incident management performance measures - I

incident management performance measures - I
INCIDENT MANAGEMENT PERFORMANCE MEASURES
By
Kevin N. Balke, Ph.D., P.E.
TransLink® Research Center Director
David W. Fenno, P.E.
Assistant Research Engineer
and
Brooke Ullman
Associate Transportation Researcher
Sponsored by: Federal Highway Administration
U.S. Department of Transportation
Contract Number: DTFH61-01-C-000182
November 2002
TEXAS TRANSPORTATION INSTITUTE
The Texas A&M University System
College Station, TX 77843-3135
DISCLAIMER
The contents of this report reflect the views of the authors, who are responsible for the facts and
the accuracy of the data presented herein. The contents do not necessarily reflect the official
view or policies of the Federal Highway Administration (FHWA). This report does not
constitute a standard, specification, or regulation. The engineer in charge was Kevin N. Balke,
Ph.D, P.E., (Texas, #66529).
iii
ACKNOWLEDGMENTS
The project team would like to acknowledge the following individuals for contributing
information and their time to this project:
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
Sgt. A.W. Bowdoin, Sheriff Office, Harris County, Texas
Jeff Galas, Traffic Systems Center Manager, Illinois Department of Transportation
Matt Volz, ITS Coordinator, Kansas Department of Transportation
Farouk Aboukar, District ITS Engineer, Ohio Department of Transportation
Ron Perry, Research Director, Fire Department, Phoenix, Arizona
Rod Mead, Operations Manager, Colorado Department of Transportation
Dan Howard, Senior Operations Manager, New York Department of Transportation
J. Thomas Bruff, Engineering Coordinator, Southeast Michigan Council of Governments
Jim Hogan, Traffic Operations Manager, New Jersey Department of Transportation
Juanita Lowe, Public Relations Coordinator, Kansas Department of Transportation
Jeanette Rash, President, Incident Management Services
B.J. Lowery, Police Department, Houston, Texas
Bill Jacobs, Incident Management Director, Tennessee Department of Transportation
Jill Greene, Operator, Arizona Department of Transportation
Charles Manuel, Emergency Management Services Coordinator, Arizona Department of
Transportation
James Mona, Supervising Engineer, Connecticut Department of Transportation
Carlton Allen, Freeway Operation Supervisor, Houston District, Texas Department of
Transportation
Pat Irwin, Director of Traffic Operations, San Antonio District, Texas Department of
Transportation
Brian Burk, Traffic Engineer, Austin District, Texas Department of Transportation
Turbell Martin, TMC Branch Chief, District 11, California Department of Transportation
Nick Thompson, Metro Division, Office of Traffic Engineering, Minnesota Department
of Transportation
Alvin Marquess, Operations Manager, Maryland State Highway Administration
Rob Stone, State Incident Management Engineer, North Carolina Department of
Transportation
Paul King, TMC Traffic Engineer, District 12, California Department of Transportation
The authors would also like to thank David Helman with the Federal Highway Administration,
U.S. Department of Transportation for his technical guidance and support throughout this
project.
iv
TABLE OF CONTENTS
LIST OF FIGURES ...................................................................................................................... vii
LIST OF TABLES........................................................................................................................ vii
SECTION 1. INTRODUCTION AND BACKGROUND............................................................. 1
OBJECTIVES ............................................................................................................................. 1
SCOPE ........................................................................................................................................ 2
METHODOLOGY ..................................................................................................................... 2
ORGANIZATION OF REPORT................................................................................................ 2
SECTION 2. REVIEW OF LITERATURE .................................................................................. 3
WHAT IS AN INCIDENT?........................................................................................................ 3
Transportation Perspective...................................................................................................... 3
Emergency Services Perspective ............................................................................................ 3
CLASSIFICATION OF INCIDENTS........................................................................................ 4
Transportation Perspective...................................................................................................... 4
Emergency Services Perspective ............................................................................................ 5
PERFORMANCE MEASURES................................................................................................. 5
Transportation Perspective...................................................................................................... 5
Emergency Services Perspective ............................................................................................ 8
COLLECTION AND STORAGE OF INCIDENT MANAGEMENT DATA .......................... 9
Transportation Perspective...................................................................................................... 9
Emergency Services Perspective ............................................................................................ 9
STANDARD OPERATING PROCEDURES .......................................................................... 15
Transportation Perspective.................................................................................................... 15
Emergency Services Perspective .......................................................................................... 18
SECTION 3. SURVEY OF INCIDENT RESPONDERS ........................................................... 19
METHODOLOGY ................................................................................................................... 19
RESPONSE RATE ................................................................................................................... 20
FINDINGS................................................................................................................................ 20
Definition of Incident............................................................................................................ 20
Classification Of Incidents.................................................................................................... 21
Information Collected Per Incident....................................................................................... 24
Collection and Retention of Incident Data............................................................................ 32
Incident Management Performance Measures...................................................................... 36
Operational Definition of Incident Management Performance Measures ............................ 36
Origins of Performance Measures ........................................................................................ 41
Costs of Generating Performance Measures......................................................................... 43
Incident Management Performance Reports......................................................................... 43
Integration of Incident Records and Information.................................................................. 47
Issues Involved in Establishing an Incident Management System ....................................... 48
Most Important Things To Be Measured in Incident Management Program....................... 53
SECTION 4. SUMMARY AND RECOMMENDATIONS....................................................... 55
SUMMARY OF FINDINGS .................................................................................................... 55
RECOMMENDATIONS.......................................................................................................... 56
SUGGESTIONS FOR FUTURE RESEARCH ........................................................................ 58
v
REFERENCES ............................................................................................................................. 60
APPENDIX A. RADIO DISPATCH CODES FROM SELECT LAW ENFORCEMENT
AND EMERGENCY RESPONSE AGENCIES .......................................................................... 62
A-1. REVISED OFFICIAL APCO TEN SIGNALS ............................................................... 63
A-2. CALIFORNIA HIGHWAY PATROL RADIO CODES ................................................ 64
A-3. DALLAS PD RADIO SIGNAL CODES ........................................................................ 65
A-4. F.D.N.Y. RADIO CODES............................................................................................... 66
APPENDIX B. SAMPLE PERFORMANCE MEASURE REPORTS PRODUCED BY
MINNESOTA DEPARTMENT OF TRANSPORTATION ........................................................ 71
B - 1 . SAMPLE OF DAILY INCIDENT MANAGEMENT PERFORMANCE REPORT
USED BY MNDOT.................................................................................................................. 72
B- 2. SAMPLE OF MONTHLY INCIDENT MANAGEMENT PERFORMANCE REPORT
PRODUCED BY MNDOT....................................................................................................... 73
B - 3. SAMPLE OF YEARLY INCIDENT MANAGEMENT PERFORMANCE REPORT
PRODUCED BY MNDOT....................................................................................................... 75
APPENDIX C. SAMPLE PERFORMANCE MEASURE REPORTS PRODUCED BY
MOTORIST ASSISTANCE PROGRAM IN HOUSTON, TX. .................................................. 78
APPENDIX D. INCIDENT MANAGEMENT PERFORMANCE MEASURE SURVEY ....... 84
vi
LIST OF FIGURES
Figure 1. Sample Incident Logging Screen from Freeway Management Software Used
in Texas................................................................................................................................ 11
Figure 2. Log Showing Typical Incident Management Information Logged by
Service Patrols. ..................................................................................................................... 13
Figure 3. Sample of Typical Operator Screen Commonly Used in Computer Aided Dispatch
Systems. ................................................................................................................................ 14
Figure 4. Incident Report Form for Logging Information in National Fire Incident Reporting
System................................................................................................................................... 14
LIST OF TABLES
Table 1. Definition of Incident by Survey Respondents.............................................................. 22
Table 2. Criteria Used to Categorize Incidents and How It Effects Incident Response .............. 25
Table 3. Information Collected About Each Incident Event........................................................ 29
Table 4. Collection and Storage Methods, Retention, and Integration Policies of Incident
Information ........................................................................................................................... 33
Table 5. Typical Performance Measure Routinely Computed by Agencies................................ 37
Table 6. Operational Definition of Performance Measures Used to Evaluate Response
Systems ................................................................................................................................. 39
Table 7. Origin of Operational Definition for Performance Measures Being Used .................... 42
Table 8. Other Performance Measures Not Currently Being Collected, but Desirable............... 42
Table 9. Estimated Cost for Collecting, Processing, and Reporting Performance Measures...... 43
Table 10. Aggregation Level of Performance Reports ................................................................ 44
Table 11. Frequency at Which Performance Measures Reported................................................ 45
Table 12. Uses for Performance Measure Reports ...................................................................... 46
Table 13. Timeliness, Usefulness, and Accuracy of Incident Management Performance
Measures ............................................................................................................................... 47
Table 14. Other Sources of Incident Information in Jurisdiction ................................................ 49
Table 15. Integration of Incident Information with Other Agencies ........................................... 50
Table 16. Issues Faced in Setting Up Incident Management System .......................................... 52
Table 17. Most Important Thing to Measure in Incident Management Program ........................ 54
vii
SECTION 1. INTRODUCTION AND BACKGROUND
Incidents continue to be a major source of congestion on freeways. Because of the significance
of incidents on traffic operations, law enforcement, emergency service providers, and
transportation agencies are banning together in many metropolitan areas in the United States to
practice “incident management.” Incident management is defined as the “systematic, planned,
and coordinated use of human, institutional, mechanical, and technical resources to reduce the
duration and impact of incidents, and improve the safety of motorist, crash victims, and incident
responders.” (1)
The level of incident management varies considerably from location to location. Many locations
in the United States use motorist assistance patrols or service patrols that roam the freeways
looking for incidents and providing necessary assistance to clear stalled or disabled vehicles off
the roadway. Other locations have built a complex traffic control system that uses video
surveillance cameras and automatic incident detection systems to monitor the status of the
freeway and detect potential problem situations. Regardless of the size and complexity of the
incident management system in operations, decision-makers and operators want to know how
well the goals and objectives of their incident management systems are currently being met.
Performance monitoring (or measurement) is the “use of statistical evidence to determine
progress toward specific defined organizational objectives.”(2) Through performance
measurement, transportation agencies and emergency response providers can accomplish the
following:
• Set goals and objectives defining how well their incident detection and response
capabilities should be in their communities;
• Detect problems with their incident management procedures in their area and identify
corrective measures for addressing these problems,
• Manage, describe, and improve the incident response in their area, and
• Document the accomplishments, benefits, and effectiveness of their response process.
In many locations throughout the United States, different agencies with different primary
missions are responsible for different elements of the incident response process. For example,
the mission of a transportation agency is to restore the normal flow of traffic on the freeway as
quickly as possible while the primary mission of emergency service providers is prevention of
further loss of life and property. During an incident event, different agencies with normally
separate (and sometime competing) missions converge. Before improvements in the response
can be discussed and identified, the different agencies have to understand each other’s
perspective.
OBJECTIVES
The goal of this task order is to begin the process of understanding the perspective of the
different response agencies. The specific objectives of the task order are as follows:
• To provide a better understanding of how agencies measure their performance in
organized traffic incident management; and
1
•
To identify the difference, if any, in the definitions of relevant measures of performance
in incident management (such as detection time, response time, clearance time, etc.).
SCOPE
The scope of this task order was limited to the preparation, execution, and reporting of the results
of a survey of transportation, law enforcement, fire, and EMS/rescue agencies as well as the
preparation, execution, and reporting of the results of the pertinent literature on the measures
used by agencies to gauge the performance of their incident management systems. The scope of
this project did not include any field studies to collect any performance measures from actual
incident management systems. The researchers relied upon the results of the survey and the
literature review to form their conclusions and recommendations.
METHODOLOGY
A two-pronged approach was used to examine the issues of incident management performance
measures. The first prong was to review the available transportation and emergency services
literature related to measuring the performance of incident management systems in the United
States. Both traditional transportation databases as well as non-traditional databases were
searched looking for pertinent literature. Most of the literature related to emergency services
was identified, however, through Internet searches.
As the second prong to the approach, TTI conducted a survey of representatives from traffic, law
enforcement, and emergency service providers with active incident management program. The
survey team asked a series of prepared questions in telephone interviews. The questions
represented the basic level of information that was to be collected from each area. The same
general questions were asked of both transportation agency and emergency service provider
representatives.
ORGANIZATION OF REPORT
The remainder of this report is divided into three sections. Section 2 presents the results of a
search of transportation and emergency provider literature, specifically focused on traffic
incident management. Section 3 presents the results of a survey of practitioners that deal with
incident management on a daily basis. Section 4 contains recommendations and suggested future
research dealing with performance measures for incident management.
2
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.
WHAT IS AN INCIDENT?
Transportation Perspective
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 (1) 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.”(3) 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 (4) 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
3
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 911dispatching 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.
CLASSIFICATION OF INCIDENTS
Transportation Perspective
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. (7) 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
4
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).(15) 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.
PERFORMANCE MEASURES
Transportation Perspective
Many transportation agencies do periodical assessments of their incident management systems.
The Traffic Incident Management Handbook (1) 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. (5)
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
5
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:
• Roadway;
• 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
6
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
7
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. (9)
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.
Many agencies that have freeway service patrols or motorist assistance programs routinely
produce performance reports. (11, 12). Generally, these reports include information on the
following:
• 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.).
8
For example, the City of Austin Fire Department has a web site in which they report their
average response times for each month. (13) 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.
COLLECTION AND STORAGE OF INCIDENT MANAGEMENT DATA
Transportation Perspective
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 (16), 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,
9
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).(18) 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.(18) 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.
10
Figure 1. Sample Incident Logging Screen from Freeway Management Software Used in
Texas
11
12
Figure 2. Log Showing Typical Incident Management Information Logged by Service
Patrols
13
Figure 3. Sample of Typical Operator Screen Commonly Used in Computer Aided
Dispatch Systems
Source: K&K Computer Solutions Website (17)
Figure 4. Incident Report Form for Logging Information in National Fire Incident
Reporting System
14
•
•
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.
STANDARD OPERATING PROCEDURES
Transportation Perspective
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
o Traffic control requirements,
o Detour routes,
o Use of emergency lights by response vehicles,
o Parking of emergency vehicles at the scene,
o Staging of incident responses,
o 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.
15
Figure 5. State of Kansas Motor Vehicle Accident Report Form.
16
Figure 6. State of Kansas Motor Vehicle Accident Report Form (continued).
17
Emergency Services Perspective
The U.S. Fire Administration (USFA) has published a Guide To Developing Effective Standard
Operating Procedures for Fire and EMS Departments. (10) 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(8), 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.
18
SECTION 3. SURVEY OF INCIDENT RESPONDERS
A survey instrument was developed to obtain information on how transportation, law
enforcement, fire, and EMS/rescue agencies measure and report incident management
performance measures in their jurisdiction. The survey instrument solicited information related
to the following issues:
• How incidents are defined by agencies in their jurisdiction;
• How information about incidents is tracked and recorded;
• What, if any, measures they are collecting, calculating, or recording regarding incidents;
• What are the cost of collecting, processing, and reporting the measurement and source
data;
• If agencies are not using any measures, why not;
• If they are planning to implement measures, why, when, and how;
• How each measure is defined and calculated or measured;
• How the measures were decided upon and by whom;
• How long performance measure data have been collected and calculated;
• To whom the measures are reported, and how often;
• With whom the measures are shared;
• What the recipients do with the measures;
• What decisions are made based on or are influenced by the measures;
• How the recipients feel about the measures (i.e. are they meaningful, are they timely, do
they provide the information necessary for effective decision-making);
• The types of data collected about incidents, and the sources of the data;
• Whether similar data exists from other sources (especially other incident management
partner agencies), whether the data from the different sources are compared to one
another, and any findings from the comparison;
• What issues exist regarding measuring incident management performance, and how they
have been dealt with;
• What are the best candidate measures, whether they are recording measures or not.
