From archives of Personal Papers ex libris Ludwig Benner, Jr.

- - - - - -Last updated on 1/11/05

[ Starline Home Page ] [ Investigation Process Research Roundtable ] [ Submit comments to Starline ]

[ Order reprint of paper here (pdf files) ]

D.E.C.I.D.E.
In Hazardous Materials Emergencies
LUDWIG BENNER, JR.
Hazardous Materials Specialist
National Transportation Safety Board

The views expressed in this paper are those of the author and do not necessarily represent
the views of the National Transportation Safety Board.

CONTENTS
(Historical note)

INTRODUCTION: EVERYONE KNOWS

EMERGENCY RESPONSE PURPOSES
THE ANATOMY OF AN HM EMERGENCY
THE “D.E.C.I.D.E.” PROCESS
1. Detecting HM Presence
2. Estimating Likely Harm Without Intervention
3. Choosing Response Objectives
4. Identifying Action Options
5. “Doing” The Best Option
6. Evaluating Progress

CONCLUSIONS
Footnotes

INTRODUCTION
EVERYONE KNOWS WHAT AN EMERGENCY IS: it is a set of sudden,
unforeseen, and urgent circumstances that demands immediate action. When
hazardous materials (HM) are involved, the possibility of the emergency escalating
into a disaster is always present, and the demand for action even more immediate.
But what action? For what purpose? By whom? When?

Some sort of response is usually mounted, either by the individuals involved or by an

emergency response team. The mounting of the response requires that discrete
decisions be made with respect to the action needed, the outcome to be achieved, the
response actions available, the methods and facilities to be employed, the timing of
the action, and numerous other factors. Because of the nature of an HM emergency,
many of these decisions are made on a “first-time” basis, usually under duress. They
involve a form of either “adaptive behavior” or “adaptive learning” in a situation
being encountered for the first time. In these circumstances, are there some
techniques that will help the decision-maker to produce “better” decisions than those
that may have resulted in so many injuries in the past?

This paper explores such techniques. It is based primarily on the analyses of HM

emergencies in transportation that have involved injury and loss of life attributable to
the behavior of HM present. It assumes that emergencies consist of a series of events
that occur in a logical sequence. This “think events” approach is amplified by
assembling the events into an “anatomy” of an HM emergency, and by treating the
decision making process in an “events” framework.

EMERGENCY RESPONSE PURPOSES

Why mount an emergency response? Answers abound. Save lives! Save the ship!
Protect the public! Prevent a disaster! It’s our duty! But there is a common purpose,
conceptually, and it can be stated simply in terms of events sequences: to favorably
influence the outcome of the events sequence that would otherwise occur.
Any emergency will stabilize in time, without any emergency response effort, as the
emergency runs its natural course. However, as they run their course, emergency
events are often accompanied by injury, and this is objectionable. The purpose of
mounting an emergency response effort, therefore, is to change the events sequence
constituting the emergency before it has run its course naturally and to minimize
harm that would otherwise occur.

In order to make an informed decision on whether to intervene, an attempt must be

made to predict what the natural outcome probably will be. If harm is predicted, then
intervention will have some purpose, provided the harm to the interveners does not
exceed the harm prevented. The “gain” achieved by reducing the naturally occurring
harm should not exceed the “cost” (harm) attributable to intervention. The difference
between the “gain” and the “cost” provides a measure of the extent to which the
emergency response effort favorably influences the outcome in any given
emergency. Although such predictions of outcomes are required, they are now
usually made intuitively — if they are made at all. Too frequently, the impulse to act
immediately overwhelms reason in crisis situations; risk-taking becomes excessive,
and losses escalate rather than decline. To forestall these wasteful losses, a structured
approach to decision making in HM emergencies seems long overdue. No such
structured approach or framework now exists for HM emergencies. Some attention
has been focused on the need for this type of approach. For example, in a report of
an HM accident in which numerous emergency personnel were injured, the National
Transportation Safety Board found that available information was inadequate for on-
scene identification and assessment of the hazards, potential injury-producing events,
and consequent response options. [1] Also, M. E. Grimes described a study of HM
emergencies that revealed problems that fire officers encountered in decision making.
[2] However, a useful method for guiding decision making throughout the emergency
is not yet available. The development of such a method would be facilitated by an
understanding of what an HM emergency is.

