Center, Indiana University School of Medicine, Indianapolis, IN, United States
| | - Guang Lin
- School of Mechanical Engineering, Purdue University, West Lafayette, IN, United States
- Department of Mathematics, Purdue University, West Lafayette, IN, United States
| | - Bradley S. Duerstock
- Weldon School of Biomedical Engineering, Purdue University, West Lafayette, IN, United States
- School of Industrial Engineering, Purdue University, West Lafayette, IN, United States
- *Correspondence: Bradley S. Duerstock,
| |
Collapse
|
|
14
|
Tan S, Unnikrishnan KP. Using Temporal Data Mining on Patient Data for Clinical Decision Making in the Care of the Sick Newborn. EC PAEDIATRICS : OPEN ACCESS 2022; 11:44-56. [PMID: 35790097 PMCID: PMC9249406] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Grants] [Subscribe] [Scholar Register] [Indexed: 06/15/2023]
Abstract
BACKGROUND In a neonatal intensive care unit, streaming healthcare data comes from many sources, but humans are unable to understand relationships between data variables. Data mining and analysis are just beginning to get utilized in critical care. We present a case study using electronic medical record data in the neonatal intensive care unit and explore possible avenues of advancement using temporal data analytics. CASE PRESENTATION Electronic medical record data were collected for physiological monitor data. Heart rate, respiratory rate, oxygen saturation and temperature data were retrospectively analyzed by temporal data mining. Three premature babies were selected and data de-identified. The first case of a urinary tract infection showed nursing ability to synthesize data streams coming from a patient. For the second case of necrotizing enterocolitis, Temporal-Data-Mining analysis of combinations of clinical events based on deviations from the mean showed specific heuristic biomarkers related to events before discovery of necrotizing enterocolitis. Specific sequences 6-event and 5-event in length were identified with nursing unease at clinical deterioration, which were 100- and 87-times unlikely to occur randomly with 99.5% confidence. No such sequences were found in the rest of the 37 days for the second case and entire 133 days of stay in the third case of an uneventful premature baby. CONCLUSION Temporal data mining is a possible clinical tool in providing useful information in the neonatal intensive care unit for diagnosis of adverse clinical occurrences such as necrotizing enterocolitis. There is the possibility of changing the clinical paradigm of episodic watchfulness to constant vigilance using real-time data gathering.
Collapse
Affiliation(s)
- Sidhartha Tan
- Department of Pediatrics, Wayne State University School of Medicine, Detroit, MI, USA
| | | |
Collapse
|
|
15
|
Solís-García G, Maderuelo-Rodríguez E, Perez-Pérez T, Torres-Soblechero L, Gutiérrez-Vélez A, Ramos-Navarro C, López-Martínez R, Sánchez-Luna M. Longitudinal Analysis of Continuous Pulse Oximetry as Prognostic Factor in Neonatal Respiratory Distress. Am J Perinatol 2022; 39:677-682. [PMID: 33075845 DOI: 10.1055/s-0040-1718877] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Track Full Text] [Journal Information] [Submit a Manuscript] [Subscribe] [Scholar Register] [Indexed: 10/23/2022]
Abstract
OBJECTIVE Analysis of longitudinal data can provide neonatologists with tools that can help predict clinical deterioration and improve outcomes. The aim of this study is to analyze continuous monitoring data in newborns, using vital signs to develop predictive models for intensive care admission and time to discharge. STUDY DESIGN We conducted a retrospective cohort study, including term and preterm newborns with respiratory distress patients admitted to the neonatal ward. Clinical and epidemiological data, as well as mean heart rate and saturation, at every minute for the first 12 hours of admission were collected. Multivariate mixed, survival and joint models were developed. RESULTS A total of 56,377 heart rate and 56,412 oxygen saturation data were analyzed from 80 admitted patients. Of them, 73 were discharged home and 7 required transfer to the intensive care unit (ICU). Longitudinal evolution of heart rate (p < 0.01) and oxygen saturation (p = 0.01) were associated with time to discharge, as well as birth weight (p < 0.01) and type of delivery (p < 0.01). Longitudinal heart rate evolution (p < 0.01) and fraction of inspired oxygen at admission at the ward (p < 0.01) predicted neonatal ICU (NICU) admission. CONCLUSION Longitudinal evolution of heart rate can help predict time to transfer to intensive care, and both heart rate and oxygen saturation can help predict time to discharge. Analysis of continuous monitoring data in patients admitted to neonatal wards provides useful tools to stratify risks and helps in taking medical decisions. KEY POINTS · Continuous monitoring of vital signs can help predict and prevent clinical deterioration in neonatal patients.. · In our study, longitudinal analysis of heart rate and oxygen saturation predicted time to discharge and intensive care admission.. · More studies are needed to prospectively prove that these models can helpmake clinical decisions and stratify patients' risks..
Collapse
Affiliation(s)
- Gonzalo Solís-García
- Department of Neonatology, Hospital General Universitario Gregorio Marañón, Madrid, Spain
| | | | - Teresa Perez-Pérez
- Department of Statistics, Universidad Complutense de Madrid, Madrid, Spain
| | | | - Ana Gutiérrez-Vélez
- Department of Neonatology, Hospital General Universitario Gregorio Marañón, Madrid, Spain
| | - Cristina Ramos-Navarro
- Department of Neonatology, Hospital General Universitario Gregorio Marañón, Madrid, Spain
| | - Raúl López-Martínez
- Information Technology Department, Hospital General Universitario Gregorio Marañón, Madrid, Spain
| | - Manuel Sánchez-Luna
- Department of Neonatology, Hospital General Universitario Gregorio Marañón, Madrid, Spain
| |
Collapse
|
|
16
|
Randall Moorman J. The principles of whole-hospital predictive analytics monitoring for clinical medicine originated in the neonatal ICU. NPJ Digit Med 2022; 5:41. [PMID: 35361861 PMCID: PMC8971442 DOI: 10.1038/s41746-022-00584-y] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Grants] [Track Full Text] [Download PDF] [Journal Information] [Subscribe] [Scholar Register] [Received: 09/28/2021] [Accepted: 02/23/2022] [Indexed: 11/17/2022] Open
Abstract
In 2011, a multicenter group spearheaded at the University of Virginia demonstrated reduced mortality from real-time continuous cardiorespiratory monitoring in the neonatal ICU using what we now call Artificial Intelligence, Big Data, and Machine Learning. The large, randomized heart rate characteristics trial made real, for the first time that we know of, the promise that early detection of illness would allow earlier and more effective intervention and improved patient outcomes. Currently, though, we hear as much of failures as we do of successes in the rapidly growing field of predictive analytics monitoring that has followed. This Perspective aims to describe the principles of how we developed heart rate characteristics monitoring for neonatal sepsis and then applied them throughout adult ICU and hospital medicine. It primarily reflects the work since the 1990s of the University of Virginia group: the theme is that sudden and catastrophic deteriorations can be preceded by subclinical but measurable physiological changes apparent in the continuous cardiorespiratory monitoring and electronic health record.
Collapse
Affiliation(s)
- J Randall Moorman
- Cardiovascular Division, Department of Internal Medicine, Center for Advanced Medical Analytics, University of Virginia, Charlottesville, VA, USA.
| |
Collapse
|
|
17
|
Loftus TJ, Tighe PJ, Ozrazgat-Baslanti T, Davis JP, Ruppert MM, Ren Y, Shickel B, Kamaleswaran R, Hogan WR, Moorman JR, Upchurch GR, Rashidi P, Bihorac A. Ideal algorithms in healthcare: Explainable, dynamic, precise, autonomous, fair, and reproducible. PLOS DIGITAL HEALTH 2022; 1:e0000006. [PMID: 36532301 PMCID: PMC9754299 DOI: 10.1371/journal.pdig.0000006] [Citation(s) in RCA: 21] [Impact Index Per Article: 7.0] [Reference Citation Analysis] [Abstract] [Grants] [Track Full Text] [Subscribe] [Scholar Register] [Indexed: 12/22/2022]
Abstract
Established guidelines describe minimum requirements for reporting algorithms in healthcare; it is equally important to objectify the characteristics of ideal algorithms that confer maximum potential benefits to patients, clinicians, and investigators. We propose a framework for ideal algorithms, including 6 desiderata: explainable (convey the relative importance of features in determining outputs), dynamic (capture temporal changes in physiologic signals and clinical events), precise (use high-resolution, multimodal data and aptly complex architecture), autonomous (learn with minimal supervision and execute without human input), fair (evaluate and mitigate implicit bias and social inequity), and reproducible (validated externally and prospectively and shared with academic communities). We present an ideal algorithms checklist and apply it to highly cited algorithms. Strategies and tools such as the predictive, descriptive, relevant (PDR) framework, the Standard Protocol Items: Recommendations for Interventional Trials-Artificial Intelligence (SPIRIT-AI) extension, sparse regression methods, and minimizing concept drift can help healthcare algorithms achieve these objectives, toward ideal algorithms in healthcare.
Collapse
Affiliation(s)
- Tyler J. Loftus
- Department of Surgery, University of Florida Health, Gainesville, Florida, United States of America
- Precision and Intelligent Systems in Medicine (PrismaP), University of Florida, Gainesville, Florida, United States of America
| | - Patrick J. Tighe
- Departments of Anesthesiology, Orthopedics, and Information Systems/Operations Management, University of Florida Health, Gainesville, Florida, United States of America
| | - Tezcan Ozrazgat-Baslanti
- Precision and Intelligent Systems in Medicine (PrismaP), University of Florida, Gainesville, Florida, United States of America
- Department of Medicine, University of Florida Health, Gainesville, Florida, United States of America
| | - John P. Davis
- Department of Surgery, University of Virginia, Charlottesville, Virginia, United States of America
| | - Matthew M. Ruppert
- Precision and Intelligent Systems in Medicine (PrismaP), University of Florida, Gainesville, Florida, United States of America
- Department of Medicine, University of Florida Health, Gainesville, Florida, United States of America
| | - Yuanfang Ren
- Precision and Intelligent Systems in Medicine (PrismaP), University of Florida, Gainesville, Florida, United States of America
- Department of Medicine, University of Florida Health, Gainesville, Florida, United States of America
| | - Benjamin Shickel
- Precision and Intelligent Systems in Medicine (PrismaP), University of Florida, Gainesville, Florida, United States of America
- Department of Medicine, University of Florida Health, Gainesville, Florida, United States of America
| | - Rishikesan Kamaleswaran
- Department of Biomedical Informatics, Emory University School of Medicine, Atlanta, Georgia, United States of America
| | - William R. Hogan
- Department of Health Outcomes & Biomedical Informatics, College of Medicine, University of Florida, Gainesville, Florida, United States of America
| | - J. Randall Moorman
- Department of Medicine, University of Virginia, Charlottesville, Virginia, United States of America
| | - Gilbert R. Upchurch
- Department of Surgery, University of Florida Health, Gainesville, Florida, United States of America
| | - Parisa Rashidi
- Precision and Intelligent Systems in Medicine (PrismaP), University of Florida, Gainesville, Florida, United States of America
- Departments of Biomedical Engineering, Computer and Information Science and Engineering, and Electrical and Computer Engineering, University of Florida, Gainesville, Florida, United States of America
| | - Azra Bihorac
- Precision and Intelligent Systems in Medicine (PrismaP), University of Florida, Gainesville, Florida, United States of America
- Department of Medicine, University of Florida Health, Gainesville, Florida, United States of America
| |
Collapse
|
|
18
|
Niestroy JC, Moorman JR, Levinson MA, Manir SA, Clark TW, Fairchild KD, Lake DE. Discovery of signatures of fatal neonatal illness in vital signs using highly comparative time-series analysis. NPJ Digit Med 2022; 5:6. [PMID: 35039624 PMCID: PMC8764068 DOI: 10.1038/s41746-021-00551-z] [Citation(s) in RCA: 10] [Impact Index Per Article: 3.3] [Reference Citation Analysis] [Abstract] [Key Words] [Grants] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 03/28/2021] [Accepted: 12/13/2021] [Indexed: 12/23/2022] Open
Abstract
To seek new signatures of illness in heart rate and oxygen saturation vital signs from Neonatal Intensive Care Unit (NICU) patients, we implemented highly comparative time-series analysis to discover features of all-cause mortality in the next 7 days. We collected 0.5 Hz heart rate and oxygen saturation vital signs of infants in the University of Virginia NICU from 2009 to 2019. We applied 4998 algorithmic operations from 11 mathematical families to random daily 10 min segments from 5957 NICU infants, 205 of whom died. We clustered the results and selected a representative from each, and examined multivariable logistic regression models. 3555 operations were usable; 20 cluster medoids held more than 81% of the information, and a multivariable model had AUC 0.83. New algorithms outperformed others: moving threshold, successive increases, surprise, and random walk. We computed provenance of the computations and constructed a software library with links to the data. We conclude that highly comparative time-series analysis revealed new vital sign measures to identify NICU patients at the highest risk of death in the next week.