METHODOLOGY
TTI used a telephone-interview type of format to collect the information from the different
transportation, law enforcement, fire, and EMS/rescue agencies. A series of questions were
developed that represented the basic level of information to be obtained from each agency. A
copy of the survey document is contained in Appendix D.
A pilot test of the survey instrument was performed prior to conducting the actual survey. The
purpose of the pilot test was to verify that the wording of the questions were clear and concise, to
fine-tune the data collection methodology, and to assess whether the questions provided
meaningful response. Based on the results of the pilot test, the survey document was revised
slightly to clarify some of the questions.
19
To conduct the survey, members of the research team initially contacted, via the telephone, each
of the identified individuals to request their participation in the survey. During this initial
contact, the researcher arranged a convenient day and time to conduct the survey or identify
alternative contacts. The researcher also obtained either a mailing address or an e-mail address
to which the survey questions could be sent. The researcher then forwarded the actual survey
questions to the respondent prior to actually conducting the survey. This was done so that the
survey respondent would have adequate time to prepare his or her responses to the questions.
At the scheduled day and time, the researcher contacted the survey respondent by telephone and
administered the survey. The researcher documented the respondent’s answers to each question.
The researcher also asked probing questions to clarify the response to survey question. The
responses were then coded into a spreadsheet to aid in analysis. This spreadsheet has been
provided to FHWA under a separate deliverable.
RESPONSE RATE
A total of 54 individuals from 30 locations were identified as potential respondents to the survey.
These individuals were identified from the following sources:
• The IEEE Incident Management Working Group,
• The ITE Traffic Incident Management Committee,
• The TRB Freeway Operations Committee,
• Personal contacts, and
• Internet searches of functioning traffic management centers.
A total of 23 individuals from 19 locations actually participated in the survey. The remainder of
the individuals originally identified either did not reply to initial inquiries about participating in
the survey, elected not to participate in the survey, or indicated that they did not have an active
incident management program in their area.
TTI planned to use representatives from the transportation agencies to identify appropriate
individuals in the law enforcement and emergency service agencies to survey. One problem with
this approach was that respondents were often unwilling to provide contact information of
representatives from other agencies that were responsible for incident management. This was
because either they did not know the correct person at the appropriate level or did not want to
increase the workload of these individuals with trying to respond to the survey. Therefore, most
of the insight into the emergency services perspective was obtained through the literature and a
limited number of survey responses.
FINDINGS
Definition of Incident
Most of the transportation agencies surveyed agree with the TMDD definition of an incident.
Most agencies define an incident as any unexpected event that causes a temporary reduction in
capacity. The term “temporary” is an important modifier because it implies that after the agency
performs some type of initial operation or response (i.e., clearing wrecked vehicles from the
travel lanes, removing a spilled load, etc.) the roadway can be reopened and normal capacity can
20
be resumed. For the most part, transportation agencies do not view highway maintenance and
reconstruction projects and non-emergency events themselves as incidents, generally, because
they are events that have planned means of accommodating traffic flow.
Most transportation agencies do not consider the long-range effects of an incident as part of the
initial incident. For example, most transportation agencies would not consider the repair of a
collapsed bridge deck, or the removal of spilled cargo that has been pushed beyond the shoulder
area as part of an incident, even though an event that they would describe as an incident was the
primary cause of the loss of capacity. This is especially true when recovery efforts extend over
multiple days. Most transportation agencies tend to classify incident events as being over once
the initial response to the incident event has left the scene and when more traditional traffic
control (i.e., work zone type traffic control) has been established at the scene.
Interestingly, many transportation agencies also classify unexpected weather events (particularly
snow and ice) as an “incident,” because they typically cause temporary reductions in capacity
(i.e., once the snow event is over and the roadways are cleared, the “incident” is over), increase
the potential for secondary events (such as crashes and stalled vehicles), and more importantly,
require a “response” from the transportation agency (dispatching of snowplows and de-icing
equipment, etc.).
Some agencies also classify events involving select sensitive users, such as school buses, railroad
crossing, etc. as incidents, primarily because these events may require special attention for
political or public welfare reasons.
Generally, events have to be on a roadway facility itself or in the right-of-way to be considered
as an incident by transportation agencies. Events that occur off the right-of-way, such as a
structure fire, are not routinely thought of as “incidents” by transportation agencies. Some
agencies do log these events in their incident management software and may broadcast messages
about these events through their motorist information systems.
Classification Of Incidents
One goal of incident management is to ensure that the appropriate response personnel and
equipment is provided at every incident. To aid in determining the appropriate level of response,
many transportation and emergency service providers have developed systems of classifying
incidents. Table 2 shows how the survey respondents replied to questions concerning methods
and criteria for classifying incidents in their local area. The table also shows how the level of
severity of the incident effects each agency’s response decisions.
21
Kansas DOT –
x
Kansas City
New Jersey
DOT
Arizona DOT
Ohio DOT Columbus
Tennessee
DOT
Phoenix Az,
Fire Dept.
Maryland
State Hwy
Admin CHART
Texas DOT Austin
Texas DOT –
San Antonio
Debris Roadway
Public Emergency
Abandoned Vehicle
On Shoulder
Hazmat Spill
Vehicle Fire
Stall on Shoulder
Abandoned vehicle
In lane
Stall in Lane
Overturned Vehicle
Agency
Collision
Table 1. Definition of Incident by Survey Respondents
Other
x
x
Only incidents requiring police accident reports are documented. Kansas DOT
is currently in the process of building a TMC. They hope to have it operational
by the end of this year to early next year. Currently, the state police and service
patrol (operated by the police) are the only incident management elements in
place. The police provide the DOT with copies of the accident reports for
accidents on their facilities.
x
x
x
Downed Utility Pole; downed signal pole; anything blocking a lane or shoulder
x
x
x
x
x
x
x
x
x
Unexpected weather change
x
x
x
x
x
Anything effecting traffic flow
x
x
x
x
x
x
x
x
Anything effecting traffic flow
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
Weather; construction; maintenance
22
Vehicle Fire
Hazmat Spill
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
North Carolina
x
DOT
Connecticut
x
DOT
Debris Roadway
Stall on Shoulder
x
Public Emergency
Abandoned vehicle
In lane
x
Abandoned Vehicle
On Shoulder
Stall in Lane
New York
DOT
Colorado DOT
Lakewood
Texas DOT Houston
Illinois DOT Chicago
Overturned Vehicle
Agency
Minnesota
DOT Minneapolis
Caltrans - San
Diego
Incident
Management
Services-Houston
Southeast
Michigan
COG - Detroit
City of
Houston Police Dept
Collision
Table 1. Definition of Incident by Survey Respondents
x
x
x
x
x
x
x
x
x
x
x
x
Assist TxDOT
x
x
x
x
x
Brush fire, pedestrian in restricted area, road work, traffic signal malfunction,
non-recurring severe congestion
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
Other
Ice on pavement, water main breaks, flooding, anything that blocks one or more
lane for 30 minutes or more, school bus involvement, railroad crossing
involvement, fatality.
Anything effecting traffic flow
23
A common classification scheme that describes the severity of the incident and/or the urgency of
the response does not exist. For the most part, transportation agencies tend to classify incidents
into two to three categories based upon the degree to which traffic is likely to be impacted
(severity) and/or the number of lanes blocked. Some of the criteria that transportation agencies
use to classify incidents include the following:
• Number of lanes blocked;
• Estimated duration of blockage;
• Severity and/or number of injuries involved;
• Time-of-day;
• Presence of hazardous materials;
• Degree of damage to vehicles and/or infrastructure;
• Type of vehicles involved (e.g., trucks, buses, etc.); and
• Number of vehicles involved.
Emergency service providers, on the other hand, typically classify events based on the potential
loss of life and/or the impact to public safety. Both of the emergency service providers use
standards that have been defined by their industry as a means of classifying incidents. These
standards take into account the presence of possible injuries or fatalities, and rely on dispatchers
soliciting correct information from the individuals reporting the incidents.
Information Collected Per Incident
One attribute of a good performance measurement system is that data to generate performance
measure be readily attainable in an economic manner.(1) This implies that in order for agencies
to develop and use performance measures, the data must be readily available through their
already existing systems. Responders are more likely to compute performance measures if they
are already collecting the data to support them. Part of this survey effort was to look at what data
is currently being collected by different agencies and how.
Table 3 shows what information many of the transportation and emergency service providers are
collecting about each incident event. Based on the survey responses, at a minimum, the
following information is recorded by most agencies:
• The roadway name where the incident occurred;
• The name of a nearby cross-street or location;
• The location of the incident in the lanes (i.e., which lanes are blocked);
• The type of incident;
• The time at which the incident was detected or reported;
• The time the first response vehicle arrived on the scene; and
• The time the incident was cleared from the scene.
24
Table 2. Criteria Used to Categorize Incidents and How It Effects Incident Response
Agency
Criteria
New Jersey Major, Minor.
DOT
Thresholds
Response Variation
Major incidents defined as those lasting more than Minor incidents -- use ITS (DMS/HAR) if applicable. For major
one hour while minor incidents defined as those
incidents, review to see if need to send IM response team. Team
lasting less than 1 hour.
consists of state trooper and DOT traffic operations person, get to
scene and try to speed clearance of incident.
Arizona
DOT
Level 1, 2, 3
OhioDOTColumbus
Severity, time-of-day,
congestion level
What changes is who gets notified and how much of a hurry we
Level 1 -- fatality; unplanned closure in one or
are to get responses from them.
both direction affecting any state route; any
incident involving HAZMAT, homicide, trains, or Level 1 -- notify Admin Major (includes ADOT Director, and
State engineer, and District Engineer).
school buses;
Level 2 -- traffic flow is restricted; requiring live Level 2 - Notify Maintenance Supervisor by pager or phone.
Level 3 -- notify supervisors via email, phone, radio.
AzDOT presence; fences cuts, livestock on
roadway, or guard rail damage presenting hazard to
motorist; red indication out / stop sign knockdown;
large dead animal in lanes; roadway damage (large
potholes, gravel on roadway); disabled vehicle
blocking flow; structural damage that does not
close hwy; threat of jumper that does not close hwy
Level 3 -- Yellow/green indication out; debris not
blocking roadway; disabled vehicle not blocking
roadway; Maintenance; anything that can be
handled at supervisor discretion; anything not
requiring immediate ADOT response
Lane blockages of more than one minute warrants Incident response plan (IRM) addresses how to handle major
activating DMS; DMS messages updated as lane incidents, stalled vehicles, debris, roadwork, congestion,
blockage changes; Service patrol will work
fire/HAZMAT, freeway diversion. For minor fender benders,
incidents expected to be under 15 minutes to clear, execute only what is helpful to motorist that doesn't cause a lot of
otherwise call for tow trucks
inconvenience. For major incidents (e.g., fatality) and EMS is on
the scene, execute full plan immediately.
Tennessee
DOT
Phoenix, Az Use universal system U.S.
Fire Dept. Fire Adm. (thru FEMA
website)
-
Long term - debriefings and updates
-
Response bases on Inc. Management System (IMS) -- developed
in California published 1985. Dispatchers - rotate
25
Table 2. Criteria Used to Categorize Incidents and How It Effects Incident Response
Agency
Maryland
State Hwy
Adm CHART
Criteria
Property damage: person
injured/fatality; Hazmat;
emergency roadwork; -- 15
items out of FHWA Data
Dictionary
Thresholds
-
Response Variation
If longer than 2 hrs shutdown, preplanned detour routes.
Dependent on magnitude of incident, different levels of
notifications is given to agencies.
Texas DOT - HCM Level of Service
Austin
Criteria; Reported vs.
verified
Compare current volume/occupancy measures to
HCM thresholds.
Texas DOT - Type of incident (I.e.,
San Antonio debris, weather, accident).
Severity of lanes closed;
Severity of accident
Severity of lanes closed -- 2 or 3 lanes closed,
TransGuide software system automatically prioritizes -- major
classified as major incident. With crash scenes,
incidents over minor incidents, minor incident in open lane.
major incident is one that requires EMS (get
System uses operator inputs (I.e., description of incidents) to
information via police). Major incident - when
driver scenario process.
demand expected to exceed capacity.
Judgment call by operator. Used past experience, Major incidents -- place motorist information system in overdrive.
type of incident, Time-of-day, expected duration of Broadcast radio messages every 10 minutes. With major incident,
incident (i.e., any road closure or any incident
use DMSs to direct motorist to tune to station and continuously
during peak period, hazmat or rollover) classified broadcast incident information. Will also call other media outlets.
as major
May pull in other operators if many going on at same time.
Minnesota Major, Minor.
DOT Minneapolis
Caltrans –
San Diego
No impact on operations -- simply informational. Emergency
services will look at speed. Haven't needed to classify incidents
(respond to all incidents). Verified vs reported -- if reported, will
look to verify with CCTV and then clear.
Use California Highway
patrol's radio call system (10
codes, 11 codes)
-
Highest level codes, Caltrans will dispatch response immediately.
With other codes, will wait until officer on-site. Will change
response or dispatch response based on officers needs.
Incident
Only respond to major
Management incident involving 18Services – wheeler rollovers/lost loads.
Houston, Tx
-
-
26
Table 2. Criteria Used to Categorize Incidents and How It Effects Incident Response
Agency
Southeast
Michigan
COG -Detroit
Criteria
No defined criteria (i.e.,
delay threshold severity).
Michigan State Police
Criminal Justice
Information Center has a
system to capture this
information called the
Automated Incident
Command System (AICS).
Thresholds
Response Variation
There are no documented thresholds that I know of The dispatcher determines the appropriate response after assessing
but there might be something defined by the State the call or by the person responding to the call once at the scene of
Police. They work by guidelines and training
the incident. Appropriate responses scenarios might also be
found in the Incident Command System (ICS).
determined through the use of ICS and CAD systems. Assistance
They also have a Computer Aided Dispatch (CAD) is provided by the Michigan Intelligent Transportation Systems
that dispatches the appropriate personnel for a
(ITS) Center if it is a freeway incident through the use of the
particular event.
cameras.
City of
Severity -- Major/Minor;
Major = major freeway blockage; Minor = minimal 90% of incidents detected by roving patrol; 6% dispatched from
Houston, Tx Location -- Moving lane of freeway blockage
TranStar; clear minor incidents alone; assist with traffic control at
Police Dept traffic (right shoulder, left
major incidents;
shoulder, lane(s) blocked - 1
23456
New York
DOT
Colorado
DOT Lakewood
Combination of severity,
anticipated duration, and
time-of-day (e.g., peak or
off-peak)
Level 1 -- no lane blocked - on shoulder;
The more severe the more they "throw" at it. They have
Level 2 -- 1 lane blocked 0-15 min (peak) 0-30 min communications with metro traffic and local media (if after metro
(off-peak);
traffic hours). Co-located in TMC with state police - get estimate
Level 3 -- 1 lane blocked 15-30 mins (peak) or 30- from trooper for duration. Level 1-2: may or may not do
60 mins (off-peak);
anything. Higher levels - At first advise metro traffic/media of
Level 4 -- 1 or more blocked 30-60min (peak) 60- problem - if worse, recommend taking alternate route (but don't
120(off-peak);
specify) - if really bad, recommend specific alternate route - more
Level 5 -- road closure, 1+blocked 60 min(peak) severe, use stronger DMS messages - use DMS to notify to tune to
60-120(off-peak)
HAR - have 1 permanent HAR and 2 portable (1 portable being
converted to permanent).
