|
|
AMIA Annu Symp Proc. 2006; 2006: 544–548. | PMCID: PMC1839501 |
Copyright This is an Open Access article: verbatim copying and redistribution of this article are permitted in all media for any purpose. The Design of a Decentralized Electronic Triage System Tammara Massey, MS,1 Tia Gao, MS,2 Matt Welsh, PhD,3 Jonathan H. Sharp,2 and Majid Sarrafzadeh, PhD1 1 University of California, Los Angeles, Los Angeles, CA 2 John Hopkins University Applied Physics Laboratory, Laurel, MD 3 Harvard University Division of Engineering and Applied Sciences, Cambridge, MA The Advanced Health and Disaster Aid Network (AID-N) project seeks to identify unmet needs of emergency response teams in the Washington, DC area during mass casualty incidents and conduct feasibility tests of technology-based solutions. The decentralized electronic triage and sensing system uses low power, electronic triage sensors to monitor the vital signs of patients and provide location tracking capabilities. The robust, decentralized location tracking software runs on a small, embedded system with limited memory and computational power that efficiently locates patients. A field study demonstrates the process of current emergency procedures and the design implications of the prototype. This field study, along with the hardware and software architecture of the electronic triage system, lay the foundation for a reliable, decentralized sensor deployment that will continuously extend network coverage during a mass casualty incident. During a mass casualty incident, many injured people must be helped in a quick and efficient manner. A useful triage system efficiently chooses the order in which individuals are sent to the hospital. An effective system also distributes limited medical resources in a manner that helps as many people as possible. At a disaster scene, it is critical that patients are correctly diagnosed, monitored, and located to ensure the preservation of the maximum number of lives. Unfortunately, the current systems use paper triage tags that inefficiently monitor and locate patients during mass casualty situations. In a mass casualty situation, initial triage is the counting and sorting of patients according to their condition. Initial triage is performed most often by first responders. In the paper triage tag system a first responder attaches a paper tag or colored ribbon to each patient. The first responder then calls the incident commander and reports the patient count. The commander tallies the patient numbers and calls for the necessary number of ambulances. Paper tags employ color codes to determine the severity of the patients’ injury. Patients classified as red are considered to need the most immediate attention, followed by patients classified as yellow. Patients classified as green are the least severely injured and patients classified as black are deceased or expected to die shortly. These tags have many obvious limitations in monitoring patients. They do not allow responders to change certain color designations. The tags also provide little room for manually recording essential information during treatment, such as the patients’ vital signs and chief complaints. Furthermore, reading the tags can be difficult because the patient information, recorded under time pressured situations, is often illegible. Paper tags also have limited visual feedback and do not aid in locating a particular patient in a sea of patients with the same triage color tags. When a commander needs to tally the number of patients triaged under a certain color, the manual count is prone to human error. Finally, paper tags do not distinguish between patients categorized under the same color. Two patients categorized as critical (red) have the same priority, even if one patient’s vital signs designate him to be much worse than the other. Upon completion of initial triage, patients are moved from the triage area to the treatment/waiting area, where they await transportation to a hospital. Secondary triage does an in-depth reassessment of the patient’s condition and is performed in the treatment area and during transport to the hospital. Secondary triage captures the patient’s demographics (age, gender), allergies, medications, chief complaint, and a description of the injury. This information is necessary for transportation officers to designate a hospital that is capable of treating the patient’s condition. During secondary triage, the vital signs of the patient, such as the heart rate, blood pressure, and respiration rate, are also assessed. Patients must be reassessed on a regular basis, every five to fifteen minutes, if transportation to a hospital is delayed. Secondary triage is a time consuming process for the paramedics. The deployed triage system has a significant impact on how the patient care resources are allocated. The system provides critical information to the pre-hospital, hospital, and alternate care sites where resources are constrained and most likely rationed. To resolve these challenges, AID-N has designed a decentralized electronic triage and sensing that contains low power embedded devices that efficiently monitors the physiological characteristics of the patients and track them through a decentralized, fault tolerant communication infrastructure. The main contributions of the AID-N electronic triage include: - An architecture that allows for more information on the vital signs and location of the patients to be obtained during initial triage.
- A system that reduces the workload of the responders during secondary triage by allowing patients’ vital status to be gathered wirelessly.