THE ANATOMY OF AN HM EMERGENCY

We can make two assumptions in describing a typical HM emergency — first, that
the emergency events sequence progresses in an orderly relationship in a given set of
circumstances, and second, that the events sequence can be generalized for
descriptive and analytical purposes.

What happens in an HM emergency? During the course of a normal activity in

which HM are involved, the HM are controlled by some method of containment or
confinement or isolation from stressing events. For an emergency to begin, the
“holding” system for the HM must be disturbed or stressed in some manner. The
system may or may not adapt to the stress. If the stress does not exceed the
capability of the “holding” system to resist or adapt to the stress, no disruption of the
activity occurs. However, if the “holding” system is overstressed beyond its
recoverable limits, some kind of failure occurs. Upon the occasion of such failure,
escape of the HM from the “holding” system can occur. This escape can take the
form of matter capable of producing harm, or of harmful energy, or some
combination of both. Once this occurs, the matter or energy can disperse until it
comes into contact with or impinges on a vulnerable exposure. Depending on the
intensity and duration, the exposure can harm the impinged resource, and possibly
cascade by overstressing and harming additional exposed “holding” systems or
resources. The emergency events sequence ends when the cascading stresses are
accommodated by the elements next impinged without injury, and the conditions
once again stabilize.

This events sequence is illustrated in Figure 1, taking into account the adaptive events
discussed earlier. The illustration uses a convention of rectangles to represent events,
and arrows to indicate a proceed/follow relationship.
In order to influence the outcome of the emergency events sequence, it can be seen
in Figure 1 that the natural events sequence must be deliberately disrupted by the
emergency response efforts. This can be indicated by displaying the
“countermeasures” on the same framework, as has been done in Figure 2. To aid in
formulating the predicted outcomes, another element has been added to Figure 1:
each general event is assigned a “stage” designation, to indicate the stage of the
emergency at which a given type of event occurs. These stages can be used to aid in
determining the events likely to occur after arrival of the decision-maker at the
scene.
THE “D.E.C.I.D.E.” PROCESS
When he becomes aware of the existence of an emergency, the decision-maker is
faced with a series of “decision events.” To affect or influence the emergency events
sequence, decisions must focus on modification of the sequence that would otherwise
occur naturally. There are six steps that can provide a framework for decision making
in a given HM emergency. These six steps are:

1. Detect HM Presence.
2. Estimate Likely Harm Without Intervention.
3. Choose Response Objectives.
4. Identify Action Options.
5. Do Best Option.
6. Evaluate Progress.

This is the D.E.C.I.D.E. process framework for HM emergency decision making.

1. Detecting HM Presence
An essential first step in any emergency is to decide whether HM are present or not.
Numerous clues usually exist when HM are present, but they must be sought out by
the decision-maker. Inductive reasoning may be required when the HM has not been
activated before the decision-maker’s arrival on the scene. This can be accomplished
by using containment principles, activity characterizations, structural or appearance
principles, etc. Numerous methods for identifying the presence of HM are described
in the literature or regulations governing the transportation or handling of HM.

2. Estimating Likely Harm Without Intervention

The next logical step, if HM are present, is to decide what the most likely
progression of events and their outcome will be, if no intervention is attempted. This
estimate of the likely harm is the most difficult step in the process, because
emergency response information now available about HM does not focus on the data
needed to support this predictive effort. Some systems claim to provide “hazard
information” indicating adverse behavior of HM, but none provide the decision
maker with information about all the variables that determine the events sequence in
a specific emergency. [3],[4] [5] [6] [7] These variables include the quantity and
quality of their predictive estimates and their resultant decisions. The communication
of such information through automated systems is envisioned.

Applying these techniques to emergency response decisions can be expected to have

additional benefits. As the events sequences are made visible by the charting
methods, pre-accident safety control opportunities will also become obvious. Thus,
losses could be reduced even further by this “preplanning” approach.

Further study of the decision processes in HM emergencies in both past accidents

and postulated emergencies is required in order to refine the concepts described, but
the benefits of the D.E.C.I.D.E. approach seem to have sufficient promise to justify
the effort. No other approaches yet proposed appear capable of resolving the
problems discussed. form of HM present, the “holding system” failure behavior,
dispersion mechanisms, dispersion rates and patterns, damage or injury mechanisms,
lethality, and other factors needed for timely on-scene events predictions. Without
such data, timely on-scene predictions of the most likely events sequence — the
timing of these events, the directions, ranges or distances of concern, and the
expected injury estimates — may contain gross uncertainties that raise the risk level
in any response.

