Elsevier

Decision Support Systems

Volume 46, Issue 2, January 2009, Pages 586-593
Decision Support Systems

Identifying RFID-embedded objects in pervasive healthcare applications

https://doi.org/10.1016/j.dss.2008.10.001Get rights and content

Abstract

The organization and delivery of pervasive healthcare have benefited much from advances in wireless systems. While wireless systems and their components have certainly enhanced the quality of pervasive healthcare administered in remote locations, their potential in other areas of healthcare cannot be underestimated. We consider Radio Frequency Identification (RFID) tags, which are increasingly being used in pervasive healthcare applications. Specifically, we study the dynamics of locating and identifying the presence of a tag in such systems. Although a tag may be present, it may not necessarily be visible to the tag reader due to various constraints or reasons. We propose and illustrate several algorithms for locating the presence of RFID tagged objects in the field of the reader and study their dynamics as well as their strengths and benefits. Our results indicate that the location accuracy of RFID tag readers can be improved through appropriate data collection as well as algorithms used for data inference.

Introduction

The primary goal of pervasive healthcare is to be able to deliver necessary quality healthcare service anytime to anyone regardless of location and other constraints. Improvements in technology have enabled the feasibility of this vision, primarily based on physical constraints, even though the actual practice lags much behind what can and should be achieved. A key player in systems facilitating pervasive healthcare is wireless technology. The role played by wireless technology in pervasive healthcare spans a wide range from basic telecommunication to identification and tracking of mobile objects. Sensors and actuators are commonly used to measure ambient conditions and then react to such conditions through intelligent information systems by appropriately instantiating necessary components. Pervasive healthcare necessarily involves the use of multiple layers of intelligent information systems technology that work together synergistically to deliver results in a seamless fashion.

The idea of incorporating intelligent information systems in the healthcare domain is not new [17]. Although the terminology used is somewhat different, the underlying principle in most of these intelligent information systems are rather similar to the extent that they all utilize knowledge in some form to enable decision making in the healthcare environment. Commonly used terms for systems used for decision support in this environment include Clinical Decision Support Systems (CDSS), Intelligent Decision Support Systems (IDSS), Healthcare Information Systems (HIS), among others. CDSS, IDSS, HIS, etc. all represent similar systems where the terminology are interchangeably used, and are therefore considered to be the same for purposes of this paper. CDSS are commonly used to aid in medical prescription, clinical laboratories, clinical surveillance, clinical education, and intensive care settings, among others. Frost & Sullivan (http://www.healthcare.frost.com) estimates that revenues for Clinical Decision Support Systems market in Europe would reach $430.7 million in 2012. While intelligent decision support systems have successfully been utilized in healthcare settings, some common reasons cited for unsuccessful scenarios generally involve their use in solving problems that were not considered to be an issue or imposing poor human interface design, reluctance or computer illiteracy of some healthcare workers, restraining or significantly modifying the overall workflow in such a setting.

Intelligent Healthcare Information Systems (HIS) have been used to critique therapy; check for drug interactions, dosage errors, or allergy to improve the quality of clinical decisions; build and utilize electronic patient record system; and improve compliance with clinical pathways and guidelines. Therapy critiquing works by looking for inconsistencies, errors, or omissions in an existing treatment plan. During physician order entry, such a system can critique the combination of patient's condition and treatment plan. For example, when a clinician enters an order for blood transfusion to a patient with haemoglobin level above the transfusion threshold, the system can prompt the clinician to justify this order such as the presence of active bleeding [21]. Order entry and results reporting systems with embedded Decision Support Systems have been shown to increase compliance with recommended clinical pathways and guidelines and reduce rates of inappropriate diagnostic tests [11].

Potential benefits that are associated with intelligent healthcare information systems include improved patient safety through reduced medication errors and adverse events, improved medication/test ordering, improved quality of care, and improved efficiency in healthcare delivery by reducing costs related to theft and shrinkage. Moreover, electronic prescribing systems have been shown to be effective in reducing errors. Given the sheer complexity of the drug-use process and the multitude of potential failure points, computerized order-entry and drug management systems are promising tools for decreasing medication errors, preventing ADEs (Adverse Drug Events), and improving drug use [6].