The overall goal of the AID-N electronic triage system is to more efficiently gather and distribute information on the vital signs and locations of patients in a manner that is extremely fault tolerant. This sets the AID-N electronic triage system apart from previous schemes that have used centralized or multi-tiered approaches with high power embedded devices resulting in shorter lifetimes than the AID-N electronic triage system [ 10]. The AID-N electronic triage system has been designed and implemented to meet the needs of the next generation of triage systems. Previous research has established standards for initial triage through the categorization of victims according to treatment urgency in the case of an explosive event or biological catastrophe [ 16][ 1][ 3]. The AID-N electronic triage system differs from these approaches because it uses an electronic triage tag instead of a paper triage tag. Previously, technology has been combined with triage through the use of barcodes, tag readers, passive RFID tags, hand-held computers, and geolocation to collect data about the mass casualty events [ 2] [ 7] [ 4] [ 9] [ 10]. Location tracking systems, such as [ 5], use active RF-ids tags in hospitals, but lack the embedded vital monitoring components of AID-N. The AID-N electronic triage system provides similar functionality as other electronic triage tags [ 10] [ 13] [ 14], but the AID-N electronic triage system is more robust due to its decentralized communication architecture and its ultra-low power embedded hardware. THE AID-N ELECTRONIC TRIAGE ARCHITECTURE The hardware in the AID-N electronic triage system provides a low power embedded system to meet the challenges of efficient triages. The system software running on the electronic triage tags is the CodeBlue system [ 12]. CodeBlue provides a robust mesh networking substrate for delivering real-time vital sign and triage status data from multiple electronic triage tags to end users. In addition, the AID-N electronic triage system also provides various sensors that detect the physiological conditions of the patient. Hardware Design: Electronic Triage Tags Our electronic triage tags replace the paper triage tags used by first responders during initial triage and address many of the limitations of the triage process described above. The electronic triage tag allows the medic to set the triage color (red/yellow/green/blue) of the patient at the push of a button. It replaces the paper triage tags that are commonly used by medics today. The button toggles the tag to shine the triage colors of: red (priority 1), yellow (priority 2), green (priority 3), and blue (priority 4). Four light emitting diodes (LEDs) represent the triage colors and a patient can only be on one level at a time. A small green LED blinks in sequence with the patient’s heartbeat and a white LED comes on if the patient is contaminated. If all the LEDs blink, then the patient should be transported to the hospital. The tag employs a lockout feature to prevent patients from triaging themselves. The electronic triage tag’s modes of operation, shown in Table 1, can be controlled directly on the device or remotely. The tag runs on a platform that has telemetry capabilities so its triage states are reported to a remote station. | Table 1Electronic Triage Tag Modes of Operation |
illustrate our current prototypes. The current prototype now only has three LEDs and one function button. Our prototype is a work in progress and does not represent the completed implementation. This mote prototype uses the MicaZ platform from Crossbow Technology [ 15]. It is powered by two AA batteries and consumes roughly 20 mA when active, resulting in a battery lifetime of 5–6 days of continuous operation. It uses a single-chip radio, with a maximum data rate of 100 kbps and practical indoor range of approximately 20–30 m. The mote is constructed to be inexpensive and light weight, and with Micro-Electro-Mechanical Systems (MEMS) manufacturing, we envision these motes to become one-time use, disposable devices [ 8]. Software and Sensors The software that runs on the triage tags allows the responders to monitor and control multiple triage tags simultaneously. The software and sensors are an extension of the CodeBlue project at Harvard University [ 12]. CodeBlue is a distributed wireless sensor network for sensing and transmitting vital signs and geolocation data. The communication vastly improves coverage and reliability with a virtually unlimited range. In order to efficiently measure the patient’s physiological conditions, various sensors are attached to the mote. A pulse oximeter will measure the amount of oxygen in the patient’s blood. There are several types of oximeter clips to read the oxygen level of a patient: a finger clip, a finger wrap, an ear clip, and a foot wrap for infants [ 12]. The BPMote has a sensor that detects and reports the blood pressure of the patient every fifteen minutes. In addition, the heart rate of the patient is detected every second through an EKG sensor [ 6]. The 2-lead EKG sensor is wireless and is attached to the patient via a chest pad. There are two types of location sensing capabilities – GPS (Global Positioning System) to provide geolocation, and the indoor location detection system to provide location where the GPS signal cannot be reached. The GPS sensor sends out a signal every five minutes to allow medics to track patients who are outdoors, e.g. at the scene of the emergency, with accuracy within three meters [ 17]. The indoor location system, based upon the MoteTrack project developed at Harvard University, requires the installation of location beacons [ 11]. This RF tracking system