This step requires the decision-maker to make a “mental movie” of the most likely
course of events after his arrival on the scene. This “movie” begins with the setting
observed upon arrival and ends with the last harmful event in the emergency. The
expected behavior of the HM in the circumstances and the kind of harm that will
probably occur constitute the plot of this “movie.” The scenario must include the
principal “actors” such as stressing agents, the HM and its packaging, and the people
or properties exposed to harm, and all the significant actions involving each actor.

For HM emergencies in transportation, two documents that partially address this

decision step are an emergency guide and a chemical hazard response information

system .[8] [10] The Environmental Protection Agency has also explored this need.
[11] None of the documents indicate the full range of events predictions and the
comprehensive decision process for which the data are supplied. Substantial study
and simplification, of the type being conducted by the US Coast Guard, [12] is
needed before timely predictions can be made with reasonable confidence in specific
emergencies aboard vessels or at the scene of accidents and spills.

3. Choosing Response Objectives

The third step, choosing the emergency response objectives, proceeds from the
predicted injury and damage estimated in step two. This decision indicates what the
emergency response effort will attempt to save. It examines the exposed elements
and attempts to distinguish between that which is irretrievably lost, that which is in
jeopardy but might be “saved” by appropriate action, and that which is not in any
jeopardy during the likely events sequence. The result of this step is an identification
of the “gain” desired, and, with the result of step two, provides valuable information
for step five. Little guidance exists for making this step.

4. Identifying Action Options

The fourth step consists of analyzing the action options plus the “gains” and “costs”
that are associated with each option. Usually more than one action option is available
that will control the emergency at hand. These options depend on the stage beyond
which the emergency is to be influenced, the predicted events selected for the
“mental movie,” the gains desired, the gains that the option is likely to achieve, the
resources available to carry out the option, and the costs of the option. The actions
that can be taken may range from a full-scale attack to an immediate withdrawal
beyond the range of effects. This step is designed to develop and select those options
that are feasible in the specific circumstances of the emergency, and will change the
outcome of the “mental movie” from step two. To arrive at these options requires
knowledge of the predicted events sequence and of methods that are available for
intervening in that sequence, if any. Fire fighting principles might provide sources for
such methods. [13] [14] [15] Additional options might be developed by considering
the stage of the emergency and the events sequence involved in the predicted
outcome. Events sequence charting and analysis of differing types of releases or
reactions for a given activity would help the decision-maker discover other response
principles. For an example of this events charting technique, see Figure 3.
These response principles could probably be broken down in a manner conducive to
automated data storage, retrieval, or display methods to assure their timely
availability in given emergencies, in a manner like that contemplated by the CHRIS
system. [16] The results of this decision must be expressed in terms compatible with
the outputs of steps two and three. This means that each option must be expressed in
terms of “gains” that each will probably achieve, and the “costs” necessary to
achieve the gains. In addition, each option must be described in terms of the events
to be influenced and the new events sequence that is anticipated. Both these
descriptions are essential for the next step.

5. “Doing” The Best Option

When the estimated harm and events sequence predictions for each option become
available, a decision to do whatever is best must now be made. This fifth decision
step will involve the weighing of factors beyond the “net gain” for each option, such
as legal requirements to respond, reputations, public expectations, personal risk-
taking propensities, etc. This is the action step — the culmination of the preparatory
steps described above by which the “mental movies” are transformed into reality. It
is in this step that technical judgments are melded with value judgments, and the
crucial decisions are made. Since the quality of the technical data influencing this
decision will affect the quality of the outcome, the degree of confidence that the
decision maker has in the prior estimates will thus have a strong influence on the
decision made and the actions taken. Therefore, measures to improve the technical
quality of the estimates will directly improve the quality of this decision. Again,
documentation of this decision step is not available for NM.

6. Evaluating Progress
Having made the action choice, the next step involves continuous observations and
decisions of a yes/no type; either the events sequence is progressing as envisioned in
the “mental movies,” based on observations, or it is not. If it is not, much of the
above process must be repeated — estimates must be revised, options reexamined,
and decisions revised. This “feedback” decision step continues until the emergency
has been stabilized. The importance of the earlier events predictions is obvious here;
the predictions provide the basis for comparing and measuring the success of the
response selected during the D.E.C.I.D.E. process. Without these predictions, an
emergency response effort has no standards for successful intervention, and the
outcome will be more a matter of luck than of sound, structured decisions. Again,
documentation of practical methods is lacking.

