Section 2. Review of Literature
This section contains a summary of the available literature related to incident management, and performance measures for incident management systems. It should be mentioned that there is very little literature from the law enforcement/emergency service providers' perspective directly related to transportation-related incident management and performance measures. Most of the information presented here for the emergency services perspective was derived or inferred from a limited number of references.
One big issue that has to resolved before incident management performance measures can be developed is what, exactly, is an incident. Transportation providers and emergency responders tend to have different definitions for what constitutes an incident. This is primarily because of the different missions that transportation and emergency service providers have in many areas.
Even within the transportation literature, transportation agencies and officials tend to define incidents differently. The Traffic Incident Management Handbook  defines an incident as "any non-recurring event that causes a reduction of roadway capacity or an abnormal increase in demand." Under this definition, events such as traffic crashes, disabled vehicles, spilled cargo, highway maintenance and reconstruction projects, and special non-emergency events (e.g., ball games, concerts, or any other event that significantly affects roadway operations) are classified as an incident. The Traffic Management Data Dictionary (TMDD), as published by ITE and AASHTO, defines an incident as "an unplanned randomly occurring traffic event that adversely effects normal traffic operations."  Developers of the TMDD distinguish incident conditions from planned activities, such as roadwork or maintenance activities by defining different data elements and message sets for both incident and planned roadway events. The 2000 Highway Capacity Manual  defines an incident as being "any occurrence on a roadway that impedes normal traffic flow." While these definitions are very similar, they tend to suggest that within the transportation community, different officials tend to define incidents slightly differently. This can lead to confusion when dealing across jurisdictional boundaries and in reporting and interpreting incident management performance measures.
Emergency Services Perspective
While there are no real clear-cut definitions of an incident, most law enforcement agency and emergency responders seem to define an "incident" as any event to which they are dispatched or requires a "response" or action by them. Generally, law enforcement and emergency responders view their mission as "public safety" and "prevention of loss of life and property." Therefore, these agencies are driven to respond to events that might be perceived as having an impact on the public safety or the potential of loss of life. Major events, such as vehicle collisions, overturned vehicles, vehicle fire, would all be classified as an incident by both law enforcement and emergency responders because the nature of these events generally requires them to respond. Less critical events, such as stalled vehicles on the shoulder, debris in the roadway, etc., may not be considered an "incident" in many locations because an action or response would not be required from a law enforcement and emergency response perspective. For example, fire departments generally do not classify stalled vehicles or debris in the roadway as an "incident" because they do not generally respond to those types of events. Again, this varies from location to location.
It should also be noted that the definition of an incident by law enforcement and emergency responders includes more than just events effecting traffic. Potential suicides, structure fires, criminal activities, and other events off the roadway are considered to be "incidents" by law enforcement and emergency responders because these events require a response from these agencies.
The definition of an incident also appears to be highly dependent upon the type of dispatching arrangements and structure of the emergency response agencies in an area. For example, in Dallas, the fire and police departments use a common 911 dispatching center. If a call comes into the dispatching center requesting both a fire and police response, both are dispatched to the scene, even though there may not be a true need for both responses. The fire unit arriving on the scene then makes the determination if their presence is truly needed. Because they have been asked to respond to the scene, the fire department would generally classify this as an incident because their equipment is in a response mode and is unavailable to respond to another event.
Because law enforcement vehicles can patrol sections of roadways, they may occasionally "happen" upon an incident scene (such as a stalled vehicle in a travel lane) and "respond" to that event without being dispatched. The decision as to whether or not classify this type of event as an incident seems to depend upon whether or not the event is a public safety concern requiring a response. For example, a stalled vehicle blocking a lane of traffic is generally viewed as a public safety issue because of the potential of the vehicle causing a secondary crash, and would generally be classified as an incident. Some law enforcement agencies may not necessarily classify a stalled vehicle on the shoulder as an "incident" requiring their response because it may not be viewed as mission critical and may not necessarily represent a public safety concern.