Mile High Courtesy patrol 3-tier system for major incidents -- total freeway Main response is public information. They have a broadcast fax
handles minor incidents.
closure or most lanes blocked
system with 300 agencies/companies signed up including media,
The TMC only responds to Level 1 -- duration less than 30 minutes;
other public agencies, trucking firms, US military, US Postal
major incidents -- duration Level 2 -- duration 30 minutes to 2 hours;
Service, visitor centers, etc. Also post information on their
is the criteria used
Level 3 -- duration over 2 hours
website
27
Table 2. Criteria Used to Categorize Incidents and How It Effects Incident Response
Agency
Criteria
Thresholds
Texas DOT - Will follow that provided by Major: One lane > 30 min (TOD dependent); Two
Houston
law enforcement
or more lanes > 15 min (TOD dependent); truck
(Fatality/Injury = major,
accidents, HazMat spills, bus accident, multiPDO = minor), as well as vehicle accidents
determining severity based Minor: Other incidents
upon lanes blocked and
duration
Illinois DOT Severity -- routine or
- Chicago
incident; Lane blockage
Response Variation
Different types of incidents require different level of response.
For example, HFS is not contacted for a minor incident, however,
HPD may be required and they are contacted the same as if it were
a major incident. They are given all details known and it is left to
them to determine their condition of response.
1 or more lane closed for 30 minutes or more; total More documentation for incidents than "routines", more public
freeway closure for 15 minutes or more; Hazmat awareness for more major incidents -- media alerts, notify DOT
personnel, DMS
28
New Jersey DOT
x
x
Arizona DOT
x
x
Ohio DOTColumbus
Tennesse DOT
x
x
Phoeniz , AZ Fire
Dept
x
x
x
x
x
x
x
x
Maryland State
Hwy Admin -CHART
Texas DOT Austin
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x,1
x
x
x
Weather condition
Injuries present
# of vehicles involved
Type of Vehicles involved
x
x
x
x
x
x
x,2
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x,3
x
x
29
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
Incident Severity (qualitative)
Light Condition
x,1
Roadway Condition
Roadway Surface Condition
Time traffic returned to normal
flow
Time incident was cleared from
scene
Time response vehicles left scene
Type of response vehicles on scene
Time response vehicle arrived on
scene
Source of incident verification
x
x
x
Time incident was verified
Time incident was detected
(reported)
x
Current status of Incident
Incident Type
x
Incident Source
Location of Lanes Blocked
Lat/Long
x
Detection Station #
x
Block Number
Location/Cross-Street Name
Agency
Kansas DOT,
Kansas City
Roadway Name
Table 3. Information Collected About Each Incident Event
Other
x Property damage; diagram;
names; vehicle makes; model,
color, plate numbers
x
x Route, direction, milepost,
type of incident (accident
with or without
injuries/death); who was
called out.
x Miler maker system location
Type of service; vehicle tag #;
direction
x Detailed info on injuries,
seatbelts, child restraints;
Trucks have live terminals
and digital cameras to collect
info
System software records time
that changes to any fields are
made, including update to
comments.
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x, 7
x
x
x
x
x
30
x
x
x
x
x
x
x
x
x
x # of lanes blocked
x
x
x
x
x
x See attachment
x
x
x
x
x
x Vehicle -- make, model,
color, year, license plate;
Driver -- male, female;
number of occupants -- driver
only, 2, 3, 4+; motorist use of
cell phone -- # called, air
time, motorist name &
signature
Other highways affected (if
any); which ITS devices
activated -- DMS, HAR
Information collected for
service patrol response to
minor incidents only. There
is currently no logging of
major incident data (level 1,
2, 3 incidents) that the TMC
responds to.
x, 5
x
x
Incident Severity (qualitative)
x
Type of Vehicles involved
x
x
# of vehicles involved
x
x
Injuries present
Colorado DOT Lakewood
x
x
x
Weather condition
x
x
x,6
Light Condition
x
x
x,4
Roadway Condition
New York DOT
x
Roadway Surface Condition
x
x
Time traffic returned to normal
flow
x
x
x
x
Time incident was cleared from
scene
x
x
Time response vehicles left scene
x
x
Type of response vehicles on scene
x
x
Time response vehicle arrived on
scene
x
x
Source of incident verification
x
x
Time incident was verified
x
Time incident was detected
(reported)
x
Current status of Incident
x
Incident Source
x
Incident Type
Minnesota DOTMinneapolis
Caltrans
San Diego
Souteast Michigan
COG - Detroit
Houston, TX -Motorist
Assistance Patrol
Location of Lanes Blocked
x
Lat/Long
x
Detection Station #
Location/Cross-Street Name
x
Agency
Texas DOT San Antonio
Block Number
Roadway Name
Table 3. Information Collected About Each Incident Event
x, 8
x
x
Other
System software records time
reported, time entered in
system, time system executed
scenario, time scenario
changed, time scenario over
(when lane back open to
traffic)
x
x
x
x
x
x
x
x
x
x
x
North Carolina
DOT
Connecticut DOT
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
1 = First on scene
2= removed from roadway altogether
3 = Individual dispatched, on scene, and benchmark points
4 = opening of lanes
5 = also record under maintenance/construction
6 = record weather at start of each shift as operator logs in
7 = time stamp when entered into MIST
8 = Not fields in software for this but try to indicate these in open comment field
31
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
Incident Severity (qualitative)
x
x
Type of Vehicles involved
x
x
# of vehicles involved
Illinois DOT –
Chicago
City of Houston,
Tx Police Dept
x, 2
Injuries present
x
Weather condition
x
Light Condition
x
Roadway Condition
x
Roadway Surface Condition
Source of incident verification
x
Time traffic returned to normal
flow
Time incident was verified
x
Time incident was cleared from
scene
Current status of Incident
x
Time response vehicles left scene
Incident Source
Type of response vehicles on scene
Incident Type
x
Time incident was detected
(reported)
Location of Lanes Blocked
x
Lat/Long
x
Detection Station #
Location/Cross-Street Name
x
Agency
Texas DOT –
Houston
Block Number
Roadway Name
Time response vehicle arrived on
scene
Table 3. Information Collected About Each Incident Event
Other
Incident date; direction of
travel; Before/After cross
street
x
HPD staffs a single console at
TranStar. While more
specific information is
collected by the officer in the
field, HPD at TranStar only
logs some general information
-- only for incidents that occur
on the freeway system
x Information only for motorist
assistance patrols
Interestingly, only eleven agencies reported that they record the time that an incident was
verified. However, in further discussion with the respondents, it was revealed that, in many
cases, time the incident was detected (or reported) and the time the incident was verified are
frequently the same time.
Thirteen agencies reported that they record the time the first incident responders arrived on the
scene. Similarly, slightly more than half of the respondents indicated that they routinely record
the time the incident response vehicles leave the scene and/or the time the incident was cleared
from the roadway. For the most part, agencies are primarily concerned with keeping track of
the time that they implement or execute their response and are not overly concerned with
recording the time that other responders perform certain functions.
Only one agency reported that they record the time that the freeway returned to normal flow. A
few common reasons cited for not recording this measure include the following:
• It is too hard to determine when “normal” flow occurs;
• The congestion resulting from an incident last so long that operators tend to forget to go
back and log when normal traffic flow occurs; and
• This time is not important to determining the effectiveness of the response.
Some respondents indicated that their software system automatically records the time (i.e., time
stamps) every time the operator makes a change to the traffic control. For example, when the
operator first initiates a message on a DMS, the time is logged by the system. If the operator
changes the message, the time the new message is implemented by the system is logged. The
advantage of this approach is that it takes the burden off the operator to log when certain changes
are made.
Collection and Retention of Incident Data
Table 4 summarizes how the respondents replied to questions concerning the collection and
storage of incident data. An approximately equal number of agencies use manual (seven of the
respondents) and automatic (eight of the respondents) means of collecting incident data. Four
agencies reported that they use a combination of manual forms and automated systems for
collecting information about incidents. In a few cases where agencies used manual data
collection means, the forms were later transferred into automated systems for further processing
and storage.
Most agencies reported that their incident information either initially or eventually ended up in a
database that could be queried. The survey also showed that information about specific incidents
was generally kept for a long-time, with most agencies retaining their incident logs for three or
more years.
Agencies were also asked if they integrated their incident reports with any of the other incident
responders. The general response was “no”; however, some agencies did state they have plans to
begin integrating their freeway management center systems with a 911 dispatching center so that
data from other agencies could be merged with incident records. This is expected to increase
both the quality and quantity of data about incidents at these locations.
32
Table 4. Collection and Storage Methods, Retention, and Integration Policies of Incident Information
How is this information
Agency
collected?
Kansas DOT Manual (1)
– Kansas City
New Jersey Automatic
DOT
Arizona DOT Automatic
Ohio DOT Columbus
Tennessee
DOT
Phoeniz, Az
Fire Dept
What format is used to store
How long is information retained?
information?
Receive paper file from state police, enter 5 years to Forever
into a queriable Oracle database. No
CCTV yet, highway patrol video for
fatality.
Queriable database
8 years
Is data integrated with other
information?
Highway patrol input accident data into
accident report database. DOT
automatically receives copy of any incident
on DOT facility No
When the police work an incident, we are
supposed to get their log number. These
are not always made available to us. We
usually enter these into the Road Condition
report and enter the HCRS# into the
documentation.
Manual (2) / Automatic
Service patrol fills out paper form, later
Not sure on the electronic files, permanent No
entered into queriable database -- Paradox. for database
DMS message logged manually to
compare accuracy of DMS electronic file
log (new)
Manual
Paper, entered into database
Since start in database (June '99). Paper
Some -- major incidents w/ multiple
not kept long term after entered into
agencies -- debrief w/ police, fire,
database
timeframe
Both: All vehicles have geo id. Paper, electronic
Paper -- 3 yrs
Yes -- police dispatch, census
Monitored by clock this tracks
time of arrivals, reposition,
leave.
Manual--Pictures;
EMS data -- handheld
computer, download later
Automatic
Oracle database
Started Feb 2000 keeping everything;
In future plans: 911 centers: ability for
before - 5yrs on-site then paper to
other agencies (police, county) to access
warehouse
software & edit incident reports eventually
Maryland
State Hwy
Admin -CHART
Texas DOT - Automatic
Austin
Queriable database
3 years
Sybase
No deletion policy has yet to be developed. No yet -- only one incident done so far but
Quarterly off-load and access through
not very detailed. Done to answer
Excel
questions about response. Ad hoc requests
-- maintenance information about
equipment failures
33
Table 4. Collection and Storage Methods, Retention, and Integration Policies of Incident Information
How is this information
Agency
collected?
Texas DOT – Automatic
San Antonio
What format is used to store
information?
Electronic files
Minnesota
DOT Minneapolis
queriable database -- Access (since 2001); Early '90
prior to '01 – paper logs
Automatic
Caltrans –San Manual
Diego
Southeast
Manual & Automatic
Michigan
COG –
Detroit
Paper files and electronic files
Less than 14 mo
Is data integrated with other
information?
System tied directly to 911 map -- don't use
one system to verify the other
Recent had FHWA intern perform big
analysis were compared police logs to
system logs. Do not routinely perform
comparison. Done on as needed basis and
when staff available. Do produce annual
volume/crash frequency report
When needed.
Data stored in both paper and electronic
SEMCOG has only just started to gather
Try to cross reference the MSP 911 data
formats. SEMCOG requests copies of the this data (over the past 5 years). Have kept with the Freeway Courtesy Patrol data
database and we query it using MS Access all of it so far
(checking to see how long abandon vehicle
have been out on the roadway after they
have been identified). Also integrate the
MSP crash data (UD10 forms/database)
with the incident database. Also integrate
the incident information with road attribute
file with includes fields like: lane, 85%ile
speed; posted speed; land use, vehicle
classification counts, traffic volume counts,
etc.
Paper file, electronic files, queriable
Data generated by MAP is compiled by
Yes. TTI compiles information and breaks
database -- Access
TTI and returned to TxDOT for storing.
numbers down to percentages.
Don't know how long they keep it
Houston, Tx
Motorist
Assistance
Patrol
New York
DOT
Manual & Automatic
Colorado
DOT Lakewood
TxDOT Houston
Automatic -- Service patrol
Oracle queriable database
calls dispatch, dispatcher enters
all info into database.
Automatic
Flat files -- queriable database
Typed into MIST
How long is information retained?
Minimum of two years
Queriable database -- Sybase
Current six months active in system (last
week of 6 months falls off each week);
burn 6 mo. Data every week to CD for
backup
Indefinitely
Service patrol logs to different system, but
if working an incident DOT is entering into
MIST, then cross-reference to service
patrol record entered.
No
Indefinitely
Not electronically. MAP files collected in
same manner but different database
34
Table 4. Collection and Storage Methods, Retention, and Integration Policies of Incident Information
Is data integrated with other
information?
Cross reference state police records; ETP
service patrol uses fill-in the dot data cards,
will soon be upgrading; the data is not
routinely compared but the capability is
there
How is this information
Agency
collected?
Illinois DOT – Manual
Chicago
What format is used to store
information?
Paper file -- shared with DOT traffic,
maintenance, and claims department
City of
Manual (3) Other (4)
Houston, TX
Police Dept
Paper. The Access database is used to
Printouts of the daily activity logs are kept
enter incidents during each shift (two shifts for 3 years.
per day). At the end of the shift, the daily
activity log is printed. The database only
retains the totals for the shift (data on
individual incidents not saved in the
database -- only on the printouts). The
database is then used to prepare the
monthly reports
Queriable database
Indefinitely (have been collecting for
No
~6yrs)
Paper and electronic
Incident reports are retained for 5 years
No
How long is information retained?
7 years
North
Manual (5)
Carolina DOT
Connecticut
DOT
(1) Accident Forms
(2) Freeway service patrol incident log form
(3) Accident reporting form filled out by officer in field, but does not go to TranStar
(4) Incident data at TranStar is manually entered into an Access database
(5) IMAP program -- called to TMC entered into database on local PC, moving to webpage to consolidate information
35
Incident Management Performance Measures
Table 5 shows the general types of performance measures that are routinely computed by the
agencies responding to the survey. Only half of the agencies responding indicated that they
routinely compute incident-related performance measures. Not surprisingly, most of the
agencies that are computing performance measures reported computing the following
performance measures:
• Incident frequency,
• Detection time,
• Response time, and
• Clearance time.
Operational Definition of Incident Management Performance Measures
Table 6 shows the operational definitions that each agency is using to compute these
performance measures. Interestingly enough, most agencies define “detection time” as the time
that they were notified of the incident (i.e., the time that the incident was reported to them in
their control center). Detection time is not defined as the time between when an incident actually
occurred and when the agency was notified of the incident (either from emergency responders,
operator observation, and direct report from citizen).
Nearly all of the respondents indicated that they define “Response Time” as the elapse time
between when the agency was first notified about an incident and when the first responder
appeared on the scene. The primary difference in the way that agencies define response time is
that emergency responders typically define response time as the time from when an incident was
reported to their dispatcher to the time when their response vehicles arrive on the scene.
