The folks of Haleyville, Alabama (population 4,000) understood in 1968 how a communications system could help during an emergency. For on February 16 of that year, the first-ever 911 phone call was placed by Alabama Speaker of the House Rankin Fite from Haleyville's City Hall to U.S. Rep. Tom Bevill at the city's police station.1
Since that day, emergency responses have become inextricably linked to the nation's telecommunication systems operationally and to the public sector organizationally. Indeed, in the almost four decades since the Haleyville call, the nation's emergency response system has evolved into a sophisticated network of dispatchers and responders and the communications system has evolved into an array of mobile and computer-aided networks. This evolution has occurred within the context of public demands for better and faster emergency medical services (EMS) to minimize the consequences of accidents and other health emergencies.
This article explores the general concept of "time-critical information services," within the specific context of EMS. While private sector-oriented information systems have focused on the critical role of information technology in achieving Just-in-Time (JIT) delivery and improved supply-chain management, our thesis is that similar attention is needed to those public sector services that are also highly time- and information-dependent. EMS represents an illustrative application domain of JIT in public services, where the adage "time is money" translates into "time is lives."
Let's begin with a fundamental question: Why is EMS a time-critical information service? In terms of time, the reason is clear: from the moment a 911 call is placed to a local public safety answering point (PSAP) and answered by an emergency operator, time is of the essenceevery moment of delay can significantly reduce an accident victim's chances of survival. For example, the proportion of fatalities from car accidents in rural areas are far greater in number than in urban areas largely due to the additional time required for resources to respond to remote locations. In rural areas the average time that lapses between a crash and the victim's arrival at a hospital is 52 minutes, compared to 34 minutes for an urban crash victim.2 The result is that fatalities on rural roadways account for approximately 60% of all vehicle fatalities in the U.S., with even higher figures in predominately rural states (for example, Maine, 90%; Minnesota, 73%; Montana, 92%).
But the service is not just time-dependent, it is information-dependent as well: this information can include a range of data on the location of a caller, the nature of the call, health condition of an injured person, or health data on biological constitution (for example, blood type) of the injured party. It can also include information used by various organizations to guide service operation and performance, such as what services have been rendered to a victim. The challenge for information and computer sciences is to devise new approaches and systems that facilitate rapid use of accurate information for time-critical use in EMS and related public services.
Over the last three years, our research team has investigated the performance of rural mobile EMS systems. This research has been conducted in multiple phases, encompassing field visits, interviews, focus groups, and performance data analysis and simulation. More recently, we have initiated a multi-tiered study (sponsored by the National Science Foundation) to refine our conceptual model and conduct empirical case studies. This analysis has led us to identify several features we think are important to time-critical information services. We expound upon these concepts here, as well as provide a summary illustration in Figure 1.
Time efficiency has been a primary objective for computer and information professionals for decades. This focus has created concepts such as JIT and business process reengineering (BPR), which have become central to private sector business operations and information technology planning. However, the issue of time is less well understood in the e-government domain, even though it can be a critical performance feature of public service. For many services, time matters in terms of days and weeks, for example, time savings possible from online tax submission. Or, in the case of determining disability or eligibility benefits, time is measured in months and years. In the case of EMS, time is measured in minutes and seconds. In short, time matters, even for government services.
As more information becomes digitized, a greater degree of precision in locating and responding to emergencies is possible.
As illustrated by the top row of Figure 1, a point of departure for examining EMS is the linear sequence of events. In our research, we have looked closely at the time-dependent course of a 911 call, tracking the "handoff" from one organization to another to understand the performance of the system . The EMS service typically begins with a consumer action (placing the emergency call), involves the private sector (telecommunications service provider) connecting and delivering the call, the public sector (PSAP) receiving and dispatching the call, the private and/or public sector (ambulance service, fire, law enforcement) providing first response, transport and health care services, and finally, either a public, private, or not-for-profit sector hospital or trauma center delivering appropriate health care services.