Collapse
Affiliation(s)
- Justin C Niestroy
- Department of Public Health Sciences, University of Virginia, Charlottesville, VA, 22947, USA
- Center for Advanced Medical Analytics, University of Virginia, Charlottesville, VA, 22947, USA
| | - J Randall Moorman
- Center for Advanced Medical Analytics, University of Virginia, Charlottesville, VA, 22947, USA.
- Department of Medicine, University of Virginia, Charlottesville, VA, 22947, USA.
| | - Maxwell A Levinson
- Department of Public Health Sciences, University of Virginia, Charlottesville, VA, 22947, USA
- Center for Advanced Medical Analytics, University of Virginia, Charlottesville, VA, 22947, USA
| | - Sadnan Al Manir
- Department of Public Health Sciences, University of Virginia, Charlottesville, VA, 22947, USA
- Center for Advanced Medical Analytics, University of Virginia, Charlottesville, VA, 22947, USA
| | - Timothy W Clark
- Department of Public Health Sciences, University of Virginia, Charlottesville, VA, 22947, USA
- Center for Advanced Medical Analytics, University of Virginia, Charlottesville, VA, 22947, USA
- School of Data Science, University of Virginia, Charlottesville, VA, 22947, USA
| | - Karen D Fairchild
- Center for Advanced Medical Analytics, University of Virginia, Charlottesville, VA, 22947, USA
- Department of Pediatrics, University of Virginia, Charlottesville, VA, 22947, USA
| | - Douglas E Lake
- Center for Advanced Medical Analytics, University of Virginia, Charlottesville, VA, 22947, USA
- Department of Medicine, University of Virginia, Charlottesville, VA, 22947, USA
- Department of Statistics, University of Virginia, Charlottesville, VA, 22947, USA
| |
Collapse
|
|
19
|
Sullivan BA, Fairchild KD. Vital signs as physiomarkers of neonatal sepsis. Pediatr Res 2022; 91:273-282. [PMID: 34493832 DOI: 10.1038/s41390-021-01709-x] [Citation(s) in RCA: 21] [Impact Index Per Article: 7.0] [Reference Citation Analysis] [Abstract] [Track Full Text] [Journal Information] [Submit a Manuscript] [Subscribe] [Scholar Register] [Received: 05/04/2021] [Revised: 07/24/2021] [Accepted: 07/27/2021] [Indexed: 02/08/2023]
Abstract
Neonatal sepsis accounts for significant morbidity and mortality, particularly among premature infants in the Neonatal Intensive Care Unit. Abnormal vital sign patterns serve as physiomarkers of sepsis and provide early warning of illness before overt clinical decompensation. The systemic inflammatory response to pathogens signals the autonomic nervous system, leading to changes in temperature, respiratory rate, heart rate, and blood pressure. In infants with comorbidities of prematurity, vital sign abnormalities often occur in the absence of infection, which confounds sepsis diagnosis. This review will cover the mechanisms of vital sign changes in neonatal sepsis, including the cholinergic anti-inflammatory pathway mediated by the vagus nerve, which is critical to the host response to infectious and inflammatory insults. We will also review the clinical implications of vital sign changes in neonatal sepsis, including their use in early warning scores and systems to direct clinicians to the bedside of infants with physiologic changes that might be due to sepsis. IMPACT: This manuscript summarizes and reviews the relevant literature on the physiological manifestations of neonatal sepsis and how we monitor and analyze these through vital signs and advanced analytics.
Collapse
Affiliation(s)
- Brynne A Sullivan
- Division of Neonatology, Department of Pediatrics, University of Virginia School of Medicine, Charlottesville, VA, USA.
| | - Karen D Fairchild
- Division of Neonatology, Department of Pediatrics, University of Virginia School of Medicine, Charlottesville, VA, USA
| |
Collapse
|
|
20
|
Reiss JD, Peterson LS, Nesamoney SN, Chang AL, Pasca AM, Marić I, Shaw GM, Gaudilliere B, Wong RJ, Sylvester KG, Bonifacio SL, Aghaeepour N, Gibbs RS, Stevenson DK. Perinatal infection, inflammation, preterm birth, and brain injury: A review with proposals for future investigations. Exp Neurol 2022; 351:113988. [DOI: 10.1016/j.expneurol.2022.113988] [Citation(s) in RCA: 3] [Impact Index Per Article: 1.0] [Reference Citation Analysis] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 06/30/2021] [Revised: 01/06/2022] [Accepted: 01/13/2022] [Indexed: 11/26/2022]
|
|
21
|
Adegboro CO, Choudhury A, Asan O, Kelly MM. Artificial Intelligence to Improve Health Outcomes in the NICU and PICU: A Systematic Review. Hosp Pediatr 2022; 12:93-110. [PMID: 34890453 DOI: 10.1542/hpeds.2021-006094] [Citation(s) in RCA: 5] [Impact Index Per Article: 1.7] [Reference Citation Analysis] [Abstract] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 12/23/2022]
Abstract
CONTEXT Artificial intelligence (AI) technologies are increasingly used in pediatrics and have the potential to help inpatient physicians provide high-quality care for critically ill children. OBJECTIVE We aimed to describe the use of AI to improve any health outcome(s) in neonatal and pediatric intensive care. DATA SOURCE PubMed, IEEE Xplore, Cochrane, and Web of Science databases. STUDY SELECTION We used peer-reviewed studies published between June 1, 2010, and May 31, 2020, in which researchers described (1) AI, (2) pediatrics, and (3) intensive care. Studies were included if researchers assessed AI use to improve at least 1 health outcome (eg, mortality). DATA EXTRACTION Data extraction was conducted independently by 2 researchers. Articles were categorized by direct or indirect impact of AI, defined by the European Institute of Innovation and Technology Health joint report. RESULTS Of the 287 publications screened, 32 met inclusion criteria. Approximately 22% (n = 7) of studies revealed a direct impact and improvement in health outcomes after AI implementation. Majority were in prototype testing, and few were deployed into an ICU setting. Among the remaining 78% (n = 25) AI models outperformed standard clinical modalities and may have indirectly influenced patient outcomes. Quantitative assessment of health outcomes using statistical measures, such as area under the receiver operating curve (56%; n = 18) and specificity (38%; n = 12), revealed marked heterogeneity in metrics and standardization. CONCLUSIONS Few studies have revealed that AI has directly improved health outcomes for pediatric critical care patients. Further prospective, experimental studies are needed to assess AI's impact by using established implementation frameworks, standardized metrics, and validated outcome measures.
Collapse
Affiliation(s)
- Claudette O Adegboro
- Department of Pediatrics, School of Medicine and Public Health, University of Wisconsin, Madison, Wisconsin
| | - Avishek Choudhury
- Division of Engineering Management, School of Systems and Enterprise, Stevens Institute of Technology, Hoboken, New Jersey
| | - Onur Asan
- Division of Engineering Management, School of Systems and Enterprise, Stevens Institute of Technology, Hoboken, New Jersey
| | - Michelle M Kelly
- Department of Pediatrics, School of Medicine and Public Health, University of Wisconsin, Madison, Wisconsin
| |
Collapse
|
|
22
|
[Predicting prolonged length of intensive care unit stay via machine learning]. BEIJING DA XUE XUE BAO. YI XUE BAN = JOURNAL OF PEKING UNIVERSITY. HEALTH SCIENCES 2021; 53. [PMID: 34916699 PMCID: PMC8695140 DOI: 10.19723/j.issn.1671-167x.2021.06.026] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Subscribe] [Scholar Register] [Indexed: 01/24/2023]
Abstract
OBJECTIVE To construct length of intensive care unit (ICU) stay (LOS-ICU) prediction models for ICU patients, based on three machine learning models support vector machine (SVM), classification and regression tree (CART), and random forest (RF), and to compare the prediction perfor-mance of the three machine learning models with the customized simplified acute physiology score Ⅱ(SAPS-Ⅱ) model. METHODS We used medical information mart for intensive care (MIMIC)-Ⅲ database for model development and validation. The primary outcome was prolonged LOS-ICU(pLOS-ICU), defined as longer than the third quartile of patients' LOS-ICU in the studied dataset. The recursive feature elimination method was used to do feature selection for three machine learning models. We utilized 5-fold cross validation to evaluate model prediction performance. The Brier value, area under the receiver operation characteristic curve (AUROC), and estimated calibration index (ECI) were used as perfor-mance measures. Performances of the four models were compared, and performance differences between the models were assessed using two-sided t test. The model with the best prediction performance was employed to generate variable importance ranking, and the identified top five important predictors were pre-sented. RESULTS The final cohort in our study consisted of 40 200 eligible ICU patients, of whom 23.7% were with pLOS-ICU. The proportion of the male patients was 57.6%, and the age of all the ICU patients was (61.9±16.5) years.Results showed that the three machine learning models outperformed the customized SAPS-Ⅱ model in terms of all the performance measures with statistical significance (P < 0.01). Among the three machine learning models, the RF model achieved the best overall performance (Brier value, 0.145), discrimination (AUROC, 0.770) and calibration (ECI, 7.259). The calibration curve showed that the RF model slightly overestimated the risk of pLOS-ICU in high-risk ICU patients, but underestimated the risk of pLOS-ICU in low-risk ICU patients. Top five important predictors for pLOS-ICU identified by the RF model included age, heart rate, systolic blood pressure, body tempe-rature, and ratio of arterial oxygen tension to the fraction of inspired oxygen(PaO2/FiO2). CONCLUSION The RF algorithm-based pLOS-ICU prediction model had a best prediction performance in this study. It lays a foundation for future application of the RF-based pLOS-ICU prediction model in ICU clinical practice.
Collapse
|
|
23
|
Hartley C. Toward personalized medicine for pharmacological interventions in neonates using vital signs. PAEDIATRIC AND NEONATAL PAIN 2021; 3:147-155. [PMID: 35372840 PMCID: PMC8937573 DOI: 10.1002/pne2.12065] [Citation(s) in RCA: 2] [Impact Index Per Article: 0.5] [Reference Citation Analysis] [Abstract] [Track Full Text] [Download PDF] [Figures] [Subscribe] [Scholar Register] [Received: 01/14/2021] [Revised: 10/22/2021] [Accepted: 11/15/2021] [Indexed: 11/17/2022]
Abstract
Vital signs, such as heart rate and oxygen saturation, are continuously monitored for infants in neonatal care units. Pharmacological interventions can alter an infant's vital signs, either as an intended effect or as a side effect, and consequently could provide an approach to explore the wide variability in pharmacodynamics across infants and could be used to develop models to predict outcome (efficacy or adverse effects) in an individual infant. This will enable doses to be tailored according to the individual, shifting the balance toward efficacy and away from the adverse effects of a drug. Pharmacological analgesics are frequently not given in part due to the risk of adverse effects, yet this exposes infants to the short‐ and long‐term effects of painful procedures. Personalized analgesic dosing will be an important step forward in providing safer effective pain relief in infants. The aim of this paper was to describe a framework to develop predictive models of drug outcome from analysis of vital signs data, focusing on analgesics as a representative example. This framework investigates changes in vital signs in response to the analgesic (prior to the painful procedure) and proposes using machine learning to examine if these changes are predictive of outcome—either efficacy (with pain response measured using a multimodal approach, as changes in vital signs alone have limited sensitivity and specificity) or adverse effects. The framework could be applied to both preterm and term infants in neonatal care units, as well as older children. Sharing vital signs data are proposed as a means to achieve this aim and bring personalized medicine rapidly to the forefront in neonatology.
Collapse
|
|
24
|
Monfredi O, Keim-Malpass J, Moorman JR. Continuous cardiorespiratory monitoring is a dominant source of predictive signal in machine learning for risk stratification and clinical decision support . Physiol Meas 2021;42. [PMID: 34580243 DOI: 10.1088/1361-6579/ac2130] [Citation(s) in RCA: 4] [Impact Index Per Article: 1.0] [Reference Citation Analysis] [Abstract] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 05/24/2021] [Accepted: 08/25/2021] [Indexed: 12/23/2022]
Abstract
Beaulieu-Jones and coworkers propose a litmus test for the field of predictive analytics-performance improvements must be demonstrated to be the result of non-clinician-initiated data, otherwise, there should be caution in assuming that predictive models could improve clinical decision-making (Beaulieu-Joneset al2021). They demonstrate substantial prognostic information in unsorted physician orders made before the first midnight of hospital admission, and we are persuaded that it is fair to ask-if the physician thought of it first, what exactly is machine learning for in-patient risk stratification learning about? While we want predictive analytics to represent the leading indicators of a patient's illness, does it instead merely reflect the lagging indicators of clinicians' actions? We propose that continuous cardiorespiratory monitoring-'routine telemetry data,' in Beaulieu-Jones' terms-represents the most valuable non-clinician-initiated predictive signal present in patient data, and the value added to patient care justifies the efforts and expense required. Here, we present a clinical and a physiological point of view to support our contention.