Clinical Decision Support System has been promoted as an enabling technology that transforms existing healthcare systems [7]. A growing body of evidence indicates that such systems may decrease error rates and improve therapy, thereby improving outcomes including survival, the length of time patients spend in dangerous conditions, hospital length of stay, and cost. After studying their dynamics in terms of health and benefits, Bates et al. [5] conclude that appropriate use of information technology in healthcare could result in process simplification and substantial improvement in patient safety. Although injuries associated with errors in healthcare are important, costs of inefficiencies related to errors that do not directly or indirectly result in injury are also important. For example, “missed dose” medication errors occur when a required medication dose is not available for administering on time, and when this delay exceeds a few hours the dose is generally not given in order to prevent interference with a subsequent dose. To alleviate such problems nurses spend a great deal of time tracking down appropriate medications so that they are delivered to the patients on time. The costs associated with such events are generally not accounted for since they are harder to assess.

Considering support tools for clinical decision making, Garg et al. [10] reviewed several studies involving systems with human interaction and concluded that systems where users were automatically prompted to use the system had better performance compared with cases where users were required to actively initiate the system. They also found that compared to manual initiation, automatic prompting by these intelligent decision support systems may result in smooth integration with practitioner workflow as well as provide better opportunities to correct inadvertent deficiencies in care.

Bates [5] recommends incorporating barcodes on medications, blood, devices, and patients to reduce errors. In the U.S., it is estimated that over 770,000 people are injured or die each year in hospitals as a result of adverse drug effects [15], of which the greatest proportion (56%) occur at the drug ordering stage. Added to this is the fact that most laboratory systems do not communicate directly with pharmacy systems. These intermittent gaps in communication among healthcare systems result in simple misunderstandings to utter chaos when different parts of the systems are not in sync.

From the above discussion, it is clear that healthcare information systems where communication among its sub-components occurs seamlessly are essential for pervasive healthcare. It is also essential to automatically capture as much data as possible and let the computer do the checking, relating, and associated book-keeping work thus releasing healthcare personnel to do their primary work in delivering quality healthcare in a pervasive environment. Hospital employees make decisions utilizing available resources including patient records and necessary specialized equipment. Healthcare information systems that have discontinuities, where their sub-components are not seamlessly integrated with one another, are not completely reliable in providing error-free and timely necessary information for the decision makers. It is therefore critical for healthcare information systems to be current in being able to provide patient records and relevant reference material in a pervasive computing environment.

Pervasive healthcare is, by its very nature, multi-faceted and requires appropriate enabling technologies depending on location, time, treatment, personnel, and the general healthcare environment. We consider RFID technology, which is increasingly becoming essential in pervasive healthcare environments [4]. RFID tags are used in scenarios where an object needs to be identified, tracked, or when ambient condition surrounding an object is captured and stored, among others [19]. Clearly, there are several modes of applications where RFID implementations are used, and each mode has its corresponding pros and cons. We consider the data generated as a result of incorporation of RFID tags. Specifically, we consider some characteristics of RFID-generated data and how they relate to overall system performance. Although it is generally assumed that data read from RFID tags are highly accurate, variations in accuracy can and do occur due to several reasons. Jeffery et al. [14] claim that over 30% of RFID tag reads are routinely dropped. Although the extent of dropped reads varies across domains and is context-specific, figures in the 60–99% range are not uncommon. We consider this problem, and propose algorithms that can be used to reduce false positive and false negative readings in such applications. It should be noted that we do not consider collisions that occur due to multiple tags simultaneously transmitting signals read by a single reader or when multiple readers are simultaneously trying to bounce off signals on a single tag. We illustrate the performance of the proposed algorithms using an example scenario. The proposed algorithms can generate more accurate tag reads, and can therefore be incorporated in an appropriate healthcare decision support system to improve its overall performance.

The remainder of this paper is organized as follows. We list general characteristics of RFID tags from a healthcare perspective in the next section. We then review and summarize some literature related to data generated by RFID tags in Section 3. We present a few algorithms that can be used to reduce false positives and false negatives, and illustrate their relative performance with examples in Section 4. We conclude with a brief discussion in Section 6.

Section snippets

RFID tags and healthcare

There are more than 3000 RFID case studies in a wide range of application areas. RFID in healthcare is growing rapidly and is forecast to become a $2.1 billion global business by 2016. One of the largest volumes of RFID application has been in the healthcare industry, where about 4.5 million tags have been used every year on Diprivan drug syringes by AstraZeneca since 1999.

