apa

See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/220579266

Information Flow During Crisis Management: Challenges to Coordination in

the Emergency Operations Center

Article  in  Cognition Technology and Work · March 2007

DOI: 10.1007/s10111-006-0059-3 · Source: DBLP

CITATIONS

73 READS

1,258

4 authors, including:

Some of the authors of this publication are also working on these related projects:

Human-in-the-Loop Simulation Studies View project

Decision support for diagnostic radiologic imaging and whole slide imaging in pathology View project

Laura G Militello

Applied Decision Science, LLC

103 PUBLICATIONS   1,517 CITATIONS   

SEE PROFILE

Emily S Patterson

The Ohio State University

226 PUBLICATIONS   4,192 CITATIONS   

SEE PROFILE

Robert Wears

University of Florida

284 PUBLICATIONS   8,300 CITATIONS   

SEE PROFILE

All content following this page was uploaded by Emily S Patterson on 16 May 2014.

The user has requested enhancement of the downloaded file.

O R I G I N A L A R T I C L E

Information flow during crisis management: challenges to coordination in the emergency operations center

Laura G. Militello Æ Emily S. Patterson Æ Lynn Bowman Æ Robert Wears

Received: 1 October 2005 / Accepted: 1 October 2006 / Published online: 12 December 2006 � Springer-Verlag London Limited 2006

Abstract This paper discusses challenges to coordina-

tion in the emergency operations center (EOC). A

county-level EOC is made up of representatives from a

range of organizations including local government, fire,

police, hospital, utility, and Red Cross representatives.

These ad hoc teams are tasked with coming together

during an emergency to obtain and deliver resources to

first responders on the scene of the disaster, as well

coordinating transportation of casualties, tracking of

fatalities, and establishment of shelters. Two county-level

exercises were observed. Themes were identified across

the two exercises. Recommendations for better sup-

porting coordination in crisis management are offered.

Keywords Emergency operations center � Crisis management � Team decision making � Team coordination � Cross agency collaboration

1 Introduction

When studying large-scale coordination, considerable

attention is given to the topic of distributed teams and

the coordination difficulties they face, focusing on

issues such as team cognition, shared mental models,

and shared situation awareness (Salas and Fiore 2004;

Endsley and Garland 2000). This paper focuses on an

individual component of large-scale coordination: the

small command and control (C2) cell tasked with the

coordination of a large-scale response to an emergency

situation. The emergency operations center (EOC) is

an ad hoc team that assembles in a central location

during an emergency to support on-scene emergency

response personnel and coordinate the response across

incident command centers, with the media, and

through EOCs at different levels of government. EOCs

exist at all levels of government, from federal to indi-

vidual municipalities. Membership comprises govern-

ment representatives, private companies (e.g., hospital

administrators, utility companies, etc.), non-govern-

mental organizations (e.g., Red Cross), and sometimes

military representatives.

Challenges associated with coordinating crisis re-

sponse are well documented via lessons learned from

real-world disasters (The 9/11 Commission 2004; Au-

brey 2002; McEntire 2002; Perry 2003; Tierney et al.

2001). Much can be learned from these retrospective

analyses. This research is intended to build on the

existing knowledge base, using ethnographic methods

to examine team coordination in disaster exercises.

Exercises offer natural opportunities for observation,

occurring more frequently than real-world disasters,

thereby offering a rich data source.

2 Methods

County-level EOC exercises were observed in two

counties in Ohio. Both were functional exercises in

L. G. Militello (&) � L. Bowman University of Dayton Research Institute, 300 College Park, Dayton, OH 45469-0158, USA e-mail: [email protected]

E. S. Patterson The Ohio State University, 210 Baker Systems, 1971 Neil Avenue, Columbus, OH 43210, USA

R. Wears University of Florida, 655 W 8th Street, Jacksonville, FL 32209, USA

123

Cogn Tech Work (2007) 9:25–31

DOI 10.1007/s10111-006-0059-3

which actual EOC representatives gathered in the

county EOC facility to respond to a simulated disaster.

Exercise control cells were used to simulate calls to and

from incident command posts, the media, hospitals,

etc. A brief description of each exercise is included

below.