Collapse
Affiliation(s)
- Oliver Monfredi
- Center for Advanced Medical Analytics, University of Virginia, United States of America.,Cardiovascular Division, Department of Internal Medicine, School of Medicine, University of Virginia, United States of America
| | - Jessica Keim-Malpass
- Center for Advanced Medical Analytics, University of Virginia, United States of America.,School of Nursing, University of Virginia, United States of America
| | - J Randall Moorman
- Center for Advanced Medical Analytics, University of Virginia, United States of America.,Cardiovascular Division, Department of Internal Medicine, School of Medicine, University of Virginia, United States of America
| |
Collapse
|
|
25
|
Mhasawade V, Zhao Y, Chunara R. Machine learning and algorithmic fairness in public and population health. NAT MACH INTELL 2021. [DOI: 10.1038/s42256-021-00373-4] [Citation(s) in RCA: 25] [Impact Index Per Article: 6.3] [Reference Citation Analysis] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 12/18/2022]
|
|
26
|
Bourcier S, Klug J, Nguyen LS. Non-occlusive mesenteric ischemia: Diagnostic challenges and perspectives in the era of artificial intelligence. World J Gastroenterol 2021; 27:4088-4103. [PMID: 34326613 PMCID: PMC8311528 DOI: 10.3748/wjg.v27.i26.4088] [Citation(s) in RCA: 24] [Impact Index Per Article: 6.0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Download PDF] [Journal Information] [Submit a Manuscript] [Subscribe] [Scholar Register] [Received: 01/28/2021] [Revised: 03/25/2021] [Accepted: 06/18/2021] [Indexed: 02/06/2023] Open
Abstract
Acute mesenteric ischemia (AMI) is a severe condition associated with poor prognosis, ultimately leading to death due to multiorgan failure. Several mechanisms may lead to AMI, and non-occlusive mesenteric ischemia (NOMI) represents a particular form of AMI. NOMI is prevalent in intensive care units in critically ill patients. In NOMI management, promptness and accuracy of diagnosis are paramount to achieve decisive treatment, but the last decades have been marked by failure to improve NOMI prognosis, due to lack of tools to detect this condition. While real-life diagnostic management relies on a combination of physical examination, several biomarkers, imaging, and endoscopy to detect the possibility of several grades of NOMI, research studies only focus on a few elements at a time. In the era of artificial intelligence (AI), which can aggregate thousands of variables in complex longitudinal models, the prospect of achieving accurate diagnosis through machine-learning-based algorithms may be sought. In the following work, we bring you a state-of-the-art literature review regarding NOMI, its presentation, its mechanics, and the pitfalls of routine work-up diagnostic exams including biomarkers, imaging, and endoscopy, we raise the perspectives of new biomarker exams, and finally we discuss what AI may add to the field, after summarizing what this technique encompasses.
Collapse
Affiliation(s)
- Simon Bourcier
- Department of Intensive Care Medicine, University Hospital of Geneva, Geneva 1201, Switzerland
| | - Julian Klug
- Department of Internal Medicine, Groupement Hospitalier de l’Ouest Lémanique, Nyon 1260, Switzerland
| | - Lee S Nguyen
- Department of Intensive Care Medicine, CMC Ambroise Paré, Neuilly-sur-Seine 92200, France
| |
Collapse
|
|
27
|
Coleman J, Ginsburg AS, Macharia WM, Ochieng R, Zhou G, Dunsmuir D, Karlen W, Ansermino JM. Identification of thresholds for accuracy comparisons of heart rate and respiratory rate in neonates. Gates Open Res 2021; 5:93. [PMID: 34901754 PMCID: PMC8630397 DOI: 10.12688/gatesopenres.13237.1] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Accepted: 06/02/2021] [Indexed: 04/05/2024] Open
Abstract
Background: Heart rate (HR) and respiratory rate (RR) can be challenging to measure accurately and reliably in neonates. The introduction of innovative, non-invasive measurement technologies suitable for resource-constrained settings is limited by the lack of appropriate clinical thresholds for accuracy comparison studies. Methods: We collected measurements of photoplethysmography-recorded HR and capnography-recorded exhaled carbon dioxide across multiple 60-second epochs (observations) in enrolled neonates admitted to the neonatal care unit at Aga Khan University Hospital in Nairobi, Kenya. Trained study nurses manually recorded HR, and the study team manually counted individual breaths from capnograms. For comparison, HR and RR also were measured using an automated signal detection algorithm. Clinical measurements were analyzed for repeatability. Results: A total of 297 epochs across 35 neonates were recorded. Manual HR showed a bias of -2.4 (-1.8%) and a spread between the 95% limits of agreement (LOA) of 40.3 (29.6%) compared to the algorithm-derived median HR. Manual RR showed a bias of -3.2 (-6.6%) and a spread between the 95% LOA of 17.9 (37.3%) compared to the algorithm-derived median RR, and a bias of -0.5 (1.1%) and a spread between the 95% LOA of 4.4 (9.1%) compared to the algorithm-derived RR count. Manual HR and RR showed repeatability of 0.6 (interquartile range (IQR) 0.5-0.7), and 0.7 (IQR 0.5-0.8), respectively. Conclusions: Appropriate clinical thresholds should be selected a priori when performing accuracy comparisons for HR and RR. Automated measurement technologies typically use median values rather than counts, which significantly impacts accuracy. A wider spread between the LOA, as much as 30%, should be considered to account for the observed physiological nuances and within- and between-neonate variability and different averaging methods. Wider adoption of thresholds by data standards organizations and technology developers and manufacturers will increase the robustness of clinical comparison studies.
Collapse
Affiliation(s)
- Jesse Coleman
- Evaluation of Technologies for Neonates in Africa (ETNA), Aga Khan University Hospital, Nairobi, Kenya
| | | | | | - Roseline Ochieng
- Department of Paediatrics, Aga Khan University Hospital, Nairobi, Kenya
| | - Guohai Zhou
- Center for Clinical Investigation, Brigham and Women's Hospital, Boston, MA, 02115, USA
| | - Dustin Dunsmuir
- Anesthesiology, Pharmacology & Therapeutics, The University of British Columbia, Vancouver, BC, V6T 1Z3, Canada
| | - Walter Karlen
- Mobile Health Systems Lab, Department of Health Sciences and Technology, ETH Zürich, Zürich, 8092, Switzerland
| | - J. Mark Ansermino
- Anesthesiology, Pharmacology & Therapeutics, The University of British Columbia, Vancouver, BC, V6T 1Z3, Canada
| |
Collapse
|
|
28
|
Baker S, Xiang W, Atkinson I. Hybridized neural networks for non-invasive and continuous mortality risk assessment in neonates. Comput Biol Med 2021; 134:104521. [PMID: 34111664 DOI: 10.1016/j.compbiomed.2021.104521] [Citation(s) in RCA: 2] [Impact Index Per Article: 0.5] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 01/24/2021] [Revised: 05/06/2021] [Accepted: 05/25/2021] [Indexed: 11/19/2022]
Abstract
Premature birth is the primary risk factor in neonatal deaths, with the majority of extremely premature babies cared for in neonatal intensive care units (NICUs). Mortality risk prediction in this setting can greatly improve patient outcomes and resource utilization. However, existing schemes often require laborious medical testing and calculation, and are typically only calculated once at admission. In this work, we propose a shallow hybrid neural network for the prediction of mortality risk in 3-day, 7-day, and 14-day risk windows using only birthweight, gestational age, sex, and heart rate (HR) and respiratory rate (RR) information from a 12-h window. As such, this scheme is capable of continuously updating mortality risk assessment, enabling analysis of health trends and responses to treatment. The highest performing scheme was the network that considered mortality risk within 3 days, with this scheme outperforming state-of-the-art works in the literature and achieving an area under the receiver-operator curve (AUROC) of 0.9336 with standard deviation of 0.0337 across 5 folds of cross-validation. As such, we conclude that our proposed scheme could readily be used for continuously-updating mortality risk prediction in NICU environments.
Collapse
Affiliation(s)
- Stephanie Baker
- College of Science & Engineering, James Cook University, Cairns, Queensland, 4878, Australia.
| | - Wei Xiang
- School of Engineering and Mathematical Sciences, La Trobe University, Melbourne, Victoria, 3086, Australia
| | - Ian Atkinson
- eResearch Centre, James Cook University, Townsville, Queensland, 4811, Australia
| |
Collapse
|
|
29
|
Sun Y, Kaur R, Gupta S, Paul R, Das R, Cho SJ, Anand S, Boutilier JJ, Saria S, Palma J, Saluja S, McAdams RM, Kaur A, Yadav G, Singh H. Development and validation of high definition phenotype-based mortality prediction in critical care units. JAMIA Open 2021; 4:ooab004. [PMID: 33796821 PMCID: PMC7991779 DOI: 10.1093/jamiaopen/ooab004] [Citation(s) in RCA: 9] [Impact Index Per Article: 2.3] [Reference Citation Analysis] [Abstract] [Key Words] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 11/07/2020] [Revised: 01/12/2021] [Accepted: 01/24/2021] [Indexed: 12/02/2022] Open
Abstract
Objectives The objectives of this study are to construct the high definition phenotype (HDP), a novel time-series data structure composed of both primary and derived parameters, using heterogeneous clinical sources and to determine whether different predictive models can utilize the HDP in the neonatal intensive care unit (NICU) to improve neonatal mortality prediction in clinical settings. Materials and Methods A total of 49 primary data parameters were collected from July 2018 to May 2020 from eight level-III NICUs. From a total of 1546 patients, 757 patients were found to contain sufficient fixed, intermittent, and continuous data to create HDPs. Two different predictive models utilizing the HDP, one a logistic regression model (LRM) and the other a deep learning long–short-term memory (LSTM) model, were constructed to predict neonatal mortality at multiple time points during the patient hospitalization. The results were compared with previous illness severity scores, including SNAPPE, SNAPPE-II, CRIB, and CRIB-II. Results A HDP matrix, including 12 221 536 minutes of patient stay in NICU, was constructed. The LRM model and the LSTM model performed better than existing neonatal illness severity scores in predicting mortality using the area under the receiver operating characteristic curve (AUC) metric. An ablation study showed that utilizing continuous parameters alone results in an AUC score of >80% for both LRM and LSTM, but combining fixed, intermittent, and continuous parameters in the HDP results in scores >85%. The probability of mortality predictive score has recall and precision of 0.88 and 0.77 for the LRM and 0.97 and 0.85 for the LSTM. Conclusions and Relevance The HDP data structure supports multiple analytic techniques, including the statistical LRM approach and the machine learning LSTM approach used in this study. LRM and LSTM predictive models of neonatal mortality utilizing the HDP performed better than existing neonatal illness severity scores. Further research is necessary to create HDP–based clinical decision tools to detect the early onset of neonatal morbidities.