enables decentralized communication at a designated auxiliary care center. Patients are admitted to an auxiliary care center if nearby hospitals have reached their occupancy capacities and cannot admit more patients. At the auxiliary care center, our system monitors the patient’s vital signs. The ability to track the location of patients indoors will aid medics in quickly locating specific patients whose conditions have deteriorated. This continuous monitoring is critical in auxiliary centers because they will likely be short staffed and overflowing with patients. The base devices are small USB compatible devices that receive data from AID-N sensors. We envision the base stations to be plugged into every ambulance, every care facility, and every disaster scene. Hence, the medic can track the general locality of the patient based upon which base station the mote is communicating with. The prototype will also alert the responders if a particular patient is walking away from the scene without an official discharge. This is accomplished when the base station recognizes wandering patients when the signal from the mote loses strength. HUMAN COMPUTER INTERACTION AID-N electronic triage must be designed so that the responders can quickly deploy the devices and retrieve useful and accurate information. To better understand the emergency response process, several researchers conducted a field study by riding along in ambulances and administering an anonymous survey. Field Study In order to concretely analyze how the system should be designed, a field study in ambulance Emergency Medical Treatment (EMT) procedure was done. The field study was conducted to gather more information on the EMT procedure and to demonstrate the patient tracking technologies to firefighters. The firefighters included three captains, two platoon chiefs, and eight paramedics. First, a hypothetical deployment scenario during mass casualty incident with electronic triage tags was described to medics. During a mass casualty incident, a medic carries mote packages and distributes them to the patients. Each mote is pre-attached with a paper triage tag with the same tag ID as the mote. The paper tag acts as a backup if the embedded system fails. When the patient is first triaged, the medic straps the mote wristband on the patient, places the finger sensor on the patient’s finger, and sets the electronic triage tag to the patient’s triage category. The BPMote, EKGMote, and GPSMote are optional modules, and the medic may put them on select patients who need the additional level of monitoring. The mote automatically starts transmitting data to the medic’s tablet PC. The tablet PC is harnessed to the first responder in a weatherproof and anti-glare casing. Six firefighters filled out an anonymous survey. At the top of the anonymous surveys, the subject was asked to write down the number of years of experience they had as paramedics. The subjects had a total of 90 years of experience as paramedics. The subjects were asked to rank various AID-N features using a range between 1 and 7. 1 was the least important or least useful; while 7 was the most important or most useful. As can be seen in , the feature that received the highest marks was electronic patient triaging. Locating the patient was ranked very low, since paramedics monitor patients in a relatively small cordoned area that involves tens instead of hundreds or thousands of casualties. AID-N currently targets the later group. In order to quantitatively analyze if the deployment can realistically triage hundreds of patients in a reasonable amount of time, a mock drill with paramedics and firefighters will measure the time it takes to apply a paper triage tag and the AID-N electronic triage tag. Design Implications There were several important design implications gathered from our field study. The system should be designed to require little setup time and operate for several days. Therefore, AID-N uses ultra-low power embedded devices and only deploys the necessary sensors. The pulse-ox and location tracking sensors are the default sensors deployed with the triage tag. If the condition of the patient requires additional information, then the EKG and blood pressure sensors are attached. Even with all the sensors attached, the battery lifetime of the overall system is approximately six hours. The blood pressure sensor is the most power hungry peripheral. Therefore when it is not used, the battery life of the overall device increases by one to two days. Also, when treating patients using a necklace like sensor when the wrist could not be used raised concern due to the fragility of the neck with spinal cord injuries. Therefore, the connector was modified so that it can also be attached to a belt clip. Packaging of the sensor is essential for its use. It must be hermetically sealed in order to go through decontamination with the patient. The medics stressed the need to make the equipment sturdy enough to withstand rough usage. The medics weighted assistance with information flow as important as or more important than vital sign monitoring. AID-N also implemented the medic’s suggestion of scanning in the patient’s name, address, date of birth, and social security number off of the driver’s license barcode. The medics also suggested the hospital be notified wirelessly when supplies became low and starting the paperwork process for replenishing controlled substance. The easy information input makes the triage system a useful day-to-day tool that can eliminate duplicate paperwork where information only had to be entered once. The system’s interaction with its users should not be ignored during the design process. The