Transportation agencies generally measure response time from when the call comes into the
TMC (or service patrol dispatcher) to when first response vehicle arrives on the scene, regardless
to which agency the vehicle belonged (i.e., this could be a fire vehicle, police vehicle, or service
patrol vehicle). The problem with defining response time this way is that often times, the
transportation agency does not have any control over when the emergency service providers are
dispatched or the priorities that are assigned to different types of incidents. In many cases, the
response time that is reported by many transportation agencies is actually the time between two
unrelated events (i.e., notification of the incident and the dispatching and arrival of the response
vehicles). This is especially true when the traffic management center (TMC) is not the first
agency notified of the incident (which is generally the case in most metropolitan areas). Without
integrating or comparing records from the dispatching agency, the response time may not
represent the true response time of the first responder to the incident, but merely the time
between unrelated events.
Clearance time is another measure that varies dramatically between freeway management
operators and emergency service providers. For the most part, transportation agencies define
clearance time as between when the first responder arrives on the scene (regardless of which
agency they work for) to when the incident is cleared from the roadway. Emergency service
providers typically define clearance time as the time between when the first of their units arrive
on the scene to when their unit leaves the scene and can be deployed elsewhere.
36
Table 5. Typical Performance Measure Routinely Computed by Agencies
What measures do you routinely compute?
Agency
Kansas DOT
Kansas DOT Kansas City
New Jersey
DOT
Arizona DOT
Ohio DOT Columbus
Tennessee DOT
City of Phoenix
Fire Dept
Maryland State
Hwy Admin –
CHART
Texas DOT Austin
Texas DOT –
San Antonio
Minnesota DOT
– Minneapolis
Caltrans –San
Diego
Southeast
Michigan COG
– Detroit
Houston, TX Motorist
Assistance
Patrol
New York DOT
Colorado DOT –
Lakewood
Do you
calculate
Performance Incident
Measures? Frequency
Yes
x (1)
Incident
Rate
Detection Response
Time
Time
Clearance
Time
Number of
Secondary
Incidents
Time to
Normal
Flow
Incident
Delay
Others
No (2)
Yes
x
x
x
Yes (3)
x
x
x
(4)
x
x
x
x
x
(7)
x
x
x
x
x
No (5)
No (6)
Yes
x
x
Yes (8)
x
x
No (9)
Error logs -- preventative maintenance
No (10)
Yes
Severity; Nature of Damage; Injuries
Delay hours; environmental impacts;
frequency by location; # of disabled
vehicles assisted
x
x
x
x
x (11)
No (12)
Yes
x
x
x
x
x
x
x
Yes (13)
x
x
x
(14)
x
x
x
No
No
37
Air quality -- pollutants (e.g., amounts
of VOC, NOx, and CO)
Types of assists provided (used to
stock supplies); location of incidents
(by corridor, by segment)
Table 5. Typical Performance Measure Routinely Computed by Agencies
What measures do you routinely compute?
Agency
Texas DOT –
Houston
Illinois DOTChicago
Do you
calculate
Performance Incident
Measures? Frequency
Incident
Rate
Detection Response
Time
Time
Clearance
Time
Number of
Secondary
Incidents
Time to
Normal
Flow
Incident
Delay
Others
No (15)
Yes
x
Other performance measures such as
response time, clearance times, and
detection time have been calculated
before but not routinely done. Only
done periodically for program
justification.
x
City of Houston,
No (16)
TX Police Dept
North Carolina
No (17)
DOT
Connecticut
Yes
x
x
x
x
x
x
DOT
(1) Use incident frequency to identify high accident locations for improvements
(2) Hope more will be done once TMC is operational
(3) These can be gotten by database query. We do not use this data, but the districts use them to rate district-wide response times
(4) Believe this is important, but they do not track it as a general rule
(5) Do not have the funding for personnel to design, implement, and update performance measures
(6) Under evaluation; Early stages through contract with University (Vanderbilt)
(7) Police do and offer to Fire, don't use
(8) University of Maryland prepares yearly report (1997 on web)
(9) Too time consuming
(10) City-wide incident management project -- visually seen 40% reduction in clearance times
(11) By type of responder
(12) Not an issue before now -- can recreate times based on logs
(13) Most incidents also depend on arrival of other agencies (I.e., ambulances, other police agencies, and other emergency equipment needed)
(14) Data collected but not currently used
(15) This is an operations staff not a research staff. There is not the time or personnel available for this function. High accident locations are identified from the information and
consideration given to these areas on a routine basis. TTI puts together an Annual Report for TranStar
(16) That information has not been required
(17) Problem is what performance measures to look at. In process of identifying for future
38
Table 6. Operational Definition of Performance Measures Used to Evaluate Response Systems
Performance
Agency
Measure
Incident
City of Phoenix Fire Dept
Frequency
Maryland State Hwy Admin – CHART
Connecticut DOT
Incident Rate City of Phoenix Fire Dept
Maryland State Hwy Admin – CHART
Detection
New Jersey DOT
Time
Arizona DOT
City of Phoenix Fire Dept
Maryland State Hwy Admin – CHART
Texas DOT – San Antonio
Caltran – San Diego
Houston Motorist Assistance Patrol
Connecticut DOT
Response Time New Jersey DOT
Arizona DOT
City of Phoenix Fire Dept
Maryland State Hwy Admin – CHART
Texas DOT – San Antonio
Minnesota DOT – Minneapolis
Caltran – San Diego
Houston Motorist Assistance Patrol
Operational Definition
Time based, incident/shift, also calculate week, month, year and compare to last year
How often occurs at a given location (mile post)
Any time there is a blockage of highway, an incident is established
# of incidents per month or year; look at each different category and calculate; use to shift response
ADT x # of incidents
When DOT finds out about the incident
Delay from the time that an incident occurs until it is reported
1st report to dispatch; if official (Police, city); ask them when they detected. Keep track of who
reported incident (official or civilian)
1st person sees to calling it in
System parameter (2 minutes) -- use 20 sec interval data with rolling average (6 cycles). System
usually 1 or so minutes after call
"Reported Time" -- time when report comes into center
Time of notification, also driver estimate of time of occurrence
The time the incident is reported to the TOC via surveillance equipment or verified phone calls
Time for DOT to get there
Starts with live voice reports receiving page and then they are responding. Ends when unit reports
they are on-scene.
Time elapse between 1st dispatch contact to 1st vehicle on-scene
Time call received until arrive on scene
System logs time every time a change or update is made to response scenario
Time detected to time responders arrived on scene; camera-based; not perfect -- only when operator
observes when respond on scene
Time when 1st responder arrive on-scene
Dispatch time and time of arrival
39
Table 6. Operational Definition of Performance Measures Used to Evaluate Response Systems
Performance
Measure
Clearance
Time
Number of
Secondary
Incidents
Agency
Connecticut DOT
New Jersey DOT
Arizona DOT
City of Phoenix Fire Dept
Maryland State Hwy Admin – CHART
Texas DOT – San Antonio
Caltran – San Diego
Houston Motorist Assistance Patrol
Connecticut DOT
Arizona DOT
City of Phoenix Fire Dept
Maryland State Hwy Admin – CHART
Caltran – San Diego
Houston Motorist Assistance Patrol
Time to
City of Phoenix Fire Dept
Normal Flow Maryland State Hwy Admin – CHART
Houston Motorist Assistance Patrol
Incident Delay Maryland State Hwy Admin – CHART
Houston Motorist Assistance Patrol
Operational Definition
The time responders arrive on scene. Arrival time and response time are calculated for state police
only out of the Bridgeport operations center coverage area. ConnDOT only contacts its internal
responders such as bridge safety, construction, maintenance, and electrical and service patrol when
required. The contact time and arrival time is then kept. Arrival time only for emergency responders
such as EMS, wrecker, fire, and environmental protections is also noted. DOT does not normally
contact these responders initially
Time between detection and incident cleared from scene
When unit reports they are clear or when operator sees all units clear. This is for when the ADOT
vehicle leaves the scene.
Time fire department declares incident over, usually as driving away from scene
How long from notification to clear, or until delays clear / all lanes open is what they use
Time 1st vehicle arrives on scene until lanes open
Time when roadway opened
Time incident ends and clearing of incident from roadway
The time the accident or debris is removed from the travel way
Accidents that occur back in queue
Count of accidents, injury, fire, hazmat each count as one not a different incident #; 1 incident with
multiple parts
Pinpoint incident is created by delay from previous incident, call by operator
Don't know how to compute
Time of notification
Set by incident commander. Wait at scene until flow returns to normal for time. Subjective.
Back to operating capacity for time-of-day
When incident clears and blockage has been removed from freeway
Length of distance (5 mile delay) Max delay (example: 10 mile backup)
Time of duration
40
Most agencies agree that the number of secondary accidents resulting from an incident was a
difficult measure to compute. In most cases, this was considered to be a subjective measure of
the operator. One agency, however, defined a secondary accident to be any accident that
occurred within a defined radius and time frame of the first incident. Both the distance and time
parameters changed by time-of-day to reflect the different levels of congestion that forms around
incidents.
Maryland defines incident delays in term of queue distance. They generally use measures such
as the length of congestion (e.g., a five-mile delay or a 10-mile backup) to help define incident
delays. Queue distance is a parameter that can be observed almost instantaneously via the
surveillance cameras while delay requires that the time it takes drivers to pass through the
congestion be measured.
Origins of Performance Measures
In Table 7, respondents were asked about the origin of the operational definitions being used to
generate the performance measures (i.e., the driving force behind the generation of the
performance measures they are currently using). Several of the respondents indicated that the
performance measure that they are currently generating were developed by FHWA and are being
used by FHWA and their local administration to monitor their performance over time.
Several other of the respondents indicated that the measures they are currently using have
evolved over time. As objectives of the control center changed or as new tasks and capabilities
were added, new performance measures were added or old ones have been modified to reflect the
new objectives of their system.
Interestingly, both of the emergency service providers that replied to the survey indicated that
they have been collecting performance measures that are standard for their industry. It appears
that these performance measures are used as a resource management tool for evaluating staffing
and asset allocations.
In an attempt to gain insight into other potential performance measures, each respondent was
asked if there were other performance measures that were not currently being generated by their
system, but would be desirable or helpful to analyzing the effectiveness of the incident response
in their area. Table 8 summarizes the responses obtained to these questions. For the most part,
agencies’ response fit into two categories. One group of agencies wants to generate more of the
traditional performance measure (such as incident frequencies, incident rates, detection time,
response time, etc.) while the other group wants to collect performance measures that relate to
administrative and institutional issues (such as operator workload, camera utilization by other
entities, web page hits, etc.). Most agencies, however, basically agree that better quality of data
needs to be entered into their systems to make the performance measures more meaningful.
41
Table 7. Origin of Operational Definition for Performance Measures Being Used
Agency
How were these operational definitions derived? By whom? What was the process for
deriving them? Were other agencies involved? If so who were they and how?
New Jersey DOT
Derived over time, FHWA and management of traffic operations at DOT have asked for it
The software developers were in-house. They actually asked the operators what they wanted. We found
out what management wanted, and told the developers how we wanted to amass the data. We kept the
screens simple and eliminated the garbage as we found we didn't use or management didn't need what the
screen or a button was offering. We also deleted things that would not work (Emergency notification
systems). Driven by available funds.
City of Phoenix Fire Labor management committee that deals with performance measures (3 union officers; 3 fire dept.
Dept
managers; shift commanders, exec. office). 1960's. Devised definitions for measures and guides,
reviewed annually
Maryland State Hwy
Work w/ FHWA over years, standard definitions
Admin - CHART
Texas DOT - Austin Developed by Traffic Operation Divisions at Headquarters
Arizona DOT
Minnesota DOT -Minneapolis
Southeast Michigan
COG – Detroit
Houston-Motorist
Assistance Patrols
Connecticut DOT
Look at data recorded to see what information can be tracked over time. Looking for trends that can be
addressed (e.g. Highway Helpers)
By SEMCOG and the Metro Detroit Incident Coordinating Committee
We are a police agency. We follow normal police data gathering according to our Department SOP
General knowledge from other agencies thru 1-95 Corridor Coalition
Table 8. Other Performance Measures Not Currently Being Collected, but Desirable
Agency
New Jersey DOT
Are there other performance measures that you are not collecting but think would be
beneficial?
Incident frequency, rate, secondary accidents, and incident delay
Tennessee DOT
Interfacing w/ police records ==> high incident rates, commuter times/speeds
Maryland State Hwy
Admin. – Chart
Texas DOT - Austin
Balance of operator workload; tow response to scene
Institutional issues ==> camera control (other agencies causing problems); web page hits (how
many people looking at cameras)
Texas DOT – San Antonio
Travel times; partial restoring of capacity (i.e., when lanes where opened)
Minnesota DOT -Minneapolis
Southeast Michigan
COG – Detroit
Better quality of information
Houston- Motorist
Assistance Patrols
New York DOT
No
Illinois DOT -- Chicago
City of Houston, TX Police
Department
Haven't really given it much thought only because we are focused on making the data better (more
accurate). For example, a call may be taken and dispatched but the officer can't locate any incident
so instead of clearing the call the record is left with no clear time or any explanation as to why the
data is missing.
Would like to collect response time, clearance time, resumption of normal flow, and times
individual lanes were open/closed. Got an estimate of $100K to upgrade MIST for these add-ons - not being pursued right now.
Detection time -- improving *999 and CCTV; Response time -- collecting data to calculate
response time but not aware of it being used.
Clearance time
42
Costs of Generating Performance Measures
One objective of this task order was to capture information about the costs associated with
collecting, processing, and reporting performance measures for incident management systems
around the United States. Almost all of the responding agencies indicated that it was impossible
to separate the costs of producing performance measure reports from their typical operating
costs. For the most part, agencies consider the cost of collecting data for producing performance
measures and performance measure reports as part of their normal operations, and the costs
associated with producing special performance reports (such as those requested on demand) are
included as part of their normal operating budgets. Table 9 summarizes a few of the responses
received from individuals when questioned about the issue of costs.
Table 9. Estimated Cost for Collecting, Processing, and Reporting Performance Measures
Agency
Arizona DOT
Maryland State
Hwy. Admin –
CHART
Caltrans – San
Diego
What would your estimate of cost to be for collecting, processing, and reporting you
performance measures?
The cost to set up the decision, notification, data collection system that is used for this was part
of the AzTech funding.
Contract with University for performance measures
Not a way to separate costs for this specific function
Incident Management Performance Reports
The respondents were also surveyed as to the type, frequency, and use of reports they produced
that documented the performance of their incident management systems. These responses can
be found in Table 10 through 13.
Only eight of the responding agencies indicated that they routinely produce reports so they could
monitor the performance of their incident management systems over time. Most of these
agencies are reporting their performance measures on a system-wide basis. Five of the agencies
also indicated that they routinely produce performance reports by roadway segment, and by
facility as well. Many of the agencies reported that their software/data management systems are
flexible enough to generate performance measure reports at any level.
Table 11 shows the frequency at which the responding agencies produce performance reports
while Table 12 summarizes the uses of the performance reports. The frequency at which
agencies produce performance reports varies greatly and seems to be a function of their use.
Almost all of the transportation agencies that responded indicated that they produce performance
reports on a monthly or quarterly basis. Monthly reports are generally used by the operations
staff to track use of resources and include such information as the number and type of incidents,
the type of responses (or assistance), the devices and/or resources used to manage the incident,
the schedules of staff, and the high incident locations. Mid-level administrative staff generally
use quarterly reports to assist in the coordination of incident responses across institutional and/or
jurisdictional boundaries.
