Risk and impact assessment in animal disease program formation

RISK AND IMPACT ASSESSMENT IN ANIMAL DISEASE PROGRAM
FORMATION










BY

SUSAN A. FERENC


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY



UNIVERSITY OF FLORIDA


1994











ACKNOWLEDGMENTS


First I wish to express my sincerest appreciation to Dr. Jim
Simpson, my major advisor, for his guidance and encouragement. I
also wish to thank him for his patience, moral support and openness
to new ideas.
Special thanks go to Dr. Bill Boggess for his contributions to
the Economic Impact Study conducted on Antigua and his moral
support during hurricane Hugo. Thanks also go to Dr. Mike Dicks and
Mr. Jim Butcher for their continued encouragement and for making
my work on Antigua possible. I also wish to acknowledge the
financial support received from AID and the USDA for the Antigua
project and my education.
Thanks are also extended to Dr. Haile, Eric Daniels, and Murat
Taner at the USDA-ARS Research laboratory in Gainesville for
adapting the Babesia transmission model for use on the Antigua data
and for allowing me to use the model in this dissertation.
I particularly wish to thank my ever-patient husband, John,
whose constant support and encouragement truly made this project
possible.











TABLE OF CONTENTS

ACKNOW LEDGEMENTS....................................................................... .......................... ii

ACRONYMS AND ABBREVIATIONS ....................................................................... vi

ABSTRACT .................................................................................. vii

INTRODUCTION............................................................................................................... 1
Problem Statement........................................................................................ 4
Background ............................................................................ .............. .................. 5
Case Study Introduction: Tropical Bont Tick Management in
the Caribbean ............................................................................ ................ 8
O objectives ...................................................................... ....................................... 11
Overview of Remaining Chapters ................................................................ 12

LITERATURE REVIEW ........................................................................................ 14
Regulations for Environmental Impact Statement
Preparation ........................................... .....---......... ................................. 14
Assessment of Environmental Impacts .................................................... 18
Impact Identification and Communication..................................... 21
Impact Measurement ................................................................................... 23
Risk A ssessm ent ................................................................................................ 27

A NETWORK FOR RISK AND IMPACT IDENTIFICATION ................................. 33
Proposed Network .............................................................................................. 35
Risk and Impact Linkages .............................................. .......................... 38
Use of the Proposed Network ........................................................................ 41

CASE STUDY: REVIEW OF A PROPOSED PROGRAM FOR
MANAGEMENT OF THE TROPICAL BONT TICK IN THE
CA RIBBEA N ............................................................................................................ 43
Development of the Tropical Bont Tick Program ................................. 43
The Tropical Bont Tick and Associated Diseases in the
Caribbean ..................................... ................................................... .... 46
Costs and Benefits of Tropical Bont Tick Eradication......... 47
A Proposed Pilot Eradication Project on Antigua............................. 50
Results of the Environmental Assessment ................................ 52
Impact on Non-Target Organisms............................................... 52






Potential Effects on Non-Target Organisms and
Mitigation ......................................................................................... 53
Availability of Alternative Nonchemical Treatments........... 53
Risk/Benefit Analysis....................................................................... 54
Areas for Improvement in the Environmental
Assessment............................................................................................ 55
Impact on Non-Target Organisms................................................. 55
Availability of Alternative Nonchemical Treatments ........... 55
Risk/Benefit Analysis....................................................................... 56
Summary................................................................................................................. 62

CASE STUDY: APPLICATION OF THE RISK FRAMEWORK,
ALTERNATIVE GENERATION, AND RISK AND IMPACT
COMPARISON ...................................................................................................... 64
Application of Risk Fram work ................................................................ 65
Evaluation of the Identified Risk and Impact Concerns................. 68
Cultural Factors......................................................................................... 68
Biological Factors..................................................................................... 72
Environm ental Factors ............................................................................ 72
Non-Target Organisms and Disruption of Endemic
Stability .................................................................................................. 74
Program Alternatives.................................................................................... 77
Eradication ................................................................................................... 77
Integrated Pest Managem ent ................................................................ 80
Control ........................................................................................................... 80
Comparison of Risks across Control and Eradication
Alternatives ................................................................................................ 83
Simulation of Eradication and Control Program Risks and
Im pacts .................................................................................................... 84
Eradication Objective........................................................................ 89
Control Objective ................................................................................ 96
Implications of Simulation Results................................................ 100
Use of an Alternative Acaricide .............................................................. 101
Summary............................................................................................................1... 02

CASE STUDY: RISK/BENEFIT ANALYSIS ........................................................ 106
Risk Analysis ................................................................................................... 106
Program Costs ........................................................................................... 106
Babesia-Related Mortality ................................................................... 108
Boophilus-Related Production Loss ................................................. 109
Benefit Analysis ............................................................................................. 111
Results .................................................................................................................. 114
Benefit/Cost Ratios ................................................................................ 114






Net Present Values and Annual Equivalent Flows..................... 11 7
Best, Worst, and Expected Cases......................................................1... 21
Economic Impact on the Antiguan Producers and the Local
Econom y ................................................................................................. 122
Sum m ary............................................................................................................1... 24

SUMMARY, CONCLUSIONS, AND RECOMMENDATIONS ................................126
Approach to the Research ...........................................................................126
Summary of Case Study Findings and Recommendations............. 1 29
Findings .......................................................................................................1... 30
Recommendations for Tropical Bont Tick Management on
A ntigua...................................................................................................1... 35
Conclusions ......................................................................................................1... 37
Recommendations .......................................................................................... 139

APPENDIX A TROPICAL BONT TICK ERADICATION COST
FA CT O RS ................................................................................................................142

APPENDIX B EXAMPLE OF BABESIA TRANSMISSION MODEL
SIM ULATION......................................................................................................... 146

APPENDIX C TROPICAL BONT TICK CONTROL COST FACTORS............148

APPENDIX D VALUATION OF SIMULATION RESULTS ...............................151

LIST OF REFERENCES ........................................................................................... 154

BIOGRAPHICAL SKETCH ......................................................................................163












AID
APHIS
CARICOM
CICP
CEQ
FAO
GDP
LW
NEPA
OECD


USDA


ACRONYMS AND ABBREVIATIONS
Agency for International Development
Animal and Plant Health Inspection Service
Caribbean Community and Common Market
Consortium for International Crop Protection
Council on Environmental Quality
Food and Agriculture Organization
Gross Domestic Product
Live Weight
National Environmental Protection Act
Organization for Economic Cooperation and
Development
United States Department of Agriculture










Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy


RISK AND IMPACT ASSESSMENT IN ANIMAL DISEASE PROGRAM
FORMATION

By

Susan A. Ferenc

August, 1994




Chairman: Dr. James R. Simpson
Major Department: Food and Resource Economics

A framework was developed to aid environmental assessment
team members in identifying potential risks and impacts associated
with the objectives and directives of some animal disease control
programs. The network methodology used leads team members from
the initial process of identifying risk characteristics and factors
that might jeopardize the successful fulfillment of program
directives through to assessing the direct and indirect impacts of
proposed program objectives and actions. Use of the network in the
environmental assessment was then demonstrated through
application to an actual animal disease control program currently
under consideration, the proposed Tropical Bont Tick Eradication


vii






Project on the island of Antigua. The network was used to identify
cultural patterns and environmental conditions characteristic of the
island which might preclude a successful conclusion to an
eradication attempt. Potential direct and indirect impacts of
program actions on the biological interrelationships indicative of
the existing livestock production system were also identified.
Finally, the framework was employed in the process of developing
and assessing feasible and appropriate alternatives to an
eradication objective. Identified risks and impacts were compared
across eradication and control alternatives.
An identified significant impact is the potential disruption of
a non-target endemic bovine tick and associated tick-borne disease
on the island; Boophilus and babesiosis. A model simulating tick
population dynamics and Babesia transmission was used to examine
the specific impacts of program actions on the tick and disease
transmission. The results of the simulations were transcribed into
economic variables and incorporated into an expanded risk/benefit
analysis of the proposed eradication and alternative control
programs. Program costs and benefits were estimated for a
simulated successful eradication program, a failed eradication
program and two control programs. The direct economic impact of
this set of program outcomes on the Antiguan producer and the local
economy were also evaluated and compared. The results of the
expanded risk/benefit analysis indicated that the projected benefits
of a Tropical Bont Tick eradication program changed dramatically
when risk and impact variables were included. Risks and impacts


viii






were mitigated to some degree by the actions of the alternative
control program.










CHAPTER 1
INTRODUCTION

The targets we hit have not always been those at which we
were firing as we discover, in the lines of Barry Commoner,
that everything is connected to everything else and the
connections often have not always been apparent. Casting
disease control programs into the context of complex
ecosystems, with interconnected parts, may permit us to
avoid, or at least anticipate, some of the unpleasant
consequences of our actions (Yuill & Kuns, 1973, page 34).




The development of regional livestock disease management
programs begins in similar fashion to agricultural projects in
general, i.e., when members of the appropriate branch of government
become convinced that a program of some kind is needed and is
feasible. Need may be established as a result of academic inquiry or
from studies initiated at the request of producers (Gittinger, 1982).
In the case of animal disease issues, public health concerns and
political interests may also play a role in prompting government
action.
Unfortunately, disease control programs exemplify what has
become known as the "single-problem track," evaluating the disease
in question as though it exists in isolation rather than as a part of a
delicately balanced, complex, and dynamic ecosystem. However,
rarely does an animal disease exist in isolation. Rosenberg et al.




2

(1979), point out that "infectious diseases occur as the consequence
of complex interactions within the realm of defined ecosystems" and
that historically, disease control campaigns have been "linear in
causal conception and with respect to corrective actions [which]
rarely took into consideration the diverse and dynamic
epidemiological characteristics of the diseases in different
ecological areas" (page 588). Hence, when a problem is detected and
the need to intervene established, the programs undertaken are
normally restricted to general, uniform, and massive campaigns such
as vaccination or vector elimination. Except for diseases with
relatively simple transmission cycles, success is rare and almost
always temporary (Rosenberg et al., 1979).
Disease management programs attempt to directly or
indirectly manipulate populations of hosts, vectors or pathogens, but
these manipulations have had some unexpected results indicating
that the solution of an immediate problem may, in fact, create new
ones. Indigenous livestock populations, particularly of developing
countries, have adapted to the presence of a variety of diseases or
pathogens via heritable immunity and genetic adaptation. Time,
breeding, and local environmental conditions have created a delicate
endemic stability between animal populations, their diseases, and
the ecosystem. Endemic stability of a disease means that it occurs
with predictable regularity in a population (Schwabe et al., 1977).
Manipulation of the interrelationships providing endemic stability
may lead to a slow, or sudden, increase in the incidence of the
disease with potentially devastating economic consequences.






Consideration of the risk of potential "non-target" effects
such as this should be, logically, an integral part of any disease
intervention program proposal. In fact, a framework for the
identification and evaluation of these types of impacts and risks
could be incorporated into the preparation of an environmental
assessment or environmental impact statement. Preparation of one
or both of these reports in the feasibility study phase of government
programs proposals has been required since the National
Environmental Policy Act (NEPA) of 1969 was passed by Congress.
However, in the scant contemporary literature available on the
preparation of animal disease proposals there is no mention of
environmental assessment's or environmental impact statement's,
much less a means for non-target effect consideration or evaluation.
This should be of concern as there are potentially high costs
associated with the indirect effects or risks in these very
specialized programs.
The nature of animal disease control programs, particularly in
developing countries, makes the potential risks of intervention
actions unique and complex. Some of these include failure of an
eradication objective leading to a rebound of the target disease or
agent, reintroduction of the target disease or agent into the now
disease-naive population, and epidemic outbreaks of a non-target
endemic disease. The losses associated with these possible
outcomes might well exceed those which had supported justification
of an intervention program initially.
What might prove useful to avoid potential problems and insure
adequate impact assessment is the incorporation of a risk







assessment protocol in conducting an environmental assessment or
environmental impact statement in the feasibility study of disease
control proposals. The function of this protocol could be to provide,
in a more formal and analytical way, a structured review and
analysis of the proposed action and alternatives. The result of this
approach should be better decision-making based on a more complete
understanding of how manipulation of delicately balanced, dynamic,
ecosystems could create rippling effects with potentially dire
consequences.