From a transportation perspective, incidents tend to be classified based upon their impact on traffic operations. Many transportation agencies have devised ranking systems for classifying incidents to assist in determining the appropriate level of responses. For example, the Chattanooga Urban Area Metropolitan Planning Organization and the Chattanooga-Hamilton County Regional Planning Agency have devised a classification system that is based on traffic flow, impact/delay, incident characteristics and types of responders.  A Level 4 incident is one that typically is causing traffic delays of less than 30 minutes where traffic is only slightly impacted and can be relatively easily routed around the incident. A Level 3 incident is one lasting more than 30 minutes but less than an hour, and a moderate impact on traffic flow. Typically a Level 3 incident involves a collision without or just minor injuries. A Level 2 incident is one lasting more than 30 minutes, but less than 2 hours. In a Level 2 incident, the impacts on the flow of traffic are significant, and the incident probably involves injuries to motorists. With a Level 2 incident, traffic management is essential and site management involves significant interagency cooperation. A Level 1 incident generally tends to be major events that close the roadway and cause major area-wide congestion.
Many other areas use similar classification systems to help agencies define the appropriate level of response in the region.
Emergency Services Perspective
While most transportation agencies tend to classify incidents based upon their impact on traffic operations, law enforcement and emergency response agencies tend to classify an incident on the number and severity of potential injuries and the number of apparatus required to affect an adequate response. Radio dispatching codes were used to gain insight into the way that different law enforcement and emergency providers classify incidents (see Appendix A for example of select radio codes). For the most part, because their level of responsibility varies from investigating potential criminal activities to maintaining law and order, law enforcement agencies generally tend to have more categories for classifying incidents than fire and emergency service responders.
Appendix A contains the model dispatching codes developed by the Association of Public-Safety Communications Officers (APCO).  Of the approximately 100 dispatch codes, 14 are related to transportation events. Ten of the 14 are used to describe different incident-related type of responses. Most police agencies use fewer numbers of dispatching codes that are used to describe or classify different incident situations.
Fire and emergency medical services generally use criteria that alert them to the number and type of apparatus that are going to be dispatched and the potential for loss of life. Dispatching codes for the New York City Fire Department are also shown in Appendix A. Relatively few dispatching codes (a total of 4) are used to describe traffic incidents.
Many transportation agencies do periodical assessments of their incident management systems. The Traffic Incident Management Handbook  reports that the most commonly used statistics in evaluating incident management programs include the following:
- The number of service patrol assists;
- The average elapsed time from incident occurrence to detection;
- The average elapsed time from the point at which the incident response team is called out until its arrival on-scene; and
- The average elapsed time to normal traffic flow restoration.
In May 2000, State Highway Administration of Maryland and the University of Maryland produced Performance Evaluation of CHART "An Incident Management Program" in 1997.  The purpose of the evaluation was to "assess the effectiveness of the Maryland CHART program with an emphasis on its ability to detect and respond to incidents on major freeways and highways" and to assess "the efficiency of the entire incident management operations along with its resulting benefits." The evaluation examined issues such as detection time, response travel time, clearance time, response time, and incident duration. The operational definitions used in the evaluation included the following:
- Detection Time – the elapse time between when an incident occurs to when it is detected
- Preparation Time – the elapse time between when an incident is detected to when the response vehicles are dispatched.
- Response Travel Time – the elapse time between when the response vehicle was dispatched and when response vehicles arrive at the incident scene.
- Clearance Time – the elapse time between when response vehicles arrive at the incident scene to when traffic completely recovers after the incident.
- Response Time – the elapse time between when an incident is detected to when the response vehicles arrive at the scene.
- Incident Duration – the elapse time between when an incident occurred to when the response vehicles depart at the scene.
The report went on to present an analysis of incident characteristics. The researchers used 12 months of incident reports from all three of the traffic operations centers and accident report data from state police for completing this analysis. The researchers use these records to examine the distribution of incidents by the following:
- Blockage duration;
- Peak and off-peak hours;
- Weekday and weekend;
- Lane blockage; and
- Location (exit ramp numbers).
The researchers indicate that this information can be used to better design incident management strategies, such as the distribution of patrol vehicles around freeway segments of a high incident frequency; assessing the impact of areas under the average and the worst incident scenarios, and identifying hazardous highway segments from both the safety and operations perspectives.