2.1 Exercise 1: tornado

The scenario used in this exercise included a tornado

that touched down twice in a largely urban and sub-

urban county. As the exercise began, participants were

told that they had been asked to report to the EOC

because of reports of severe weather with the possi-

bility of tornados in the area. Over the next 2 h, par-

ticipants received and placed calls to obtain

information about the unfolding situation and to field

resource requests from various emergency responders

at the scene as well as other emergency support orga-

nizations (simulated in the emergency control cell). As

EOC participants learned of a tornado touching down

in the local area, key events began to unfold, including

the collapse of a Sears department store in a large

shopping mall resulting in seven fatalities. An ele-

mentary school several miles away also collapsed,

trapping 72 children. Debris in the roadways sur-

rounding the school frustrated efforts to begin search

and rescue or even adequately assess the situation at

the school. Other events posed more challenges to the

EOC participants, such as a damaged veterinary clinic

resulting in the release of 50 dogs into the area, as well

as mobs of concerned parents at local hospitals and the

elementary school. The magnitude of this event re-

sulted in requests for National Guard assistance and

additional medical support from the state.

2.2 Exercise 2: train derailment

The second exercise simulated a terrorist attack in a

largely rural county. The scenario began with a loud

explosion outside a nearby plant. As the scenario un-

folded, it became evident that a bomb had been placed

on railroad tracks leading to the derailment of two rail

cars carrying hazardous material. Flying debris caused

a secondary incident involving a tanker truck also

carrying a hazardous substance. Bystanders were in-

jured by flying shrapnel, and a school bus was struck by

debris, killing the driver and a student. The need to

barricade a 1-mile perimeter, decontaminate those

exposed to hydrochloric acid from the rail cars, contain

the hydrogen gas released from the tanker truck, and

transport and treat the injured represented a significant

challenge to the rural county. As a result, additional

fire and police support were requested and mutual

support agreements with outlying hospitals were in-

voked. Resources such as cranes and barricades were

located, rented/hired, transported, and managed.

2.3 Participants

In both exercises, participants were recruited via a

request to local EOC representatives inviting them to

participate in a training exercise. Twenty-three people

participated in the tornado exercise, and 15 people

participated in the train derailment exercise. In both

exercises participants included an EOC director, hos-

pital coordinator, coroner, county engineer, and public

information officer (PIO), as well as representatives

from fire, police, and county administrative offices. The

tornado exercise also included animal control, SBC

911, a local bus company, and utility companies. The

train derailment exercise included a representative

from the local school district, the environmental pro-

tection agency, and an amateur radio organization. In

both exercises, the EOC representatives had little if

any previous experience working with their fellow

EOC colleagues.

2.4 Facility description

In both exercises, participants assembled in a dedicated

EOC facility, consisting of a room with large tables and

shared displays on the walls. The tables contained

nameplates and workstations assigned to each partici-

pant. Depending on the participant role, the worksta-

tion included a computer and a radio. All workstations

were equipped with telephones. Shared displays in-

cluded large-screen monitors for displaying television

news and projected computer images (i.e., maps).

There were also whiteboards mounted on walls to aid

in tracking events and key information (i.e., fatalities,

injured people, etc.). In the train derailment exercise, a

paper map was mounted on the wall and annotated

with post-it notes.

2.5 Data collection

The primary data source was observation data. During

the tornado exercise, four observers were present

throughout the exercise. Each focused on a particular

portion of the room. All were asked to note:

– The roles and responsibilities that the participants

took.

– The key decisions that the participants made, and

how they came to these decisions.

26 Cogn Tech Work (2007) 9:25–31

123

– The information that the participants required but

did not have, and how they obtained this informa-

tion.

– The type of information that the participants needed

to disseminate, and the organizations that they sent

the information to.

– The issues that were unresolved at the end of the

exercise.

– Any difficulty or unique challenge in coordination or

information flow.

Time hacks were used so that exercise notes could

be integrated across observers. Due to space con-

straints in the train derailment exercise, only one ob-

server was allowed in the EOC. A second observer was

located in the exercise control cell. Part way through

the exercise, the observers switched locations so that

each observed first-hand the EOC and the exercise

control cell.

Handwritten notes were typed and shared across the

research team for analysis. A series of meetings to

identify and discuss themes that emerged from the

observations were held. These themes allowed con-

clusions to be drawn and recommendations to be made

to aid in future cross agency coordination.

