Distributed data-driven platform for urgent decision making in cardiological ambulance control

https://doi.org/10.1016/j.future.2016.09.017Get rights and content

Highlights

  • A conceptual architecture to diverse data integration within clinical DSS was proposed.

  • An urgent computing platform for data processing was developed to support reactive data-driven support for decision making.

  • An application of DSS to support ambulance and emergency medical service within a cases of acute coronary syndrome (ACS) was described.

  • Experimental study was conducted to show applicability of the proposed approach and to estimate performance characteristics.

Abstract

This paper presents ongoing research aimed at developing a data-driven platform for clinical decision support systems (DSSs) that require integration and processing of various data sources within a single solution. Resource management is developed within a framework of an urgent computing approach to address changing requirements defined by the incoming flow of patients with urgent diseases. This work presents DSS for support of ambulance and emergency medical service management for patients with acute coronary syndrome (ACS) as a working example with integration distributed streaming data sources as well as data storages (containing electrocardiography (ECG) data, electronic medical records (EMR), real-time monitoring of medical facilities, and schedules of hospitals within a network). This DSS has been developed in collaboration with the Federal Almazov North-West Medical Research Centre in Saint Petersburg.

Introduction

According to a World Health Organization report  [1], cardiovascular diseases are the major cause of death in the world. Many types, such as acute coronary syndrome (ACS) or stroke, require treatment within several hours, and delays significantly decrease the probability of a favorable outcome  [2]. Usually, patients who suffer such acute diseases are served by emergency medical services: visited and transferred by ambulance, have a surgery or intensive therapy at the hospital. In the prehospital setting, patient evaluation, triage (care strategy and hospitalization pathway selection), and early treatment are essential for positive outcomes. Considering differences in equipment, a load of hospitals and ambulances, transportation difficulties in urban traffic, and possible variations in the urgency for a particular patient and multiple ambulance services, the issue of coordination and decision support for such services becomes especially important.

In this paper, we present an ongoing project aimed at developing a clinical decision support system (DSS) for managing emergency medical services in cases of acute cardiac disease using an ACS patient as a working example. The goal of the developed DSS will be centralized coordination of ambulance cars and emergency medical services in a set of hospitals with special facilities for cardiological healthcare to respond to ACS treatment requirements. The DSS has been developed in cooperation with the Federal Almazov North-West Medical Research Centre (FANWMRC)1 in Saint Petersburg.

Today, development of DSS for ambulance dispatching and routing is a widely covered area (see e.g.,  [3], [4], [5]) with a strong theoretical background in field operational research and management science  [6]. Nevertheless, practical implementation of DSS for ambulance control for the case considered here requires the management of many specific issues. First, all the mentioned issues and specific aspects of particular disease treatment should be taken into account and resolved within the considered scenarios. Second, proper implementation of DSS requires integration of diverse data and information sources (sensors, public data services, and electronic medical records) with variations in formats, access policies, protocols, and so on. As a result, when developing a DSS, one is faced with problems of data and information fusion  [7] and issues related to the integration of various data sources and data processing methods. Third, enhanced decision making requires a set of models (e.g., traffic model, patients queue mode) and data analysis components (e.g., search for precedents, classification of cases), to be integrated within a composite application to assess a situation, predict its development, and optimize the management scenario. Finally, decision support requires a set of resources to be managed in accordance with the emergency needs and patient inflow in an urgent way with manageable priority elevation and task scheduling.

This paper presents a general approach to building a data-driven platform for urgent DSS development to enable management of data sources, computations, and other resources within different scenarios with changing priorities, multiple stakeholders and decision makers, and urgent requests. The presented platform is now being developed to serve as a technological basis for DSS for the ambulance and emergency medical service coordination and management using ACS cases as a working example where urgent decision making is critical.

Section snippets

Clinical backgrounds

ACS, a typical difficulty of coronary illness, is related to more than 2.5 million hospitalizations worldwide every year  [8]. It is a circumstance where the blood supplied to the heart muscle is suddenly blocked. The blockage can be sudden and complete, or it can go back and forth, clotting, breaking open, and then clotting again. In either case, the heart tissue is dying, regardless of the possibility that it is only a couple of cells or an entire huge area of the heart  [9], [10]. Usually,

Previous works and problem statements

Today the development of clinical DSS is a well-developed area with a long history  [18]. Still, it has multiple open issues, and there is a multitude of papers aimed at various aspects of clinical DSS development from basic principles of development  [19], [20] to psychological  [21] and sociological  [22] issues. Within our current research, we are focused on the development of a conceptual and technological basis for clinical DSS building within P4 medicine approach  [23] as well as DSS

Distributed data-driven infrastructure

To fulfill the conceptual needs described in Section  3, the following distributed data-driven infrastructure was developed and experimentally studied in Section  5.2. The main principles were taken from the formerly developed platform CLAVIRE  [28], with adapted principles for urgent computing  [29]. However, for the new platform, the basic accent shifted from an HPC-oriented system to a data-driven infrastructure adapting advanced Big Data technologies. High-performance computing, in its

Queueing in hospitals

To analyze the process of ambulance patients with ACS being served by the hospital network, a queueing model was developed. It describes the activity of a hospital with several PCI facilities in cases of defined inflow of patients. Simulations performed by the developed model provide extended information on patients’ inflow processing, e.g., analysis of time spent in a queue by patients can be used to generate a predictive estimation of mortality rates.