Using the incident data, the researchers also evaluated the effectiveness and efficiency of their incident detection system. The researchers used two primary measures of effectiveness in this evaluation:
- Incident response rate
- Distribution of detection sources
For this evaluation, the researchers defined the incident response rate to be "the ratio between the total number of traffic incidents reported to the CHART control center and those managed by the CHART incident response team." Not surprisingly, the researchers reported response rates at the three TMCs to be 99%, 94.7%, and 92.3%. The researcher noted that no reasons were given in those incidents when the incident team did not respond. The researchers recommended that CHART operators "should clearly document such incident scenarios, and detail the reasons for those incidents to be handled by police alone." In discussions with the CHART operators, the researchers found that in some of those incidents, the response team was unable to respond because of "equipment limitations or manpower shortage."
The researchers also conducted an analysis of incident response efficiency specifically addressing the following:
- The time it took for an incident response unit to reach the reported incident site after the control center was informed
- The average travel distance for incident response units to reach the identified incident site.
- The approximate reduction in the incident blockage time due to the operations of CHART's incident response program.
As noted above, the researchers defined response time as the "elapsed duration from the moment the control center received a reported incident to the physical presence of the incident management team at the target incident site."
In looking at the reduction in incident duration, the researchers noted that there are two ways of doing this. The first way is to perform a "before and after" comparison where response times to incidents before and after the system is operational. The researchers rightfully noted that in most locations, incident response time data prior to actual operations of a center is sparse, at best. They suggested that another way to examine the reduction in incident duration is to compare incident durations when the incident management team responded to incident durations when the incident management team did not respond. One drawback to this, however, is that data from when no response occurred may be limited in many centers.
The report included information estimating benefits of incident management system. The researchers indicated that "despite well perceived benefits from an efficient incident management system, most state highway agencies, including MSHA, are facing the pressing need to justify their system investment and operating costs, especially in view of diminishing resources and increasing demand for infrastructure renovation." The researchers indicated "to ensure the quality of analysis under the data limitations as well as resource constraints, the benefit assessment of CHART was focused only on those [measures] either directly measurable or quantifiable from the given data." Therefore, the researcher focused on the following performance measures:
- The number of assistance request from drivers;
- The reduction in secondary incidents;
- The reduction in driver delay time;
- The reduction in vehicle operating hours;
- The reduction in fuel consumption; and
- The reduction in vehicle emissions.
In their analysis, the researchers defined assistance requests as an event where the driver asked for assistance such as flat tire, shortage of gas, or some mechanical problem. The researchers noted that "according to CHART staff, its response teams actually responded to many more assistance requests from drivers" than was used in the analysis, but because "most of the unreported driver assistance [requests] did not need major efforts or equipment from the response unit," no data were recorded on these events. This suggests several issues that must be addressed in assessing the performance of incident management systems:
- It is important to define the measures that are going to be used to evaluate the performance of the system PRIOR to analysis period so you know what data to collect.
- It is important to have the mechanisms in place to ensure that all the data that will be used to evaluate your system is collected.
The researchers also used the reduction in the number of secondary incidents in their assessment of the benefits of the CHART system. For the purposes of their evaluation, the researchers defined "secondary incidents" to be "any incidents occurring within two hours after a major incident and within a two mile range of a reported incident." In looking at the "reduction in secondary incidents," the researchers estimated the number of secondary incident without CHART by factoring up the number of observed number of incidents by the percent reduction in average incident duration. The researchers used simulation to quantify the reductions in driver delay, fuel consumption, and vehicle emissions.
While some agencies undertake performance assessments similar to that performed for the CHART system (i.e. a before-and-after comparison performed by an outside agency), other agencies produce performance reports on a more routine basis. For example, the Minnesota Department of Transportation (MnDOT) routinely produces performance reports that summarize the performance of their incident management system on a daily, monthly, or annual basis.  Samples of these reports are contained in Appendix B. These reports generally contain information on the following:
- The number and type of incident occurring;
- The number and type of vehicle involved;
- The number of times different agencies responded to incidents; and
- The average response times by each responding agency.