43
Both of the fire and police agencies that responded to the survey indicated that they generally
produce daily reports of the “incidents” (not just those related to traffic operations) that they
work. Watch commanders generally use these reports to assess the workload and readiness of
the various units to respond to other types of incidents.
Table 10. Aggregation Level of Performance Reports
By
By
SystemOther
Agency
Facility Segment Wide
Kansas DOT –
Kansas City
Accident frequency can be on any of these levels
New Jersey DOT
x
Arizona DOT
x (1)
Ohio DOT –
Columbus
Tennessee DOT
City of Phoenix, AZ
x
x
x
Fire Department
Maryland State
Hwy. Admin –
CHART
x
x
x
Upon request
Texas DOT -Austin
x
x
x
Monthly reports on LCU failures; communications errors
Texas DOT –San
Antonio
Minnesota DOTMinneapolis
-
-
-
Everytime something is changed, system documents time;
therefore, have complete "history" of response
-
-
-
By responder on monthly basis; also produce annual
crash/volume report, by location
By incident
Caltrans –San Diego
Southeast Michigan
x
x
x
COG – Detroit
Houston, Tx –
Motorist Assistance
x
x
x
Patrols
New York DOT
Colorado DOT –
Lakewood
Texas DOT –
Houston
Illinois DOT –
x
Chicago
City of Houston, TX
Police Department
North Carolina DOT
Connecticut DOT
(1) Think they are generated system-wide, but know they are grouped by Districts and ORGS (small operating
units). Districts then examine the reports specific for their area.
44
Table 11. Frequency at Which Performance Measures Reported
Agency
How often are they produced?
New Jersey DOT Monthly
Arizona DOT
Quarterly
City of Phoenix,
AZ - Fire Dept. Daily (Captain gets his last shift & last shift before he arrived)
Maryland State Monthly-- # of incidents by reg; assists; use of devices (monthly meetings); Annually -- big
Hwy. Admin – picture by University, legislature, other agencies
CHART
Texas DOT –
Austin
Quarterly
Texas DOT –
As Needed basis -- have done 2 system wide evaluations; also use on-line survey on homepage to
San Antonio
gauge motorist responses (subjective)
Minnesota DOT–
Minneapolis
Monthly and yearly -- incidents by type and response; special days (e.g., snow days)
Caltrans-San
Diego
As needed basis -- some annual (accidents); monthly -- for meeting purposes
Southeast
Monthly (for operators); quarterly (coordinating committee); and annually (program evaluation)
Michigan COG –
Detroit
Houston,Tx Motorist
Assistance Patrol Quarterly
Colorado DOT–
Lakewood
Monthly
Illinois DOT –
Chicago
Annually
City of Houston,
TX Police Dept Daily; monthly
Connecticut
DOT
As needed basis; monthly
All of the agencies indicated that they also produce annual reports for their systems. These
annual reports generally provide an overall summary of the performance of the system and give a
“big picture” view of the effectiveness of the system. High-level administrators typically use
these annual reports to provide justification for continued operation or expansion of their
incident management programs. These reports are also used to identify high incident or “hot
spot” locations.
Several agencies indicated that they would occasionally produce performance measure reports on
individual or specific incidents. These reports are generally produced on an “as needed” basis
and are used to critique the performance of the response agencies and to address problems with
the responses to specific incidents. Generally, transportation agencies use these reports as a
mechanism for improving coordination between response agencies.
45
Table 12. Uses for Performance Measure Reports
Agency
How are these measures generally used in your system?
New Jersey DOT Feds look at it, not really used by DOT though
City of Phoenix, 1) Response planning; 2) Budget planning; 3) Quality Assurance (10% detailed check); 4) Internal
AZ – Fire Dept. Assessment - by command officers, mostly fire side
Maryland State
Hwy Admin –
To get funding (big picture report); identify "hot spots"
CHART
Texas DOT –
Access queries through Sybase
Austin
Minnesota DOT– Generally tracking trends; in past month or two started generating reports to track operators; use
w/ media for political support
Minneapolis
Caltrans—
Automatically by the system software
San Diego
Southeast
They are provided to the Incident Management Coordinating Committee, MDOT, and the FCP
Michigan COG – operators. They are also provided to the MSP, as requested, for selective enforcement. MDOT
Detroit
uses the information for determining the benefit of the FCP program and to obtain additional
funding for expansion.
Colorado DOT –
Statistics, program justification
Lakewood
Illinois DOT – Incident frequency/rate used in justification of service patrol, used to determine locations for
Chicago
safety improvements
City of Houston,
Not sure how they are used
TX Police Dept
Connecticut
Can be used to evaluate staffing schedules, determine high accident locations, and evaluate
DOT
effective response time and performance.
Arizona DOT
We use them to prove we are achieving our goals
Texas DOT –
Justify giving less money to ITS
San Antonio
Houston, TX –
To determine success of program and deputy performance ratings.
Motorist
Assistance Patrol
Respondents were also asked to indicate whether they thought these performance reports were
timely, useful, and accurate. Table 13 summarizes these responses. While most of the
respondents generally felt the reports were timely and provided decision-makers with the
appropriate level of information they need, a few questioned the usefulness (particularly from the
viewpoint of the operators) and the accuracy of the information. Several respondents indicated
that they did not exactly know how the higher-level administrators in their agencies actually used
the information.
46
Table 13. Timeliness, Usefulness, and Accuracy of Incident Management Performance
Measures
In general, do you think the information in these reports
or the performance measures themselves to be …
Agency
Timely?
Yes
No(3)
Useful?
Yes(1)
Yes
Accurate?
Yes
No(4)
Provide the information
necessary for effective
decision-making?
New Jersey DOT
No(2)
Arizona DOT
Yes
City of Phoenix, AZ –
Fire Dept
Yes
Yes (5)
Yes
Yes
Maryland State Hwy
Admin – CHART
Yes
Yes
Yes
Yes
Texas DOT - Austin
No
Yes
Yes
Yes
Minnesota DOT–
Minneapolis
Yes
Yes (6)
No (7)
Yes
Caltrans-San Diego
Yes
Yes
Yes
Southeast Michigan
COG – Detroit
Yes
Yes
Yes
Yes
Houston, TX.
Motorist Assistance
Patrol
Yes
Yes
Yes
Yes
Colorado DOT–
Lakewood
Yes
Yes
Yes
Yes
Illinois DOT –
Chicago
Yes
Yes
Yes
Yes
City of Houston, TX
Police Dept.
Yes
Not sure
Yes
Not sure
Connecticut DOT
Yes
Yes
Yes
Yes
(1) Somewhat -- not enough “meat” to be really useful, just break down number of incidents over and under one
hour, by type, monthly average incident duration, etc.
(2) Don't know enough to capture enough
(3) Quarterly reports are up to 3 months behind today
(4) It depends on where you get the data -- somehow different people can find different numbers
(5) For targeted audience
(6) Over time
(7) Based on operators view - not as good as could be
Integration of Incident Records and Information
Agencies were also asked about the kinds of incident information other agencies kept and their
efforts to use this other information to supplement data used to develop incident management
performance measures. Their responses are summarized in Tables 14 and 15.
Although many agencies are aware of other sources of incident records (such as 911 dispatching
logs), relatively few agencies indicated that they routinely integrate response information about
incidents with other agencies (such as fire and police). Several agencies mentioned, however,
that efforts were underway in their areas to integrate police and fire computer-aided dispatching
(CAD) systems with their freeway management systems. These agencies anticipated that
47
integrating 911 CAD dispatching with their systems should greatly enhance response and recordkeeping capabilities.
Several agencies indicated that they do combine information (or harmonize information) with
police and/or emergency response agencies on an “as needed” basis. Generally, this involves
taking information for the transportation agency’s logs and matching them with information on
the police or fire incident report forms. In those few cases when this is done, it is generally done
as part of a debriefing effort between agencies after a major incident or as part of the preparation
for litigation. Generally, when this is done, agencies find the exercise to be fruitful in helping to
establish a timeline of response events to a specific incident, which, in turn allows them to more
readily identify problems or bottlenecks in the response process.
Issues Involved in Establishing an Incident Management System
Table 15 shows how various agencies responded to questions concerning the issues faced when
establishing an incident management system. Common issues cited include the following:
• Bringing agencies together to work in a coordinated and integrated fashion;
• Expanding the system to meet new objectives or added functionality with limited
resources;
• Being the “new guy on the block” and having to establish a good working relationship
with other response agencies;
• Providing consistent training for all agencies responsible for responding to incidents;
• Working with emergency services to strike a balance between providing a safe work
environment for responders and maintaining traffic flow past the incident;
• Maintaining security of the system and confidentiality of data without effecting
performance or response;
• Getting accurate information entered into databases without overburdening operators
with too many data entry screens;
• Asking operations centers to do too much with too little resources; and
• Involving private towing industry in development of system.
48
Table 14. Other Sources of Incident Information in Jurisdiction
Agency
Do other agencies (such as fire, police, DOT, etc.) keep similar information about incidents
in your jurisdiction?
State Police, Service Patrol
Kansas DOT–
Kansas City
New Jersey DOT Police and fire keep information like number of incidents, but only part of the same information
that the DOT collects
Arizona DOT
No. They cover different aspects of the incident
Ohio DOT –
Yes -- police, service patrol
Columbus
Tennessee DOT 911 center log - no interaction
City of Phoenix, Yes -- other fire departments in valley (outside jurisdiction)
AZ. – Fire Dept.
Maryland State Police and fire keep accident reports. All police reports go to DOT to look at for traditional
Hwy Admin –
statistics of accidents.
CHART
Texas DOT –
Have project to integrate ATMS with CAD system -- automatically generate reports -- operator
Austin
will verify incident
Texas DOT –
San Antonio
Police -- incident report on call, keep when they arrive on scene and when cleared; Fire -- own
method of notification, on file at district
Minnesota DOT– No. Now have CAD linked to State Patrol
Minneapolis
Caltrans – San No. Other do, but haven't tried to integrate
Diego
Southeast
Yes, I assume so but probably not to the degree SEMCOG does (with all the integrated data).
Michigan COG –
Detroit
Houston, TX. – Yes, TxDOT
Motorist
Assistance
Patrols
New York DOT State police use incident cards. Fire, EMS keeps records of dispatch, arrival, departure times
but no traffic incident information.
Colorado DOT
Texas DOT Houston
No
Please contact those agencies. Three law enforcement agencies, City and County Traffic and
METRO the local transit authority are also housed at TranStar. They have access to the incident
database as well as access to input data. To the best of our knowledge they do not do so.
State police, service patrol
Illinois DOT –
Chicago
City of Houston, Yes -- TxDOT, MAP
TX Police Dept
North Carolina Police reports
DOT
Connecticut
Yes
DOT
49
Table 15. Integration of Incident Information with Other Agencies
Agency
Kansas DOT –
Kansas City
New Jersey DOT
Arizona DOT
Ohio DOT –
Columbus
Do you integrate or
compare information
with other agencies?
No
If so, …..
When?
How Often?
How ?
-
-
-
What are generally your
findings when this
occurs?
-
Share information with Delaware regional planning organization, DOT planning unit for congestion management program
No. They cover different Partnering sessions
Quarterly
Given as a presentation
Does not change the state
aspects of the incident
between DPS and state
with report as supporting of how things are handled.
documentation
Haven't compared yet -- requested that information six months ago and just now receiving data from City of
Columbus public safety and police department to compare with service patrol, hope to show reduction in
accident rates due to service patrol and TMC
City of Phoenix,
AZ. – Fire Dept.
Yes
Maryland State
Hwy Admin –
CHART
Texas DOT –
Austin
Starting to look at this w/
police and 911 centers
Texas DOT – San
Antonio
Minnesota DOT–
Minneapolis
Hope to integrate with
Police CAD system
No. Now have CAD link
to State patrol
Yes
January
Annual formally;
informally more often
(phone)
-
As needed
As Needed
Accident reports w/
highway patrol MinnDOT
compare to State -- on as
needed basis
50
Across all 26 cities in
agreement, written copies
to chiefs
-
-
Hardcopy - TMT response Information similar -to specific incidents
similar time stamps, when
responders showed up on
scene. Records state
change in TCD response
-
-
-
-
-
Generally good. Lot of
incident not accidents. See
crashes that don’t have
accident reports. Stalls are
big incident source.
Table 15. Integration of Incident Information with Other Agencies
Agency
Caltrans – San
Diego
Do you integrate or
compare information
with other agencies?
Yes
Southeast Michigan Yes
COG – Detroit
New York DOT
Yes
Texas DOT Houston
City of Houston,
TX Police Dept.
North Carolina
DOT
-
No
Yes
If so, …..
When?
How Often?
How ?
For specific reason – may Infrequently, rare
debrief after major
incident; serve in court case
-
What are generally your
findings when this
occurs?
-
Whenever we can
-
Using GIS
Can find out from state
police (co located). Time
incident came in -- can use
to enter more accurate
detection time than time
stamp from MIST when
entered (for major
incidents)
-
-
May get CAD system in
future, be able to query
other agency activities.
-
-
Law enforcement does not
share information readily
with the DOT
-
-
-
-
Varies -- regular meeting in Monthly
areas to critique incident
management
51
Still being determined.
Meeting of interagency
Committee
-
Depends on area. Don't
want to point fingers in
area. Good information for
improving response.
Table 16. Issues Faced in Setting Up Incident Management System
Agency
What kinds of issues were faced when setting up the system and how were they resolved?
Kansas DOT– Current system is incident management manual. Manual is posted on website
Kansas City (www.kdot1.kfdot.org/public/kdot/kcmetro/kcindex). Website also includes press release, lane
closures, etc.; before, had problems with police/fire unnecessarily blocking lanes (e.g., fire block 2
lanes to extinguish brush fire, police not clearing lanes fast enough; before, multiple agencies may
respond to major incidents. No way to notify media, because each agency might want to use
different diversion route. Now 30 cities, 12 counties, 2 states cooperate, use incident manual Juanita
developed. She talked to each agency before developing manual to get input, then again after
created to explain need for prompt response and clearance. Manual has planned diversions for
specific locations, list of contacts, and also describes what agencies cover what, and when to notify
other agencies including other states and federal agencies. Manual is updated 2 times/year. All
agencies receive e-mail to notify of manual updates.
New Jersey
Have problems trying to expand. Feds are behind expansion 100 percent as is the MPO, but design
DOT
wants to spend money for paving, etc.
Arizona DOT We went from a Phoenix-only based operation to a statewide center. Created institutional barriers
within the state DOT as local employees started to handle statewide system issues. Financial
barriers were encountered in the form of communications needs. Operations were found to be nonuniform across the state. Training for the handling of incidents was found to be inconsistent.
Creation of standards for training.
Ohio DOT–
It is going to take some time to develop a real collaborative effort with all of us to understand that
Columbus
we work for the same employer -- the taxpayer. City police work real well on freeway, understand
the importance of quick removal of lane blocking incidents. Have problems with the fire
department blocking too many lanes (e.g., blocking three lanes for a one lane blocking incident).
Had a recent event where multiple units on the side of the freeway with the incident blocked extra
lanes. An additional fire unit arrived on the other side of the freeway and blocked the inside lane,
they were not needed but remained on scene in the vehicle. Police did not make them clear the area.
Have heard fire agencies in other areas act similarly, may need Washington to act to change. Need
better communication system between agencies, currently using cell phones.
Tennessee
DOT
They are the "new guy". Initially, had warm welcome at scene. Has greatly improved over years.