CONCLUSIONS
An approach and method for structuring the HM emergency response decision-
making process has been presented here, and the needs for developing data to
support this process are described. Weaknesses in present HM emergency
information systems can be seen, and an approach is suggested for resolving these
weaknesses by the use of events charting methods. Within the “think events”
framework presented, information to support a structured decision-making process in
HM emergencies can be developed and made available to on-scene decision makers
to improve the quality of their predictive estimates and their resultant decisions. The
communication of such information through automated systems is envisioned.

Applying these techniques to emergency response decisions can be expected to have

additional benefits. As the events sequences are made visible by the charting
methods, pre-accident safety control opportunities will also become obvious. Thus,
losses could be reduced even further by this preplanning” approach.

Further study of the decision processes in HM emergencies in both past accidents

and postulated emergencies is required in order to refine the concepts described, but
the benefits of the D.E.C.I.D.E. approach seem to have sufficient promise to justify
the effort. No other approaches yet proposed appear capable of resolving the
problems discussed .
Source: Fire Journal Vol. 69 No. 4 July 1975

Footnotes
[1] Derailment of Missouri Pacific Railroad Company’s Train 94 at Houston, Texas,
October 19, 1971 , Accident Report RAR 72—6, (Washington, DC: National
Transportation Safety Board).

[2] M. E. Grimes, “Hazardous Materials Transportation Accidents,” Fire Command!,

Vol. 41, No. 4 (April 1974), p. 11.

[3] ChemCard Manual (Washington, DC: Manufacturing Chemists’ Association,

1965).

[4] Fire Protection Guide on Hazardous Materials, Fifth Ed. (Boston: NEPA, 1973).

[5] Notice of Proposed Rulemaking No. 73—10, Federal Register 39 FR 3164

(Washington, DC: United States Department of Transportation, Hazardous Materials
Regulations Board, 1974).

[6] B E Pamphlet No. I Hazardous Materials Emergency Guide (Washington, DC:

Bureau of Explosives, Association of American Railroads, 1973).

[7] Handling Guide for Potentially Hazardous Materials (Chicago, Ill.: Railway
Systems Management Association, 1972).

[89] Emergency Services Guide for Selected Hazardous Materials (Washington, DC:
United States Department of Transportation, Office of Hazardous Materials, 1972).

[10] A. D. Little, Inc. ., Preliminary System Development, Chemical Hazard

Response In formation System (CHRIS) (Washington, DC: US Coast Guard, 1972).

[11] Episode Manual , Contract No. 68—02—0029 (Research Triangle Park, NC:
United States Environmental Protection Agency, 1972).

[12] US Department of Transportation, “Vulnerability Model,” Contact DOT

CG33377A (Washington, DC: US Coast Guard, 1974).

[13] C. V. Walsh, Modern Guidelines for Fire Control (Brooklyn, NY: Theo. Gaus’
Sons, Inc., 1972).

[14] F. L. Brannigan and C. 5. Miles, Living with Radiation, No. 2 Fire Service
Problems (Washington, DC) United States Atomic Energy Commission, 1963)

[15] C. W. Bahme, Fire Officer's Guide to Dangerous Chemicals (Boston: NEPA,

1972).

[16] See Footnote 10.

Return to Table of Contents

Historical note

The paper, first published in the July 1975 "Fire Journal" summarized the
results of a 3 year effort to develop training for hazmat responders that
would produce better results than were then being achieved. It began one
day in 1972 when Frank Brannigan, who was in charge of a fire service
curriculum at Montgomery College Maryland, expressed his consternation
at the criticisms of fire service performance by the NTSB, for which I was
largely responsible. His challenge: "If you are so damn smart, why don't you
tell us how to do it better." My response to his challenge was to agree to
develop and teach a course for him, and with the help of the students, a
better program evolved. The first description of the D.E.C.I.D.E. process
concepts and approach was in the publication "HAZARDOUS
MATERIALS EMERGENCIES, developed for the course and posted
elsewhere. This article was the first dissemination of the program in the fire
service media.

The D.E.C.I.D.E. process seems to have withstood the test of time, having
been applied in other fields as well. This is the original paper.

Reprints should be requested from the NFPA. If not available there, Click
here to order a reprint of this paper ( Item R071)
© 1973 by Ludwig Benner, jr.