New technologies provide technical platforms to enhance and alter this sequential process. For example, the rise of mobile telecommunications has become an important means to deliver faster emergency service. At mid-year 2006 there were more than 219 million wireless users making about 240,000 911 calls a day across the U.S.3 A wireless 911 phone call can shave valuable minutes from the time otherwise required for a caller (or motorist aid) to find a conventional phone. However, there are still multiple technical issues to resolve. For example, location information is still primarily transmitted by voice despite the large push toward advanced E-911 systems to provide geospatial data along with a mobile 911 call. As will be noted in the Virginia example here, the opportunity now exists to integrate location and other data in a manner that alters the traditional sequential handoffs by allowing the data to be available to all parties in the emergency response chain at any time in the process.
Related to the timeliness of information is the quality of the information. Even in the early days of the 911 system, it was important to get accurate information about the emergency. Conveying voice information on the location and nature of a distress call proved valuable to response agencies to help prepare for the unique circumstances of each individual incident. As more information becomes digitized, a greater degree of precision in locating and responding to emergencies is possible. For example, GPS-enabled vehicles can provide dispatchers with up-to-the-second visual data about their location. However, Pettersson and colleagues  note how such emergency information needs to be visible "at a glance" and in a manner that facilitates collaborative problem solving by dispatchers amidst an array of available data. In this sense, information quality is about having the right information at the right time in the right format.
One of the key problems of governmental services is that information typically travels serially and sequentially, from one processing unit to the next, often with time-consuming feedback loops when incomplete or inaccurate information is detected. New technologies, such as computer-aided dispatch, can establish technical interorganizational linkages, but do not ensure a comprehensive and reliable relationship between organizations during dynamic emergency response situations . For this, a socio-technical approach is needed that integrates an organizational perspective into technical solutions. For EMS, this perspective needs to emphasize the linkages across organizations.
A time-critical information service is typically inter-organizational in naturethere is a handoff from one agency/organization to another. For private sector businesses, JIT, BPR, and related concepts have for many years been extended beyond the boundaries of a single organization. Information systems that extend across organizational boundaries have become both a complex technical undertaking as well as a challenging interorganizational phenomenon as it includes a network of information systems and organizations, across supply chains, to improve business processes with partners, vendors, suppliers, distributors, and sales channels. Over two decades of experience has demonstrated that such complex undertakings necessitate a great deal of organizational understanding and change .
For emergency response, new technologies can be developed and implemented to enhance organizational and interorganizational cooperation, but this does not negate the need to understand unique and varying characteristics of response organizations. Our research has examined, for example, interorganizational handoffs between state patrol agencies, departments of transportation, fire agencies, emergency medical service providers, and health care facilities . This examination revealed how cultural differences between agencies can affect cooperation. For example, we found state patrol agencies to be more service-oriented and more hierarchical with their information flows than transportation departments, and this difference required ongoing attention to ensure information sharing.
A key finding at our recent expert workshop was the need to focus on the multilevel linkages between cooperating EMS organizations (see Figure 1, second row from the top). More specifically, the workshop participants noted how inter-linkages occur at the operational, organizational, and governance levels.4 In a subsequent case study, a representative from a northern California EMS agency told us about their use of cross-organizational management controls to monitor the performance (response times) of the subcontracted ambulance service provider. This agency's review of performance data revealed that the subcontracted ambulance service had reduced staffing levels, which had resulted in longer response times than were contractually allowed. The situation was corrected and response time performance returned to normal. In this case, cross-organizational information sharing of performance data affected management and governance decision making, which directly impacted the timeliness of emergency services.
For time-critical information services, total performance is essential to the operation (see Figure 1, third row from the top). It makes little difference for an operator to dispatch quickly if the ambulance takes a very long time to arrive, goes to the wrong location, or if a receiving hospital has too few physicians on duty to see an incoming patient. The critical descriptor here is "end-to-end." Measuring performance across service processes and organizations is essential to understanding how public services are delivered to the public, the level of service (timeliness, quality) with which they are delivered, and how the service network can be improved to deliver better services under time- and information-critical circumstances.