3 Findings and implications

Observation data were examined across the two exer-

cises for common themes. Three challenges to EOC

coordination were identified including: (1) asymmetric

knowledge and experience, (2) barriers to maintaining

mutual awareness, and (3) uneven workload distribution

and disrupted communication. Each is discussed in turn.

3.1 Asymmetric knowledge and experience

One important element to consider is that many

counties employ a very small number of full-time

emergency management personnel. When an emer-

gency occurs, it is likely that the majority of the people

reporting to the EOC will not have extensive experi-

ence in crisis management. With the obvious excep-

tions of police, fire, and the Red Cross, most are

accustomed to working within an organization that

provides a service with relatively predictable day-to-

day operations. This disparity in experience level is

visible on several levels. In addition to differences in

experience level with regard to crisis management,

participants have varying levels of familiarity with

emergency response tools and procedures, and often

divergent knowledge of the roles and functions of

individual EOC members. This asymmetry creates a

coordination challenge as issues of trust and cohesion

become increasingly important during high-intensity

disasters.

In the tornado exercise, more than half of the par-

ticipants had very little experience, reporting that they

had participated in either zero or one functional

exercise previously. Six participants were moderately

experienced with between two and four previous

functional exercises, and three highly experienced

participants reported that they had participated in be-

tween 4 and 12 exercises previously. In this context

observers noted that most experienced participants

such as the police, fire, coroner, and Red Cross took on

leadership roles, while less-experienced personnel dis-

engaged from the exercise at times. More experienced

personnel seemed to know what tasks needed to be

done and understood the priorities necessary to man-

age the incident. For example, the coroner, who was

not overloaded with telephone calls, began to help the

less-experienced hospital representative track casual-

ties. The Red Cross representative anticipated the

need to set up centers for tornado victims and began

making calls to begin the process before an official

need was articulated. Interestingly, those with the most

experience seemed to be the most engaged in the

exercise, suggesting that they may obtain the greatest

training benefit. In the context of the exercise, partic-

ipants were quite patient with each other. It seems

likely, however, that in a real-world incident, with time

pressure and life risk involved, issues of trust in team

members’ abilities will arise more as experienced per-

sonnel find that they must rely on (or compensate for)

less-experienced personnel.

Coordination in the EOC also occurs across differ-

ent echelons of emergency management. Perhaps most

obvious to observers is the information flow inside the

EOC. However, it is also important to consider that

each EOC member is responsible for coordinating with

people outside the EOC. Police and fire have direct

communication with first responders at the scene of the

disaster. The hospital representative must coordinate

with local hospitals to direct the movement of patients

depending on equipment and specialties available at

individual hospitals, as well as overload conditions that

result in diversions to outlying hospitals. Local gov-

ernment officials must communicate the situation to

representatives up and down the chain of com-

mand—who may be located in another EOC and must

also manage aspects of the disaster from a different

role and perspective. Representatives tend to have a

high-level of expertise within their own specialty,

making them particularly well-suited for this type of

Cogn Tech Work (2007) 9:25–31 27

123

county-wide, specialty-specific coordination. Only the

most experienced in emergency response, however,

have a solid understanding of the roles others in the

EOC play. This asymmetry in knowledge and experi-

ence is multiplied across multiple levels of command.

3.2 Barriers to maintaining mutual awareness

In order for the coordination to go smoothly, it is

important that all participants in the EOC have a

shared understanding of the situation, current priori-

ties, and actions being taken. The amount of informa-

tion sharing that must happen within and outside of the

EOC is quite daunting. Although members of the EOC

are co-located, they are coordinating actions across a

county—and in real life, often across multiple counties.

Information comes in primarily via telephone and

radio communications. Simple communications are

easily distorted. For example, during the tornado

exercise observers noted that participants in the EOC

identified the collapsed elementary school by three

different names including ‘‘Orchard Park,’’ ‘‘Oak

Park,’’ and ‘‘Oakview,’’ creating significant confusion

over which elementary school was in fact damaged.

Overloaded telephone lines made confirmation of

information difficult, and noisy conditions increased

the likelihood that words would be misheard.

As communication channels were overloaded,

information flow quickly disintegrated. Some partici-

pants found work-a-rounds, such as moving closer to

the core group in order to overhear. This strategy,

however, was met with limited success. In the tornado

exercise, participants were worried about missing

important telephone calls. The telephones were wired

to each work station, so they were not mobile. Further,

the telephones all had identical rings and gave no vis-

ible indication regarding which phone was ringing.