Quick clearance issues w/ fire dept. Trying to add this to fire training; Memorandum of
understanding with TennDOT and local
City of
System very old, built like snowball (began in 1945 with chiefs meeting and sharing; 1960 expanded
Phoenix, AZ – kept information; 1971 began paramedics; 1977 HAZMAT); At each expansion, obstacles were City
Fire Dept.
Manager asking why greater funds; labor sees this as extra added to their job -- collecting was a pain
-- automation has minimized this.
Maryland State Hard to get code that is user (operator) friendly from contractor (off-the-shelf) -- want to create
Hwy Admin – custom software
CHART
Texas DOT – How do we use the system -- when/how do we pull information from the system
Austin
Texas DOT – Security (keeping the system safe so someone can't corrupt the system) and confidentiality
San Antonio (displaying accidents without notify family, police need more detailed personal information than
traffic)
Caltrans – San Too much to do; too little resources
Diego
52
Table 16. Issues Faced in Setting Up Incident Management System
Agency
Houston, TX –
Motorist
Assistance
Patrols
Colorado DOT
– Lakewood
What kinds of issues were faced when setting up the system and how were they resolved?
Funding -- type of vehicles to use, type of services to offer; Funding -- created a public/private
partnership; Vehicles -- Carrying capacity and safety of vehicle; Services -- determined type of
incidents that might occur while driving.
Getting accurate information to database, increased training; Response/clearance times reduced now
through cooperation with police. DOT has provided police units with courtesy patrol radios, so
courtesy patrol can contact police directly from the scene if police involvement needed.
Texas DOT Houston
When the integrated incident management database was developed, input was requested of all
TranStar partner agencies. This included Law Enforcement and Transit. There were features
requested by Law Enforcement that have never been used because they choose not to get involved in
inputting data. However incorporating these features expanded the database GUI beyond what was
needed by TxDOT causing operators to have to sift through more functions than were required.
However, it was deemed that too much was better than too little.
IlDOT
Private towing industry complaints when starting up service patrol, those issues were ironed out
over time. Some opposition to using tax dollars for service patrol, but have showed that the peak
periods are shorter with the patrol than without. Been in the incident management business for 40
years, none of those guys left to talk to.
North Carolina Turf battles between agencies -- face-to-face talks
DOT
Most Important Things To Be Measured in Incident Management Program
As a final question in the survey, respondents were asked what were the most important things to
be measured in an incident management program, whether or not they were currently collecting
the particular performance measures. Their responses are contained in Table 17.
Almost all of the agencies agreed that monitoring time-related performance measures was
important for gauging the success of an incident management program. Important time-related
performance measures to the monitored include the following:
• Response time,
• Duration on scene,
• Clearance times, and
• Detection times.
Many also cited the need to have performance measures that relate to the quality of the service
being provided, or to quantify the ability of the system to monitor and effect a change in the
traffic control. Several performance measures that agencies mentioned along these lines include
the following:
• The amount of delay caused by incidents in the system;
• The road user costs associated with congestion caused by incidents;
• The reduction in the overall delay caused by incidents;
• The reduction in the total duration of the incident (how long lanes were blocked); and
• The reduction in driving time of the public through incident scenes.
53
Table 17. Most Important Thing to Measure in Incident Management Program
Agency
New Jersey DOT
Arizona DOT
Ohio DOT – Columbus
Tennessee DOT
In your opinion, what are the most important things to be measured, whether or not
you are currently collecting?
Delay caused by incidents; road user costs, B/C -- how incident duration is reduced by ITS
Notification, detection, response time, on-scene time, clear time, and closing of incident
It differs from urban area to urban area. The incident managers need to define their worst
enemy, e.g., Hazmat, roadway geometries, weather, etc. and collect data before and after
program implemented to show reduction in performance measures for program
justification.
Time of clearance -- moved to shoulder or exit; # of response units -- make sure isn't people
there that don't need to be
City of Phoenix, AZ –- Time related measures; quality (of performance) related measures; info to tie performance
Fire Dept.
to specific budget expenditures
Maryland State Hwy
Admin. – CHART
Texas DOT - Austin
More data you have, better off you are
Texas DOT – San
Antonio
Incident detection time; power of system that allows you to make changes in system; ability
of system to monitor system and recommend changes; quality of information (data) -direct impact on response; good PR program
Minnesota DOT –
Minneapolis
Response time; clearance time -- when they arrive, when they are out of lanes, and when
total clear; on-site measures to ensure scene safety
Caltrans – San Diego
Southeast Michigan
COG – Detroit
What decision-makers are doing; when is significant to people and decision-makers
Clear times, time it takes to return to free flow conditions, time and locations of
occurrences, location of abandoned vehicles
Response time; traffic control device changes; when response is provided, who/how many
need -- right now, we are more interested in did we do something, and not necessarily when
we did something; finding information and making sure public has access to it.
Houston, TX. Motorist Services offered, reduction in delays in driving time for the public due to traffic incidents
Assistance Patrols
New York DOT
Response time; clearance time; resumption to normal flow; times individual lanes
opened/closed; secondary accidents -- can reduce if get the work out quickly of existing
incidents
Texas DOT - Houston Accident: location, frequency, time of day, surface conditions; Detection: time, method
frequency; Response time; Clearance time; time required to dissipate the queue.
Quantitative differences in these areas by type of incident
Illinois DOT – Chicago Cause and effect of incident; Incident type vs. congestion factor; Will be upgrading
computers and software -- new database should improve information data collection and
reporting.
City of Houston, TX
Time incident occurred; location - street and intersection; response time; clearance time;
Police Dept.
lane closure information
North Carolina DOT
Incident duration; response by agencies; effectiveness of response
54
SECTION 4. SUMMARY AND RECOMMENDATIONS
SUMMARY OF FINDINGS
The following represents a summary of the major finding from research conducted as part of this
task order:
• Transportation agencies define incidents differently than emergency service providers.
Transportation agencies typically define an incident to be any unexpected event that
causes a temporary reduction in the traffic carrying ability (i.e., capacity) of a facility.
Emergency service providers use the word “incident” to describe any event to which they
have to respond, whether it is on the roadway or not. Usually these events involve
situations where there is the potential for loss of life, possible injuries, property damage,
or potential criminal activities.
• While the actual measures vary slightly from location to location and between agencies,
most transportation and emergency service providers are currently using performance
measures to assess how well their incident management systems are functioning.
• Both transportation and emergency response providers recognize the need for collecting
and storing information about incidents. Transportation agencies generally collect
information about all aspects of traffic incidents (such as the arrival and departure times
of all response vehicles). Emergency service providers generally collect information only
related to their agency (i.e., the response time of fire trucks to the incident scene).
• Transportation agencies generally use performance measure to quantify the effectiveness
of the overall incident management process, while emergency service providers generally
use the information as a resource management tool to justify additional staffing and
equipment.
• Most transportation agencies use the following measures to assess the performance of
their incident management systems:
• Number (or frequency) of incidents;
• Detection time;
• Response time; and
• Clearance time.
• For the most part, emergency service providers use “response time” and time spent on
scene. Measures such as the number of secondary incidents and the time to normal flow
are difficult to define and collect without using operator judgment.
• While most transportation agencies indicated that they define “detection time” as the time
differential between when an incident occurred and when it was first detected or reported
to any official response agency, most only record “detection time” as the time of day at
which the incident was reported to the TMC.
• Both transportation agencies and emergency service providers use “response time” as a
critical performance measure; however, the operational definition of this measure varies
significantly. Transportation agencies generally define “response time” as the time
differential between when an incident was reported to the TMC to when the first
responder from any official response agency arrived on-scene. Emergency service
55
•
•
•
•
•
•
•
providers generally define “response time” as the time differential between when a call
was received by their dispatcher to when their first response vehicle arrived on-scene.
The operational definition of “clearance time” also varies considerably between
transportation agencies and emergency service providers. Transportation agencies
typically define “clearance time” as the time differential between when the first
responders arrive on the scene to when the capacity of the facility has been fully restored
(i.e., when the incident has been removed from the travel lanes). Emergency service
providers define clearance time as the time when all or most of the response equipment is
again ready to respond to another event at another location.
Emergency service providers define incident duration (or total time spent at the scene) as
the time differential between when they first received a request for service (i.e., issued an
alarm) to when they have been cleared to leave an incident scene. Transportation
agencies generally define incident duration as the time from when a TMC is alerted of an
incident until when the incident has been cleared from the roadway.
The performance measures (and the way that they are defined) used by emergency
service providers are fairly standard across their industry. National reporting database
(such as the National Fire Incident Reporting System) have caused emergency service
providers to adopt common terminology and collect data in a consistent manner. For
transportation agencies, the type and manner in which performance measures are defined
are local decisions.
Many transportation agencies are currently producing performance reports routinely.
Reports are frequently produced on a monthly, quarterly, or annual basis. Mid-level
administrators are generally using monthly and quarterly reports to assist in managing
assets and resources. Higher-level administrators use annual reports.
While most agencies are willing to share incident information and performance measures
with other agencies, this is rarely done, except on an as needed basis to evaluate a
response or address a specific problem that has occurred at a particular incident.
At some locations, emergency service providers and transportation agencies are
beginning to work towards integrating dispatching and incident management recording
keeping systems. This should allow for more accurate and better quality data from which
to develop incident management performance measures.
Most transportation agencies use a combination of automated and paper-based systems to
gather performance measure data, but one common complaint about these systems is that
the quality of information in their databases needed to be improved significantly.
RECOMMENDATIONS
First, incident management officials need recognize that having a “one size fits all” approach for
incident management performance measures may not be possible. The same set of performance
measures that are used to evaluate the more routine types of traffic incidents (such as an twovehicle collision, or a stalled vehicle) cannot be used to assess the performance of the system
during complex, major events (such as a multiple vehicle collision involving multiple fatalities
and/or serious injuries with major structural damage). It is recommended, however, that all
agencies reconstruct and review the timeline of response events that occur with such incidents to
identify and resolve potential problems with the responses prior to another major event.
56
For the more “routine” type of incidents, there seems to be a need for two sets of performance
measures. The first set would be used to describe the overall effectiveness and responsiveness of
the incident management process in a region. Administrators in the various response agencies
could use this first set of performance measures to identify mechanisms for improving response
and coordination between agencies. This first set would include measures such as the following:
• Incident Notification Time – This would represent the time it takes for all the appropriate
response agencies to become aware of an incident. It would be computed by taking the
time differential between when the first detection/report of an incident to any agency
(whether it be fire, police, 911-dispatch, or TMC) to when the other response agencies
also receive notification of the incident. This performance measure would need to be
computed separately for each of the official response agencies.
• First-Responder Response Time – This would represent what many transportation
agencies and emergency service responders are calling “response time”. This
performance measure would be the time differential between the first report of an
incident to any agency to when the first official responder from any agency arrived on the
scene.
• Incident Assessment Time – This time would represent the duration it takes the first
responder to determine what needs to be done to clear the incident and when capacity of
the roadway is first partially restored. This performance measure would be defined as the
time differential between when the first responder arrived on the scene and when the first
action is taken to fully or partial restore capacity (for example, opening one previous
blocked lane of traffic).
• Total Blockage Duration – This time would represent the total amount of time that
freeway capacity is reduced. This performance measure would be defined as the time
differential between when the first responder arrived on the scene to when the freeway
capacity was fully restored (i.e., all lanes opened).
• Total Incident Duration – This time would represent the total amount of time that the
incident had an effect on traffic operations. This performance measure would be defined
as the time differential between when the event was first reported to any official response
agency until when the last official response vehicle left the scene.
Other statistics that agencies may want to collect include the following:
• The frequency (or percentage of total incidents) at which each official response agency
was the “first detector.”
• The frequency (or percentage of total incidents) at which each official response agency
was the “first responder.”
• The frequency (or percentage of total incidents) where capacity was partially restored.
• The frequency (or percentage of total incidents) at which each official response agency
was the last to leave the scene.
Obviously, this evaluation becomes more feasible and practical for locations where recording
keeping systems from all the response agencies are integrated and coordinated. Being able to
perform this type of analysis requires that the evaluator have the capabilities for constructing a
complete timeline across agencies for every incident. Recognizing its complexity, it is
recommended that this type of evaluation occur annually in most regions.
57
The other set of performance measures that agencies may want to consider collecting would be
those that are directly related to their own specific mission in the incident management process.
An example of this type of performance measure would include the “response time” that most
emergency service providers and service patrol operations are currently collecting. These types
of performance measures would be generally geared toward helping agencies track the use of
resource or to assess an agency’s performance towards a specific objective (i.e., the fire
department’s objective is to have a 3 minute response time to all alarms).
In most locations in the United States, the role of the transportation agencies (with the exception
of service patrols) is one of support and demand management. For the agency specific
performance measures, transportation agencies, and in particular TMCs, need to develop
objectives and performance measures that more directly related to their specific mission in the
incident response process. Examples of these types of performance measure might include the
following:
• The time lag between when an incident was reported to a TMC and when devices were
activated on the roadway;
• The average delay to motorists through an incident site;
• The average queue length associated with different incident types;
• The average amount of diversion generated by the traffic control devices used in
managing an incident.
How to actually measure these performance measures directly in the field and how they relate to
the objectives of a region’s incident management process is the subject of future research.
SUGGESTIONS FOR FUTURE RESEARCH
Historically, transportation research has focused on identifying techniques and strategies for
improving the “response” side of the equation (i.e., how do we detect incidents quicker, how can
we get police and fire agencies to respond quicker to incidents, how can we clear the incident
faster, etc.). While this reducing response times and restoring capacity is critical to managing an
incident, it is only half of the equation and, to a large degree, out of the direct control of the
transportation agency. While coordinating responses with emergency service providers is
essential and perhaps can provide the greatest order of magnitude reduction in congestion,
transportation agencies cannot assert much influence over how quickly emergency service
providers response and clear incidents. Because most of the response process is out of the
control of a transportation agency, we believe that the research emphasis needs to drift away
from looking at what transportation agencies can do to reduce detection and response times to
incidents and focus more on the harder questions of how incident management systems can be
used to influence the “demand” side of the equation. Examples of the types of questions that
need to be explored through additional research include the following:
• What are agencies trying to accomplish with their incident management systems? By
activating traffic control and motor information systems in response to incidents, what
kind of impact are agencies trying to affect on traffic operations? What are agencies
hoping to accomplish?
• How effective are the response techniques (the DMSs, the ramp metering system, the
lane control signals, etc.) at reducing the amount of delay caused by motorists,
58
•
•
encouraging diversion, etc.? How do agencies measure the effectiveness of these devices
and strategies in real-time?
How do we need to change our detection and surveillance systems to be able to measure
the effectiveness of our incident management strategies?
What are the incremental impacts of combining traffic control devices (e.g., lane control
signals coupled with DMS signs, the systematic use of ramp meters, etc.) during incident
conditions?
59
REFERENCES
1. Traffic Incident Management Handbook. Prepared for Federal Highway Administration,
Office of Travel Management. P.B. Farradyne. November 2000.
(http://www.itsdocs.fhwa.dot.gov/jpodocs/rept_mis/@9201!.pdf)
2. Transportation Performance Measures Web Page. Federal Highway Administration,
Operations Core Business Unit, Office of Travel Management.
http://www.ops.fhwa.dot.gov/Travel/Deployment_Task_Force/perf_measures.htm
3. Traffic Management Data Dictionary (TMDD) and Message Sets for External Traffic
Management Center Communications (MS/ETMCC) Website. Institute of Transportation
Engineers. (http://www.ite.org/tmdd)
4. Highway Capacity Manual. Transportation Research Board, National Research Council,
Washington, D.C., 2000.