The challenge is how to implement this "end-to-end" concept within and across emergency provider organizations. We are taking a step in this direction by creating an ontology and knowledge base to collect performance information about each step in the end-to-end process. In our early case study research in a rural region of Minnesota, we were able to collect incident data from multiple organizations and disparate information systems to account for a year (2002) of performance. We began with an estimated 21,745 mobile 911 phone calls, producing 7,215 EMS responses to 3,325 automobile crashes involving 71 deaths. We found the average ambulance response interval to be 60.5 minutes, (from EMS notification, to arrival at the scene, to arrival to a hospital), which is significantly above the state average for both urban and rural areas .
Our work in this area has made clear that a need exists for public agencies to first agree on how to measure performance across agencies and then build performance tracking across new or existing public agency information systems. While public sector information systems are often implemented to address separate silos of a governmental process, the end-to-end nature of time-critical information services facilitates or at least allows for information systems that can report on overall system performance. The real challenge then becomes working across very different organizational cultures (for example, departments of transportation, law enforcement, fire, and ambulance) to achieve a holistic understanding and use of performance information.
It is said that a crisis can define character; in EMS, a crisis can define performance.
It is said that a crisis can define character; in EMS, a crisis can define performance. End-to-end performance is not only a function of system processes but also a function of exogenous occurrences such as storms, natural disasters, or terrorist attacks. Research focused solely on a system's normal behavior cannot fully characterize the full range of possible system dynamics. Extreme events provide unique opportunities to examine systems, including how cooperative organizations function when exogenous variables cause a range of systemwide "stress" issues, including increased service demand that can lead to system overload or even collapse.
That is, extreme events can be a pivotal test of overall system management capability. As Perrow proclaimed "...it is no overstatement to suggest that humanity's future will be shaped by its capacity to anticipate, prepare for, respond to, and, when possible, even prevent extreme events ." For this reason, the National Critical Infrastructure Report devotes considerable attention to the use of modeling and simulation of crisis events to learn about system performance and response . In a similar vein, we are using process simulation software to examine end-to-end performance of emergency response systems, including naturally occurring spikes in emergency service requests (such as during holiday periods) . By examining these naturally occurring spikes, a more comprehensive understanding of the variable performance of emergency response systems could be achieved; including the role of new, innovative information systems during normal and peak conditions (see Figure 1, bottom row).
One example of a next-generation system can be found in the Blue Ridge Mountains of Virginia. Some 38 years after the call from Haleyville, a next-generation data call was made during a test of the Virginia E-Safety Network. The E-Safety Network is an emergency Web services information architecture that facilitates linkages between the various information systems used by law enforcement, dispatch (PSAPs), ambulance, fire, transportation, health care systems, and emergency management. It is based on open, non-proprietary standards in order to facilitate interoperability between disparate systems and agencies and to encourage the development of new applications by private companies competing for a national market .
Located in the greater Winchester, VA area, this system was designed based on a distributed computing (service-oriented) architecture utilizing XML-based standards and protocols. It was created to receive and distribute various emergency incident-related XML messages via standard interfaces. The intention from the outset was to support several time-critical functions as illustrated in Figure 2. For example, a series of notification services would selectively "push" new incident alerts to pre-identified and pre-authorized recipients on the basis of geospatial and other message content through an Intelligent Message Broker (IMB). The IMB integrates message alerts with geographic information system (GIS) data about an incident and sends them to registered users and associated applications. The IMB tracks the emergency response and facilitates real-time information exchange between response organizations. Authorized users could also "pull" data to their systems or become data sources to provide critical information to the network. A system that has for many years been completely reliant upon serially transmitting emergency information using voice over two-way radio has now been designed with the capability to transfer data to and from multiple organizations dynamically.
The E-Safety design not only enables timely transmission of key information, it also supports interorganizational coordination and information sharing. This can be organized through a shared registry of authorized emergency response and associated agencies and organizations called the Emergency Provider Access Directory (EPAD). The purpose of EPAD is to facilitate registration with the network as well as data access levels for each registered and authorized organization. Of course, the development and use of this EPAD directory hinges on interorganizational participation. In this regard, the E-Safety network has been expanded over the past few years to include the 13 hospitals and six counties that make up the Northern Virginia Hospital Alliance.