- The number of assists performed annually, quarterly, or per month,
- The types of assists encountered,
- The types of services rendered,
- The time of the assists (e.g., Morning, Afternoon, Evening)
- The average duration of assists.
Sample reports from the motorist assistance program in Houston, TX are shown in Appendix C.
Emergency Services Perspective
In many respects, emergency service providers are much more cognizant of the benefits of performance measures. Many emergency service providers routinely monitor and produce reports that show their average response times. Historically, emergency service providers have used response times for justifying adding new equipment and staffing, and for strategic planning purposes (such as determining when new fire stations need to be added and where, etc.).
For example, the City of Austin Fire Department has a web site in which they report their average response times for each month.  Response times are summarized separately based on calls that come into the fire department dispatch and calls that go into the 911 dispatch center. In producing these reports, the City defines response time as the time "from the moment a call is received by the Fire Department Dispatch [or the 911 center] to the moment when an engine or truck company arrives on the scene." The definition of response time used by the Fire Department seems to the representative of most emergency response systems.
Many locales use their freeway management system software as the primary means of collecting and storing information about incidents on the freeway networks. Through various input screens, information about incidents is entered either by the operator or, at some locations, automatically by the system itself. The general type of information logged by most systems included the following:
- The roadway on which the incident occurred;
- The location (cross-street, mile point, or incident reference system) of the incident;
- The number of vehicles involved;
- The severity of the incident (stalled vehicle, property-damage only, possible injuries, etc.);
- The source reporting the incident;
- The number of lanes blocked; and
- The potential duration of the blockage.
Figure 1 shows an example of two incident management data input screens employed in Texas.
Another source of incident information is motorist assistance or service patrol logs. These logs are kept either by the responding officer in the field or by the dispatcher located in the control center. These logs generally contain the same information as the incident management software system, but are collected by the response individual. In most locations, service patrols are responsible for responding to minor incidents (such as stalled vehicles); therefore, the patrol logs are used more to keep track of what resources (such as fuel, etc.) are used in a response rather than as a mechanism for measuring performance such as response times, and response durations. Figure 2, which shows the type of information logged in a service patrol in Ohio, serves as a typical example of the type of information collected by most service patrol systems.
Emergency Services Perspective
Many law enforcement and emergency service providers (either through their combined E911 dispatching centers or through their own dispatching centers) use Computer-Aided Dispatching (CAD) systems. According to Dispatch Monthly Magazine , 56% of local police departments with their own communication center and 70% of the sheriff departments with their own communications center use CAD to assist them in their dispatching. The numbers grow considerably when 911 and E911 dispatching centers are also incorporated. CAD systems were originally intended to speed-up the process of dispatching roving patrol officers to a scene; thus, reducing response time. However, these systems generally have the capability for logging and storing large quantities of data that can be used to develop response performance measures.
There are literally hundreds of different types of CAD software systems available on the market, but they generally log similar types of information about responses — most notably, the time that a request for assistance (or call) was entered in the dispatching system, the time response was dispatched, the time the response arrived on the scene, and the time the response vehicle "cleared" the call (or was available to receive another call). Some CAD systems have been integrated with automatic vehicle locating systems so that the location of vehicles is constantly monitored and event times such as vehicle arrival times and vehicle clear times are logged automatically by the CAD system. Figure 3 shows a screen capture of one version of a CAD system and illustrates the type of information that is captured in most CAD systems.
The U.S. Fire Administration (USFA), part of the Federal Emergency Management Agency (FEMA), maintains a National Fire Data Center (NFDC) that collects, analyzes, and publishes statistical information about fires and fire responses. To gather this information, the NFDC established the National Fire Incident Reporting System (NFIRS).  Participating local fire departments fill out an Incident and Casualty Report as the fires occur. They then forward the completed forms to their state office where the data are validated and consolidated into a single database. A blank Incident and Casualty Report form is shown in Figure 4.
One function of the Incident and Casualty Report is to serve as a model for the type of records that fire departments around the country should keep.  The type of data collected for each fire response includes the following:
- The day, date, and time of each fire event,
- The type of situation found when the responders arrived on the scene,
- The type of actions taken upon arrival (i.e., extinguished fire, provided first aid, etc.)