Collapse
Affiliation(s)
- Yao Sun
- Division of Neonatology, Department of Pediatrics, University of California San Francisco, San Francisco, California, USA
| | - Ravneet Kaur
- Research and Development, Child Health Imprints (CHIL) Pte. Ltd., Singapore
| | - Shubham Gupta
- Research and Development, Child Health Imprints (CHIL) Pte. Ltd., Singapore
| | - Rahul Paul
- Research and Development, Child Health Imprints (CHIL) Pte. Ltd., Singapore
| | - Ritu Das
- Research and Development, Child Health Imprints (CHIL) Pte. Ltd., Singapore
| | - Su Jin Cho
- Department of Pediatrics, College of Medicine, Ewha Womans University Seoul, Seoul, Korea
| | - Saket Anand
- Department of Computer Science, Indraprastha Institute of Information Technology, New Delhi, India
| | - Justin J Boutilier
- Department of Industrial and Systems Engineering, University of Wisconsin-Madison, Wisconsin, USA
| | - Suchi Saria
- Machine Learning and Healthcare Lab, Johns Hopkins University, Baltimore, Maryland, USA.,Department of Applied Math & Statistics, Johns Hopkins University, Baltimore, Maryland, USA.,Department of Health Policy & Management, Johns Hopkins University, Baltimore, Maryland, USA
| | - Jonathan Palma
- Department of Pediatrics, Stanford University, Stanford, California, USA
| | - Satish Saluja
- Department of Neonatology, Sir Ganga Ram Hospital, New Delhi, India
| | - Ryan M McAdams
- Department of Pediatrics, University of Wisconsin School of Medicine and Public Health, Madison, Wisconsin, USA
| | - Avneet Kaur
- Department of Neonatology, Apollo Cradle Hospitals, New Delhi, India
| | - Gautam Yadav
- Department of Pediatrics, Kalawati Hospital, Rewari, India
| | - Harpreet Singh
- Research and Development, Child Health Imprints (CHIL) Pte. Ltd., Singapore
| |
Collapse
|
|
30
|
Ray A, Chaudhuri AK. Smart healthcare disease diagnosis and patient management: Innovation, improvement and skill development. MACHINE LEARNING WITH APPLICATIONS 2021. [DOI: 10.1016/j.mlwa.2020.100011] [Citation(s) in RCA: 6] [Impact Index Per Article: 1.5] [Reference Citation Analysis] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 01/13/2023] Open
|
|
31
|
Affiliation(s)
- F. Sessions Cole
- grid.4367.60000 0001 2355 7002Division of Newborn Medicine, Edward Mallinckrodt Department of Pediatrics, Washington University School of Medicine and St. Louis Children’s Hospital, St. Louis, MO USA
| |
Collapse
|
|
32
|
Singh H, Kusuda S, McAdams RM, Gupta S, Kalra J, Kaur R, Das R, Anand S, Pandey AK, Cho SJ, Saluja S, Boutilier JJ, Saria S, Palma J, Kaur A, Yadav G, Sun Y. Machine Learning-Based Automatic Classification of Video Recorded Neonatal Manipulations and Associated Physiological Parameters: A Feasibility Study. CHILDREN-BASEL 2020; 8:children8010001. [PMID: 33375101 PMCID: PMC7822162 DOI: 10.3390/children8010001] [Citation(s) in RCA: 2] [Impact Index Per Article: 0.4] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Download PDF] [Figures] [Subscribe] [Scholar Register] [Received: 11/18/2020] [Revised: 12/15/2020] [Accepted: 12/18/2020] [Indexed: 11/16/2022]
Abstract
Our objective in this study was to determine if machine learning (ML) can automatically recognize neonatal manipulations, along with associated changes in physiological parameters. A retrospective observational study was carried out in two Neonatal Intensive Care Units (NICUs) between December 2019 to April 2020. Both the video and physiological data (heart rate (HR) and oxygen saturation (SpO2)) were captured during NICU hospitalization. The proposed classification of neonatal manipulations was achieved by a deep learning system consisting of an Inception-v3 convolutional neural network (CNN), followed by transfer learning layers of Long Short-Term Memory (LSTM). Physiological signals prior to manipulations (baseline) were compared to during and after manipulations. The validation of the system was done using the leave-one-out strategy with input of 8 s of video exhibiting manipulation activity. Ten neonates were video recorded during an average length of stay of 24.5 days. Each neonate had an average of 528 manipulations during their NICU hospitalization, with the average duration of performing these manipulations varying from 28.9 s for patting, 45.5 s for a diaper change, and 108.9 s for tube feeding. The accuracy of the system was 95% for training and 85% for the validation dataset. In neonates <32 weeks’ gestation, diaper changes were associated with significant changes in HR and SpO2, and, for neonates ≥32 weeks’ gestation, patting and tube feeding were associated with significant changes in HR. The presented system can classify and document the manipulations with high accuracy. Moreover, the study suggests that manipulations impact physiological parameters.
Collapse
Affiliation(s)
- Harpreet Singh
- Child Health Imprints (CHIL) Pte. Ltd., Singapore 048545, Singapore; (S.G.); (J.K.); (R.K.); (R.D.)
- Correspondence: ; Tel.: +65-91-9910861112
| | - Satoshi Kusuda
- Department of Pediatrics, Kyorin University, Tokyo 181-8612, Japan;
| | - Ryan M. McAdams
- Department of Pediatrics, University of Wisconsin School of Medicine and Public Health, Madison, WI 53726, USA;
| | - Shubham Gupta
- Child Health Imprints (CHIL) Pte. Ltd., Singapore 048545, Singapore; (S.G.); (J.K.); (R.K.); (R.D.)
| | - Jayant Kalra
- Child Health Imprints (CHIL) Pte. Ltd., Singapore 048545, Singapore; (S.G.); (J.K.); (R.K.); (R.D.)
| | - Ravneet Kaur
- Child Health Imprints (CHIL) Pte. Ltd., Singapore 048545, Singapore; (S.G.); (J.K.); (R.K.); (R.D.)
| | - Ritu Das
- Child Health Imprints (CHIL) Pte. Ltd., Singapore 048545, Singapore; (S.G.); (J.K.); (R.K.); (R.D.)
| | - Saket Anand
- Department of Computer Science and Engineering, Indraprastha Institute of Information Technology, New Delhi 110020, India;
| | - Ashish Kumar Pandey
- Department of Mathematics, Indraprastha Institute of Information Technology, New Delhi 110020, India;
| | - Su Jin Cho
- College of Medicine, Ewha Womans University Seoul, Seoul 03760, Korea;
| | - Satish Saluja
- Department of Neonatology, Sir Ganga Ram Hospital, New Delhi 110060, India;
| | - Justin J. Boutilier
- Department of Industrial and Systems Engineering, College of Engineering, University of Wisconsin, Madison, WI 53706, USA;
| | - Suchi Saria
- Machine Learning and Healthcare Lab, Johns Hopkins University, 3400 N. Charles St, Baltimore, MD 21218, USA;
| | - Jonathan Palma
- Department of Pediatrics, Stanford University, Stanford, CA 94305, USA;
| | - Avneet Kaur
- Department of Neonatology, Apollo Cradle Hospitals, New Delhi 110015, India;
| | - Gautam Yadav
- Department of Pediatrics, Kalawati Hospital, Rewari 123401, India;
| | - Yao Sun
- Division of Neonatology, University of California, San Francisco, CA 92521, USA;
| |
Collapse
|
|
33
|
Datta S, Loftus TJ, Ruppert MM, Giordano C, Upchurch GR, Rashidi P, Ozrazgat-Baslanti T, Bihorac A. Added Value of Intraoperative Data for Predicting Postoperative Complications: The MySurgeryRisk PostOp Extension. J Surg Res 2020; 254:350-363. [PMID: 32531520 PMCID: PMC7755426 DOI: 10.1016/j.jss.2020.05.007] [Citation(s) in RCA: 21] [Impact Index Per Article: 4.2] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 12/05/2019] [Revised: 03/19/2020] [Accepted: 05/05/2020] [Indexed: 12/14/2022]
Abstract
BACKGROUND Models that predict postoperative complications often ignore important intraoperative events and physiological changes. This study tested the hypothesis that accuracy, discrimination, and precision in predicting postoperative complications would improve when using both preoperative and intraoperative data input data compared with preoperative data alone. METHODS This retrospective cohort analysis included 43,943 adults undergoing 52,529 inpatient surgeries at a single institution during a 5-y period. Random forest machine learning models in the validated MySurgeryRisk platform made patient-level predictions for seven postoperative complications and mortality occurring during hospital admission using electronic health record data and patient neighborhood characteristics. For each outcome, one model trained with preoperative data alone; one model trained with both preoperative and intraoperative data. Models were compared by accuracy, discrimination (expressed as area under the receiver operating characteristic curve), precision (expressed as area under the precision-recall curve), and reclassification indices. RESULTS Machine learning models incorporating both preoperative and intraoperative data had greater accuracy, discrimination, and precision than models using preoperative data alone for predicting all seven postoperative complications (intensive care unit length of stay >48 h, mechanical ventilation >48 h, neurologic complications including delirium, cardiovascular complications, acute kidney injury, venous thromboembolism, and wound complications), and in-hospital mortality (accuracy: 88% versus 77%; area under the receiver operating characteristic curve: 0.93 versus 0.87; area under the precision-recall curve: 0.21 versus 0.15). Overall reclassification improvement was 2.4%-10.0% for complications and 11.2% for in-hospital mortality. CONCLUSIONS Incorporating both preoperative and intraoperative data significantly increased the accuracy, discrimination, and precision of machine learning models predicting postoperative complications and mortality.
Collapse
Affiliation(s)
- Shounak Datta
- Department of Medicine, University of Florida, Gainesville, Florida; Precision and Intelligent Systems in Medicine (Prisma(P)), University of Florida, Gainesville, Florida
| | - Tyler J Loftus
- Precision and Intelligent Systems in Medicine (Prisma(P)), University of Florida, Gainesville, Florida; Department of Surgery, University of Florida, Gainesville, Florida
| | - Matthew M Ruppert
- Department of Medicine, University of Florida, Gainesville, Florida; Precision and Intelligent Systems in Medicine (Prisma(P)), University of Florida, Gainesville, Florida
| | - Chris Giordano
- Department of Anesthesiology, University of Florida, Gainesville, Florida
| | | | - Parisa Rashidi
- Precision and Intelligent Systems in Medicine (Prisma(P)), University of Florida, Gainesville, Florida; Department of Biomedical Engineering, University of Florida, Gainesville, Florida
| | - Tezcan Ozrazgat-Baslanti
- Department of Medicine, University of Florida, Gainesville, Florida; Precision and Intelligent Systems in Medicine (Prisma(P)), University of Florida, Gainesville, Florida
| | - Azra Bihorac
- Department of Medicine, University of Florida, Gainesville, Florida; Precision and Intelligent Systems in Medicine (Prisma(P)), University of Florida, Gainesville, Florida.
| |
Collapse
|
|
34
|
Bernert RA, Hilberg AM, Melia R, Kim JP, Shah NH, Abnousi F. Artificial Intelligence and Suicide Prevention: A Systematic Review of Machine Learning Investigations. INTERNATIONAL JOURNAL OF ENVIRONMENTAL RESEARCH AND PUBLIC HEALTH 2020; 17:E5929. [PMID: 32824149 PMCID: PMC7460360 DOI: 10.3390/ijerph17165929] [Citation(s) in RCA: 88] [Impact Index Per Article: 17.6] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Grants] [Track Full Text] [Download PDF] [Figures] [Subscribe] [Scholar Register] [Received: 07/20/2020] [Accepted: 07/28/2020] [Indexed: 12/12/2022]
Abstract
Suicide is a leading cause of death that defies prediction and challenges prevention efforts worldwide. Artificial intelligence (AI) and machine learning (ML) have emerged as a means of investigating large datasets to enhance risk detection. A systematic review of ML investigations evaluating suicidal behaviors was conducted using PubMed/MEDLINE, PsychInfo, Web-of-Science, and EMBASE, employing search strings and MeSH terms relevant to suicide and AI. Databases were supplemented by hand-search techniques and Google Scholar. Inclusion criteria: (1) journal article, available in English, (2) original investigation, (3) employment of AI/ML, (4) evaluation of a suicide risk outcome. N = 594 records were identified based on abstract search, and 25 hand-searched reports. N = 461 reports remained after duplicates were removed, n = 316 were excluded after abstract screening. Of n = 149 full-text articles assessed for eligibility, n = 87 were included for quantitative synthesis, grouped according to suicide behavior outcome. Reports varied widely in methodology and outcomes. Results suggest high levels of risk classification accuracy (>90%) and Area Under the Curve (AUC) in the prediction of suicidal behaviors. We report key findings and central limitations in the use of AI/ML frameworks to guide additional research, which hold the potential to impact suicide on broad scale.