Problem Statement

The current process for the formation of large-scale public
programs to control or eradicate animal diseases tends to be linear,
or single-focused, in conception, approach and particularly,
evaluation. As a result, this process may lead to flawed decision-
making.
When the need for disease management is established, a public
program is sought. It is at this point that the "single-problem" track
or linear process begins. From here on, the disease complex in
question is evaluated as though it were in a vacuum, separate from
the complex social and biological ecosystem in which it actually
prevails. The regulations set down for preparation of an
environmental impact statement or environmental assessment
indicate that these issues should, in fact, be identified and
evaluated. The proposed action itself, not merely the measures
used, should be assessed for potential adverse impacts. Although






the environmental impact statement and environmental assessment,
by design, are meant to evaluate biophysical and socio-economic
impacts of program actions, indications are that these reports fall
short of this goal (Whitney & Maclaren, 1985).


Background

Regional programs for disease control are developed and
implemented by the government and funded largely or entirely by
society (Hanson & Hanson, 1983). At top decision levels, the choice
is usually between funding a disease control program or some
totally unrelated program (Gittenger, 1982; McGregor, 1973) because
administrators confronted only with alternative disease programs
are not likely to fund any of them. If justification for an animal
disease control program is narrowly defined, the program is also
unlikely to be funded (Hanson & Hanson, 1983). Bringing about
improvements through disease control or eradication programs
requires recognition that the programs themselves affect the social
and economic conditions of society.
In developing a case for the control of an animal disease, two
questions must be answered: 1) who suffers from the presence of
the disease; and 2) who benefits from its control? When these
programs have to compete for funding against totally unrelated
programs, the animal disease control program can become a matter
of priority if the disease is shown to affect the availability of food
(Horsfall, 1975). Animal diseases can have a devastating impact on
the economic viability of a single farm, a state, a country or even an







entire continent. The massive African rinderpest panzootic of the
late 1800's was so great that entire nations were reduced to
starving remnants and none of the Rift Valley civilizations were
able to survive it as self-governing peoples (Mack, 1970; McKelvey,
1973). Today, the rationale for animal disease control is purely
economic; however, the response is often politically based.
There are two classes of disease control or eradication
programs. One is an emergency effort in response to a newly
introduced pathogen or the sudden explosion of a pathogen already
present. The other is a preplanned, nonemergency program against
an established pathogen. Experienced administrators know the same
public that will accept, even demand, immediate action against a
new and or explosive disease will resist initiating a program
against a long-established one. Livestock owners are not greatly
concerned over a disease that does not exert an immediately
apparent economic effect on their operations (Schnurrenberger et
al., 1987).
Some livestock disease eradication programs have been
successful, particularly in developed countries (Riemann &
Bankowski, 1973; Schnurrenberger et al., 1987) and there is no
evidence in the literature that any of these programs have been
directly responsible for an acute outbreak of a non-target disease.
However, programs have been severely compromised by having
ignored the risks associated with failure to achieve the program
objective--if eradication was the chosen course of action--or to
evaluate fully the dynamic social and biological ecosystem in which
the affected animals existed (Nawathe & Lamorde, 1982). An






example is an early rinderpest eradication program in sub-Saharan
Africa which suffered a setback when, just prior to full eradication
in the late 1970's, several of the program's financial donors
withdrew funding support. The program had no contingency plan for
such an event. By the time the eradication effort reconvened, a
quarter of a billion dollars of production loss had ensued as the
disease spread, epidemically, back across the areas from which it
had been eliminated (Sollod, 1992).
In another example, failure to assess the potential
socioeconomic impacts of an eradication program had costly
consequences. Eradication of African swine fever from Haiti in the
1980's required the total depopulation of native Creole pigs. The
pigs "donated" to replace the Creole species were totally unsuited
for tropical habitats. The white-haired, pink-skinned Yorkshires
from Canada suffered from sunburn and malnutrition when
introduced onto Haiti and quickly succumbed to the harsh
environmental conditions. Depopulation of the native pigs had
secondary, though major, socioeconomic consequences. Over the
years, the Creole pig had adapted to a very meager diet and
supplemented its nutritional needs with human excrement. As the
replacement pigs were not suited to this, Haiti was soon facing a
major problem with human waste build-up. In a "band-aid" attempt
to temporarily deal with the problem, 2,300 chemical toilets were
flown down to the island (Brown, 1992).
It is unknown whether the potential risks associated with
program failure or non-target effects were ever considered in the
developmental stages of past programs such as these. However, if







impact assessments were conducted, these potential effects were
apparently not anticipated and contributed, at least in part, to
program failure.

Case Study Introduction: Tropical Bont Tick Management in the
Caribbean

This dissertation contains a case study examining tick
management policy on the Caribbean island of Antigua. The study
will demonstrate the need for proper impact assessment in the
development of animal disease programs and its value toward
improved decision-making.
Eradication of the Tropical Bont Tick from the Caribbean has
been under consideration for at least 7 years (Consortium for
International Crop Protection, 1987). The tick, currently found on
16 islands and slowly spreading, is associated with two livestock
diseases of severe economic impact, heartwater and
dermatophilosis. Spread of the tick and its associated diseases to
uninfested islands and neighboring continents is considered to pose
a serious threat, particularly to the US.
In 1987, a task force was formed to develop a feasibility study
proposal for the management of the Tropical Bont Tick in the
Caribbean basin. As a result of the study, the task force
recommended eradication rather than control as the preferred
strategy for dealing with the tick. Eradication was believed to be
the lowest cost method of management and it eliminated both the
long-term costs to producers and the environmental burden of
insecticides.






Intensive review of the feasibility study raised major
questions concerning the costs of eradication. Several of the
component studies of the eradication proposal and the final review
analysis raised concerns about the regional presence of another
livestock disease complex; the Boophilus tick and the cattle disease
it transmits, babesiosis. Contributing experts for the feasibility
study, including the participating agricultural economist,
acknowledged the possibility of disruption of the endemic stability
of this often-costly disease (Ristic & Krier, 1981). No methodology
or framework for dealing with such an externality was in use at that
time and no attempt was made to evaluate the risks. Estimation of
the potential costs and quantification of the threat of endemic
disruption were deemed impossible. One continuing question has
been whether the ultimate objective of Tropical Bont Tick
management should be eradication or control and, further, whether
eradication is even feasible given the environmental constraints and
existing social patterns.
A pilot eradication project and economic impact study were
recommended for Antigua, a highly infested island, as a means of
collecting the critical information needed for making decisions
about future Caribbean-wide Bont tick eradication. Even though
$3.49 million was allocated for the Antigua pilot eradication
attempt, the stated acknowledgment was that, at the conclusion of
the project, eradication might not be achieved. Only passing concern
was given to the disrupted disease state in which the animal
population would be left should this eradication attempt fail. No
impact assessment was attempted.







The decision was made to use amitraz as the acaricide in the
Antigua pilot project as a result of a Tropical Bont Tick eradication
campaign in Puerto Rico, and an environmental assessment was
prepared for use of this compound. However, in the interim another
acaricide surfaced which might have had an effect on program
efficacy. A separate environmental assessment was conducted to
examine the potential use of this alternative acaricide, flumethrin.
This second environmental assessment compared the two acaricides
for efficacy, potential impacts on the physical environment, and
cost-effectiveness.
The pilot project environmental assessment included a
risk/benefit analysis incorporating "corrected" cost and benefit
estimates to address some of the doubts surrounding the cost and
benefit values estimated in the original feasibility study. However,
the environmental assessment team did not attempt to 1) measure
or assess the impact of the consequential altered immune status of
the animals (to either target or non-target diseases), 2) make any
predictions of the potential for owner noncompliance or estimations
of the economic consequences of this possibility, or, 3) attempt to
evaluate the risk and consequences of failure of the eradication
objective if the program were implemented.
An examination of how these risks or impacts may vary across
program alternatives was not conducted, nor were a thorough
evaluation and comparison of program alternatives. These failings
are likely the result of inadequacies in the scoping process, the lack
of adequate impact assessment protocols, and insufficient
information and technical capability to measure and evaluate






impacts (i.e., personal computers and adequate software were not
readily available at the time). In fact, the risk-benefit analysis
conducted for the environmental assessment report (required in all
project appraisals) included only estimates for the standard cost
and benefit economic variables associated with direct program
effects (Dicks, 1993, pers. comm.). True "risk" impact parameters
received minimal verbal attention and were not compared across
program alternatives.


Objectives

The overall goal of this dissertation is to encourage and
promote a more holistic approach to decision-making with regard to
large-scale animal disease management programs through
refinement of the environmental assessment and/or environmental
impact statement. This will be accomplished through establishment
of a protocol for the preparation of a risk assessment within the
environmental assessment or environmental impact statement.
Research objectives are as follows:
1. Review the current formats and regulations for the
preparation of environmental impact statement's and environmental
assessment's in the development of government programs.
2. Propose and evaluate refinements and/or improvements in
preparation of the statements) to allow for more complete impact
assessment consistent with the spirit and letter of NEPA
regulations.






3. Design a generic framework for preparing an environmental
impact statement/environmental assessment risk assessment
component in animal disease pest control programs. (This will
assist agencies/preparers to a) identify potential impacts and/or
risks of the proposed action and alternatives, b) assess,
qualitatively or quantitatively, the identified impacts, c) examine
how risks and impacts vary across program alternatives, and d)
present qualitative or quantitative results in comparative form to
assist decision-making.)
A case study will be used to
4. investigate and demonstrate how non-target effects can be
systematically recognized and evaluated within the environmental
assessment or environmental impact statement,
5. investigate and demonstrate how local social and biological
system interrelationships have impact on the potential for program
success, how these may be evaluated in an environmental
assessment/environmental impact statement, and how the expected
outcome of course of action alternatives may change as a result of
consideration of these impacts,
6. demonstrate how analytical risk analysis methods and
procedures can be used in an environmental assessment and/or
environmental impact statement empirical risk/benefit analysis.


Overview of Remaining Chapters

Chapter 2 contains a review of the literature dealing with
impact identification in public programs, and the methods used in






risk assessment and impact measurement. This includes a review of
the tools used to identify risks and incorporate uncertainty in the
analytical methods of program evaluation. A framework for the
identification of potential direct and indirect effects of animal
disease programs involving vector control is the topic of Chapter 3.
Chapter 4 examines the 1987 Antigua pilot eradication project
environmental assessment and identifies areas for improvement in
the environmental assessment preparation process. Use of the
framework developed in Chapter 3 is demonstrated in Chapter 5 in a
refined impact assessment for Tropical Bont Tick management on
Antigua. The new assessment includes the development and analysis
of an alternative to eradication. The results of an economic impact
study conducted on Antigua in 1989-90 provide the information
needed to quantify and evaluate risks and uncertainties present in
the proposed eradication project. Reasonable estimates are used for
purposes of demonstration where there is a lack of necessary
information. Estimates are only proposed for variables that could be
either evaluated scientifically and within reasonable time and
monetary constraints or estimated through expert opinion. A
detailed risk/benefit analysis is conducted in Chapter 6
incorporating quantifiable risks into standard benefit-cost analysis
methodology. The summary section in Chapter 6 examines and
discusses the results of the refined impact and risk assessment of
the proposed TBT eradication program and its alternatives.
Chapter 7 will summarize this research and propose areas for
future work in impact identification and measurement in disease
program formation and appraisal.