buzzer is useful for medics who may need to quickly locate a patient in a sea of patients. However, the buzzing should be used with caution, as it could be a source of stress for the patient whose mote is buzzing. In a study of ICU patients, patients whose monitor started to buzz experienced elevated heart rates and stress levels. Hence, we provide an alternative non-intrusive mode of light emitting diodes (LEDs) blinking for the medic to locate the patients. These tags must be designed with consideration for colorblind medics. The LED colors are placed in order of priority with a priority number labeled next to it. Therefore, the medic has three modes for identifying the priority level: color, position, and label. An instruction card is on the back of the mote for medics who are called to duty, but are unfamiliar with the devices (). Also, misuse of the system should be prevented. Medics expressed concerns that once the patients figured out that the red light patients get preferential treatment, the patients may change their own colors on their tags. The password button is put in place to prevent patients from triaging themselves. When a new technology is introduced into the emergency response arena, it is important to note its limitations as well as its capabilities. Due to the chaotic nature of emergencies, our system faces the difficulty of operating in situations that challenge instrumentation designed for use in the controlled environment or clinical situations. Pulse oximeter readings have limited accuracy in the presence of methemoglobin, carboxyhemoglobin, nail polish, nail fungus, fluorescent light, and motion. In conclusion, we have presented a system that relieves responders from the burden of manually recording vital signs onto hardcopy pre-hospital care reports. AID-N electronic triage operates under very low power constraints and uses twenty times less energy than previous triage embedded systems [ 15][ 10]. Furthermore, AID-N electronic triage uses a decentralized location tracking system that is extremely resilient to node failures that are likely to occur at chaotic disaster sites [ 11]. A field study uncovered several important design implications that should be considered when constructing our system. The next step in AID-N electronic triage research is to deploy this system and verify its resilience in a real setting. Indoor location beacons are currently being installed at a designated auxiliary care center near Washington, DC. In this real life testbed, we will do additional testing of resilience under failure and also network load testing to determine how much data the system can efficiently route. AID-N plans to incorporate lightweight security into the AID-N electronic triage system into the hospital wireless network to meet the Health Insurance Portability and Accountability Act (HIPAA) requirement. HIPAA requires that medical data only be disclosed to the patient or medical personnel processing claims or using the data for services to improve the patient’s health. 1. Benson M, Koenig KL, Schultz CH. Disaster Triage: START, Then SAVE—A New Method of Dynamic Triage for Victims of a Catastrophic Earthquake. Prehospital and Disaster Medicine. 1996 Apr–Jun;11(2) 2. Bouman J, Schouwerwou R, Van der Eijk K, van Leusden A, Savelkoul T. Computerization of Patient Tracking and Tracing During Mass Casualty Incidents. Eur J Emerg Med. 2000;7(3):211–6. [PubMed] 3. Burkle F, Mass M. Casualty Management of a Large-Scale Bioterrorist Event: An Epidemiological Approach That Shapes Triage Decision. Emergency Med Clin N Am. 2002;20:409–36. 4. Chang P, Hsu Y, Tzeng Y, Sang Y, Hou I, Kao W. The Development of Intelligent, Triage-Based Mass-Gathering Emergency Medical Service PDA Support Systems. J Nurs Res. 2004;12(3):227. [PubMed] 5. Dempsey Mike. Analyzing the Return of Investment for Indoor Positioning Systems. Radianse Inc; White paper: 2004. 6. Fulford-Jones T, Wei G, Welsh M. A Portable, Low-Power, Wireless Two-Lead EKG System; Proceedings of the 26th IEEE EMBS Annual International Conference; San Francisco. September 2004. 7. Hamilton J. Automated MCI Patient Tracking: Managing Mass Casualty Chaos Via the Internet. Jems. 2003;28(4):52–6. [PubMed] 8. Hill J, Szewczyk R, Woo A, Hollar S, Culler D, Pister K. System Architecture Directions for Networked Sensors; Proceeding of 9th Int’l Conf. Architectural Support for Programming Languages and Operating Systems (ASPLOS 2000); ACM Press. 2000. pp. 93–104. 9. Lauraent C, Beaucourt L. Instant Electronic Patient Data Input During Emergency Response in Major Disaster Setting. Stud Health Technol Inform. 2005;111:290–3. [PubMed] 10. Lenert L, Palmar D, Chan T, Rao R. An Intelligent 802.11Triage Tag for Medical Response to Disasters. AMIA 2005 Symposium. 2005 11. Lorincz K, Welsh M. MoteTrack: A Robust, Decentralized Approach to RF-Based Location Tracking; Proceedings of Personal and Ubiquitous Computing, Special Issue on Location and Context-Awareness; Springer-Verlag. 2006. 12. Malan D, Fulford-Jones T, Welsh M, Moulton S. CodeBlue: An Ad Hoc Sensor Network Infrastructure for Emergency Medical Care; Proceedings of the MobiSys 2004 Workshop on Applications of Mobile Embedded Systems (WAMES 2004); Boston, MA. Jun, 2004. 13. McGrath S, Grigg E, Wendelken S, Blike G, De Rosa M, Fiske A, Gray R. ARTEMIS: A Vision for Remote Triage and Emergency Management Information Integration. Dartmouth University; Novermber. 2003. 17. Tollefsen W, Pepe M, Myung D, Gaynor M, Welsh M, Moulton S. iRevive, a Pre-hospital Mobile Database for Emergency Medical Services. International Journal of Healthcare Technology and Management (IJHTM). 2004
|
N ELECTRONIC TRIAGE ARCHITECTURE