Participants, therefore, were somewhat tethered to

their individual work stations. In the train derailment

exercise, the room was too crowded to allow for easy

movement. To encourage information sharing, the

EOC director requested frequent updates from each

participant. These verbal updates were potentially very

useful in establishing and maintaining a shared

understanding of the situation. However, noise level in

the room and the constant stream of telephone calls

into the EOC were significant barriers to this type of

information sharing.

Shared displays and information log forms have the

potential to aid in information sharing and a more even

distribution of workload. However, our observations

indicate that as workload increased, updating shared

displays and logging telephone calls were some of the

first tasks to be shed. This tendency was exacerbated by

unfamiliar forms and unfamiliar software. Because

emergencies do not occur routinely, less-experienced

participants must re-learn these tools to some extent

each time they are used. In the tornado exercise,

observers noted a cluster of 3–4 people around one

terminal trying to update the map during one portion

of the exercise. While one person sat at the keyboard,

others advised as to which menu to access, etc. In

addition, it was not always clear who was responsible

for updating the wall displays, making it difficult for

EOC members to know to whom they should com-

municate new information and to whom they should

direct questions. As a result, shared displays contained

incomplete or inaccurate information at times—infor-

mation that was available in the EOC, but had not

made its way into the display.

3.3 Uneven workload distribution and disrupted

communication

Varying levels of experience, multiple levels of coor-

dination, noisy communication channels, and incom-

plete shared displays combined to create a situation in

which information did not flow smoothly throughout

the EOC. While some members of the EOC received

frequent updates directly from confederates simulating

personnel onsite at the disaster scene, this information

was not systematically shared throughout the EOC. An

interesting pattern emerged across the two exercises.

This pattern is illustrated in Figs. 1 and 2. Figure 1

depicts information flow as it is expected to happen.

Fig. 1 Information generally flows directly from the community to a core team, in this case the EOC director, fire, and police. In theory, information would flow outward to the rest of the EOC, and from there to specialists in the community

28 Cogn Tech Work (2007) 9:25–31

123

The bulk of information about the disaster itself came

in to the EOC director, police, and fire representatives.

Depending upon the type of disaster, these three are

often the most experienced people in the room with

handling crises. Further, they are accustomed to

working together and are able to smoothly coordinate

amongst themselves. In theory, a shared understanding

of the situation would move from this core team to the

rest of the EOC, allowing each member to coordinate

within his/her area of specialty, all working as part of a

coordinated effort.

In the exercises we observed, however, information

did not flow systematically out from the core team who

were busy fielding telephone calls and problem solving

(Fig. 2). Instead information flowed asymmetrically

throughout the room, resulting in fragmented situation

awareness. Some participants were overloaded, while

others were underutilized. As the workload for this

core group increased, they had less time to share

information and offload tasks. At the same time, others

in the room became more removed from the situation

as they did not have access to the frequent updates

from first responders. Those outside of the core group

had a difficult time maintaining enough awareness of

the situation to identify leverage points and offer help.

Those inside the core group did not have time to bring

the others up to speed and delegate tasks. The few

exceptions were those sitting close to the core group or

those who moved close enough to overhear the core

group. By simply overhearing, some participants were

able to offer relevant resources and become involved

in the disaster response. Others deliberately sought

information. For example, in the tornado exercise the

PIO received many requests from local TV stations to

confirm or deny rumors, or merely update the public

on the situation. The PIO had to be quite persistent to

get the attention of this core group as they were

dealing with a number of pressing issues throughout

the exercise. Examples of the types of information she

sought included:

– Confirmation or denial that martial law was being

established based on reports of military troops

sighted in the area.

– Information regarding the status of the Sears depart-

ment store and the collapsed elementary school,

including reported injuries and fatalities.

– The location of shelters established for tornado

victims.

A debrief session held after the tornado exercise

revealed that some participants could have provided

resources, but were unaware that they were needed.

Participants seemed to be frustrated during the exer-

cise, wanting to be more engaged, but unsure as to

what was needed.