Collapse
Affiliation(s)
- Rebecca A. Bernert
- Stanford Suicide Prevention Research Laboratory, Department of Psychiatry and Behavioral Sciences, Stanford University School of Medicine, Stanford, CA 94304, USA
| | - Amanda M. Hilberg
- Stanford Suicide Prevention Research Laboratory, Department of Psychiatry and Behavioral Sciences, Stanford University School of Medicine, Stanford, CA 94304, USA
| | - Ruth Melia
- Stanford Suicide Prevention Research Laboratory, Department of Psychiatry and Behavioral Sciences, Stanford University School of Medicine, Stanford, CA 94304, USA
- Department of Psychology, National University of Ireland, Galway, Ireland
| | - Jane Paik Kim
- Stanford Suicide Prevention Research Laboratory, Department of Psychiatry and Behavioral Sciences, Stanford University School of Medicine, Stanford, CA 94304, USA
| | - Nigam H. Shah
- Department of Medicine, Center for Biomedical Informatics Research, Stanford University School of Medicine, Stanford, CA 94304, USA
- Informatics, Stanford Center for Clinical and Translational Research, and Education (Spectrum), Stanford University, Stanford CA 94304, USA
| | - Freddy Abnousi
- Facebook, Menlo Park, CA 94025, USA
- Yale University School of Medicine, New Haven, CT 06510, USA
| |
Collapse
|
|
35
|
Tseng PY, Chen YT, Wang CH, Chiu KM, Peng YS, Hsu SP, Chen KL, Yang CY, Lee OKS. Prediction of the development of acute kidney injury following cardiac surgery by machine learning. CRITICAL CARE : THE OFFICIAL JOURNAL OF THE CRITICAL CARE FORUM 2020; 24:478. [PMID: 32736589 PMCID: PMC7395374 DOI: 10.1186/s13054-020-03179-9] [Citation(s) in RCA: 237] [Impact Index Per Article: 47.4] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Download PDF] [Figures] [Subscribe] [Scholar Register] [Received: 04/26/2020] [Accepted: 07/14/2020] [Indexed: 12/14/2022]
Abstract
Background Cardiac surgery–associated acute kidney injury (CSA-AKI) is a major complication that results in increased morbidity and mortality after cardiac surgery. Most established prediction models are limited to the analysis of nonlinear relationships and fail to fully consider intraoperative variables, which represent the acute response to surgery. Therefore, this study utilized an artificial intelligence–based machine learning approach thorough perioperative data-driven learning to predict CSA-AKI. Methods A total of 671 patients undergoing cardiac surgery from August 2016 to August 2018 were enrolled. AKI following cardiac surgery was defined according to criteria from Kidney Disease: Improving Global Outcomes (KDIGO). The variables used for analysis included demographic characteristics, clinical condition, preoperative biochemistry data, preoperative medication, and intraoperative variables such as time-series hemodynamic changes. The machine learning methods used included logistic regression, support vector machine (SVM), random forest (RF), extreme gradient boosting (XGboost), and ensemble (RF + XGboost). The performance of these models was evaluated using the area under the receiver operating characteristic curve (AUC). We also utilized SHapley Additive exPlanation (SHAP) values to explain the prediction model. Results Development of CSA-AKI was noted in 163 patients (24.3%) during the first postoperative week. Regarding the efficacy of the single model that most accurately predicted the outcome, RF exhibited the greatest AUC (0.839, 95% confidence interval [CI] 0.772–0.898), whereas the AUC (0.843, 95% CI 0.778–0.899) of ensemble model (RF + XGboost) was even greater than that of the RF model alone. The top 3 most influential features in the RF importance matrix plot were intraoperative urine output, units of packed red blood cells (pRBCs) transfused during surgery, and preoperative hemoglobin level. The SHAP summary plot was used to illustrate the positive or negative effects of the top 20 features attributed to the RF. We also used the SHAP dependence plot to explain how a single feature affects the output of the RF prediction model. Conclusions In this study, machine learning methods were successfully established to predict CSA-AKI, which determines risks following cardiac surgery, enabling the optimization of postoperative treatment strategies to minimize the postoperative complications following cardiac surgeries.
Collapse
Affiliation(s)
- Po-Yu Tseng
- Institute of Clinical Medicine, School of Medicine, National Yang-Ming University, No. 155, Section 2, Li-Nong Street, Beitou District, Taipei, 11221, Taiwan.,Stem Cell Research Center, National Yang-Ming University, Taipei, Taiwan.,Division of Nephrology, Department of Internal Medicine, Taipei City Hospital, Heping Fuyou Branch, Taipei, Taiwan
| | - Yi-Ting Chen
- Muen Biomedical and Optoelectronics Technologies Inc., New Taipei City, Taiwan
| | - Chuen-Heng Wang
- Muen Biomedical and Optoelectronics Technologies Inc., New Taipei City, Taiwan
| | - Kuan-Ming Chiu
- Division of Cardiovascular Surgery, Cardiovascular Center, Far Eastern Memorial Hospital, New Taipei City, Taiwan.,Department of Electrical Engineering, Yuan Ze University, Taoyuan City, Taiwan
| | - Yu-Sen Peng
- Division of Nephrology, Department of Internal Medicine, Far Eastern Memorial Hospital, New Taipei City, Taiwan.,College of Electrical and Communication Engineering, Yuan Ze University, Taoyuan City, Taiwan.,Department of Applied Cosmetology, Lee-Ming Institute of Technology, New Taipei City, Taiwan
| | - Shih-Ping Hsu
- Division of Nephrology, Department of Internal Medicine, Far Eastern Memorial Hospital, New Taipei City, Taiwan
| | - Kang-Lung Chen
- Division of Cardiovascular Surgery, Cardiovascular Center, Far Eastern Memorial Hospital, New Taipei City, Taiwan.,Division of Cardiovascular Surgery, En Chu Kong Hospital, New Taipei City, Taiwan
| | - Chih-Yu Yang
- Institute of Clinical Medicine, School of Medicine, National Yang-Ming University, No. 155, Section 2, Li-Nong Street, Beitou District, Taipei, 11221, Taiwan. .,Stem Cell Research Center, National Yang-Ming University, Taipei, Taiwan. .,Division of Nephrology, Department of Medicine, Taipei Veterans General Hospital, Taipei, Taiwan. .,Center for Intelligent Drug Systems and Smart Bio-devices (IDS2B), Hsinchu, Taiwan.
| | - Oscar Kuang-Sheng Lee
- Institute of Clinical Medicine, School of Medicine, National Yang-Ming University, No. 155, Section 2, Li-Nong Street, Beitou District, Taipei, 11221, Taiwan. .,Stem Cell Research Center, National Yang-Ming University, Taipei, Taiwan. .,China Medical University Hospital, Taichung, Taiwan.
| |
Collapse
|
|
36
|
Megjhani M, Kaffashi F, Terilli K, Alkhachroum A, Esmaeili B, Doyle KW, Murthy S, Velazquez AG, Connolly ES, Roh DJ, Agarwal S, Loparo KA, Claassen J, Boehme A, Park S. Heart Rate Variability as a Biomarker of Neurocardiogenic Injury After Subarachnoid Hemorrhage. Neurocrit Care 2020; 32:162-171. [PMID: 31093884 PMCID: PMC6856427 DOI: 10.1007/s12028-019-00734-3] [Citation(s) in RCA: 20] [Impact Index Per Article: 4.0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 01/05/2023]
Abstract
BACKGROUND The objective of this study was to examine whether heart rate variability (HRV) measures can be used to detect neurocardiogenic injury (NCI). METHODS Three hundred and twenty-six consecutive admissions with aneurysmal subarachnoid hemorrhage (SAH) met criteria for the study. Of 326 subjects, 56 (17.2%) developed NCI which we defined by wall motion abnormality with ventricular dysfunction on transthoracic echocardiogram or cardiac troponin-I > 0.3 ng/mL without electrocardiogram evidence of coronary artery insufficiency. HRV measures (in time and frequency domains, as well as nonlinear technique of detrended fluctuation analysis) were calculated over the first 48 h. We applied longitudinal multilevel linear regression to characterize the relationship of HRV measures with NCI and examine between-group differences at baseline and over time. RESULTS There was decreased vagal activity in NCI subjects with a between-group difference in low/high frequency ratio (β 3.42, SE 0.92, p = 0.0002), with sympathovagal balance in favor of sympathetic nervous activity. All time-domain measures were decreased in SAH subjects with NCI. An ensemble machine learning approach translated these measures into a classification tool that demonstrated good discrimination using the area under the receiver operating characteristic curve (AUROC 0.82), the area under precision recall curve (AUPRC 0.75), and a correct classification rate of 0.81. CONCLUSIONS HRV measures are significantly associated with our label of NCI and a machine learning approach using features derived from HRV measures can classify SAH patients that develop NCI.
Collapse
Affiliation(s)
- Murad Megjhani
- Department of Neurology, Columbia University Irving Medical Center, 177 Fort Washington Ave, 8 Milstein-300 Center, New York, NY, 10032, USA
| | - Farhad Kaffashi
- Case School of Engineering, Case Western Reserve University, Cleveland, USA
| | - Kalijah Terilli
- Department of Neurology, Columbia University Irving Medical Center, 177 Fort Washington Ave, 8 Milstein-300 Center, New York, NY, 10032, USA
| | - Ayham Alkhachroum
- Department of Neurology, Columbia University Irving Medical Center, 177 Fort Washington Ave, 8 Milstein-300 Center, New York, NY, 10032, USA
| | - Behnaz Esmaeili
- Department of Neurology, Columbia University Irving Medical Center, 177 Fort Washington Ave, 8 Milstein-300 Center, New York, NY, 10032, USA
| | - Kevin William Doyle
- Department of Neurology, Columbia University Irving Medical Center, 177 Fort Washington Ave, 8 Milstein-300 Center, New York, NY, 10032, USA
| | - Santosh Murthy
- Department of Neurology, Weill Cornell Medical College, New York, USA
| | - Angela G Velazquez
- Department of Neurology, Columbia University Irving Medical Center, 177 Fort Washington Ave, 8 Milstein-300 Center, New York, NY, 10032, USA
| | - E Sander Connolly
- Department of Neurosurgery, Columbia University Irving Medical Center, New York, USA
| | - David Jinou Roh
- Department of Neurology, Columbia University Irving Medical Center, 177 Fort Washington Ave, 8 Milstein-300 Center, New York, NY, 10032, USA
| | - Sachin Agarwal
- Department of Neurology, Columbia University Irving Medical Center, 177 Fort Washington Ave, 8 Milstein-300 Center, New York, NY, 10032, USA
| | - Ken A Loparo
- Case School of Engineering, Case Western Reserve University, Cleveland, USA
| | - Jan Claassen
- Department of Neurology, Columbia University Irving Medical Center, 177 Fort Washington Ave, 8 Milstein-300 Center, New York, NY, 10032, USA
| | - Amelia Boehme
- Department of Neurology, Columbia University Irving Medical Center, 177 Fort Washington Ave, 8 Milstein-300 Center, New York, NY, 10032, USA
| | - Soojin Park
- Department of Neurology, Columbia University Irving Medical Center, 177 Fort Washington Ave, 8 Milstein-300 Center, New York, NY, 10032, USA.
| |
Collapse
|
|
37
|
McMahon AW, Cooper WO, Brown JS, Carleton B, Doshi-Velez F, Kohane I, Goldman JL, Hoffman MA, Kamaleswaran R, Sakiyama M, Sekine S, Sturkenboom MCJM, Turner MA, Califf RM. Big Data in the Assessment of Pediatric Medication Safety. Pediatrics 2020; 145:peds.2019-0562. [PMID: 31937606 DOI: 10.1542/peds.2019-0562] [Citation(s) in RCA: 6] [Impact Index Per Article: 1.2] [Reference Citation Analysis] [Abstract] [Track Full Text] [Journal Information] [Submit a Manuscript] [Subscribe] [Scholar Register] [Accepted: 11/13/2019] [Indexed: 11/24/2022] Open
Abstract
Big data (BD) in pediatric medication safety research provides many opportunities to improve the safety and health of children. The number of pediatric medication and device trials has increased in part because of the past 20 years of US legislation requiring and incentivizing study of the effects of medical products in children (Food and Drug Administration Modernization Act of 1997, Pediatric Rule in 1998, Best Pharmaceuticals for Children Act of 2002, and Pediatric Research Equity Act of 2003). There are some limitations of traditional approaches to studying medication safety in children. Randomized clinical trials within the regulatory context may not enroll patients who are representative of the general pediatric population, provide the power to detect rare safety signals, or provide long-term safety data. BD sources may have these capabilities. In recent years, medical records have become digitized, and cell phones and personal devices have proliferated. In this process, the field of biomedical science has progressively used BD from those records coupled with other data sources, both digital and traditional. Additionally, large distributed databases that include pediatric-specific outcome variables are available. A workshop entitled "Advancing the Development of Pediatric Therapeutics: Application of 'Big Data' to Pediatric Safety Studies" held September 18 to 19, 2017, in Silver Spring, Maryland, formed the basis of many of the ideas outlined in this article, which are intended to identify key examples, critical issues, and future directions in this early phase of an anticipated dramatic change in the availability and use of BD.