Time distribution for the main parameters

Conclusion and future works

The paper presents an ongoing study aimed at developing a data-driven platform for clinical decision support systems. The architecture of the platform integrates diverse data sources including streaming and stored data of various formats and nature. The important aspect of the architecture is the urgent management of data sources and computational resources to meet the needs of emergency clinical situations. Being implemented as a general-purpose solution, the architecture is primarily aimed to

Acknowledgments

The modeling and simulation results, as well as basic decision making solutions for emergency medical services approaches, were developed within a project supported by Russian Scientific Foundation, grant #14-11-00823. The high-performance distributed infrastructure for data processing was developed and investigated within a project sponsored by Russian Foundation for Basic Research, grant 15-29-0703416.

Sergey V. Kovalchuk is a senior researcher at eScience Research Institute at ITMO University (Saint-Petersburg, Russia). In 2008 he defended a Ph.D. in computer science on development of high-performance software system for extreme metocean events simulation. He participates in major R&D projects of eScience Research Institute, gives lectures and supervises Master students. His current research area covers complex system simulation, high-performance and distributed computing, workflow

References (38)

  • D.K. Moser et al.

    Reducing delay in seeking treatment by patients with acute coronary syndrome and stroke: A scientific statement from the American heart association council on cardiovascular nursing and stroke council

    Circulation

    (2006)
  • T. Andersson et al.

    Decision support tools for ambulance dispatch and relocation

    J. Oper. Res. Soc.

    (2006)
  • N. Meinzer, S. Storandt, Decision support in emergency medical systems: New strategies for dynamic ambulance...
  • P.J.H. Hulshof et al.

    Taxonomic classification of planning decisions in health care: a structured review of the state of the art in OR/MS

    Health Syst.

    (2012)
  • P.H. Foo et al.

    High-level information fusion: An overview

    J. Adv. Inform. Fusion

    (2013)
  • Acute coronary syndrome: a leading cause of mortality, (n.d.)....
  • J.W. Warnica, Acute Coronary Syndromes (ACS), (2013)....
  • A.C. Fanaroff et al.

    Acute coronary syndrome

    J. Am. Med. Assoc.

    (2015)
  • G.-Y. Cho et al.

    An optimized compression algorithm for real-time ECG data transmission in wireless network of medical information systems

    J. Med. Syst.

    (2015)
  • Cited by (37)

    • KNNOR: An oversampling technique for imbalanced datasets[Formula presented]

      2022, Applied Soft Computing
      Citation Excerpt :

      Machine learning techniques can help in extraction of useful knowledge from large, complex, heterogeneous and hierarchical data. When it comes to classification, various methods like neural network, naïve bayes, k Nearest Neighbors(KNN), Support Vector Machines(SVM) and several others have achieved good performance [1–3]. As these state-of-the-art models emphasize more on classification accuracy than the imbalanced nature of the input data, it is possible for a classifier to achieve high classification accuracy and not predict a single minority class correctly in case of an imbalanced dataset.

    • Detection-based prioritisation: Framework of multi-laboratory characteristics for asymptomatic COVID-19 carriers based on integrated Entropy–TOPSIS methods

      2021, Artificial Intelligence in Medicine
      Citation Excerpt :

      A clear demonstration of such a problem may cast doubt because the existing published works failed to produce a comprehensive patient dataset considering the laboratory criteria for this subject. Detecting health conditions by using the most important criteria with respect to the appropriate weight allocated for each criteria is considered a multi-criteria decision matrix that can describe the specific problems of prioritisation [24,25]. Different weights are often given for various criteria which further increase the complexity of the task [26].

    • A new emergency decision support system: the automatic interpretation and contextualisation of events to model a crisis situation in real-time

      2020, Decision Support Systems
      Citation Excerpt :

      It refers to a dedicated event ontology and performs event trust analysis: it runs an anomaly detection algorithm on the location of complex (output) events. Finally, Kovalchuk et al. [24] simulate queues of patients to predict mortality rates and support planning decisions in hospitals. They also analyse streams of ECG events to trigger an alert if a threshold is exceeded.

    View all citing articles on Scopus

    Sergey V. Kovalchuk is a senior researcher at eScience Research Institute at ITMO University (Saint-Petersburg, Russia). In 2008 he defended a Ph.D. in computer science on development of high-performance software system for extreme metocean events simulation. He participates in major R&D projects of eScience Research Institute, gives lectures and supervises Master students. His current research area covers complex system simulation, high-performance and distributed computing, workflow management systems, knowledge-based eScience technologies.

    Evgeniy Krotov is a master student of double-degree master’s program at ITMO University (Saint-Petersburg) and UvA (Amsterdam). He got his bachelor’s Degree in higher mathematics in the Tomsk State University in 2015. His research studies and interests include big data technologies and machine learning algorithms with applications in medicine and healthcare.

    Pavel A. Smirnov is a research engineer at eScience Research Institute at ITMO University. Doing research in field of high-performance computing, cloud and Big Data infrastructure. In 2014 defended a Ph.D. devoted to workflow-based applications design via Virtual Simulation Objects concept and technology. His topics of interest: data streaming, workload scaling, resource allocation, scheduling, metrics monitoring, QoS.

    Denis A. Nasonov is a researcher at eScience Research Institute in ITMO University. His major research interests are cyberinfrastructure development, technologies for high-performance and distributed computing, grid computing, cloud computing, technologies for processing BigData.

    Alexey N. Yakovalev MD, is head of scientific laboratory of acute coronary syndrome, associate professor of department of anesthesiology and intensive care of Federal Almazov North-West Medical Research Centre (Saint-Petersburg, Russian Federation), and practicing physician of intensive care unit. The topic of thesis for Ph.D. degree in medicine (2008) was the acute treatment strategies of acute coronary syndrome (ACS). Now his current research interests include the organization of regional and local systems of care for ACS patients and information technologies in acute medical care.

    View full text