3.4 Recommendations for improving coordination

Although many potential recommendations for

improving coordination in the EOC were generated,

we have focused primarily on interventions that are

available in the near term and likely to have high

payoff with limited resource investment. Table 1

summarizes our findings and the corresponding rec-

ommendations described in this section.

To better support a team with asymmetric knowl-

edge and experience, low cost—often paper-base-

d—tools can provide considerable support. In both

exercises notebooks, whiteboards, and telephone

books provided important tools for communication. In

the train derailment exercise, a paper wall map was

easily and efficiently updated. These low cost, manual

tools can be used for tracking data and sharing infor-

mation. Well-designed tools can also serve as job aids

by providing key information about roles, procedures,

and resources for less-experienced personnel. These

easy-to-learn artifacts also represent important redun-

dancy for situations in which power and internet con-

nectivity may be interrupted.

While electronic coordination tools and information

displays offer great potential in the EOC, onsite sup-

port may be necessary to facilitate effective use during

crises. It is unrealistic to assume that EOC members

will easily remember (or learn for the first time) how to

operate software tools that are used infrequently while

in the midst of a rapidly developing crisis.

Fig. 2 In both exercises observed, the EOC director, police, and fire coordinated smoothly and efficiently. Information flow to the rest of the EOC was sporadic and unsystematic

Cogn Tech Work (2007) 9:25–31 29

123

Perhaps the most powerful strategy for supporting a

team with asymmetric knowledge and experience is to

provide periodic exercises with representatives from all

member organizations. While high-fidelity functional

exercises are clearly the most beneficial, even low

fidelity exercises provide opportunities to interact, ex-

plore each other’s roles and resources, develop trust

relationships, become familiar with forms and tools

available, etc. It may be possible to design exercises

that focus specifically on issues identified here such as

managing a team with asymmetric knowledge, main-

taining shared situation awareness, and avoiding un-

even workload distribution.

Overcoming challenges to maintaining mutual

awareness can be addressed in a number of ways.

Dedicated communication systems with flexible chan-

nels that do not rely on vulnerable infrastructure such

as cell phone towers or the internet are an important

component. Being able to maintain communication

between the EOC and the incident command posts, as

well as other EOCs is critical, as these channels are

used continuously throughout the incident.

Re-designing the EOC facility is another means to

encourage a shared understanding of the situation as it

unfolds. In the two exercises we observed, the layout of

the room had an impact on how information moved

throughout the room. The train derailment exercise

took place in a crowded EOC in which it was difficult for

participants to move around. This limited their ability to

move into small groups to problem solve, to stand near

shared displays to study the situation, or even move to

the shared displays to update them. In contrast, the

tornado exercise took place in a larger room, which

facilitated movement and communication between

members who were not seated close to each other.

We would recommend designing a room that would

encourage movement and communication. This might

include configurable workstations that would allow

EOC members to form small groups to work on a

specific problem. There should be space around the

shared displays, allowing room for small groups to

examine the displays and discuss the situation. Noise

damping materials that would reduce the amount of

ambient noise, but allow small groups and individual

telephone calls to be heard would likely reduce the

sense of confusion in the room. The room configura-

tion should also take into account the need for regular

verbal briefings.

Likewise, shared displays are a potentially powerful

tool for sharing information. If displays are updated as

information becomes available, team members can

refer to them as needed, reducing the need to search

the room for the individual with the most recent

information on a specific topic. Accurate and up-to-

date displays can serve to greatly reduce the level of

confusion and chaos in the room. Whether this display

is projected on the wall for group problem-solving or

viewed on individual, networked terminals, it is critical

that specific personnel are assigned to update and

maintain shared displays. Making this role visible to all

team members will ensure that all participants know to

whom new information should be delivered as well as

to whom questions should be addressed regarding the

accuracy/recency of information on the display.