CHAPTER 2
LITERATURE REVIEW

Regulations for Environmental Impact Statement Preparation

In 1969, the Senate and the House of Representatives enacted
the National Environmental Policy Act (NEPA). The purpose of the
act is "to declare a national policy which will encourage productive
and enjoyable harmony between man and his environment; to promote
efforts which will prevent or eliminate damage to the environment
and biosphere and stimulate the health and welfare of man; to enrich
the understanding of the ecological systems and natural resources
important to the Nation" (CEQ, 1978, page 34).
Section 102(2) of Title I of the Act directs that all agencies of
the federal government must utilize a systematic, interdisciplinary
approach to insure the integrated use of the natural and social
sciences and the environmental design arts in planning and in
decision-making which might impact on the environment.
Regulations were set down for the preparing and writing of an
Environmental Impact Statement (EIS) to implement Section 102(2).
The primary purpose of an EIS is to serve as an action-forcing device
to insure that the policies and goals defined in NEPA are
incorporated into ongoing programs and actions of the Federal
government. Preparation of an EIS must begin as close as possible
to the time the agency is developing, or is presented with, a






proposal so that preparation can be completed in time for the final
statement to be included in any recommendation or report on the
proposal. The EIS must be prepared early enough so that it can serve
practically as an important contribution to the decision-making
process and is not to be used to rationalize or justify decisions
already made.
A standard format for an EIS has been established and includes
among its sections 1) purpose of and need for action, 2) alternatives
including proposed action and, 3) environmental consequences.
The alternatives section is the very heart of the EIS (CEQ,
1978). This section is to present the environmental impacts of the
proposal and the alternatives in comparative form, sharply defining
the issues and providing a clear basis for choice among options by
the decision-makers. The statement preparers are to rigorously
explore and objectively evaluate all reasonable alternatives and for
alternatives eliminated from detailed study, briefly discuss the
reasons for their having been eliminated. They must also devote
substantial treatment to each alternative considered in detail
including the proposed action so that reviewers may evaluate their
comparative merits. The environmental consequences section forms
the scientific and analytic basis for the comparisons made in the
alternatives section (CEQ, 1978). Discussions of the direct and
indirect effects of the alternatives, including the proposed action,
and their significance are to be included in this section. The literal
definition of effects includes 1) direct effects, which are caused
by the action and occur at the same place and time; 2) indirect
effects, which are caused by the action and are later in time and






further removed, but are reasonably foreseeable" (CEQ, 1978, page
28). In sum, "effects includes ecological (such as the effects on
natural resources and on the components, structures, and functioning
of affected ecosystems), aesthetic, historic, cultural, economical,
social, or health, whether direct, indirect, or cumulative" (CEQ,
1978, page 28).
Not all agencies nor all program proposals require an EIS be
prepared. Categorically, certain types of proposals automatically
require an EIS; other types specifically do not require an EIS. If the
proposed action is not covered by one of these two specifications,
then the agency must prepare an environmental assessment (EA) and
based on this EA, make its determination whether to prepare an EIS.
A finding of "no significant impact" on an EA eliminates the need for
an EIS. Among other functions, the EA serves to aid an agency's
compliance with NEPA and facilitate preparation of an EIS when one
is necessary. It should include brief discussions of the need for the
proposal, of alternatives, of the impacts of the proposed action and
alternatives, and a listing of agencies and persons consulted. An EA
may be prepared at any time by an agency for purposes of decision-
aiding in cases where there is no requirement for one.
An EIS is to be "analytical rather than encyclopedic" (CEQ,
1978, page 9). However, it is acknowledged in the regulations that
when evaluating significant adverse impacts, there may be gaps in
relevant information or scientific uncertainty. If it is neither
technically feasible nor cost-effective to gather this information,
the statement preparers must "weigh the need for action against the
risk and severity of possible adverse impacts were the action to






proceed in the face of uncertainty" and "it shall include a worst case
analysis and an indication of the probability or improbability of its
occurrence" (CEQ, 1978, page 15). The Council on Environmental
Quality (CEQ) published (51 FR 15618, April 25, 1986) a regulation
for dealing with incomplete or unavailable information when
preparing an EIS. The regulation calls for disclosing when vital
information is incomplete or unavailable, explaining its relevance,
and summarizing the existing credible data. The regulation also
calls for using the best available information, methods, and
theoretical approach for estimating impacts. This constitutes
conducting, when feasible, a formal risk assessment on the
identified significant impacts. The overall intent of the regulation
is to provide decision-makers with the best available professional
assessments on potential or likely impacts (Marcot & Salwasser,
1991).
Despite the appeal of formal risk assessment for impact
evaluation, it is rarely done, and even when conducted is often
ignored. Reasons for avoiding risk analysis include the difficulty of
structuring an analysis, lack of easily available data (probabilities,
likelihood's, future costs and payoffs), lack of time to carry out an
analysis before a decision must be made, and, probably most
important, a lack of understanding about risk analysis by both
proposal analysts and decision-makers. Many program decisions are
governed by momentum from earlier decision-making which might
have been based on rules-of-thumb, blanket decisions based on
averages, and bureaucratic preprogrammed decisions (Beuter, 1991;
Rosenberg, 1979). Decision-makers are often doubtful when risk







analysis supports decision responses contrary to those made in the
past and they lack confidence in likelihood's and expected values
generated by risk analysis. Some decision-makers are even
reluctant to accept the possibility of bad outcomes made visible by
risk analysis (Beuter, 1991). Nevertheless, both governmental and
nongovernmental agencies are increasingly challenged to recognize
and evaluate the risks associated with government programs and
policy. Risk analysis has enhanced decision-making by allowing for
quantification of the chances of adverse outcomes (Lave, 1987). The
challenge is to incorporate risk analysis into the preparation of an
EA or EIS, in keeping with the intent of the NEPA and CEQ
regulations.

Assessment of Environmental Impacts

The term "environment" is broadly defined as the natural and
social conditions surrounding all mankind and including future
generations for the purpose of conducting any form of environmental
assessments (EA's), whether referred to as environmental impact
statement (EIS), EA, or environmental impact assessment (EIA)
(World Bank, 1991). Environmental assessments prepared for the
Bank must take sociocultural and health effects and the impacts on
cultural property and indigenous peoples into account as well as
evaluating impacts on the "natural" environment. In other words, the
environment is a combination of all natural and physical
surroundings and the relationship of people with that environment,
which includes aesthetic, historic, cultural, economic, and social






aspects. All these elements should be considered in an EIA (Jain et
al., 1981).
Most government agencies take a very broad view of the
environment in establishing EA guidelines and require the analysis
of economic and social impacts in addition to physical impacts. In
most current planning, meaningful environmental impact assessment
involves just about everything (So et al., 1986). "Impacts" is an
equally broad term and, according to the letter of NEPA regulations
noted above, should include direct, indirect and cumulative effects
on almost any imaginable "environmental" parameter. However, in
general, an EIA can be described as a process having the ultimate
objective of providing decision-makers with an indication of the
likely consequences of their actions (Wathern, 1988).
Impact assessment is a technical process concerned with the
identification, investigation, and refinement of alternative options
and has particular utility in evaluating alternatives and selecting a
plan (So et al., 1986). It aims at avoiding, reducing or mitigating
any adverse effects of program or plan implementation (So et al.,
1986). Wathern (1988) advocates the further development of this
capability as the most logical step in the evolution of the EA
process. There exists a duality in EA preparation as it includes both
art and science. Merkhofer (1987) finds that alternative generation
is an art, limited only by the biases and focus of the team members.
The science lies in the methods employed in impact assessment and
measurement.
There are many books and articles on the activities involved in
the impact assessment process. Because of the complexity of






environmental systems and the specialized functions of the various
public agencies involved in the EA process, it is unlikely that one
universal method will ever be developed or would even be
appropriate in all cases (Rau & Wooten, 1980). However, some
common threads exist as do some common criticisms.
A starting point in developing an EA is the "scoping phase."
This is the process of identifying a number of priority issues to be
addressed in the EA from a broad range of potential problems
(Wathern, 1988; Beanlands, 1988). It is an attempt to focus the
assessment studies on the most significant potential effects.
Social and economic values are major factors used to narrow the
range of ecosystem components which should be considered
(Beanlands, 1988). In addition, there are fiscal and time realities to
limit how sweeping an EA can be (Kennedy, 1988). The time and
money needed for an EA vary with the size, type and location of the
project itself. Factors determining these include the amount of
information readily available versus that which must be gathered,
whether the assessment is done in-house, and when the assessment
is conducted in the proposal/feasibility study process (Kennedy,
1988). Boundaries are determined for the extent of impact
evaluation within the scoping and conducting of an EA.
There are two basic kinds of impacts, direct (or primary) and
indirect (or secondary) (So et al., 1986). NEPA regulations include a
third category, often ignored in impact assessment literature, that
of cumulative impacts. Primary impacts are those directly related
to the location, construction and operation of projects or programs.
They are important, but secondary impacts may be even more






important (So et al., 1986). These indirect impacts arise from
subtle, often long-term changes in the biophysical and economic
system brought about by program actions. It is these secondary or
indirect impacts which may be of most concern in animal disease
programs. Direct manipulation of biological system components,
such as species elimination, while having fairly easily measured
direct effects such as increased production and reduced veterinary
costs, may have indirect impacts much more difficult to evaluate,
such as reduced disease immunity. It is the complex nature of
disease control programs that has led to their being included on the
list of development aid projects most in need of EA by the OECD
member countries (Kennedy, 1988).
The major elements included in the preparation of an EA are
the identification, measurement, interpretation and communication
of impacts (Wathern, 1988; Rau & Wooten, 1980).

Impact Identification and Communication

Various methodologies have been developed and standardized
for impact identification and communication; the most commonly
practiced follow.
Ad Hoc. These methods provide minimal guidance for total
impact assessment while suggesting the broad areas of possible
impacts and the general nature of these possible impacts (Rau &
Wooten, 1980). These statements are qualitative and could be based
on subjective or intuitive assessments, or could be qualitative
interpretations of quantitative results. Simple identification of the






nature of an impact may be related to decision-makers as no effect,
problematic, short- or long-term, and reversible or irreversible.
Impact Checklists. These are found in one form or another in
nearly all EIA methods. A checklist combines a list of potential
impact areas that need to be considered in the EA process with an
assessment of the individual impacts (Wathern, 1988; So et al.,
1986; Rau & Wooten, 1980; Munn, 1979; Canter, 1977). Four broad
categories of checklists can be defined (Canter, 1977). Simple
checklists consist of a list of parameters; however, no guidelines
are provided for measurements or interpretation. Descriptive
checklists build on the former by providing guidelines for parameter
measurement. Scaling checklists, similar to descriptive, add
information basic to subjective scaling of parameter values.
Scaling-weighting checklists build on scaling lists with information
provided as to subjective evaluation of each parameter with respect
to every other parameter. A drawback to the checklist methods is
the "tunnel vision" that results from an established or
predetermined impact list for reference.
Matrices. These function as both checklists and display
relationships (Wathern, 1988; So et al., 1986; Rau & Wooten, 1980;
Canter, 1977). Matrices basically incorporate a list of project
activities or actions with a checklist of environmental conditions or
parameters that might be affected. Combining these lists as
horizontal and vertical axes in a matrix allow for the identification
of cause-effect relationships--lacking in the checklist methods--
between specific activities and impacts. The entry in any cell can
be either a quantitative or qualitative estimate of these cause-






effect relationships. The latter are in many cases combined into a
weighting scheme leading to a total "impact score". A weakness of
this method, shared with checklists, is its restricted use in
comparing alternatives (So et al., 1986). Though ideally-suited for
impact identification, matrices are weak at identifying indirect
impacts (Wathern, 1988).
Networks or Flow Diagrams. These methods start with a list
of project activities or actions, then generate cause-effect impact
networks (Wathern, 1988; So et al., 1986; Rau & Wooten, 1980;
Canter, 1977). This type of method is basically an attempt to
recognize that a series of impacts may be triggered by a program
action. Networks give exposure to indirect effects and show both
joint and individual impacts (So et al, 1986). The idea is to start
with a project activity and identify the types of impacts that would
initially occur, then to select each impact and identify the impacts
which may be induced as a result. This process is repeated until all
possible impacts have been identified. It is easy to imagine that if
all impacts are detailed and all possible interrelationships are
included, the resulting impact networks could become too extensive
and complex to be truly useful (Rau & Wooten, 1980; Munn, 1979).
Another drawback is that networks are usually unidirectional
without feedback loops (So et al., 1986).