Collapse
Affiliation(s)
- Ann W McMahon
- Office of Pediatric Therapeutics, US Food and Drug Administration, Rockville, Maryland;
| | - William O Cooper
- Departments of Pediatrics and Health Policy, Vanderbilt University Medical Center, Nashville, Tennessee
| | - Jeffrey S Brown
- Population Medicine, Harvard Medical School and Harvard Pilgrim Healthcare Insititute, Boston, Massachusetts
| | - Bruce Carleton
- Division of Translational Therapeutics, Department of Pediatrics, Faculty of Medicine, University of British Columbia, Vancouver, Canada
| | - Finale Doshi-Velez
- Paulson School of Engineering and Applied Sciences, Harvard University, Boston, Massachusetts
| | - Isaac Kohane
- Departments of Biomedical Informatics, Pediatrics, and
| | - Jennifer L Goldman
- Divisions of Pediatric Infectious Diseases and Clinical Parmacology, Department of Pediatrics, and
| | - Mark A Hoffman
- Departments of Biomedical Informatics, Pediatrics, and Emergency Medicine, School of Medicine, Emory University, Atlanta, Georgia
| | | | - Michiyo Sakiyama
- Office of New Drug IV, Pharmaceuticals and Medical Devices Agency, Tokyo, Japan.,Department of Epidemiology, Julius Center Research Program Cardiovascular Edpidemiology, Utrecht University Medical Center, Utrecht, Netherlands
| | - Shohko Sekine
- Institute of Translational Medicine, University of Liverpool, Liverpool, United Kingdom; and
| | - Miriam C J M Sturkenboom
- Division of Cardiology, Department of Internal Medicine, School of Medicine, Center for Health Science, Duke Clinical Research Institute, Duke University, Durham, North Carolina
| | - Mark A Turner
- Institute of Translational Medicine, University of Liverpool, Liverpool, United Kingdom; and
| | - Robert M Califf
- Division of Cardiology, Department of Internal Medicine, School of Medicine, Center for Health Science, Duke Clinical Research Institute, Duke University, Durham, North Carolina
| |
Collapse
|
|
38
|
Khan FA, Mullany LC, Wu LFS, Ali H, Shaikh S, Alland K, West Jr KP, Labrique AB. Predictors of neonatal mortality: development and validation of prognostic models using prospective data from rural Bangladesh. BMJ Glob Health 2020; 5:e001983. [PMID: 32133171 PMCID: PMC7042570 DOI: 10.1136/bmjgh-2019-001983] [Citation(s) in RCA: 5] [Impact Index Per Article: 1.0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Grants] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 09/10/2019] [Revised: 11/10/2019] [Accepted: 12/10/2019] [Indexed: 12/23/2022] Open
Abstract
Objective To assess the extent to which maternal histories of newborn danger signs independently or combined with birth weight and/or gestational age (GA) can capture and/or predict postsecond day (age>48 hours) neonatal death. Methods Data from a cluster-randomised trial conducted in rural Bangladesh were split into development and validation sets. The prompted recall of danger signs and birth weight measurements were collected within 48 hours postchildbirth. Maternally recalled danger signs included cyanosis (any part of the infant's body was blue at birth), non-cephalic presentation (part other than head came out first at birth), lethargy (weak or no arm/leg movement and/or cry at birth), trouble suckling (infant unable to suckle/feed normally in the 2 days after birth or before death, collected 1-month postpartum or from verbal autopsy). Last menstrual period was collected at maternal enrolment early in pregnancy. Singleton newborns surviving 2 days past childbirth were eligible for analysis. Prognostic multivariable models were developed and internally validated. Results Recalling ≥1 sign of lethargy, cyanosis, non-cephalic presentation or trouble suckling identified postsecond day neonatal death with 65.3% sensitivity, 60.8% specificity, 2.1% positive predictive value (PPV) and 99.3% negative predictive value (NPV) in the development set. Requiring either lethargy or weight <2.5 kg identified 89.1% of deaths (at 39.7% specificity, 1.9% PPV and 99.6% NPV) while lethargy or preterm birth (<37 weeks) captured 81.0% of deaths (at 53.6% specificity, 2.3% PPV and 99.5% NPV). A simplified model (birth weight, GA, lethargy, cyanosis, non-cephalic presentation and trouble suckling) predicted death with good discrimination (validation area under the receiver-operator characteristic curve (AUC) 0.80, 95% CI 0.73 to 0.87). A further simplified model (GA, non-cephalic presentation, lethargy, trouble suckling) predicted death with moderate discrimination (validation AUC 0.74, 95% CI 0.66 to 0.81). Conclusion Maternally recalled danger signs, coupled to either birth weight or GA, can predict and capture postsecond day neonatal death with high discrimination and sensitivity.
Collapse
Affiliation(s)
- Farhad A Khan
- Department of International Health, Johns Hopkins Bloomberg School of Public Health, Baltimore, Maryland, USA
| | - Luke C Mullany
- Department of International Health, Johns Hopkins Bloomberg School of Public Health, Baltimore, Maryland, USA
| | - Lee F-S Wu
- Department of International Health, Johns Hopkins Bloomberg School of Public Health, Baltimore, Maryland, USA
| | - Hasmot Ali
- Department of International Health, Johns Hopkins Bloomberg School of Public Health, Baltimore, Maryland, USA
| | - Saijuddin Shaikh
- Department of International Health, Johns Hopkins Bloomberg School of Public Health, Baltimore, Maryland, USA
| | - Kelsey Alland
- Department of International Health, Johns Hopkins Bloomberg School of Public Health, Baltimore, Maryland, USA
| | - Keith P West Jr
- Department of International Health, Johns Hopkins Bloomberg School of Public Health, Baltimore, Maryland, USA
| | - Alain B Labrique
- Department of International Health, Johns Hopkins Bloomberg School of Public Health, Baltimore, Maryland, USA
| |
Collapse
|
|
39
|
Continuous vital sign analysis for predicting and preventing neonatal diseases in the twenty-first century: big data to the forefront. Pediatr Res 2020; 87:210-220. [PMID: 31377752 PMCID: PMC6962536 DOI: 10.1038/s41390-019-0527-0] [Citation(s) in RCA: 71] [Impact Index Per Article: 14.2] [Reference Citation Analysis] [Abstract] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Submit a Manuscript] [Subscribe] [Scholar Register] [Received: 04/19/2019] [Revised: 07/20/2019] [Accepted: 07/25/2019] [Indexed: 01/03/2023]
Abstract
In the neonatal intensive care unit (NICU), heart rate, respiratory rate, and oxygen saturation are vital signs (VS) that are continuously monitored in infants, while blood pressure is often monitored continuously immediately after birth, or during critical illness. Although changes in VS can reflect infant physiology or circadian rhythms, persistent deviations in absolute values or complex changes in variability can indicate acute or chronic pathology. Recent studies demonstrate that analysis of continuous VS trends can predict sepsis, necrotizing enterocolitis, brain injury, bronchopulmonary dysplasia, cardiorespiratory decompensation, and mortality. Subtle changes in continuous VS patterns may not be discerned even by experienced clinicians reviewing spot VS data or VS trends captured in the monitor. In contrast, objective analysis of continuous VS data can improve neonatal outcomes by allowing heightened vigilance or preemptive interventions. In this review, we provide an overview of the studies that have used continuous analysis of single or multiple VS, their interactions, and combined VS and clinical analytic tools, to predict or detect neonatal pathophysiology. We make the case that big-data analytics are promising, and with continued improvements, can become a powerful tool to mitigate neonatal diseases in the twenty-first century.
Collapse
|
|
40
|
Plate JDJ, van de Leur RR, Leenen LPH, Hietbrink F, Peelen LM, Eijkemans MJC. Incorporating repeated measurements into prediction models in the critical care setting: a framework, systematic review and meta-analysis. BMC Med Res Methodol 2019; 19:199. [PMID: 31655567 PMCID: PMC6815391 DOI: 10.1186/s12874-019-0847-0] [Citation(s) in RCA: 11] [Impact Index Per Article: 1.8] [Reference Citation Analysis] [Abstract] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 03/05/2019] [Accepted: 10/10/2019] [Indexed: 12/22/2022] Open
Abstract
BACKGROUND The incorporation of repeated measurements into multivariable prediction research may greatly enhance predictive performance. However, the methodological possibilities vary widely and a structured overview of the possible and utilized approaches lacks. Therefore, we [1] propose a structured framework for these approaches, [2] determine what methods are currently used to incorporate repeated measurements in prediction research in the critical care setting and, where possible, [3] assess the added discriminative value of incorporating repeated measurements. METHODS The proposed framework consists of three domains: the observation window (static or dynamic), the processing of the raw data (raw data modelling, feature extraction and reduction) and the type of modelling. A systematic review was performed to identify studies which incorporate repeated measurements to predict (e.g. mortality) in the critical care setting. The within-study difference in c-statistics between models with versus without repeated measurements were obtained and pooled in a meta-analysis. RESULTS From the 2618 studies found, 29 studies incorporated multiple repeated measurements. The annual number of studies with repeated measurements increased from 2.8/year (2000-2005) to 16.0/year (2016-2018). The majority of studies that incorporated repeated measurements for prediction research used a dynamic observation window, and extracted features directly from the data. Differences in c statistics ranged from - 0.048 to 0.217 in favour of models that utilize repeated measurements. CONCLUSIONS Repeated measurements are increasingly common to predict events in the critical care domain, but their incorporation is lagging. A framework of possible approaches could aid researchers to optimize future prediction models.
Collapse
Affiliation(s)
- Joost D J Plate
- Division of Surgery, University Medical Centre Utrecht, Heidelberglaan 100, Utrecht, 3584, CX, the Netherlands.
| | - Rutger R van de Leur
- Julius Center for Health Sciences and Primary Care, University Medical Center Utrecht, Utrecht, the Netherlands
| | - Luke P H Leenen
- Division of Surgery, University Medical Centre Utrecht, Heidelberglaan 100, Utrecht, 3584, CX, the Netherlands
| | - Falco Hietbrink
- Division of Surgery, University Medical Centre Utrecht, Heidelberglaan 100, Utrecht, 3584, CX, the Netherlands
| | - Linda M Peelen
- Julius Center for Health Sciences and Primary Care, University Medical Center Utrecht, Utrecht, the Netherlands.,Departments of Anesthesiology and Intensive Care Medicine, University Medical Center Utrecht, Utrecht, the Netherlands
| | - M J C Eijkemans
- Julius Center for Health Sciences and Primary Care, University Medical Center Utrecht, Utrecht, the Netherlands
| |
Collapse
|
|
41
|
Adhikari L, Ozrazgat-Baslanti T, Ruppert M, Madushani RWMA, Paliwal S, Hashemighouchani H, Zheng F, Tao M, Lopes JM, Li X, Rashidi P, Bihorac A. Improved predictive models for acute kidney injury with IDEA: Intraoperative Data Embedded Analytics. PLoS One 2019; 14:e0214904. [PMID: 30947282 PMCID: PMC6448850 DOI: 10.1371/journal.pone.0214904] [Citation(s) in RCA: 58] [Impact Index Per Article: 9.7] [Reference Citation Analysis] [Abstract] [MESH Headings] [Grants] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 05/29/2018] [Accepted: 03/18/2019] [Indexed: 12/12/2022] Open
Abstract
BACKGROUND Acute kidney injury (AKI) is a common complication after surgery that is associated with increased morbidity and mortality. The majority of existing perioperative AKI risk prediction models are limited in their generalizability and do not fully utilize intraoperative physiological time-series data. Thus, there is a need for intelligent, accurate, and robust systems to leverage new information as it becomes available to predict the risk of developing postoperative AKI. METHODS A retrospective single-center cohort of 2,911 adults who underwent surgery at the University of Florida Health between 2000 and 2010 was utilized for this study. Machine learning and statistical analysis techniques were used to develop perioperative models to predict the risk of developing AKI during the first three days after surgery, first seven days after surgery, and overall (after surgery during the index hospitalization). The improvement in risk prediction was examined by incorporating intraoperative physiological time-series variables. Our proposed model enriched a preoperative model that produced a probabilistic AKI risk score by integrating intraoperative statistical features through a machine learning stacking approach inside a random forest classifier. Model performance was evaluated using the area under the receiver operating characteristic curve (AUC), accuracy, and Net Reclassification Improvement (NRI). RESULTS The predictive performance of the proposed model is better than the preoperative data only model. The proposed model had an AUC of 0.86 (accuracy of 0.78) for the seven-day AKI outcome, while the preoperative model had an AUC of 0.84 (accuracy of 0.76). Furthermore, by integrating intraoperative features, the algorithm was able to reclassify 40% of the false negative patients from the preoperative model. The NRI for each outcome was AKI at three days (8%), seven days (7%), and overall (4%). CONCLUSIONS Postoperative AKI prediction was improved with high sensitivity and specificity through a machine learning approach that dynamically incorporated intraoperative data.