Yet another strategy for encouraging mutual

awareness in the EOC is the use of groupware tools

(Grudin 1994). Given the high workload in the EOC,

there is a need for tools that support passive informa-

tion sharing as well as deliberate activities such as

verbal briefings. NASA mission control uses voice

Table 1 Summary of findings and recommendations for improving coordination in the EOC

Findings Recommendations

Asymmetric knowledge and experience Crisis management Tools and procedures Roles of EOC members

Job aids for less-experienced personnel Easy-to-learn artifacts On-site support for tool use Frequent exercises

Barriers to maintaining mutual awareness Overloaded communication channels One EOC too crowded to move around One EOC tethered to telephones Noise level limits overhearing Diffused responsibility for shared displays

Communication systems that do not rely on vulnerable infrastructure

Facility layout re-design Increase ability to overhear via facility design

and groupware tools Assign role to update displays

Uneven workload distribution and disrupted communication

Core team high workload Non-core team unable to contribute Core team too busy to share info

Assign role to field info requests to and from core team

Assign role to ensure periodic updates Make job functions and resources of each

member known in EOC

30 Cogn Tech Work (2007) 9:25–31

123

loops to allow team members to listen in on commu-

nications in which they are not directly involved

(Patterson et al. 1999). Adapting this technology for

use in the EOC would allow participants to hear

communications between the EOC director, police,

and fire representatives, reducing the frequency with

which these high workload positions need to engage in

deliberate information sharing activities.

With regard to uneven workload distribution and

disrupted communication, re-engineering the EOC to

better support the core team pattern seen in both

exercises would result in more effective information

flow and delegation. For example, assigning specific

EOC personnel the task of fielding information re-

quests to and from the core team would likely aid

communication flow. Assigning someone outside of the

core team to the role of periodically ensuring that

information updates are occurring and displays are

updated would serve as an additional check during

intense, highly dynamic situations. Key to the success

of any re-engineering would be to make the roles clear

to all EOC members, so that each individual knows

who to send information to and who to seek informa-

tion from.

4 Discussion

Perhaps the most significant observation from these

two exercises is the issue of ‘‘stovepiping’’ between the

core team and other members of the EOC. Although

this phenomenon has been described in terms of nat-

ural fault lines, we are hesitant to use that label here.

The term natural fault line suggests somewhat stable

groupings, rather than movable lines. We suspect that

membership in the core group will vary somewhat

depending on the nature of the incident and the

experience level of the participants in a given EOC.

Despite this variability, it seems likely that in nearly

every crisis, a core group will receive large amounts of

information, as well as tasking, from those on scene at

the disaster. Without adequate tools, procedures, and

training, those in this core group will become increas-

ingly busy while those outside the core group will be-

come increasingly removed from the situation.

Given the many challenges the EOC faces, the dis-

parate experience levels within the EOC, and the very

unpredictable nature of disasters, there is a tendency to

simply accept that coordination in this setting will be

somewhat chaotic and inefficient. Certainly, this is not

a situation in which communication paths and roles can

be easily scripted. However, our observations indicate

that we can provide the tools, facilities, procedures,

and training that will facilitate smooth coordination,

even for an ad hoc team facing unpredictable and

highly variable disasters.

Acknowledgments This work was funded in part by contract # F33615-99-D-6001 and was sponsored by the Air Force Research Laboratory, Logistics Readiness Branch (AFRL/HEAL).

References

Aubrey L (2002) The hard lessons of September. Mod Healthc 32(4):50

Endsley MR, Garland DJ (2000) Situation awareness analysis and measurement. Lawrence Erlbaum Associates, Mahwah

Grudin J (1994) Computer-supported cooperative work: history and focus. IEEE Comput 27(5):19–27

Hackman JR (1990) Groups that work (and those that don’t): creating conditions for effective teamwork. Jossey-Bass, San Francisco

McEntire DA (2002) Coordinating multi-organizational re- sponses to disaster: lessons from the March 28, 2000, Fort Worth tornado. Disaster Prev Manag 11(5):369–379

Patterson ES, Watts-Perotti J, Woods DD (1999) Voice loops as coordination aids in space shuttle mission control. Comput Support Coop Work 8(4):353–371

Perry RW (2003) Emergency operations centers in an era of terrorism: policy and management functions. J Contingen- cies Crisis Manag 11(4):151–159

Salas E, Fiore SM (2004) Team cognition: understanding the factors that drive process and performance. American Psychological Association, Washington

The 9/11 Commission (2004) Final report of the national commission on terrorist attacks upon the United States. Retrieved September 2004 from http://www.gpoaccess.gov/ 911/

Tierney KJ, Lindell MK, Perry RW (2001) Facing the unex- pected: disaster preparedness and response in the United States. Joseph Henry, Washington

Cogn Tech Work (2007) 9:25–31 31

123

View publication statsView publication stats