Impact Measurement

The tools described above assist statement preparers initially
in identifying impacts and ultimately, in organizing their
presentation to decision-makers. After identification, the






measurement of impacts is the next step in the assessment process.
Ideally, all impacts should be translatable into common units of
measurement; however, this is not usually possible. Due to the
diverse nature of impacts, measurement units tend not to be
commensurate. Major difficulties in impact measurement and
comparison include lack of information and quantification
techniques (Rau & Wooten, 1980). Therefore, most EA's make use of
both qualitative and quantitative measurement methodologies.
There is no doubt among researchers that EA's still need a healthy
injection of scientific rigor, particularly where impact prediction is
concerned (Bisset, 1988; Whitney & Maclaren, 1985).
Impact prediction is a central element of the impact
assessment process. Known effects and anticipated impacts are the
vanguard for alternative comparison and evaluation. However, this
phase of EA receives little attention and remains relatively
underdeveloped (So et al., 1986). Culhane et al. (1987), find that
most EIS's suffer from varying degrees of imprecision; most often
represented by qualitative assertions with no clear statement of
either the likelihood or significance of the impacts. EIS's usually
offer only vague generalizations about possible impacts. Such
predictions are of little value to decision-makers because of their
ambiguous nature (Bisset, 1988). It wasn't until the 1980's
discussions on the scientific content of EA's that "uncertainty" was
first mentioned as an important issue. This lead to recognition of
the need for development and appropriate use of scientifically
defensible impact prediction techniques, including methods that






would yield a range of predictions and associated probabilities for
those predictions to occur (de Jongh, 1988).
Whereas problems of uncertainty have only recently been
addressed in EA's, in other related fields, such as risk assessment,
the management of uncertainty is well established. There are many
methods available for reducing uncertainty in prediction. However,
constraints on time and money for EA preparation are sometimes
severe and it may not be possible to develop mathematical models
for prediction (de Jongh, 1988). Often, a system is not sufficiently
understood for a reasonable model to be developed or resources for
building or improving predictions are simply not available.
Two commonly used methods for handling uncertainty in input
data and prediction in EA's and risk assessment are sensitivity
analysis and Monte Carlo simulation (de Jongh, 1988; Whitney &
Maclaren, 1985; Munn, 1979). Sensitivity analysis is used to
identify those inputs which contribute most to uncertainty in
prediction. Inputs can then be ranked according to their priority for
further research to improve the accuracy of predictions. One method
developed to identify high-priority inputs relies on the creation of a
sensitivity index (Boggess & Amerling, 1983). In this method, key
inputs are identified by determining the percent change in the model
output, such as net present value, given a percent change in an input
parameter. Inputs with the highest sensitivity index value have the
greatest influence on model predictions. The sensitivity index of
each input and the cost of improving the accuracy of each would then
be considered in assigning priorities for research effort. The Monte
Carlo simulation technique is used to predict the probability






distribution of possible outcomes under different scenarios taking
into account input uncertainty. To conduct the simulation,
probability density functions (PDF) are specified for chosen
uncertain input parameters. A large number of input sets are
randomly selected and each is used to make a deterministic
prediction.
In many cases, particularly in the biological and health
sciences, the generation of objective PDF's is not possible. Methods
of probability encoding have been developed for eliciting expert
opinion on uncertain variables to generate subjective PDF's. The use
of a Delphi survey, an iterative interview technique, has been well
established in these fields (de Jongh, 1988; So et al., 1986; Jain et
al., 1981; Schoenbaum et al., 1976). Once PDF's for uncertain
variables are generated, easy-to-use software packages for personal
computers such as @Risk1, can be used to run simulations
incorporating them. These programs have made Monte Carlo
simulation a fast, flexible, and inexpensive method of uncertainty
estimation. They permit easy manipulation of parameter values and
examination of how risks and uncertainties impact on decision
criteria, such as net present value, and how these can be reduced in
alternative action comparison. The results can be plotted to give a
PDF of prediction outcomes.






1Risk Analysis and Simulation Add-In for Microsoft Excel, Palisade
Corporation, Newfield, New York. 1991.








Risk Assessment

Environmental impact assessment and risk assessment (RA)
have evolved as parallel and sometimes overlapping procedures for
rational reform of policy making. As with other forms of policy
analysis, such as benefit-cost and cost-effectiveness analysis, they
share the common conviction that the decision processes in policy
formation can be improved with the inclusion of explicit analysis
and documentation (Andrews, 1988). In practice, however, EIA and
RA have been nurtured by different disciplinary and professional
communities in largely separate policy contexts. However, risk
assessment procedures employ many of the same techniques as
impact assessment for managing uncertainty in analyses, most
notably sensitivity analysis and Monte Carlo simulation (Jasanoff,
1993; Thompson et al., 1992; Burmaster & von Stackelberg, 1989).
Many researchers now advocate consolidation of the two closely
related fields of risk and impact assessment as the only way to
produce anything like a comprehensive accounting of the nature and
extent of risk in a technological society (Jasanoff, 1993; Andrews,
1988).
Risk assessment and analysis techniques have been developed
to organize complex risk information to lead to better decision-
making (Lave, 1987; Russell & Gruber, 1987). As an example,
classical risk assessment protocols are widely used for evaluation
of nuclear power installations (Okrent, 1987) and for the use of
hazardous chemicals in the environment. In these applications the






hazard is often self-evident and can be defined in terms of specific
dose-response relationships. The classic formula for characterizing
risk in these types of applications is; Risk = Probability x Severity
(Wilson & Crouch, 1982). When actions incur risks, the
straightforward approach to evaluate the risk is to break down each
action or event into its constituent components, to evaluate the risk
measure for each, and then to sum them to obtain the resultant
measure of overall risk for the original action or event (Wilson &
Crouch, 1982).
The majority of formal risk assessment applications in the
public arena lie in the evaluation of risk to human health. However,
there has been a steadily-increasing awareness that risks to
ecosystems and the environment as a result of major actions must
also be considered. Although the need to recognize, evaluate, and
manage such risks has been acknowledged in several agricultural and
related fields, direct application of established risk protocols to the
more complex biological issues facing government decision-making
today is proving difficult. Identification and measurement of non-
target effects and ecosystem disruption are just beginning to be
addressed in government program and policy formation.
In the rapidly expanding field of biotechnology, Parry (1991)
points out that problems arise in evaluating risk due to the lack of
an established risk assessment protocol. According to Parry, the
idea of a rational framework to evaluate non-target effects of
current technologies is relatively new and the absence of such a
framework frequently leads to confusion by consideration of all
theoretically possible impacts and even improbable speculation.






Within the USDA, risk assessment has always been a
formalized aspect of National Forest Service planning against
wildfires. Current methods use stochastic modeling with expert
opinion to determine optimum strategies for fighting wildfires that
escape first attack. Studies have demonstrated how decision trees,
Bayesian decision analysis, Markov processes, and game theory can
be used to estimate maximum expected values for forest resource
decisions (Beuter, 1991). The Forest Service now recognizes that
the existence of large tracts of national forest is being increasingly
challenged and there is a need to expand risk assessment within a
holistic ecosystem management approach that provides for
sustaining whole ecosystems in relation to all sources of change and
risk (Beuter et al., 1991). They acknowledge that this challenge is
extremely complex and burdened by lack of scientific data on
ecosystem relationships and management interactions.
Risk assessment is also being explored in the planning and
management of wildlife species and habitats within our national
forests. The Forest Service finds that the less-than-perfect
information on wildlife species and their habitats is difficult to
handle within the EIS framework. In most wildlife viability risk
assessments, risk factors are qualitatively ranked from low to high
based on how well each possible management alternative performs.
A ranking of likelihood's is made in ordinal-scale categories rather
than quantified as ratio-scale probabilities due to lack of
understanding of how environmental risk factors compound (Marcot
& Salwasser, 1991). Ordinal scale rankings have been used in EIS
and analysis documents to evaluate efficacy of management plans






for the northern spotted owl (Strix occidentalis caurina) (Marcot &
Salwasser, 1991).
Formal risk analysis for chemical use to control native plant
pest species is well established. However, the current trend is
toward nonchemical methods including the use of native or exotic
biological control agents. When the classical formula for
environmental chemical risk is applied to the release of exotic
organisms, this model may greatly over- or under-emphasize risks
because the biological interactions are complex. When the chemical
model is used for ecological studies, the approach tends to lead to
quantification of a few well-studied interactions rather than
identification of likely interactions. The reason for this is, in part,
the model's over-reliance on quantifiable measures. Since much of
the data on biological interactions is qualitative and doesn't fit well
into the model, it may be overlooked or underutilized.
Medley & Payne (1991) propose that a more-correct EA for the
release of an exotic biological organism contain one or more risk
components. The environmental analysis in this case might be
characterized as containing a biological assessment in a specific
ecological framework. Their proposed framework is comprised of
questions posed more formally than in the past so as to provide a
structured review analysis within the EA of the organism and
potential positive or negative impacts that may occur from its
introduction. Examples of the questions Medley and Payne propose
are as follows: 1) What is the organism? 2) What is the biology of
the organism in its natural environment? 3) What are the
characteristics of the new environment? 4) What is the organism






intended to control in the new environment? 5) What other
organisms is it expected to significantly interact with in the new
environment and what are the potential outcomes of this
interaction? and, 6) What procedures or controls are available to
mitigate any negative impacts identified? It is important to note
that the questions are of a form that would allow the use of valid
qualitative information as well as quantitative data to provide the
answers. The authors purport that use of the component in the
review process provides an appropriate structure for informed and
responsible decision-making.
According to the World Bank (1991), the proper response to
risk for the purpose of project appraisal is to count it as a cost in
expected value functions. In practice, the way risk and uncertainty
are included in EA's is through sensitivity analysis, which evaluates
how the net present value, internal rate of return, or other decision
criterion, is impacted by different variables (World Bank, 1991).
The Bank recognizes that "external factors" have been neglected in
the past, but these should now be internalized to the extent possible.
At this point in time, impact assessments in animal disease
programs are narrow in scope. Though they must include an
economic analysis, namely benefit-cost analysis, the variables
incorporated are limited to standard cost and benefit estimates of
program objectives (Dicks, 1992, pers. comm.). Sensitivity analysis
is the only currently practiced method for evaluating uncertainty in
these "direct effect" variables for an EA in disease programs. No
indication can be found in the literature that probability analysis, by
Monte Carlo simulation or any other technique, has been attempted in






this field, although these methods have been used for uncertainty
estimation in public programs since the 1970's (Reutlinger, 1970;
Pouliquen, 1970). Benefit-cost and risk-benefit analyses must be
pushed to the limit to incorporate risk and uncertainty for sound
decision-making (World Bank, 1991).
The following chapters present, as a case study, the history of
one livestock program proposed for the Caribbean basin, and through
expansion of the environmental assessment prepared for this
program, present a new approach for responding to the increasing
concern over recognizing and evaluating the indirect effects of
program actions.