Brown and Hagel  have advanced the case for Web services and service-oriented architectures to produce new levels of supply chain performance. Similar to the literature on BPR and JIT, discussions of Web services tend to focus on private sector applications. But the Virginia E-Safety network suggests that such architectures can be of great use in developing time-critical information services for public agencies and services. There have been some recent efforts to develop XML-based standards for the collection, sharing, and reporting of EMS performance data, including a national coalition that has recently tested new data structures in various locales . While these technical efforts are moving forward, there is a parallel need to derive methods and analytic devices that are geared toward developing interorganizational collaboration. In this sense, there is a need to devise "org-ware" as well as software.
Fortunately, there is a burgeoning research effort that promises to shed much needed light on the value and functioning of time-critical information services. Beyond our own modest efforts, Turoff, Van de Walle and colleagues  have led a series of workshops and literature on the emergency and crisis response topic (see the special section in this issue). Funding agencies such as the National Science Foundation and the Department of Homeland Security have begun to sponsor integrated testing and assessments of emergency response systems. These recent technology-focused efforts benefit from a bevy of prior emergency response studies, a portion of which has been noted in our time-critical concept explication. Moreover, there are a variety of methods that have been applied to better understand dimensions of time-critical services, ranging from test beds  to usability testing  to field ethnographies . These add to the qualitative and case study-based research that has formed a strong foundation for studying interorganizational cooperation in EMS, such as we have undertaken.
In conclusion, the post-September 11 and Hurricane Katrina environment is one in which practitioner and research communities have become more concerned and involved with emergency and crisis management. Due to advances in XML and related open systems, the technical basis now exists to develop time-critical interorganizational information services for the public. The EMS application domain provides an emblematic case where such implementation has the possibility to achieve demonstrable public good in terms of lives saved. Despite the challenges ahead, this is an opportunity for computer and information professionals to innovate through new systems, applications, and research. The factors raised in our framework are but a starting point.
2. Cai, G., Bolelli, L., MacEachren, A.M., Sharma, R., Fuhrmann, S., and McNeese, M. GeoCollaborative crisis management: Using maps to mediate EOC-mobile team collaboration. In Proceedings of the 5th Annual NSF Digital Government Conference (Los Angeles, CA, May 2326, 2004).
4. Horan, T., McCabe, D., Burkhard, R. and Schooley, B. Performance Information Systems for Emergency Response: Field Examination and Simulation of End-To-End Rural Response Systems. J. Homeland Security and Emergency Management 2, 1 (2005).
7. Midkiff, S.F. and Bostian, C.W. Rapidly deployable broadband wireless networks for disaster and emergency response. In Proceedings of the First IEEE Workshop on Disaster Recover Networks, (June 24, 2002), New York, NY; accessed June 27, 2005; www.cwt.vt.edu/research/detail/disaster_response/Midkiff_Bostian_DIREN02.pdf
8. Perrow, C. Extreme events: A framework for organizing, integrating and ensuring the public value of research. Paper prepared for Extreme Events: Developing a Research Agenda for the 21st Century (Boulder, CO., July 79, 2000).
11. Turoff, M., Chumer, M., Van de Walle, B., Yao, X. The design of a dynamic emergency response management information system. Journal of Information Technology Theory and Application 5, 4 (2004), 136.
2National Center for Statistics and Analysis (NCSA), U.S. Department of Transportation. Fatality Analysis Reporting System (FARS) Web-Based Encyclopedia. Accessed Oct. 24, 2006; www-fars.nhtsa.dot.gov.
This work is supported in part by the National Science Foundation (Grant Nos. 508938 and 535273). The authors also gratefully acknowledge financial support from the U.S. Department of Transportation through the ITS Institute and State and Local Policy Program, University of Minnesota, as well as research support from Richard Burkhard, Ugur Kaplancali, Michael Marich, Denise McCabe, and David Aylward.
©2007 ACM 0001-0782/07/0300 $5.00
Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. To copy otherwise, to republish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee.
The Digital Library is published by the Association for Computing Machinery. Copyright © 2007 ACM, Inc.
No entries found