- The type of property involved (including automobiles),
- The source or cause of the fire,
- Information about the property (address, owner, etc.), and
- Information about the type of response provided (i.e., number and type of responders).
Several fields on this form illustrate the type of data that many fire and emergency medical service providers routinely collect. These fields are the Alarm Time, the Arrival Time, and the Time in Service. Each of these data entry fields are described as follows:
- Alarm Time — This is the exact time of day (hour and minute) when an alarm is received by a fire department alarm center. It is important for three reasons: (1) as a legal requirement for recording the precise time of an incident, (2) as information for determining the frequency of particular types of incidents by time period, and (3) as the starting time for going into action on an incident, which can be compared with Arrival Time to determine the length of time necessary to arrive at an incident [transportation agencies typically think of this as "Response Time"] and Time In Service to determine the total amount of time spent at the incident.
- Arrival Time — This is the actual clock time when the first responding units arrive at the incident scene. This time is valuable to department management because it reflects the actual time spent in traveling to the scene of the incident. It is useful in determining the actual time spent at an incident and would indicate any delay between alarm and arrival.
- Time In Service — Although each fire department generally has their own operational definition for "time in service," it is usually defined as the time when all or most of the equipment is again ready for response to another alarm, as determined by the officer in charge at the scene. This entry is generally in 24-hour clock time and is necessary along with Arrival Time for calculating the total time spend on an incident.
Several law enforcement agencies (Kansas, and Houston HPD) that participated in the survey indicated that their primary means of collecting information about an incident was the standard accident investigation form. A sample accident investigation form used in Kansas is shown in Figure 5. Generally, these forms have fields where officers can fill-in when the accident occurred, when they were notified, and when they arrived on the scene (see upper right-hand quadrant of the form). Notice, however, there is not a field to indicate when the officer left the scene.
Many agencies have developed Incident Management Response Manuals. [7, 19] These manuals define the roles and responsibilities of agencies when responding to incidents, outline the general procedures to follow when responding to and clearing incidents, and identify the available resources and capabilities of each agency. These manuals are generally developed using input from both transportation agencies and emergency response providers. Some of the special items included in many of these manuals include the following:
- Goals and objectives of the incident management program,
- A listing of the agencies involved in incident management in an area,
- General procedures for responding to incidents
- Procedures for responding to incidents, including
- Traffic control requirements,
- Detour routes,
- Use of emergency lights by response vehicles,
- Parking of emergency vehicles at the scene,
- Staging of incident responses,
- Establishment of command posts,
- Procedures for removing disabled vehicles,
- Procedures for handling hazardous materials,
- Procedures for investigating fatalities and felony incidents,
- Procedures for notifying the public about incidents,
- Use of video surveillance cameras,
- Listing of contacts within response agencies,
- Listing of available equipment and resources within each response agency.
Emergency Services Perspective
The U.S. Fire Administration (USFA) has published a Guide To Developing Effective Standard Operating Procedures for Fire and EMS Departments.  The guide is designed to "assist emergency service managers in establishing effective standard operating procedures (SOPs)" that "clearly spell out what is expected and required of personnel during emergency response and non-emergency activities." This guide specifically states that the standard operating procedures should not tell firefighters how to do their jobs (i.e., technical skills) but describe a department's rules for doing a job (i.e., procedural guidelines). It suggests that one important item that should be included in an agency's SOP is how responders should operate on the roadway. While the guide does not provide any specific recommendations on how to do it, it does recommend to fire departments that the SOP cover such items as the following:
- Operations near moving traffic,
- Traffic control procedures,
- Use of warning devices,
- Vehicle/scene stabilization,
- Coordination with law enforcement personnel,
- Standard procedures and precautions, and
- Special situations (e.g., downed power lines)
USFA has also produced a Hazardous Materials Guide for First Responders , which provides a generalized approached for handling hazardous material spills and incidents. The guide gives first responders information about how to approach a potential hazardous material spill, what to look for, where to set up command posts, where to park vehicles, etc. It also provides information on regulatory considerations, training, and operations in and around hazardous material spills.
Neither of these guides contain information on what performance measures fire and emergency response system should be computing or how.Previous | Next