CHAPTER 3
A NETWORK FOR RISK AND IMPACT IDENTIFICATION
When nations are confronted with diseases of severe
consequence, whether animal or human, there is little choice but for
governments to act. In the case of human diseases, massive
campaigns, which sometimes cross national borders, are often
undertaken. These campaigns may be based on control measures to
limit the incidence or severity of disease, such as childhood
vaccination for polio, or on eradication measures to completely
eliminate the vectors or agents of disease, such as the successful
eradication of smallpox or the eradication of malaria from the US by
eliminating the mosquito vectors of transmission. However, animal
disease programs are more constrained by cost-efficiency measures.
They are not based on reducing animal suffering, but rather on
economic parameters such as producer and consumer benefits, food
security and self-sufficiency, and improving national income. There
are few agencies involved in the funding and implementation of
these national or regional programs. For example, the USDA has
almost exclusive domain over the intervention programs waged
against livestock diseases within or of threat to the US and has been
conducting animal disease eradication campaigns since 1884
(Hanson & Hanson, 1983). The published USDA stance on the
preferred course of action is to eradicate a disease when it is
technically feasible to do so (Animal and Plant Health Inspection






Service, 1971). Technical feasibility of eradication refers to the
identification of means to attack a disease pathogen or agent, and a
target stage in which to fatally disrupt its transmission cycle. In
general, if the benefits of a program, i.e. increased production,
decreased production loss due to disease, increased food security,
and increased producer income outweigh the costs of
implementation, the project is undertaken.
Livestock projects, in general, are diverse in nature. However,
in many cases, the major animal disease programs undertaken by
public agencies involve the eradication or control of the vectors of
disease. In particular, ticks and their associated tick-borne
diseases (TBD) are a major focus of developmental programs in
livestock improvement and disease control throughout the world.
Ticks play a major role in disease transmission and economic
hardship in many countries. Approximately 80 percent of the world's
cattle population of 1,281 million are at risk from ticks and tick-
borne diseases (Pegram et al., 1993). In 1979, McCosker estimated
global costs of control and productivity losses associated with ticks
to be about $7 billion annually. Ticks and TBD's are considered
major constraints to livestock development in developing countries,
and in particular, the tropics. Past and current programs aimed at
tick management have focused on total eradication as a primary
objective and feasibility studies concentrate on methods available
to obtain this goal. These programs provide a unique, though
recurrent, case for a melding of risk and impact assessment
techniques.






Proposed Network

Due to the prominence that tick control programs have in
livestock development, and the availability of a specific case study
involving the control of the Tropical Bont Tick, a network for impact
assessment is proposed in Figure 1. It must be understood that it is
limited in use to the development of tick management programs. In
all likelihood, no one universal generic format for risk and impact
identification can be created which will adequately identify all
potentially significant effects in all types of disease programs.
However, the rationale used for creating the cause-effect linkages
highlighted in the proposed network could be adapted for use in other
types of disease control programs as well. The types of risk factors
and impacts included in this network have a strong historical basis.
Study of past disease control programs lends support to advocating
that in the planning phase of these programs attention be paid to the
factors that could affect program success, to the potential for
program failure, and to the possible consequences of either of these
outcomes. There is evidence from past programs that a tunnel-
vision view of technical feasibility, supported by a cost-
effectiveness criterion, may have been responsible for the failure or
protraction of some disease programs. Past programs have failed or
had unexpected negative economic consequences in part as a result
of their failure to consider fully the types of linkages depicted in
the proposed impact and risk network.





























z z z z z z z z z I z I
I I- I- I- SI5-=5" Zy
01Z | Z 1 Z Z
< < < < < < < < < < < <
o U z (D z C z (
05 n z ( n zV z -l z


I II
z
UJ
.1 z
0 0 *0
u (A 0


0
U U
FO
..-4 L4.











0
l LU u- (A Cl


< a 2 >) o z
z .) V) u < ]
Z 0Z 0
z 0 (n 0
LU 0 o c
o ,- z ( 0










0 .ua =): ZL w D w'
0 0 --
-j u U] .4 _









z -' io o
D ex U OU 5 0 W
0 z M 0
UL z LL 0 uj ui L ]








o o- -5
0 0 0 0




0 0 <
Su Z LU 0 LU
LU LU LU H 0.7 0 --


-J Z '7 _J


> z ) u a. ui 0.

*'~~~U a SS '
ixs z (D Ia
0D <4 D 0 <


hi' 0 -j < l ^ a ^ 7 a'
s ~ 0 m tz 0 0 ^ l"fc a
o~ F- 0 m z f -


LU ,
0 0 w
H 0
t/\ UJ UnM n L J Z (
> I- H u! 0 H
0 .4 0 :D.


= = Z aJ u -





< cr 0- wc Es LL i D n L
a.~C LL0 c nU


0 < LU 4- J
>U Qi Z .4U] L Z
.4Ln o QL J LU
a -J UUz ~ 4 u uj

L U L_ L 0 .





u z
< 4 ~ 0 LU0








UJ-



LUU
0i 0














zz
Lu t 0j31. Z
UJ a <4 0


FU-





















I



Q.
0~









Risk and Impact Linkages

The proposed network illustrates the linkages, some of which
are causal, between potential biophysical and socioeconomic risks
to program objectives, potential program outcomes, and the possible
direct and indirect impacts of program actions and outcomes.
Application of this tool to a particular tick management project
leads to the identification or delineation of linkages specific to that
program.
The first types of linkages identified in the network relate to
risk characteristics and factors that impact on the potential for
successful completion of the stated program objective. In several
past programs impediments to the successful fulfillment of program
directives had economic backlashes that substantially reduced or
negated projected program benefits and, at the same time, increased
program costs. The rinderpest eradication attempt, mentioned
previously, is one example. Political upheavals interrupted the flow
of funding for the program just prior to conclusion and the result
was a complete breakdown of the program effort. This led to an
epidemic outbreak of the disease which would be identified in the
network as an indirect impact of a failed program objective, before
the program could be reestablished. The epidemic occurrence of a
disease can be caused by a combination of the disruption of the
immune status of the indigenous livestock population, reintroduction
or resurgence of the pathogen or agent, and the presence of carrier-
state animals. When the disease recurs, production and mortality






losses may be high. In the rinderpest case, the losses to local
producers resulting from the epidemic backlash were estimated to
be in the millions of dollars, much more than would have been
realized by uninterrupted occurrence of the disease (Sollod, 1992).
The rinderpest program is now considered to be nearing a successful
completion, much later and at a much higher social and resource cost
than ever anticipated. In another example, an intensive cattle-
dipping program in Zimbabwe was disrupted by war in the 1970's and
the resulting epidemic backlash claimed the lives of approximately
one-sixth of the cattle population (Lawrence et al., 1980). These
examples are an indication of how factors that interfere with
program implementation can cause consequences of severe economic
impact. In these case's the failure to achieve program objectives on
schedule had repercussions due solely to the nature of program
actions, i.e., manipulation of key elements of mature, stable
biological ecosystems. Most public programs, such as highway
construction, are not likely to initiate comparable rebound impacts,
especially ones that would directly impact program "beneficiaries".
A final case example for the factors affecting program
success is the long-standing USDA effort to eradicate brucellosis
from US dairy herds. Brucellosis eradication first became
compulsory on a state-wide basis in Wisconsin after the
announcement that all milk marketed in Chicago after January 1,
1955, must be from herds free of the disease. The program fell
behind schedule very early when not enough veterinarians were
available to bleed all the cattle in the state within the required
three-month time frame. Eradication of brucellosis from Wisconsin






was eventually accomplished at a cost of $46 million. However,
when the same implementation techniques were applied to
eradication efforts in southern states, the program began to fail
again and major changes had to be made. While success appears to
be possible in the foreseeable future, the program has gone
drastically over-budget and taken far longer than projected (Hanson
& Hanson, 1983).
These case examples depict how program failure (or
protraction) can initiate a chain of events with dramatic economic
consequence; however program success can as well. Such
possibilities are also highlighted in the risk network and can include
impacts on non-target diseases as well as target diseases if the
methods used to combat a pathogen or agent affect other disease
complexes. There is precedent for advocating evaluation of the
socioeconomic and biophysical roles played by target and affected,
non-target species. There are several examples in the literature on
disease control where indirect, unanticipated impacts had dramatic
consequences. For example, in southeast Asia, DDT sprayed into
thatch-roofed houses to control malaria vectors killed predacious
insects that kept thatch-consuming insects under control;
consequently, the roofs deteriorated (Cheng, 1963). Another
example is the case of African swine fever on Haiti, mentioned
previously, where depopulation of the native Creole pig resulted in
the costly and dangerous build-up of human waste.








Use of the Proposed Network

The proposed risk and impact identification tool is based on
the network or flow diagram impact identification methodology
summarized in the previous chapter. The network tool proposed here
differs somewhat from that described in Chapter 2 in that it
includes not only impacts triggered by program actions but it also
attempts to identify risks to program success. The addition of this
function results in a network that combines the cause-condition-
effect linkages of standard network methods of impact assessment
and decision-tree methods of risk assessment. The network
represents a conceptual framework, highlighting how interdependent
are the economic, social, biological and environmental factors of
animal diseases affecting agricultural production. The network is
not designed to function as a quantitative model as is sometimes
offered in decision tree diagrams. Indeed, it is likely that many of
the risks and impacts involved in disease control programs cannot be
quantified. However, the magnitude of potential risks or impacts
identified with the network can be qualitatively described for
comparison across the program alternatives within the EA and for
presentation to decision-makers.
Environmental assessment's are usually prepared by
multidisciplinary teams. These teams normally include human
health, veterinary health, economic, and environmental
professionals. The proposed network is designed to assist the EA
team members in focusing their areas of expertise on some of the






unique aspects of tick control programs. The tool is proposed as an
adjunct to the standard impact identification techniques already in
use. These standard techniques typically direct impact evaluation
toward preestablished categories of environmental concern. In the
case of tick control these categories include pesticide toxicity
levels, chemical handling, storage and disposal, and potential off-
site impacts.
The network is meant to be used initially in the scoping phase
of statement preparation when the specific tasks of the statement
are being defined and set down by the team and funding agency. The
team will have already had an opportunity to examine the diagram
and, in doing so, evaluate which risks and impacts warrant further
examination when conferring with agency representatives. Early use
of the network may also lead team members to generate program
alternatives for consideration in the EA based on the perceived
magnitude of identified risks or impacts. The network can be
referred to again during the course of EA preparation, as
alternatives are developed, to examine how risks or impacts may be
reduced--the linkages weakened or eliminated--with program
objective alternatives.
Although some methodologies used in impact identification are
suitable for direct presentation to decision-makers, network
diagrams typically are not. However, significant risks and impacts
identified by use of the diagram can be presented to decision makers
in a summarized and comparative form. Use of the risk network is
demonstrated in this dissertation in the case study of chapters 4, 5,
and 6.











CHAPTER 4
CASE STUDY: REVIEW OF A PROPOSED PROGRAM FOR MANAGEMENT OF
THE TROPICAL BONT TICK IN THE CARIBBEAN
In this chapter a case study is begun which examines the
development of the joint AID and USDA program, "Management of the
Tropical Bont Tick (TBT) and Associated Diseases in the Caribbean".
This case serves to demonstrate the current process used by the US
government and other funding agencies in developing and assessing
an animal disease management program. The process is critiqued
through use of this specific program. The history of the TBT
program with regard to feasibility studies, environmental
assessments (EA), and risk/benefit/cost analyses is presented and
areas for improvement in assessment of the program are identified.
The next chapter continues the case study by demonstrating how the
proposed risk network can be used to help identify potential problem
areas in the TBT program and how some risks can be quantified and
reflected in the economic risk/benefit analysis of an environmental
assessment.