5. Performance Evaluation of CHART — An Incident Management Program — in 1997.
Prepared by Dr. Gang-Lan Chang, Deepak Shrestha of the University of Maryland, College
Park and Jean Yves Point-Du-Jour of the State Highway Administration of Maryland. May
2000.
6. Summary of Incident Management Practices in the I-95 Coalition.
7. Chattanooga Urban Area Highway Incident Management Plan. Prepared by the
Chattanooga Urban Area Metropolitan Planning Organization and the Chattanooga-Hamilton
County Regional Planning Agency. (http://www.chcrpa.org)
8. Hazardous Materials Guide. Federal Emergency Management Agency, U.S. Fire
Administration. (http://www.usfa.fema.hazmat)
9. MDOT’s Freeway Courtesy Patrol 2001 Activities Report. Southeast Michigan Council of
Governments, Detrioit, MI. March 2002.
10. Guide To Developing Effective Standard Operating Procedures for Fire and EMS
Departments. Federal Emergency Management Agency, U.S. Fire Administration.
(http://usfa.fema.gov/downloads/pdf/publications/fa-197.pdf)
11. Houston Motorist Assistance Program. Quarterly Report. Period July 2000-September
2000. Texas Transportation Institute. Texas A&M University System, College Station, TX.
12. Personal communication with Nick Thompson, Minnesota Department of Transportation.
March 2002.
60
13. City of Austin-Austin Fire Department Website. (http://www.ci.austin.tx.us/fire/fdavg.htm)
14. International Association of Chief of Police Website. (http://www.theiacp.org/index.htm)
15. International Association of Public-Safety Communications Officials Website.
(http://www.apcointl.org)
16. Dispatch Monthly Magazine. (http://www.911dispatch.com/cad/cadsoftware.html).
17. K&K Computer Solutions Website. (http://kk-police.epagecity.com/site/
epage/2113_196.htm)
18. National Fire Incident Reporting System Website. Federal Emergency Management
Agency, U.S. Fire Administration. (http://www.usfa.fema.gov/dhtml/inside-usfa/nfirs.cfm)
19. Kansas City Incident Management Program Manual. Developed by the Kansas Department
of Transportation, and the Missouri Department of Transportation. Spring 1999.
61
APPENDIX A. RADIO DISPATCH CODES FROM SELECT LAW
ENFORCEMENT AND EMERGENCY RESPONSE AGENCIES
APPENDIX A-1. REVISED OFFICIAL APCO TEN SIGNALS
APPENDIX A-2. CALIFORNIA HIGHWAY PATROL RADIO CODES
APPENDIX A-3. DALLAS PD RADIO SIGNAL CODES
APPENDIX A-4. F.D.N.Y. RADIO CODES
62
A-1. REVISED OFFICIAL APCO TEN SIGNALS
10-1
10-2
10-3
10-4
10-5
10-6
10-7
10-8
10-9
10-10
10-11
10-12
10-13
10-14
10-15
10-16
10-17
10-18
10-19
10-20
10-21
10-22
10-23
10-24
10-25
10-26
10-27
10-28
10-29
10-30
10-31
10-32
10-33
10-34
10-35
10-36
10-37
10-38
10-39
10-40
10-41
10-42
10-43
10-44
10-45
10-46
10-47
10-48
10-49
10-50
10-51
10-52
10-53
10-54
10-55
10-56
10-57
10-58
10-59
10-60
10-61
10-62
10-63
10-64
10-65
10-66
10-67
10-68
10-69
10-70
10-71
Unable To Copy Re-Locate
Signals Good
Stop Transmitting
Acknowledgement
Relay
Busy Stand-By
Out Of Service
In Service
Repeat
Fight In Progress
Dog Case
Stand By (stop)
Weather & Road Report
Report Of Prowler
Civil Disturbance
Domestic Trouble
Meet Complainant
Complete Assgn. Quickly
Return To --------Location
Call----By Telephone
Disregard
Arrived At Scene
Assignment Completed
Report In Person To---Detaining Subject, Expid
Drivers License Info.
Vehicle Registration
Check Records For Want
Illegal Use Of Radio
Crime In Progress
Man With Gun
Emergency
Riot
Major Crime Alert
Correct Time
Inves. Susp. Vehicle
Stopping Susp. Vehicle
(give complete discript)
Urgent (light/siren)
Silent Run
Beginning Tour Of Duty
Ending Tour Of Duty
Information
Request Permission To Leave
Patrol---For--Animal Carcass In Road
10-72
10-73
10-74
10-75
10-76
10-77
10-78
10-79
10-82
10-84
10-85
10-87
10-88
10-90
10-91
10-93
10-94
10-96
10-98
10-99
Source: http://www.bearcat1.com/radioco.htm
63
Assist Motorist
Emerg. Road Repairs Needed
Traffic Standard Repair
Traffic Light Out
Traffic Accident-F, PI, PD
Wrecker Needed
Ambulance Needed
Road Blocked
Livestock On Highway
Intoxicated Driver
Intoxicated Person
Hit & Run--F, PI, Pd
Direct Traffic
Convoy Or Escort
Squad In Vicinity
Personnel In Area
Reply To Message
Prepare To Make Written Cpy.
Message For Local Del.
Net Message Assgn.
Message Cancellation
Clear To Read Net Msg.
Dispatch Information
Message Received
Fire Alarm
Advise Nature Of Fire(size, type,
contents of bldg.)
Report Progress On Fire
Smoke Report
Negative
In Contact With
En Route
ETA
Need Assistance
Notify Coroner
Reserve Lodging
If Meeting----Advise ETA
Will Be Late
Pick Up Checks For Dist.
Advise Telephone # Of---Bank Alarm
Unnecessary Use Of Radio
Blockade
Drag Racing
Mental Subject
Prison/Jail Break
Records Indicate Want/Stolen
A-2. CALIFORNIA HIGHWAY PATROL RADIO CODES
104 = MESSAGE RECEIVED
106 = BUSY
1013 = ADVISE ROAD OR WEATHER CONDITIONS
1014 = PROVIDE ESCORT
1020 = LOCATION REQUESTED
1021 = TELEPHONE _________
1022 = DISREGARD
1023 = STANDBY
1031 = ATTEMPTED SUICIDE
1039 = MESSAGE OR ITEM DELIVERED
1097 = ON SCENE
1098 = ASSIGNMENT COMPLETED
1110 = TAKE A REPORT
1124 = ABANDONED VEHICLE
1125 = TRAFFIC HAZARD
1126 = DISABLED VEHICLE
1141 = AMBULANCE REQUIRED
1142 = PARAMEDICS REQUIRED
1144 = POSSIBLE FATALITY
1148 = PROVIDE TRANSPORTATION
1166 = DEFECTIVE TRAFFIC SIGNALS
1179 = ACCIDENT - AMBULANCE RESPONDING
1180 = ACCIDENT - MAJOR INJURY
1181 = ACCIDENT - MINOR INJURY
1182 = ACCIDENT - PROPERTY DAMAGE
1184 = TRAFFIC CONTROL
1185 = TOW TRUCK REQUIRED OR REQUESTED
1186 = BOMB THREAT
1187 = BOMB FOUND
1198 = MEET
Source: http://cad.chp.ca.gov/body_glossary.htm
64
A-3. DALLAS PD RADIO SIGNAL CODES
Signal Description
------ ----------DH
Drug House
2
Witness
3
Hang up call
4
911 Hang up
6
Disturbance
6G
Random Gunfire
6X
Major Disturb.
7X
Major Accident
8
Drunk
9
Theft
11
Burglary
12
Burglar Alarm
13
Prowler
14
*Cutting
15
*ASSIST OFFICER
16
Injured Person
18
Fire Alarm
19
*Shooting
20
Robbery
21
Holdup Alarm
22
Animal Complaint
23
Parking Viol.
24
Abandonded Prop.
25
Criminal Assault
26
Missing Person
27
Dead Person
28
Sick Person
29
Open Building
30
Prisoner
31
Crim. Mischief
32
Sus. Person
33
Poisoning
|Signal Description
|------ ----------|34 Suicide
|35 *Emergency Blood Transfer
|36 Abandon Child
|37 Street Blockage
|38 Meet Complainant
|39 Racing, Speeding
|41-40 Kidnapping in Progress
|41 Felony
|42 Pursuit
|44 *Person in Danger
|50 Eat
|51 Coffee
|52 City Court
|53 County Court
|54 Escort
|55 Traffic Violation
|56 Out to Station
|57 Out to Garage
|58 Routine Investigation
|59 Follow-Up Investigation
|60 Special Assignment
|61 Foot Patrol
|62 Public Service
|63 Cover Element
|64 Radio Shop
|65 Use Telephone
|66 End Duty Tour
|67 Monitor Radio
|41-20 *Robbery in Progress
|41-25 *Criminal Assault in Progress
|41-40 *Kidnapping in Progress
* Automatically dispatched as a Code 3 call
---------------------------------------------------------Communication Codes are as follows:
Code 1:
Code 3:
Code 4:
Code 5:
Code 6:
Code 10:
Code 10C:
Code 10W:
Code 10X:
Normal Response (no lights or sirens)
Emergency Lights & Sirens
Disregard
En Route
Arrived
Known Offender
Known Dangerous Offender
Felony Warrant
Stolen Vehicle
Source: http://www.policescanner.com/dalcodes.html
65
A-4. F.D.N.Y. RADIO CODES
10-1
CALL YOUR QUARTERS OR OTHER UNIT
10-2
RETURN TO QUARTERS
10-3
CALL DISPATCHER BY TELEPHONE
10-4
ACKNOWLEDGEMENT
10-5
REPEAT MESSAGE
10-6
STAND BY
10-7
VERIFY ADDRESS/LOCATION
10-8
IN-SERVICE BY RADIO
This signal is also used when the AT/SP is not in service and a unit is leaving quarters.
CODE 1 Used only by a Division or Battalion to indicate it is in-service by radio when leaving a
quarters other than its own at which it had been off the air and to which alarms had
been routed.
CODE 2 Used by any unit to indicate it is on the air outside its response area. When the unit
returns to its response area, the dispatcher must be notified again using 10-8.
10-9
OFF THE AIR
A unit will be out of radio contact. (State the reason: entering tunnel; visiting quarters; at a fire or
emergency, etc.) A 10-8 is to be transmitted, with code if necessary, when contact is reestablished.
10-10
UNIT LOCATION
A request is made for a unit's location. The unit is to acknowledge by giving its present location.
10-11
RADIO TEST COUNT
10-12
PRELIMINARY REPORT
A request by a chief officer or dispatcher for a preliminary report by the first arriving unit. The
dispatcher shall relay the report to the responding units.
10-13
FIRE MARSHALL NEEDS ASSISTANCE
Urgent Assist by NYPD.
10-14
ROSTER STAFFED ENGINE COMPANY
Use by roster staffed Engine Companies when acknowledging a structural response.
10-18
RETURN ALL UNITS, EXCEPT 1 ENGINE AND 1 LADDER
Transmitted for a fire or emergency, when in the judgment of the officer in command, conditions
indicate that 1 Engine and 1 Ladder company are required. No further assistance is necessary. If
the required unit(s) has not arrived, they will be notified by the dispatcher to continue responding
to the location. Other responding units shall return to quarters or previous activity. The Battalion
Chief need not continue to the scene after a 10-18 has been transmitted unless he deems it
necessary. This decision must be based, in part, on the experience level of the officer transmitting
the signal. Battalion Chiefs are to notify the dispatcher that they are 10-8 when not continuing in
on a 10-18.
10-19
RETURN ALL UNITS, EXCEPT FOR 1 ENGINE OR LADDER
66
Transmitted for a fire or emergency, when in the judgment of the officer is command, conditions
indicate that 1 Engine or 1 Ladder Company is required. No further assistance is necessary. If the
required unit(s) has not arrived they will be notified by the dispatcher to continue responding to
the location. Other responding units shall return to quarters or other activity. The Battalion Chief
need not continue to the scene after a 10-19 has been transmitted unless he deems in necessary.
This decision must be based, in part, on the experience level of the officer transmitting the signal.
Battalion Chiefs are to notify the dispatcher that they are 10-8 when not continuing in on a 10-19.
10-20
PROCEED AT REDUCED SPEED
No warning devices are to be used and all traffic regulations are to be observed.
10-21
BRUSH FIRE
10-22
OUTSIDE RUBBISH FIRE
10-23
ABANDON/DERELICT VEHICLE FIRE (ADV)
A fire in a vehicle which has no value other than salvage and no owner can be located.
10-24
AUTO FIRE
A fire in a vehicle with plates or in any vehicle having a value greater than that of salvage.
10-25
MANHOLE or TRANSFORMER VAULT FIRE
CODE 1 Fire has extended from the manhole or conduit into a building.
CODE 2 Fire has blown one or more manhole covers, or smoke is issuing from a manhole under
pressure.
CODE 3 Smoke is seeping from a manhole.
10-26
FOOD ON STOVE
10-27
COMPACTOR FIRE
Fire has not extended from compactor or shaft
10-28
SUBWAY OR RAILROAD SYSTEM - FIRE, EMERGENCY OR SMOKE CONDITION
(CODE REQUIRED)
CODE 1 NYCTA.
CODE 2 Other than NYCTA.
10-31
CLOGGED INCINERATOR
Fire has not extended from shaft.
10-32
DEFECTIVE OIL BURNER
Fire has not extended from fire box.
10-33
ODOR OF SMOKE
A smoke condition caused by a nearby working fire or fires such as barbecues, salamanders, etc.
10-34
SPRINKLER SYSTEM EMERGENCY
CODE 1 Defective sprinkler device or system (defective alarm valve, broken pipe, etc.)
CODE 2 Unwarranted sprinkler alarm. Not defective (surge in pressure, people working on
system, etc.)
CODE 3 Sprinkler has been activated by heat source not associated with an accidental fire.
10-35
ALARM SYSTEM EMERGENCY
Other than a sprinkler system.
CODE 1 Defective alarm device or system.
67
CODE 2
CODE 3
Unwarranted alarm. Not defective (accidentally activated by cigarette smoke; low
battery, etc.)
Recorded alarm.
10-36
AUTOMOBILE EMERGENCY
Any type of automobile accident or washdown of a fuel spill.
CODE 1 Washdown
CODE 2 No injury or washdown
CODE 3 Injury
CODE 4 Extrication
10-37
ASSIST CIVILIAN
First Aid or other call for assistance.
CODE 1 Victim deceased
CODE 2 Victim not breathing
CODE 3 Victim injured
10-38
CARBON MONOXIDE RESPONSE
CODE 1 DETECTOR ACTIVATION - Defective, low battery, unwarranted.
CODE 2 CO INCIDENT - Readings from 1 - 9 PPM.
CODE 3 CO EMERGENCY - Readings over 9 PPM.
CODE 4 No detector activation during incident or emergency.
10-40
GAS OR ELECTRICAL EMERGENCY
CODE 1 Gas Emergency. (Gas main leak; gas leak in structure; efective gas appliance; etc.)
CODE 2 Electrical emergency. (Wires down; sparking fixture; short circuit; etc.)
CODE 3 Water condition.
CODE 4 Steam leak.
NOTE: Do not use 10-40 where the emergency causes a structural fire.
10-41
SUSPICIOUS FIRE (CODE REQUIRED)
Fire Marshall investigation is required.
CODE 1 Occupied Structure or Vehicle. A Structure (commercial, residential, public), or
vehicle (car, bus or train) which is occupied at the time of the fire. This also includes a
vacant apartment in an occupied building, or a store with a dwelling above.