Collapse
Affiliation(s)
- Lasith Adhikari
- Division of Nephrology, Hypertension and Renal Transplantation, Department of Medicine, University of Florida, Gainesville, FL, United States of America
- Precision and Intelligent Systems in Medicine (PrismaP), University of Florida, Gainesville, FL, United States of America
| | - Tezcan Ozrazgat-Baslanti
- Division of Nephrology, Hypertension and Renal Transplantation, Department of Medicine, University of Florida, Gainesville, FL, United States of America
- Precision and Intelligent Systems in Medicine (PrismaP), University of Florida, Gainesville, FL, United States of America
| | - Matthew Ruppert
- Division of Nephrology, Hypertension and Renal Transplantation, Department of Medicine, University of Florida, Gainesville, FL, United States of America
- Precision and Intelligent Systems in Medicine (PrismaP), University of Florida, Gainesville, FL, United States of America
| | - R. W. M. A. Madushani
- Division of Nephrology, Hypertension and Renal Transplantation, Department of Medicine, University of Florida, Gainesville, FL, United States of America
- Precision and Intelligent Systems in Medicine (PrismaP), University of Florida, Gainesville, FL, United States of America
| | - Srajan Paliwal
- Division of Nephrology, Hypertension and Renal Transplantation, Department of Medicine, University of Florida, Gainesville, FL, United States of America
- Precision and Intelligent Systems in Medicine (PrismaP), University of Florida, Gainesville, FL, United States of America
| | - Haleh Hashemighouchani
- Division of Nephrology, Hypertension and Renal Transplantation, Department of Medicine, University of Florida, Gainesville, FL, United States of America
- Precision and Intelligent Systems in Medicine (PrismaP), University of Florida, Gainesville, FL, United States of America
| | - Feng Zheng
- Precision and Intelligent Systems in Medicine (PrismaP), University of Florida, Gainesville, FL, United States of America
- Department of Electrical and Computer Engineering, University of Florida, Gainesville, FL, United States of America
| | - Ming Tao
- Division of Nephrology, Hypertension and Renal Transplantation, Department of Medicine, University of Florida, Gainesville, FL, United States of America
- Precision and Intelligent Systems in Medicine (PrismaP), University of Florida, Gainesville, FL, United States of America
| | - Juliano M. Lopes
- Precision and Intelligent Systems in Medicine (PrismaP), University of Florida, Gainesville, FL, United States of America
| | - Xiaolin Li
- Precision and Intelligent Systems in Medicine (PrismaP), University of Florida, Gainesville, FL, United States of America
- Department of Electrical and Computer Engineering, University of Florida, Gainesville, FL, United States of America
| | - Parisa Rashidi
- Precision and Intelligent Systems in Medicine (PrismaP), University of Florida, Gainesville, FL, United States of America
- Biomedical Engineering Department, University of Florida, Gainesville, FL, United States of America
| | - Azra Bihorac
- Division of Nephrology, Hypertension and Renal Transplantation, Department of Medicine, University of Florida, Gainesville, FL, United States of America
- Precision and Intelligent Systems in Medicine (PrismaP), University of Florida, Gainesville, FL, United States of America
| |
Collapse
|
|
42
|
Leiserson MDM, Syrgkanis V, Gilson A, Dudik M, Gillett S, Chayes J, Borgs C, Bajorin DF, Rosenberg JE, Funt S, Snyder A, Mackey L. A multifactorial model of T cell expansion and durable clinical benefit in response to a PD-L1 inhibitor. PLoS One 2018; 13:e0208422. [PMID: 30596661 PMCID: PMC6312275 DOI: 10.1371/journal.pone.0208422] [Citation(s) in RCA: 13] [Impact Index Per Article: 1.9] [Reference Citation Analysis] [Abstract] [MESH Headings] [Grants] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 07/26/2018] [Accepted: 11/17/2018] [Indexed: 01/05/2023] Open
Abstract
Checkpoint inhibitor immunotherapies have had major success in treating patients with late-stage cancers, yet the minority of patients benefit. Mutation load and PD-L1 staining are leading biomarkers associated with response, but each is an imperfect predictor. A key challenge to predicting response is modeling the interaction between the tumor and immune system. We begin to address this challenge with a multifactorial model for response to anti-PD-L1 therapy. We train a model to predict immune response in patients after treatment based on 36 clinical, tumor, and circulating features collected prior to treatment. We analyze data from 21 bladder cancer patients using the elastic net high-dimensional regression procedure and, as training set error is a biased and overly optimistic measure of prediction error, we use leave-one-out cross-validation to obtain unbiased estimates of accuracy on held-out patients. In held-out patients, the model explains 79% of the variance in T cell clonal expansion. This predicted immune response is multifactorial, as the variance explained is at most 23% if clinical, tumor, or circulating features are excluded. Moreover, if patients are triaged according to predicted expansion, only 38% of non-durable clinical benefit (DCB) patients need be treated to ensure that 100% of DCB patients are treated. In contrast, using mutation load or PD-L1 staining alone, one must treat at least 77% of non-DCB patients to ensure that all DCB patients receive treatment. Thus, integrative models of immune response may improve our ability to anticipate clinical benefit of immunotherapy.
Collapse
MESH Headings
- Adult
- Antibodies, Monoclonal/therapeutic use
- Antibodies, Monoclonal, Humanized
- B7-H1 Antigen/antagonists & inhibitors
- B7-H1 Antigen/immunology
- Biomarkers, Pharmacological/analysis
- Biomarkers, Tumor/analysis
- Carcinoma, Transitional Cell/drug therapy
- Carcinoma, Transitional Cell/immunology
- Carcinoma, Transitional Cell/pathology
- Cell Proliferation/drug effects
- Cell Proliferation/genetics
- Clonal Evolution/drug effects
- Clonal Evolution/genetics
- Female
- Humans
- Immunotherapy/methods
- Lymphocytes, Tumor-Infiltrating/drug effects
- Lymphocytes, Tumor-Infiltrating/physiology
- Male
- Models, Statistical
- Mutation
- Protein Kinase Inhibitors/therapeutic use
- Risk Assessment
- T-Lymphocytes/drug effects
- T-Lymphocytes/physiology
- Treatment Outcome
- Urinary Bladder Neoplasms/drug therapy
- Urinary Bladder Neoplasms/immunology
- Urinary Bladder Neoplasms/pathology
Collapse
Affiliation(s)
- Mark D. M. Leiserson
- Microsoft Research New England, Cambridge, MA, United States of America
- University of Maryland, College Park, Department of Computer Science, College Park, MD, United States of America
| | - Vasilis Syrgkanis
- Microsoft Research New England, Cambridge, MA, United States of America
| | - Amy Gilson
- Microsoft Research New England, Cambridge, MA, United States of America
| | - Miroslav Dudik
- Microsoft Research New York, New York, NY, United States of America
| | - Sharon Gillett
- Microsoft Research New England, Cambridge, MA, United States of America
| | - Jennifer Chayes
- Microsoft Research New England, Cambridge, MA, United States of America
- Microsoft Research New York, New York, NY, United States of America
| | - Christian Borgs
- Microsoft Research New England, Cambridge, MA, United States of America
| | - Dean F. Bajorin
- Department of Medicine, Memorial Sloan Kettering Cancer Center, New York, NY, United States of America
- Department of Medicine, Weill Cornell Medical College, New York, NY, United States of America
| | - Jonathan E. Rosenberg
- Department of Medicine, Memorial Sloan Kettering Cancer Center, New York, NY, United States of America
| | - Samuel Funt
- Department of Medicine, Memorial Sloan Kettering Cancer Center, New York, NY, United States of America
- Department of Medicine, Weill Cornell Medical College, New York, NY, United States of America
| | - Alexandra Snyder
- Department of Medicine, Memorial Sloan Kettering Cancer Center, New York, NY, United States of America
- Adaptive Biotechnologies, Seattle, WA, United States of America
| | - Lester Mackey
- Microsoft Research New England, Cambridge, MA, United States of America
- * E-mail:
| |
Collapse
|
|
43
|
Sullivan BA, Wallman-Stokes A, Isler J, Sahni R, Moorman JR, Fairchild KD, Lake DE. Early Pulse Oximetry Data Improves Prediction of Death and Adverse Outcomes in a Two-Center Cohort of Very Low Birth Weight Infants. Am J Perinatol 2018; 35:1331-1338. [PMID: 29807371 PMCID: PMC6262889 DOI: 10.1055/s-0038-1654712] [Citation(s) in RCA: 34] [Impact Index Per Article: 4.9] [Reference Citation Analysis] [Abstract] [MESH Headings] [Grants] [Track Full Text] [Journal Information] [Submit a Manuscript] [Subscribe] [Scholar Register] [Indexed: 10/14/2022]
Abstract
BACKGROUND We previously showed, in a single-center study, that early heart rate (HR) characteristics predicted later adverse outcomes in very low birth weight (VLBW) infants. We sought to improve predictive models by adding oxygenation data and testing in a second neonatal intensive care unit (NICU). METHODS HR and oxygen saturation (SpO2) from the first 12 hours and first 7 days after birth were analyzed for 778 VLBW infants at two NICUs. Using multivariate logistic regression, clinical predictive scores were developed for death, severe intraventricular hemorrhage (sIVH), bronchopulmonary dysplasia (BPD), treated retinopathy of prematurity (tROP), late-onset septicemia (LOS), and necrotizing enterocolitis (NEC). Ten HR-SpO2 measures were analyzed, with first 12 hours data used for predicting death or sIVH and first 7 days for the other outcomes. HR-SpO2 models were combined with clinical models to develop a pulse oximetry predictive score (POPS). Net reclassification improvement (NRI) compared performance of POPS with the clinical predictive score. RESULTS Models using clinical or pulse oximetry variables alone performed well for each outcome. POPS performed better than clinical variables for predicting death, sIVH, and BPD (NRI > 0.5, p < 0.01), but not tROP, LOS, or NEC. CONCLUSION Analysis of early HR-SpO2 characteristics adds to clinical risk factors to predict later adverse outcomes in VLBW infants.
Collapse
Affiliation(s)
- B A Sullivan
- Department of Pediatrics, University of Virginia, Charlottesville, Virginia
| | - A Wallman-Stokes
- Department of Pediatrics, Columbia University, New York, New York
| | - J Isler
- Department of Pediatrics, Columbia University, New York, New York
| | - R Sahni
- Department of Pediatrics, Columbia University, New York, New York
| | - J R Moorman
- Department of Medicine, University of Virginia, Charlottesville, Virginia
| | - K D Fairchild
- Department of Pediatrics, University of Virginia, Charlottesville, Virginia
| | - D E Lake
- Department of Medicine, University of Virginia, Charlottesville, Virginia
| |
Collapse
|
|
44
|
Chisholm KM, Norton ME, Penn AA, Heerema-McKenney A. Classification of Preterm Birth With Placental Correlates. Pediatr Dev Pathol 2018; 21:548-560. [PMID: 29759046 DOI: 10.1177/1093526618775958] [Citation(s) in RCA: 30] [Impact Index Per Article: 4.3] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Submit a Manuscript] [Subscribe] [Scholar Register] [Indexed: 01/01/2023]
Abstract
Premature birth lacks a widely accepted classification that unites features of the clinical presentation with placental pathology. To further explore associations between the clinical categories of preterm birth and placental histology, 109 infants with gestational age <34 weeks and birth weight <2000 g were selected and, based on electronic records, were classified into preterm birth categories of preterm labor, prelabor premature rupture of membranes, preeclampsia, indicated preterm birth for maternal factors (other than preeclampsia), indicated preterm birth for fetal factors, and the clinical diagnosis of abruption. Corresponding placentas were analyzed for gross and microscopic variables, with findings grouped into categories of amniotic fluid infection, lymphocytic inflammation, maternal vascular malperfusion, and fetal vascular malperfusion. Placental features of maternal vascular malperfusion were pervasive in all preterm birth categories and were commonly associated with amniotic fluid infection and lymphocytic inflammation. Features of maternal vascular malperfusion were significantly associated with preterm birth due to preeclampsia, and amniotic fluid infection was highly associated with prelabor preterm rupture of membranes. Findings of lymphocytic inflammation were significantly increased in cases of abruption. Laminar decidual necrosis was present in all cases of abruption. Placentas from multiple gestations had significantly less histologic findings compared to singletons. Given that 75% of placentas demonstrated at least 1 feature of maternal vascular malperfusion despite different clinical presentations, seemingly different pathologies such as ascending amniotic fluid infection or lymphocytic inflammation may be mechanistically related to processes established early in pregnancy. The concept of "uterine ischemia" may be too simplistic to account for all of the changes attributed to maternal vascular malperfusion in the preterm placenta.
Collapse
Affiliation(s)
- Karen M Chisholm
- 1 Department of Pathology, Stanford University School of Medicine, Stanford, California
| | - Mary E Norton
- 2 Department of Obstetrics and Gynecology, Division of Maternal-Fetal Medicine, Stanford University School of Medicine, Stanford, California
| | - Anna A Penn
- 3 Division of Neonatal and Developmental Medicine, Department of Pediatrics, Stanford University School of Medicine, Stanford, California
| | - Amy Heerema-McKenney
- 1 Department of Pathology, Stanford University School of Medicine, Stanford, California
| |
Collapse
|
|
45
|
Abstract
Analog-to-digital data conversion has created massive amounts of historical and real-time health care data. Costs associated with neonatal health issues are high. Big data use in the neonatal intensive care unit has the potential to facilitate earlier detection of clinical deterioration, expedite application of efficient clinical decision-making algorithms based on real-time and historical data mining, and yield significant cost-savings.