Development of the Tropical Bont Tick Program

In response to recommendations and resolutions made by Chief
Veterinary Officers and Ministers of Agriculture of CARICOM
countries, the Inter-American Commission on Animal Health, the
Third FAO Expert Consultation on Research on Tick-borne Diseases






and their Vectors, the National Cattlemen's Association, and the US
Animal Health Association in the early 1980's, a task force met to
outline a feasibility study proposal for management of the TBT in
the Caribbean. The task force produced an outline for the feasibility
proposal, identified members of a study group needed to produce the
feasibility proposal, and developed budgetary requirements to
accomplish the task. Financial support for the study group was
provided by AID, USDA, and FAO. The study group first met in 1986
and the "Management of the Tropical Bont Tick and Associated
Diseases in the Caribbean: A Feasibility Proposal" (FP) was finalized
in 1987. The proposal provided an economic evaluation of TBT
infestation, consequences of its spread, and benefits derived from
its control or eradication. The report also outlined several options
for management of the tick as follows: 1) eradication of TBT from
all islands, 2) eradication from islands with isolated populations of
the tick combined with control measures on islands where the tick
is widespread, and 3) a program similar to the previous one but with
the possible program conversion from control to eradication after
several years, based on accrued information and appraisal of
feasibility. The main concern, as indicated by the FP, was that the
TBT and its associated diseases represent an economic threat to the
surrounding continental regions that have significant livestock
industries. Of primary importance is the threat of spread of the tick
to the continental US and the subsequent impact it would have on the
livestock industry in southeastern and southwestern coastal states
where the climate may favor tick survival. The document suggests
that if heartwater and the TBT were introduced into Florida, the






cost to the cattle industry would be "tremendous" though no
empirical analysis of the risk of introduction or of spread given
ecological and husbandry conditions has been attempted on this
issue. In contrast to the perceived severity of the economic threat
to continental regions, the threat to Caribbean countries was
minimized; "the economic impact of the TBT to each individual
Caribbean country is minimal given the minor importance of the
livestock industry in the Caribbean" (Consortium for International
Crop Protection, 1987, page 1).
Many questions were subsequently raised, however, with
regard to the estimated program costs presented in the proposal, the
feasibility of an eradication objective, and the appropriateness of
program action. As a result, a demonstration eradication, or pilot
eradication project (PEP), was proposed in 1987 to be conducted on
Antigua, an island on which the tick is widespread. The project was
to develop the information needed by AID, other US governmental
agencies, governments in the region, technical organizations, and
other donors, to formulate policies, strategies, and operating
procedures required to effectively manage the TBT and associated
diseases in the Caribbean. The proposal was "designed to assure
that necessary entomological, veterinary, economic, social, and
environmental data are derived from the demonstration project"
(Agency for International Development, 1987, page 10).
Authorization was given for the pilot project with funding not to
exceed $3,499,000. As part of the PEP proposal process, an
environmental assessment was prepared in 1987 and a second was
prepared in 1989.






The results of the economic analysis conducted for the PEP
proposal varied considerably with the analysis results of the
original FP. A third analysis, a risk/benefit analysis, was conducted
in the first PEP EA with results differing from both previous
analyses. Yet another analysis was conducted in the second PEP EA
to compare the estimated project costs of using an alternative to
amitrazl, the acaricide chosen for the eradication.
To date, the only projects actually undertaken with regard to
the PEP are an economic impact study and a wildlife study conducted
in 1989-90. In 1991, the PEP was canceled. In 1993 the project
was on again but without US involvement and with the evaluated
alternative acaricide, flumethrin2, slated for use (Drummond, pers.
comm.). In early 1994 the PEP was again canceled (Butcher, pers.
comm.) but funding for the Caribbean-wide program is currently
being sought by CARICOM and FAO for CARICOM countries and by
France for its territories (Food and Agriculture Organization, 1994).

The Tropical Bont Tick and Associated Diseases in the Caribbean

It is thought the TBT, Amblyomma variegatum, was introduced
to the Caribbean in 1830 when infested cattle were imported to
Guadeloupe from Senegal. The tick serves as the biological vector of
the causative agent of heartwater, a potentially fatal disease of
cattle causing mortality in up to 50% of previously unexposed adult
animals. The presence of the tick is also associated with a high



1 TAKTIC; NOR-AM Chemical Company
2Bayticol; Bayer Australia Limited






incidence of dermatophilosis, a debilitating skin disease caused by
Dermatophilus congolensis.
The initial islands infested with the tick were Guadeloupe,
Marie Galante, and Antigua, and the tick is not known to have spread
for over 150 years. During the next 38 years the tick spread to 11
other islands. However, it still has not spread to 10 Lesser Antilles
islands intermingled with the 3 island's first infested. Once
identified on an island, the rate of spread of the tick among the
island's herds and flocks has varied considerably with each of the
islands infested. In general however, spread is slow. As a means of
comparison, on Puerto Rico the TBT spread approximately 10 miles
during a four-year period whereas Boophilus microplus, the cattle
fever tick, spread the full 100 miles across the island in only two to
three years. On every island where the TBT has appeared the
incidence (the number of new cases each year) of dermatophilosis
has risen dramatically from an average background level of one
percent, although no direct relationship between the tick and the
disease agent can be found. There is no clinical evidence that the
causative agent of heartwater, which is directly transmitted from
tick to host, has spread with the tick from the initial islands of
infestation.

Costs and Benefits of Tropical Bont Tick Eradication

The cost of a Caribbean-wide eradication effort was projected
in the FP to be between $26 and $35 million where, it was noted,
effective control could be attained for about one-half this cost. No
specific information on the costs and benefits of a control program






was offered and as "control does not remove the threat to the United
States and other surrounding countries, and implies the need to
continue funding for control into the indefinite future" (Consortium
for International Crop Protection, 1987, page 1), the conservative
alternatives to eradication--temporary or permanent control--were
discarded. All subsequent empirical analyses conducted to evaluate
benefit/cost indicators, traditionally of primary importance in
animal disease control decision-making, included only the estimated
costs and benefits of eradication measures. The "benefits" in all
cases were the avoidance of production losses and producer control
costs associated with the TBT, heartwater, and dermatophilosis for
Caribbean producers. All estimated costs were to be incurred by the
United States and supporting agencies.
In order to project the economic consequences of taking no
action against the tick, assumptions were made about the expected
rate of spread to additional islands and the rate of spread among an
island's flocks and herds once introduced (Alderink & McCauley,
1988). Islands were classified by perceived risk of infestation into
four risk classes: Class I, free of the tick and at low risk; Class II,
free of the tick and at higher risk; Class III, tick established but
with limited distribution; Class IV, tick widespread. In projecting
the rate of spread, it was determined "plausible" that in the next
five years, the TBT would be introduced to five of the islands in
Class II. It was presumed that at the end of 12 years, all islands in
this class would be infested. For both Class II and III islands,
spread of the tick within flocks and herds was estimated to be at a
rate such that one-fifth of the animals would become infested in






year one, another one-fifth in year two, etc., until all herds could be
expected to be infested at the end of year five. In the executive
summary of the FP, the management plan called for eradication of
the tick from Class III islands. For Class IV islands, the plan was
more conservative and called for voluntary TBT treatment that
could, at a later date, lead to a mandatory eradication effort. This
is the conservative approach that was dropped; eradication became
the stated objective.
In the FP economic benefit/cost (B/C) analysis, the cost of
eradication and/or prevention of introduction and spread of the tick
was weighed against potential livestock production losses avoided
for islands in Risk Classes II, III, and IV separately and together
over a 20 year period. An average milk price of US$.60/kg was used
to evaluate tick and disease impacts on dairy production across the
Caribbean. For the impacts on beef production, two prices were
evaluated: 1) a border value for beef of similar quality in the US,
estimated at US$.60/kg live weight (LW); and 2) a weighted average
island price, obtained through questionnaires, of US$2.52/kg LW.
Goat and mutton were also valued at their border and weighted
average prices of US$.80/kg LW and US$2.55/kg LW, respectively. As
part of the analysis, alternative B/C ratios were calculated using
both sets of prices and three discount rates. The calculated B/C
ratio for Caribbean-wide eradication based on a 15% discount rate
and the US$.60/kg price was 0.73, representing a total program loss
of US$6,666,000. The most attractive B/C ratio calculated, 3.92,
came from using the averaged island price for the Caribbean,
reportedly better reflecting actual producer benefits and incentives,






and a six percent discount rate. The estimated present value of the
production losses associated with TBT was presented for each
island in each risk class but eradication program costs were
presented only for each risk class as a whole. When all Class IV
islands, those with widespread TBT, were considered together, the
B/C ratio's of avoided production loss and eradication costs were
less than 1 (negative net present value) at all discount rates. The
values used to estimate benefits were static and did not take into
account changes due to technical and economic responses to a
program that may occur in the livestock production and marketing
system. Also, the benefits were estimated to accrue only to
producers, who presently have affected animals or would have after
ticks are introduced and spread, not to the local economy.

A Proposed Pilot Eradication Project on Antigua

Serious questions were raised in response to the FP concerning
the estimated program costs and benefits, and the rate of spread of
the tick on and between islands. Also, questions remained as to the
feasibility and desirability of an eradication objective given the
diverse political, cultural and environmental nature of the islands.
The TBT Pilot Eradication Project (PEP) was designed to address
some of these issues and it was specifically noted that a major
component of the work would be a study of the economic feasibility
of TBT eradication. The resulting B/C analysis would become "an
important element in subsequent decisions on management of the
Bont tick problem on other Caribbean islands" (Agency for
International Development, 1987, page 16).






As a requisite of the PEP, an environmental assessment (EA)
was prepared in 1987 by a team of experts. The stated purpose of
the report was to review and assess 1) the environmental impacts of
the proposed pesticide, amitraz, 2) the economic risks and benefits
of use of the pesticide, and 3) current and alternative control
programs. The team was further directed to outline the research
needed to improve the PEP and any Caribbean-wide effort, including
alternatives, and to outline the research needed to better understand
the role of A. variegatum and the transmission of dermatophilosis in
as much as it influences the evaluation of alternatives.
Various assumptions were made for the EA based on previously
accrued information. For the risk/benefit analysis it was assumed
that 1) eradication of TBT would result in the complete elimination
of dermatophilosis and heartwater, 2) on average, droughts occur
four out of every ten years, 3) the level of management, quantity of
animals and standing livestock numbers remain unchanged over time,
4) almost one-half million pounds of meat and meat products are
lost annually due to dermatophilosis (US$180,000), and 5) the size
of the standing national herd, based on the latest Census of
Agriculture (1984) for Antigua and Barbuda, is approximately 15,200
cattle, 10,300 sheep, 14,200 goats, 5,300 pigs, and 2,000 donkeys
and other equines.










Results of the Environmental Assessment

The following is a partial summarization of the findings and
recommendations of the EA team report (Consortium for
International Crop Protection, 1987) that pertain to this study:

Impact on Non-Target Organisms
Findings: "Because a successful or partially successful
eradication program would result in [endemic instability], any future
reinfestation of heartwater-infected TBT will likely result in
substantial mortality and/or morbidity due to both heartwater and
dermatophilosis. Similarly, if cattle ticks are eradicated or
suppressed as a result of the PEP, the calves born during and after
the PEP will not be exposed to Babesia and possibly Anaplasma
infections. Without these challenging infections, animals will not
develop immunity and their susceptibility could result in substantial
morbidity and mortality if infected vectors return" (page v).
Recommendations: "Surveillance for TBT, the cattle tick, and
tick borne diseases on Antigua, prior, during, and after the PEP is
recommended to alert personnel for necessary action against tick
reinfestation and required disease management as a result of
induced endemic instability. Appropriate personnel and laboratory
facilities will need to be established" (page v).









Potential Effects on Non-Target Organisms and Mitigation
Findings: "There are serious limits to the degree to which
Antigua's environment is representative of the larger Caribbean
environment; one which spans 1,200 miles north to south and 1,500
miles east to west" (page vii).
Recommendations: "As per Congressional direction, mitigation
should include identification and feasibility determination for
various people, land and wildlife management techniques that would
maintain and/or improve biological diversity and long-term
productivity" (page viii).

Availability of Alternative Nonchemical Treatments
Findings: Five options were identified by team members. The
options included variations on and combinations of integrated pest
management (IPM) techniques and TBT eradication. The IPM options
were evaluated on the IPM definition that two or more techniques
are integrated through a management strategy to reduce the losses
to one or more pest species to an acceptable economic threshold.
IPM options short of eradication would include strategic acaricide
treatments. "It is important to emphasize that the TBT and related
diseases would remain resident on the island, requiring continual
production inputs for tick control and disease management" (page x).
Recommendations: None were made.