Asset tracking is a prime candidate for RFID applications. A typical hospital is unable to locate about 15–20% of its

Related literature on improving RFID tag identification accuracy

While preserving the arrival sequence of tags, Bai et al. [2] propose means to filter and clean data streams from RFID applications that contain false (e.g., false positive, false negative) readings and duplicates. False positive readings or noise could be a result of the reader's detection field being spread wider than necessary that signals from tags farther than its intended scope are captured. Duplicate readings occur when tags remain in the reader's detection field for a longer duration,

Proposed methods

Identifying the presence of RFID tags is not always necessarily as straight-forward as it seems. False positives and false negatives can be a problem in RFID-embedded systems, especially when signal from a given tag is blocked by an impenetrable object (e.g., metal shielding) or when corrupted signal is read. We propose means to reduce false positives and false negatives through a variation of triangulation where we consider the presence/absence of a related tag. An example of this scenario

Discussion

The complexities associated with seamlessly integrating different related components to deliver pervasive healthcare necessitate peak performance of the individual components of the system both by themselves and also in concert with others. Although RFID tags are increasingly being used in such pervasive healthcare environments, their performance and therefore the resulting overall system performance can be improved further. Specifically, we considered the detection problem where a tagged

Yu-Ju Tu is a graduate student in Information Systems at the University of Illinois at Urbana-Champaign. His research interests include RFID systems.

References (25)

  • J.M. Corchado et al.

    Intelligent environment for monitoring Alzheimer patients, agent technology for healthcare

    Decision Support Systems

    (2008)
  • E.W.T. Ngai et al.

    Mobile commerce integrated with RFID technology in a container depot

    Decision Support Systems

    (2007)
  • S. Piramuthu

    Protocols for RFID tag/reader authentication

    Decision Support Systems

    (2007)
  • S. Agarwal et al.

    A pervasive computing system for the operating room of the future

    Mobile Networks and Applications

    (2007)
  • Y. Bai et al.

    Efficiently Filtering RFID Data Streams, CleanDB

    (2006)
  • Y. Bai et al.

    RFID data processing with a data stream query language

  • D.W. Bates et al.

    Reducing the frequency of errors in medicine using information technology

    Journal of the American Medical Informatics Association

    (2001)
  • D.W. Bates et al.

    Effects of computerized physician order entry and a team intervention on prevention of serious medication errors

    Journal of the American Medical Association

    (1998)
  • E.S. Berner

    Diagnostic decision support systems: why aren't they used more and what can we do about it?

  • G. Borriello et al.

    Reminding about tagged objects using passive RFIDs

    Ubicomp

    (2004)
  • B. Carbunar et al.

    Redundancy and coverage detection in sensor networks

    ACM Transactions on Sensor Networks

    (2006)
  • A.X. Garg et al.

    Effects of computerized clinical decision support systems on practitioner performance and patient outcomes

    Journal of the American Medical Association

    (2005)
  • Cited by (126)

    • Safety risks in underground operations: management and assessment techniques

      2022, Risk, Reliability and Sustainable Remediation in the Field of Civil and Environmental Engineering
    • Success factors in developing iHeart as a patient-centric healthcare system: A multi-group analysis

      2018, Telematics and Informatics
      Citation Excerpt :

      The success factors, which are explored by this paper, assist us to develop a patient-centric healthcare system. Location management can add value to patient-centric healthcare by offering a location-based healthcare services using location tracking technologies such as global positioning systems (GPS), cellular networks, wireless local area networks (LANs), and radio frequency identification (RFID) (Tu et al., 2009; Varshney, 2003). Various remote devices are utilized in wireless patient-centric healthcare systems such as wireless sensors for blood oxygen saturation monitoring (Asada et al., 2003; Sokwoo et al., 1998), infrared and radio-based locator badges (Stanford, 2002), using wearable devices for long-term health monitoring (Jovanov et al., 2002), remote heart monitoring systems (Ernest et al., 2011; Ross, 2004), healthcare monitoring for patients with hypertension and arrhythmia (Keikhosrokiani et al., 2012), providing emergency services based on patient’s current location (El-Masri and Saddik, 2012; Keikhosrokiani et al., 2012), wireless telemetry systems for EEG epilepsy monitoring (Liszka et al., 2004), real-time monitoring of patients in the home environment (Korhonen et al., 2003), and lastly, using GlucoBeep device for checking and transmitting the patient’s diabetic information (Jansà et al., 2006).

    View all citing articles on Scopus

    Yu-Ju Tu is a graduate student in Information Systems at the University of Illinois at Urbana-Champaign. His research interests include RFID systems.

    Wei Zhou received his Ph.D. in Information Systems from the University of Florida in 2008. His research interests include RFID-enabled item-level information visibility, Internet advertising, and knowledge-based learning systems. His work has appeared in European Journal of Operational Research, IEEE Transactions on Geosciences and Remote Sensing, International Journal of Electronic Commerce, and Optical Engineering.

    Selwyn Piramuthu is Associate Professor in the Information Systems and Operations Management Department at the University of Florida. His research interests include RFID systems, pattern recognition and its application in supply chain management, computer-aided manufacturing, and financial credit-risk analysis.

    View full text