CODE 2 Unoccupied Structure. A structure (commercial, residential, public) normally occupied
which is unoccupied at the time of the fire.
CODE 3 Unoccupied Vehicle. A vehicle (car, bus or train) with or without plates which is
unoccupied a the time of the fire.
CODE 4 Vacant Structure. A vacant building or building under construction or demolition. (If
there are squatters or workmen in structure, transmit a CODE 1.)
10-44
PUBLIC AMBULANCE
A request for a public ambulance. Specify the reason.
10-45
D.O.A. OR SERIOUS INJURY
Transmitted IMMEDIATELY upon the discovery of a fatality or serious injury at a fire or
emergency. This shall be followed as soon as possible with the appropriate Code and the number
of victims.
CODE 1 Victim Deceased
CODE 2 Victim suffering serious injury. (Apparently life threatening)
CODE 3 Victim suffering serious injury. (Apparently NOT life threatening.)
NOTE: Do not transmit this signal for minor injuries.
68
10-47
POLICE RESPONSE
Police assistance is needed for crowd or traffic control, security, apprehension, etc. (specify
reason)
10-48
POLICE RESPONSE FOR HARASSMENT
Firefighters are being harassed an police assistance is needed immediately.
10-51
CANCELLATION OF OUTSIDE ACTIVITIES
Transmitted when outside activities are to be canceled. When the conditions that caused
suspension of outside activities, the following message will be transmitted by radio, voice alarm
and teleprinter: "All units shall resume outside activity forthwith."
10-60
MAJOR EMERGENCY RESPONSE
Transmitted for a collapse, airplane crash (except airport crash boxes 0037-LaGuardia Airport, and
0269-Kennedy Airport) train derailment, or similar emergency with the potential for multiple
casualties. The following will respond. 3 Engines, 2 Ladders, 3 Rescue Companies (including
Res3cuE with the Collapse Unit) 5 Battalion Chiefs, 1 Tactical Service Unit, HazMat, Field
Communications Unit, Squad Company 1 with Technical Response Vehicle, 1 Deputy Chief, 1
Safety Battalion.
10-70
WATER RELAY REQUIRED
A notification that the first arriving engine has no positive water source and a water relay is
required.
10-75
NOTIFICATION OF A FIRE OR EMERGENCY
A notification signal transmitted when, in the judgment of the officer in command, conditions
indicate a fire or emergency that requires a total response of the following units: 4 Engines, 2
Ladders, 2 Battalion Chiefs, 1 Rescue Company and Squad Company. Officers transmitting a 1075 shall also state if it is for a fire or emergency and if a building is involved along with the type
of building.
10-76
NOTIFICATION OF A FIRE IN A HIGH-RISE BUILDING
A notification signal transmitted when, in the judgment of the officer in command, conditions
indicate a fire in a high-rise building that requires a total response of the following units: 5 engines
(1 is CFR), 5 Ladders (1 is fast), 3 Battalion Chiefs,Engine 3 and High Rise Unit, 1 Deputy Chief,
Field Communications Unit, 1 Rescue Company, Mask Service Unit, PIO, Command Post
Company, 1 Squad company, 1 Tac unit. (restricted use of Citywide frequency, for operation of
the High-Rise repeater.)
10-77
HIGH-RISE MULTIPLE DWELLING FIRE
High-rise multiple dwelling fire (Response of 5 engines, 5 ladders, 3 battalion chiefs, 1 deputy, 1
rescue, 1 squad, the Special Operations battalion chief, a safety coordinator, transmitted by the
Incident Commander after size up and the Safety Operating battalion chief.)
10-80
HAZARDOUS MATERIALS INCIDENT
The Initial notification by field units of a hazardous materials incident. Responders are to proceed
with caution to avoid entering a restricted area. Code will transmitted by the Incident Commander
after size up and evaluations of the incident.
CODE 1 An incident confined to a small area and which does not pose an immediate threat to
life or property. Can be controlled by a unit or units up to and including: 3 Engines, 2
Ladders, and 2 Battalion Chiefs.
CODE 2 An incident involving a greater hazard or larger area which posses a potential threat to
life or property. Following units will respond: 3 Engines, 2 Ladders, 2 Battalion
Chiefs, 1 Deputy Chief, HazMat Company 1, 1 Safety Battalion and Field
Communications Unit.
69
10-84
UNITS ARRIVED AT SCENE
All units shall IMMEDIATELY transmit a 10-84 by Radio or MDT when arriving at the box of
the incident. Other signals should not be given with the 10-84 except where the situation is
apparent on arrival. This signal must be followed within 5 minutes by a preliminary report
including the appropriate radio code signal or additional information by first arriving unit. Fire
commanders and Chiefs within their command are to take necessary steps to insure that units are
transmitting the 10-84 signal for responses. Chief officers shall transmit a 10-84 upon their arrival
at an alarm.
10-85
FIRE MARSHALL REQUIRES ADDITIONAL UNITS
Used for additional units (other than NYPD - See 10-13)
10-86
FOAM OPERATION
Transmitted for a fire or emergency requiring any type of foam concentrate in addition to that
carried by units on the scene. The following are to respond: 2 Foam Carriers, 1 Satellite Hose
Wagon, 1 Foam Coordinator (Batt.Chief).
CODE 1 Maximum amount of Flouroprotein foam required. In addition to units on the 10-86
the remainder of the Foam Carriers and all Bulk Foam Units will respond.
CODE 2 Flouroprotein foam required. In addition to units on 10-86 the remainder of the Foam
Carriers will respond.
CODE 3 High Expansion Foam required. The High Expansion Foam Unit will respond. Units
on 10-86 will also respond.
NOTE: On all Special Calls for foam (Foam carrier or High Expansion Foam) the associated
Engine Company will respond with both pieces of apparatus and all members.
10-91
EMERGENCY; FD NOT REQUIRED.
10-92
MALICIOUS FALSE ALARM
Indicates that a false alarm was transmitted with malicious intent.
10-99
UNITS WILL BE OPERATING FOR A LEAST 30 MINUTES
An operating unit or all units at an incident will be unavailable for at least 30 minutes. The unit(s)
is to state the reason it will not be available.
Source: http://www.nyfd.com/radio.html
70
APPENDIX B. SAMPLE PERFORMANCE MEASURE
REPORTS PRODUCED BY MINNESOTA DEPARTMENT
OF TRANSPORTATION
APPENDIX B - 1 . SAMPLE OF DAILY INCIDENT MANAGEMENT
PERFORMANCE REPORT USED BY MNDOT.
APPENDIX B- 2. SAMPLE OF MONTHLY INCIDENT MANAGEMENT
PERFORMANCE REPORT PRODUCED BY MNDOT.
APPENDIX B - 3. SAMPLE OF YEARLY INCIDENT MANAGEMENT
PERFORMANCE REPORT PRODUCED BY MNDOT.
71
B - 1 . SAMPLE OF DAILY INCIDENT MANAGEMENT PERFORMANCE
REPORT USED BY MNDOT.
72
B- 2. SAMPLE OF MONTHLY INCIDENT MANAGEMENT PERFORMANCE
REPORT PRODUCED BY MNDOT.
73
B – 2 (Continued). SAMPLE OF MONTHLY INCIDENT MANAGEMENT
PERFORMANCE REPORT PRODUCED BY MNDOT.
74
B - 3. SAMPLE OF YEARLY INCIDENT MANAGEMENT PERFORMANCE
REPORT PRODUCED BY MNDOT.
75
B - 3 (CONTINUED). SAMPLE OF YEARLY INCIDENT MANAGEMENT
PERFORMANCE REPORT PRODUCED BY MNDOT.
76
B - 3 (CONTINUED). SAMPLE OF YEARLY INCIDENT MANAGEMENT
PERFORMANCE REPORT PRODUCED BY MNDOT.
77
APPENDIX C. SAMPLE PERFORMANCE MEASURE
REPORTS PRODUCED BY MOTORIST ASSISTANCE
PROGRAM IN HOUSTON, TX.
78
79
80
81
82
83
APPENDIX D. INCIDENT MANAGEMENT
PERFORMANCE MEASURE SURVEY
84
INCIDENT MANAGEMENT PERFORMANCE MEASURES
AGENCY SURVEY
Contact Person:________________________
Telephone Number: _________________________
Agency: ______________________________
Date/Time of Survey: ________________________
Position:_______________________________
Fax Number__
_____ ________________
Duties related to the system: (operations, management, etc.)_____________________________________
Hello. My name is _______________________ and I am with the Texas Transportation Institute.
We are currently working on a project for the Federal Highway Administration dealing with performance
measures for incident management systems. The purpose of this project is to obtain a better understanding
of how agencies measure the performance of their organized incident management systems, and to identify
the difference, if any, in the definitions of relevant measures of performance of their incident management
systems. As part of this project, we are conducting a survey of several locations in the United States that
have active incident management programs and I would like to ask you to participate in this survey.
I have a series of questions that I would like to ask you concerning how you measure the performance of
your systems and how these performance measures are generated. The survey takes about 20-30 minutes to
complete. Some of the questions have predefined responses while others are open-ended. We used
predefined responses in some questions only to speed up the data collection process. If one or more of the
predefined responses does not fit your situation, please feel free to add others. Occasionally, I may ask you
some follow-up questions so that I’m sure I understand your response.
Again, the survey takes about 20-30 minutes to complete. Is now a convenient time or would you prefer
that I call you back at a later time?
Call back
When? (set date and time)_____________________________________
85
1. DEFINITIONS -- In looking at the literature, it appears that different agencies
define what an incident is differently. In the first series of questions, we are trying to
understand how different agencies define incidents and how this might effect their
response.
1.1.
From your agencies perspectives, what events affecting traffic does you agency
define as an “incident”?
Collisions
Vehicle on Fire
Overturned vehicles
HAZMAT Spill
Stalled/Disabled vehicle in a travel
Abandoned vehicle on shoulder
lane
Public Emergency
Abandoned vehicle in a travel lane
Debris on roadway
Stalled vehicle on the shoulder
All the above
Any others? (please identify)
1.2.
Does your agency have a system for classifying incidents?
No
GO TO SECTION 2
Yes
1.3.
What criterion is (are) used (e.g., severity, duration of blockages, etc.)?
1.4.
What are the thresholds for each classification level?
1.5.
How is this classification system used? In other words, how does your response differ
based upon the classification of the incident?
86
2. INFORMATION COLLECTED PER INCIDENT – Different agencies and different
systems collect incident data differently. With these questions, we are trying to get a handle
on what information about incidents different agency collect, how they do it, how long they
keep incident information, etc.
2.1.
Does your agency keep a permanent or semi-permanent log of events for each type of
incident?
No. Why not?
GO TO SECTION 3!
Yes
Continue below
2.2.
What information is collected about each incident?
Roadway Name
Location/Cross –Street Name
Block Number
Detector Station #
Geographic Location (lat/long)
Location of Lanes Blocked
Incident Type
Incident Source (Detected by system or Reported by cell phone, courtesy patrol, etc.)
The current status of the incident i.e., whether it has been Detected, Verified, Canceled,
etc.)
Time incident was detected
Time incident was verified
Source of incident verification
Time response vehicles arrived on scene (Do you record each individual vehicle arrivals
or collectively?)
Type of response vehicles on scene
Time response vehicles left scene
Time incident was cleared from scene (What is your definition of clearance – moved to
shoulder, response vehicles departs, removed from roadway altogether, other?
Time traffic returned to normal flow
Roadway Surface Condition
Roadway Condition (Wet, Dry, etc)
Light Condition (Daylight, Nighttime, Dawn, Dusk, etc.)
Weather Conditions
Injuries Present
# of Vehicle Involved
Type of Vehicle Involved
Incident severity (qualitative)
Others (Please Specify)
87
2.3.
How is this information collected?
Manual forms -- Can I get a copy of your incident logging forms?
Automatically through freeway management software -- Can I get a screen capture of
your logging screen?
Other:
2.4.
In what format is this information stored (paper file, electronic file, queriable database)?
2.5.
How long to you generally retain this information?
2.6.
Are other sources of incident information ever integrated with yours to cross-reference or
verify your information (i.e. police logs, accident reports, courtesy patrol records, etc.)?
If so, what sources?
2.7.
What would you estimate the cost to be for collecting, processing, and reporting your
incident measures?
88
3. PERFORMANCE MEASURES
3.1.
Do you calculate different performance measures from the information you routinely
collect about each incident (e.g., incident duration, response times, etc.)?
Yes
Continue Below
No
Why not?
3.2.
What measures do you routinely compute to assess the performance of your incident
management program?
Incident Frequency
Incident Rate
Detection Time
Response Time
Clearance Time
Number of Secondary Incidents
Time to Normal Flow
Incident Delay
Others:
3.3.
What are your operational definitions for each performance measure (i.e., when does the
clock start and stop for each performance measure)
Incident Frequency
Incident Rate
Detection Time
Response Time
Clearance Time
89
Number of Secondary Incidents
Time to Normal Flow
Incident Delay
Others:
3.4.
How are these reports generated?
By facility
System Wide
By Segment
Other:
3.5.
How were these operational definitions derived? By whom? What was the process for
deriving them? Were other agencies involved? If so, who were they and how?
3.6.
Are there other performance measures that you are not collecting, but you think would be
beneficial for you to know as they relate to the performance of your incident
management system? If so, what are they and how would you measure it?
90
3.7.
How long have you been collecting and calculating these performance measures?
3.8.
What would you estimate the cost to be for collecting, processing, and reporting your
incident mgmt. measures?
91
4. USE OF PERFORMANCE MEASURES
4.1.
Do you commonly generate any reports, tables, summary statistics, etc. that use these
performance measures?
Yes
Request Copy of typical report and continue
No. Do you have any plans?
o No
GO TO SECTION 5!
o Yes
Continue below
What kinds of reports/tables/summary statistics?
4.2.
When do expect to start producing them?
4.3.
How are you planning to produce them?
4.4.
Why are you going to start producing them?
4.5.
How are these performance measures generally used in your system?
4.6.
How often are they produced?
On an as needed basis
Daily
Weekly
Bi-weekly
Other (Please specify)
Monthly
Quarterly
Semi-Annually
Annually
92
4.7.
With whom are these performance measures shared (within agency, other agencies,
public)?
4.8.
How does your agency use the information in these reports? What decisions are made
based on or are influenced by these measures?
4.9.
In general, do you think the information in these reports or the performance measures
themselves to be:
4.9.1. Timely
Yes
No. Why?
4.9.2. Useful
Yes
No. Why?
4.9.3. Accurate
Yes
No. Why?
4.9.4. Provide the information necessary for effective decision-making?
Yes
No. Why?
93
5. INSTITUTIONAL ISSUES
5.1.
Do other agencies (such as fire, police, DOT, etc.) keep similar information about
incidents in your jurisdiction?
5.2.
Do you integrate or compare your information with other agencies?
When?
How often?
How?
5.3.
What are generally your findings when this occurs?
5.4.
What kind of issues did you face when you set up your system and how did you deal with
them?
5.5.
In your opinion, what are the most important things to be measuring, whether or not you
currently collecting?
94
6. CONTACTS IN OTHER AGENGIES
As part of this project, we would also like to ask these same questions to other agencies that
are active in your incident management program. Would it be possible for you to give me
the name and telephone number of your contacts in the other agencies that participate in local
incident management program?
STATE DOT:
CITY DOT(s):
POLICE:
FIRE:
EMS:
Others:
95
Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Download PDF

advertisement