Risk/Benefit Analysis
Findings: "Livestock contributes 2-3% of the GDP (versus 6%
for total agriculture). Livestock, however, are very important to the
Antiguan as a store of wealth and tradition. The impacts of the PEP
are as follows:
Direct Impacts: A total annual savings to the livestock owners
of approximately US$221,000 (or almost $6 million over 50 years at
a 6% discount rate).
Indirect Impacts: Avoidance of economic loss to the Antiguan
economy of US$145,000 per year (or $2.3 million over 50 years at a
6% discount rate).
Total economic loss to Antigua, because of dermatophilosis, is
estimated at US$360,000 per year (or a net present value of US$5-6
million over 50 years at a 6% discount rate). Benefits to Antigua
from the PEP are less than 0.1% of GDP. However, the economic
impact on the livestock industry would be a more significant 2%,
with a benefit/cost ratio of 2.8 (based on the estimated project cost
of US$2 million)" (page xiii).
Recommendations: "Eradication of TBT may result in the
eradication of the cattle tick and the diseases associated with that
tick. The existing immunity to these diseases may be lost such that
any reinfestations of the cattle tick will subject susceptible
animals to substantial losses. Provisions should be made to
document and quantitate the costs of these events, and managing







disease outbreaks should they occur, in order to better approximate
the economic impacts of the Project and Program" (page xiv).


Areas for Improvement in the Environmental Assessment

There are several areas in the EA report which could be
improved to better address the stated purpose of the report and to
better serve the spirit of the Congressional directives set forth for
EIS and EA preparation:

Impact on Non-Target Organisms
Disruption of the established endemic stability of both target
and non-target organisms can have potentially severe consequences.
This concern was raised prior to the PEP EA. Nevertheless, no
further development of this issue was done in the EA nor was there
an attempt to evaluate the magnitude of this risk. No new
information was presented to decision makers on how to weigh or
consider this risk in determining the best course-of-action for TBT
management on Antigua or for the greater Caribbean basin. Although
the team indicated that some measures should be taken to recognize
tick reinfestation and manage subsequent disease outbreaks, no
implementation strategies were offered, nor were costs, feasibility
and appropriateness of after-the-fact mitigation discussed.

Availability of Alternative Nonchemical Treatments
Nonchemical alternatives to eradication were the only
alternatives considered. This was in keeping with the outlined
scope of the EA, but fell short of the broader stated purpose of the






PEP EA and of the more global directives for EIS and EA preparation
set down by Congressional directive. Development and discussion of
other forms of alternative action were possible at that time and
could have been beneficial to decision makers. The development of
alternatives could include a discussion of the degree to which
identified risks may or may not be mitigated through program
alternatives. Discussion of the impacts of these risks and
alternatives could represent an improvement in the information
given to decision makers for Caribbean-wide action. No particular
importance was given to the possible impacts or repercussions of a
failed eradication attempt, nor to the equity issues of this outcome
even though the authors of the PEP proposal acknowledged that
failure was a distinct possibility.



Risk/Benefit Analysis
The three economic analyses performed for estimating the
benefits and costs of TBT management on Antigua yielded
inconsistent results and incorporated estimated values from
previous studies apparently without assessing their validity. The
annual value of production losses, including producer control costs,
were estimated for Antigua in the FP but the costs associated with
an eradication effort on the island were not. The costs associated
with TBT eradication on Antigua were first presented in the PEP
paper but no new estimation of benefits was conducted. The
expected benefits to the Antiguan producer and the local economy
were much more fully developed in the subsequent PEP EA as






compared to all previous analyses. However, the estimated
benefit/cost ratios were derived using a program cost estimation
from an unspecified previous study. The AID estimated program
costs cited in the PEP, presented in Table 1, were developed using
animal census figures inconsistent with those used in the EA benefit
estimations. The AID projected costs of eradication were based on a
total standing herd of 25,600 cattle, sheep, and goats. The benefits
were estimated using more recent census figures of 39,700 cattle,
sheep, and goats. A comparison of discounted benefits and costs
based on the estimates presented in the FP, PEP proposal and PEP EA
are presented in Table 2. In order to develop eradication
alternatives and evaluate economic indices in the next chapter, it is
first necessary to adjust the estimated program costs to reflect the
more current animal census figures used in the EA benefit
estimation. The amended program costs are presented in Table 3 and
a reevaluation of discounted benefits and costs, and B/C ratio's are
presented in Table 4. A breakdown of the adjustments made in
program costs is given in Appendix A.
No empirical estimation of the economic impact of risks to the
fulfillment of the eradication objective or resulting from
eradication program actions was attempted. No estimation of the
risks, costs, or benefits of alternatives was done. Finally, no
evaluation of the distributional aspects of risks, costs, and benefits
of an unsuccessful eradication attempt or for feasible alternatives
was conducted.








Table 1. Summary of estimated program costs for the Tropical Bont
Tick pilot eradication project on Antigua based on an animal census
figure of 25000 (US$ in 1000's).




Program Year

Factor 1 2 3 4

I. Demonstration Eradication Component
A. Acaricide and Treatment
1. Acaricide (a) 84 332 84
2. Surveillance Crews 104 72 104
3. Treatment Crews (a) 152 288 152
4. Laboratory Technicians 11 11 11
5. Maintenance Staff 18 18 18
B. Program Management/Administration
1. Project Director support 60 60 60
2. Travel & per diem for USDA teams 50 100 50
3. Project Director Counterpart 10 10 10
4. Local Administrative Officer 8 8 9
5. Data Clerk 5 5 5
6. Administrative Clerk 5 5 5
7. Secretary 5 5 5
8. Compliance Officer 7 7 7
C. Commodities
1. Vehicles (a) 250
2. Spray Rigs (a) 98
3. Livestock Handling Equipment 44
4. Other Equipment 16
5. Protective Clothing 15 12 3
D. Equipment Maintenance
1. Vehicle 40 44 32
2. Spray Rig 6 15 6
3. Office 8 8 8
E. Information and Public Relations
1. Printing, Radio, TV 30 30 30
2. Training Materials 20 20 15
II. Information and Evaluation Component
A. Production Loss Study 50 50 10
B. Field Expenses Study 15 20 30
C. Wildlife Study 30 40 30
IIIll. Strategy and Policy Formulation
A. Studies 20
B. Conferences 30

Total Proaram:
Sub-total 1141 1160 684 50
Contingencies 155 155 154
Total 1296 1315 838 50


(a) Cost factors most sensitive to change in the estimated animal census figure.






Table 2. Discounted benefits, discounted costs, net present values,
and benefit/cost ratio's based on estimates from the 1987
Feasibility Proposal and Tropical Bont Tick Project Environmental
Assessment for Tropical Bont Tick eradication on Antigua (US$ in
1000's).



Discount Rate Discounted Discounted Net present Benefit/cost
(%) benefits costs (a) values ratios

6 1211 (b) 3136 -1925 0.39
9 964 (b) 2978 -2014 0.32
12 789 (b) 2834 -2045 0.28

6 4114 (c) 3136 982 1.31
9 3274 (c) 2978 299 1.10
12 2679 (c) 2833 -152 0.95


(a) Present values of discounted "Tropical Bont Tick" eradication program
costs based on an animal census figure of 25000.
(b) Present values of discounted benefits (20 years) estimated in the
1987 "Feasibility Proposal" for management of the Tropical Bont Tick.
(c) Present values of discounted benefits (20 years) estimated in the 1987
"Tropical Bont Tick Project" Environmental Assessment based on an animal
census figure of 39700.




60


Table 3. Summary of estimated program costs for the Tropical Bont
Tick pilot eradication project on Antigua based on an amended
animal census figure of 39700 (US$ in 1000's).




Program Year

Factor: 1 2 3 4

I. Demonstration Eradication Component
A. Acaricide and Treatment
1. Acaricide (a) 129 513 129
2. Surveillance Crews 160 104 160
3. Treatment Crews (a) 224 432 224
4. Laboratory Technicians 11 11 11
5. Maintenance Staff 18 18 18
B. Program Management/Administration
1. Project Director support 60 60 60
2. Travel & per diem for USDA teams 50 100 50
3. Project Director Counterpart 10 10 10
4. Local Administrative Officer 8 8 9
5. Data Clerk 5 5 5
6. Administrative Clerk 5 5 5
7. Secretary 5 5 5
8. Compliance Officer 7 7 7
C. Commodities
1. Vehicles (a) 380
2. Spray Rigs (a) 154
3. Livestock Handling Equipment 62
4. Other Equipment 16
5. Protective Clothing 23 18 5
D. Equipment Maintenance
1. Vehicle 61 67 48
2. Spray Rig 9 24 9
3. Office 8 8 8
E. Information and Public Relations
1. Printing, Radio, TV 30 30 30
2. Training Materials 20 20 15
II. Information and Evaluation Component
A. Production Loss Study 50 50 10
B. Field Expenses Study 15 20 30
C. Wildlife Study 30 40 30
III. Strategy and Policy Formulation
A. Studies 20
B. Conferences 30

Total Program:
Sub-total 1550 1555 878 50
Contingencies 206 206 205
Total 1756 1761 1083 50


(a) Cost factors most sensitive to change in animal census figure.






Table 4. Discounted benefits, discounted costs, net present values,
and benefit/cost ratio's based on estimates from the 1987
Feasibility Proposal and Tropical Bont Tick Project Environmental
Assessment for Tropical Bont Tick eradication on Antigua, and
amended animal census figures (US$ in 1000's).



Discount Rate Discounted Discounted Net present Benefit/cost
(%) benefits costs values ratios

6 1211(a) 4173 (b) -2962 0.29
9 964 (a) 3965 (b) -3001 0.24
12 789 (a) 3774 (b) -2985 0.21

6 4118 (c) 4173 -55 0.99
9 3277 (c) 3965 -688 0.83
12 2682 (c) 3774 -1092 0.71


(a) Present values of discounted benefits (20 years) estimated in the 1987
"Feasibility Proposal" for management of the Tropical Bont Tick.
(b) Present values of discounted "Tropical Bont Tick" eradication program
costs based on an animal census figure of 39700.
(c) Present values of discounted benefits (20 years) estimated in the 1987
"Tropical Bont Tick Project" Environmental Assessment based on an animal
census figure of 39700.









Summary

In 1987, a feasibility proposal for management of the Tropical
Bont Tick in the Caribbean was prepared only to have the findings
heavily debated with regard to the estimated program costs and
benefits, and the appropriateness of an eradication objective. The
decision was made to conduct a pilot eradication project on Antigua,
a highly infested island, to address the debated issues and provide
information to assist in the decision process for future Caribbean-
wide efforts. As part of the proposal process for the pilot project,
an environmental assessment (EA) was prepared to examine the
potential impacts of acaricide use, evaluate nonchemical program
alternatives, and provide more comprehensive benefit estimates of
the eradication objective.
The EA report noted that the acaricide chosen might affect a
specific non-target tick specie and consequently disrupt the
endemic stability of a potentially economically important cattle
disease transmitted by the tick. However, no attempt was made to
fully evaluate this risk. For alternative program evaluation, the EA
considered only nonchemical means by which to eradicate or control
the TBT. No control alternatives to eradication were evaluated even
though this might better serve the informational purposes of the
pilot project in considering Caribbean-wide program objectives. The
risk/benefit analysis conducted in the report evaluated the direct
and indirect economic impacts of TBT eradication on the Antiguan
producer and the local economy but did not adjust previously




63

estimated program costs based on inconsistent animal census
figures. The analysis failed to consider the economic impacts of a
failed eradication attempt, both with regard to the TBT and to the
non-target tick and its associated disease.