Collapse
Affiliation(s)
- Lynn E Bayne
- Christiana Care Health System, 4755 Ogletown-Stanton Road, Newark, DE 19716, USA; Alfred I. duPont Hospital for Children, 1600 Rockland Road, Wilmington, DE 19803, USA.
| |
Collapse
|
|
46
|
Lundberg SM, Nair B, Vavilala MS, Horibe M, Eisses MJ, Adams T, Liston DE, Low DKW, Newman SF, Kim J, Lee SI. Explainable machine-learning predictions for the prevention of hypoxaemia during surgery. Nat Biomed Eng 2018; 2:749-760. [PMID: 31001455 PMCID: PMC6467492 DOI: 10.1038/s41551-018-0304-0] [Citation(s) in RCA: 780] [Impact Index Per Article: 111.4] [Reference Citation Analysis] [Abstract] [MESH Headings] [Grants] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 10/11/2017] [Accepted: 08/31/2018] [Indexed: 12/21/2022]
Abstract
Although anaesthesiologists strive to avoid hypoxemia during surgery, reliably predicting future intraoperative hypoxemia is not currently possible. Here, we report the development and testing of a machine-learning-based system that, in real time during general anaesthesia, predicts the risk of hypoxemia and provides explanations of the risk factors. The system, which was trained on minute-by-minute data from the electronic medical records of over fifty thousand surgeries, improved the performance of anaesthesiologists when providing interpretable hypoxemia risks and contributing factors. The explanations for the predictions are broadly consistent with the literature and with prior knowledge from anaesthesiologists. Our results suggest that if anaesthesiologists currently anticipate 15% of hypoxemia events, with this system's assistance they would anticipate 30% of them, a large portion of which may benefit from early intervention because they are associated with modifiable factors. The system can help improve the clinical understanding of hypoxemia risk during anaesthesia care by providing general insights into the exact changes in risk induced by certain patient or procedure characteristics.
Collapse
Affiliation(s)
- Scott M Lundberg
- Paul G. Allen School of Computer Science and Engineering, University of Washington, Seattle, WA, USA
| | - Bala Nair
- Department of Anesthesiology and Pain Medicine, University of Washington, Seattle, WA, USA
- Center for Perioperative and Pain initiatives in Quality Safety Outcome, University of Washington, Seattle, WA, USA
- Harborview Injury Prevention and Research Center, University of Washington, Seattle, WA, USA
| | - Monica S Vavilala
- Department of Anesthesiology and Pain Medicine, University of Washington, Seattle, WA, USA
- Center for Perioperative and Pain initiatives in Quality Safety Outcome, University of Washington, Seattle, WA, USA
- Harborview Injury Prevention and Research Center, University of Washington, Seattle, WA, USA
| | - Mayumi Horibe
- Veterans Affairs Puget Sound Health Care System, Seattle, WA, USA
| | - Michael J Eisses
- Department of Anesthesiology and Pain Medicine, University of Washington, Seattle, WA, USA
- Seattle Children's Hospital, Seattle, WA, USA
| | - Trevor Adams
- Department of Anesthesiology and Pain Medicine, University of Washington, Seattle, WA, USA
- Seattle Children's Hospital, Seattle, WA, USA
| | - David E Liston
- Department of Anesthesiology and Pain Medicine, University of Washington, Seattle, WA, USA
- Seattle Children's Hospital, Seattle, WA, USA
| | - Daniel King-Wai Low
- Department of Anesthesiology and Pain Medicine, University of Washington, Seattle, WA, USA
- Seattle Children's Hospital, Seattle, WA, USA
| | - Shu-Fang Newman
- Department of Anesthesiology and Pain Medicine, University of Washington, Seattle, WA, USA
- Center for Perioperative and Pain initiatives in Quality Safety Outcome, University of Washington, Seattle, WA, USA
| | - Jerry Kim
- Department of Anesthesiology and Pain Medicine, University of Washington, Seattle, WA, USA
- Seattle Children's Hospital, Seattle, WA, USA
| | - Su-In Lee
- Paul G. Allen School of Computer Science and Engineering, University of Washington, Seattle, WA, USA.
| |
Collapse
|
|
47
|
Rinta-Koski OP, Särkkä S, Hollmén J, Leskinen M, Andersson S. Gaussian process classification for prediction of in-hospital mortality among preterm infants. Neurocomputing 2018. [DOI: 10.1016/j.neucom.2017.12.064] [Citation(s) in RCA: 5] [Impact Index Per Article: 0.7] [Reference Citation Analysis] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 10/18/2022]
|
|
48
|
Wu S, Liu S, Sohn S, Moon S, Wi CI, Juhn Y, Liu H. Modeling asynchronous event sequences with RNNs. J Biomed Inform 2018; 83:167-177. [PMID: 29883623 PMCID: PMC6103779 DOI: 10.1016/j.jbi.2018.05.016] [Citation(s) in RCA: 22] [Impact Index Per Article: 3.1] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 06/26/2017] [Revised: 05/10/2018] [Accepted: 05/26/2018] [Indexed: 12/14/2022]
Abstract
Sequences of events have often been modeled with computational techniques, but typical preprocessing steps and problem settings do not explicitly address the ramifications of timestamped events. Clinical data, such as is found in electronic health records (EHRs), typically comes with timestamp information. In this work, we define event sequences and their properties: synchronicity, evenness, and co-cardinality; we then show how asynchronous, uneven, and multi-cardinal problem settings can support explicit accountings of relative time. Our evaluation uses the temporally sensitive clinical use case of pediatric asthma, which is a chronic disease with symptoms (and lack thereof) evolving over time. We show several approaches to explicitly incorporating relative time into a recurrent neural network (RNN) model that improve the overall classification of patients into those with no asthma, those with persistent asthma, those in long-term remission, and those who have experienced relapse. We also compare and contrast these results with those in an inpatient intensive care setting.
Collapse
Affiliation(s)
- Stephen Wu
- Department of Health Sciences Research, Mayo Clinic, Rochester, MN, United States.
| | - Sijia Liu
- Department of Health Sciences Research, Mayo Clinic, Rochester, MN, United States
| | - Sunghwan Sohn
- Department of Health Sciences Research, Mayo Clinic, Rochester, MN, United States
| | - Sungrim Moon
- Department of Health Sciences Research, Mayo Clinic, Rochester, MN, United States
| | - Chung-Il Wi
- Department of Pediatrics, Mayo Clinic, Rochester, MN, United States
| | - Young Juhn
- Department of Pediatrics, Mayo Clinic, Rochester, MN, United States
| | - Hongfang Liu
- Department of Health Sciences Research, Mayo Clinic, Rochester, MN, United States
| |
Collapse
|
|
49
|
Terrill PI, Dakin C, Edwards BA, Wilson SJ, MacLean JE. A graphical method for comparing nocturnal oxygen saturation profiles in individuals and populations: Application to healthy infants and preterm neonates. Pediatr Pulmonol 2018; 53:645-655. [PMID: 29575753 DOI: 10.1002/ppul.23987] [Citation(s) in RCA: 11] [Impact Index Per Article: 1.6] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Submit a Manuscript] [Subscribe] [Scholar Register] [Received: 10/31/2017] [Accepted: 02/24/2018] [Indexed: 01/09/2023]
Abstract
STUDY OBJECTIVES Pulse-oximetry (SpO2 ) allows the identification of important clinical physiology. However, summary statistics such as mean values and desaturation incidence do not capture the complexity of the information contained within continuous recordings. The aim of this study was to develop an objective method to quantify important SpO2 characteristics; and assess its utility in healthy infant and preterm neonate cohorts. METHODS An algorithm was developed to calculate the desaturation incidence, depth, and duration. These variables are presented using three plots: SpO2 cumulative-frequency relationship; desaturation-depth versus incidence; desaturation-duration versus incidence. This method was applied to two populations who underwent nocturnal pulse-oximetry: (1) thirty-four healthy term infants studied at 2-weeks, 3, 6, 12, and 24-months of age and (2) thirty-seven neonates born <26 weeks and studied at discharge from NICU (37-44 weeks post-conceptual age). RESULTS The maturation in healthy infants was characterized by reduced desaturation index (27.2/h vs 3.3/h at 2-weeks and 24-months, P < 0.01), and increased percentage of desaturation events ≥6-s in duration (27.8% vs 43.2% at 2-weeks and 3-months, P < 0.01). Compared with term-infants, preterm infants had a greater desaturation incidence (54.8/h vs 27.2/h, P < 0.01), and these desaturations were deeper (52.9% vs 37.6% were ≥6% below baseline, P < 0.01). The incidence of longer desaturations (≥14-s) in preterm infants was correlated with healthcare utilization over the first 24-months (r = 0.63, P < 0.01). CONCLUSIONS This tool allows the objective comparison of extended oximetry recordings between groups and for individuals; and serves as a basis for the development of reference ranges for populations.
Collapse
Affiliation(s)
- Philip I Terrill
- School of Information Technology and Electrical Engineering, The University of Queensland, Brisbane, Queensland, Australia
| | - Carolyn Dakin
- The Canberra Hospital, Garran, Australian Capital Territory, Australia
| | - Bradley A Edwards
- Department of Physiology, Monash University, Melbourne, Victoria, Australia.,School of Psychological Sciences and Monash Institute of Cognitive and Clinical Neurosciences, Monash University, Melbourne, Victoria, Australia
| | - Stephen J Wilson
- School of Information Technology and Electrical Engineering, The University of Queensland, Brisbane, Queensland, Australia
| | - Joanna E MacLean
- Faculty of Medicine and Dentistry, Division of Respiratory Medicine, Department of Pediatrics, University of Alberta, Edmonton, Alberta, Canada
| |
Collapse
|
|
50
|
Lee J, Sun J, Wang F, Wang S, Jun CH, Jiang X. Privacy-Preserving Patient Similarity Learning in a Federated Environment: Development and Analysis. JMIR Med Inform 2018; 6:e20. [PMID: 29653917 PMCID: PMC5924379 DOI: 10.2196/medinform.7744] [Citation(s) in RCA: 61] [Impact Index Per Article: 8.7] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 03/29/2017] [Revised: 09/12/2017] [Accepted: 01/06/2018] [Indexed: 12/14/2022] Open
Abstract
Background There is an urgent need for the development of global analytic frameworks that can perform analyses in a privacy-preserving federated environment across multiple institutions without privacy leakage. A few studies on the topic of federated medical analysis have been conducted recently with the focus on several algorithms. However, none of them have solved similar patient matching, which is useful for applications such as cohort construction for cross-institution observational studies, disease surveillance, and clinical trials recruitment. Objective The aim of this study was to present a privacy-preserving platform in a federated setting for patient similarity learning across institutions. Without sharing patient-level information, our model can find similar patients from one hospital to another. Methods We proposed a federated patient hashing framework and developed a novel algorithm to learn context-specific hash codes to represent patients across institutions. The similarities between patients can be efficiently computed using the resulting hash codes of corresponding patients. To avoid security attack from reverse engineering on the model, we applied homomorphic encryption to patient similarity search in a federated setting. Results We used sequential medical events extracted from the Multiparameter Intelligent Monitoring in Intensive Care-III database to evaluate the proposed algorithm in predicting the incidence of five diseases independently. Our algorithm achieved averaged area under the curves of 0.9154 and 0.8012 with balanced and imbalanced data, respectively, in κ-nearest neighbor with κ=3. We also confirmed privacy preservation in similarity search by using homomorphic encryption. Conclusions The proposed algorithm can help search similar patients across institutions effectively to support federated data analysis in a privacy-preserving manner.
Collapse
Affiliation(s)
- Junghye Lee
- School of Management Engineering, Ulsan National Institute of Science and Technology, Ulsan, Republic Of Korea.,Department of Biomedical Informatics, University of California San Diego, San Diego, CA, United States.,Department of Industrial and Management Engineering, Pohang University of Science and Technology, Pohang, Republic Of Korea
| | - Jimeng Sun
- College of Computing, Georgia Institute of Technology, Atlanta, GA, United States
| | - Fei Wang
- Division of Health Informatics, Department of Healthcare Policy and Research, Weill Cornell Medical College, Cornell University, New York City, NY, United States
| | - Shuang Wang
- Department of Biomedical Informatics, University of California San Diego, San Diego, CA, United States
| | - Chi-Hyuck Jun
- Department of Industrial and Management Engineering, Pohang University of Science and Technology, Pohang, Republic Of Korea
| | - Xiaoqian Jiang
- Department of Biomedical Informatics, University of California San Diego, San Diego, CA, United States
| |
Collapse
|