CHAPTER 5
CASE STUDY: APPLICATION OF THE RISK FRAMEWORK, ALTERNATIVE
GENERATION, AND RISK AND IMPACT COMPARISON
This chapter demonstrates how the Environmental Assessment
(EA) or Environmental Impact Statement (EIS) can be broadened in
scope and preparation to include impact analysis with regard to
biological system disruption. The Antigua pilot eradication project
(PEP) reviewed in the previous chapter is used as a case study for
integrating the proposed risk assessment framework and expanded
risk/benefit analysis into the program development process.
To demonstrate the informational benefits to decision makers
of broadening the EA process, the Antigua PEP EA is expanded here to
include 1) identification of risks through application of the risk
framework developed in Chapter 3, 2) full development of an
eradication alternative, 3) evaluation and comparison of the
magnitude of identified risks with respect to alternatives, and 4) in
Chapter 6, reevaluation of the risk/benefit analysis to incorporate
identified, quantifiable risks and examine distributional aspects of
risks, costs, and benefits.
A result of the PEP proposal process was the funding and
conducting of an economic impact (El) study as a means to obtain
baseline data with which to compare the pre- and post-PEP
situations. The purpose of the El study, conducted over 8 months
from 1989-1990, was to determine with more empirical precision
the relationship between tick burden, dermatophilosis and economic






parameters such as live weight gain, weight and condition at
slaughter. Portions of the data collected from the study were used
to support, though not direct, the development of risk and
alternative action discussion put forth in this work.
This chapter will first demonstrate use of the proposed risk
framework, then develop program alternatives, and finally, compare
identified risks and impacts across the program alternatives.


Application of Risk Framework

Based on information provided from previous program
documentation on the PEP, a preliminary application of the risk
framework by the EA team should identify several risk
characteristics and factors which are insignificant or do not
directly pertain to this particular program. These include funding
sources, political stability, gender issues and temperature
extremes. However, other risk characteristics are pertinent to this
program and based on information already accrued and available,
should signal concern. These characteristics include producer
motivation and attitudes, weather conditions including availability
of natural resources, accessibility of all hosts of the targeted life-
cycle phase of the tick, adequate understanding of tick biology, and
efficacy of available acaricide. Also, there are direct impacts from
program actions which could be quickly eliminated in this case.
These include a possible significant change in mortality rates, and
the presence of alternative species to serve as vectors for target






disease/agent transmission. Other potential impacts in this case do
bear some examination.
Based on the information provided to the EA team members and
their areas of expertise, there are some risks and impacts that
might be identified as warranting a more complete assessment and
analysis. These are as follows:
1) Cultural Factors owner participation
2) Biological Factors presence of wild or other host species
3) Environmental Factors drought conditions and inclement
weather
4) Non-Target Organisms impact on other tick species of
health and economic consequence
5) Disruption of Endemic Stability both target and non-target
organisms currently participate in endemic disease cycles
The reasons for identifying these particular risks and impacts
as important in this case are briefly described below. They are
developed more fully in the next section of this chapter.
The first three categories represent risks to the successful
completion of an eradication objective. With eradication as the
stated objective, 100% owner participation is required to ensure
that all animals will be treated every two weeks. Also, all animals
which can serve as a final host for this three-host tick must be
included in the campaign. The biology of the TBT is such that, if an
animal serving as a final host misses a single spray treatment,
female ticks may mate, feed to repletion, drop off the animal and lay
eggs. The resulting offspring may not need to return to livestock to
feed for 18-46 months (Barre & Garris, unpublished).






Weather conditions present considerable risk for several
reasons. Amitraz, the acaricide selected for use on nonequine
livestock, is a water soluble, broad-spectrum product which can be
applied either in dip or spray form. The water-diluted spray route of
application was chosen for the PEP. This requires the use and
guaranteed supply of thousands of gallons of water per day, five
days every week for a minimum of two years. However, drought and
heavy seasonal rainfall characterize Antigua's climate. The water
soluble nature of amitraz renders it ineffective if applied during
heavy rain.
The remaining two identified risk areas concern the impact on
the disease status of the livestock population as a result of any
program action. Over time, the indigenous breed of cattle on
Antigua, Bos taurus x Bos indicus, has adapted to the continuing
presence of both Amblyomma and Boophilus ticks and the diseases
they are associated with, dermatophilosis, heartwater, and
babesiosis. The cycle of heritable immunity, natural challenge and
tick density that exerts a natural control over disease incidence,
mortality losses, and tick burdens is delicately balanced and easily
disrupted. Both of these tick species have been present on Antigua
for a sufficient period that this endemic balance has been
established and is maintained. Various factors in the livestock
production and marketing system in addition to these natural factors
reflect this tentative equilibrium. Animal numbers and breed-types,
import and export levels, self-sufficiency, and prices are a result of
the continued presence of these endemic and hence, low-impact
disease vectors and agents. The risk here is that success or failure






of an eradication attempt on the island could disrupt the endemic
stability of the tick-disease cycle.
The ramifications of endemic disruption are that livestock
populations, rendered naive to ticks and tick-borne diseases within
nine months to two years of nonexposure, are fully susceptible to
full-blown acute cases of tick infestation and tick-borne diseases if
reintroduction or resurgence occurs (Ristic & Krier, 1981). Tick
burdens are controlled by host resistance factors, generated through
breeding and continued tick presence. Without host resistance and
other biological factors restricting tick population growth, the
population growth rate of reintroduced ticks may surpass that rate
which existed under the limiting, endemic conditions (Ristic & Krier,
1981). As a result of endemic disease disruption, tick burdens per
animal, disease transmission, acute cases of disease, and hence
production and mortality losses could surpass preproject endemic
levels. In brief, if disease vectors and agents are effectively
eliminated from Antigua, the island's livestock become as
economically and physically at risk as that of the continental US's.

Evaluation of the Identified Risk and Impact Concerns

The five categories of risk outlined in the last section have
potentially significant impact on project feasibility, costs and
benefits, and desirability. Thus they warrant further examination.

Cultural Factors

There exists the potential for cultural and social attitudes to
endanger program success in virtually any program in a less






developed country requiring compliance by local farmers. A standard
task of the social soundness component of program proposals is to
identify cultural and social attitudes which might impact program
feasibility and to present suggestions on how these impacts can be
mitigated. The social soundness study for the PEP proposal was
conducted simultaneously with the environmental assessment in
June and July of 1987. The study report summarized the Antiguan
livestock sector as follows.
Cattle, sheep and goats are the main agricultural activity on
the island. It is an activity that people from various socio-economic
levels can participate in. People with or without land have found
that ownership of livestock, grazing mainly on government land at no
cost, allows them to accumulate savings that can be converted into
cash in times of need. The livestock management philosophy is
generally the same for all types of livestock owners on the island;
accumulate and maintain livestock at minimal cost rather than to
pursue profit maximization. Consequently, the Antiguan livestock
system that has evolved is based on a low-input/low-output system
in which the growth of total numbers in the herd has priority over
high production returns for marketing the animals. The typical
Antiguan livestock owner is a male, in full or part-time off-farm
employment, who looks after his cattle before and after work hours
and on Sundays. In fact, 74% of the landless livestock owners and
49% of the landed owners have a second occupation.
Accepted census figures for livestock numbers and ownership,
based on several studies, as cited in the PEP environmental
assessment (Consortium for International Crop Protection, 1987)






indicate that there are, on average, 15,200 cattle, 10,300 sheep,
14,200 goats, 5,300 pigs and 2,000 horses and donkeys. Landless
farmers, with few animals per capital, are estimated to own 12,000,
or 80% of the cattle. The remaining 20% are owned by farmers
owning or leasing land, and by the government. Only 10% of all
livestock are kept on private estates with fenced, improved
pastures. Medium sized livestock owners, holding 25-30% of the
livestock have less than 20 acres each and animals may be tethered
or untethered. The remaining 60-65% of livestock (80% of cattle),
owned by the landless, are tethered and grazed on government lands.
Landless farmers are estimated to own 57% of the sheep and goats
but these animals are typically unrestrained and often become feral.
Ownership of sheep and goats is difficult to ascertain. There is a
great deal of variation in the quality of care of livestock in those
animals not held on the large private estates. Over-grazing and
drought conditions lead to above-average death loss in livestock
from poor nutrition, starvation and dehydration. Some animals are
left tethered for days at a time, unable to get even water.
Antiguans are independent people and are accustomed to little
government control or coercion in terms of their livestock.
Livestock owners are accustomed to treating their animals at their
convenience and close to home and may be reluctant to move their
animals biweekly to a designated treatment site. Movement of
animals is a cost, in time and investment, to the owners. The
concept of mandatory participation and enforcement of such
demands is not within their present understanding of pest
management. Short-term incentives for gaining and maintaining






owners' compliance include reducing inputs into livestock and
avoiding exposure to pesticides, though not all farmers treat their
animals. Long-term incentives for compliance are not clearly
understood. Tick eradication probably will not increase total
numbers of livestock, but rather the quality; which currently is not a
concern. It remains uncertain if treating animals on the premise of
having healthier, larger, more productive and efficient livestock is a
culturally appropriate incentive.
The impact of less than 100% owner compliance on the
eradication effort is two-fold. Based on census figures, landless
farmers and those with holdings less than one acre, own an average
of five head of cattle, seven sheep and eight goats each. The social
soundness team of the PEP proposal indicated that this group of
farmers would be the least amenable to any governmental
intervention in farming practices and therefore the most likely to
not comply with program directives. Although it is not known how
many of these producers own more than one specie of livestock, it
can be seen that a single producer in this category, missing a single
scheduled treatment, may be responsible for 5-20 animals on which
ticks may complete their life cycle.
A cohort study, conducted as a part the economic impact study,
contracted landless farmers to present their cattle for treatment
once every two weeks for six months. Twenty percent of the
treatments were missed. The ramifications of this, in terms of the
potential magnitude of ticks surviving to complete their life cycle,
is evident. Compounding this is the fact that ticks hatched from
eggs laid as a result of missed treatments late in the program may






not be seen on livestock for upwards of two years after program
completion.

Biological Factors

Antigua has no wild deer or other identified wild species
known to serve as final hosts for the TBT. However, domestic dogs
have been found harboring adult TBT, including attached females
(Barre et al., 1988; Barre & Garris, unpublished). The TBT
eradication program on Puerto Rico, used as a model for the
organizational design and cost estimations of the PEP, included the
spraying of dogs (Garris et al., 1989). Although dogs play a minor
role as final hosts under normal circumstances, it is not known how
this role might change in response to the heavy pressure of repeated
spraying of the preferred host species. There are significant
numbers of feral dogs on Antigua and it would be impossible under
the current animal control situation to eliminate them, much less to
round them up for repeated treatment. The widespread presence of
these potential final hosts poses a considerable risk to successful
interruption of the tick life cycle.

Environmental Factors

The main risks addressed in the PEP EA concerned the potential
impacts) of the chosen pesticide on the local environment, as is
standard practice in EA and EIS preparation. However, in this case,
environmental factors also pose risk to successful fulfillment of an
eradication objective. The island has a short, mild rainy season in
early spring and a more intense, longer rainy season from July






through about November (Dicks et al., 1992). It has been known to
rain for several days literally nonstop, rendering dirt roads
impassable. In addition to limiting program personnel transit,
amitraz is ineffective if administered in heavy rain. The potential
for less than efficacious acaricide treatment therefore, is highest
during these periods of heavy rainfall. Compounding this, the season
of heavy rains corresponds with the period of highest TBT burdens on
livestock. The PEP program was designed such that approximately
10% of the standing herd must be treated on a single day to remain
on schedule. Using the more recent census figures, this amounts to
approximately 4,000 head/day, five days a week, 52 weeks a year. A
single missed day, due to inclement weather conditions, could
account for the survival of thousands of ticks. There is no
contingency plan in the program design for the rescheduling of
missed spray sessions. The current level of communication
technology is such that it would be almost impossible to contact all
the producers to arrange an alternative treatment site or time.
Amitraz is a water-soluble compound which must be diluted
for use. The application rate of diluted compound is one gallon/head
for sheep and goats and two gallons/head for cattle. For application,
100 and 200 gallon truck-mounted sprayers were slated for use. The
application schedule requires the daily availability of approximately
5,500 gallons of fresh water/spray day to correctly dilute the
acaricide. Antigua is known to suffer droughts four out of every 10
years. When this occurs, the availability of fresh water is
drastically reduced. There is only one large reservoir to provide the
entire island with fresh water, and supplemental water must be






collected in household cisterns. It is not possible to guarantee that
the large volume of water necessary for spraying will be available if
a drought occurs.

Non-Target Organisms and Disruption of Endemic Stability

The risk with the most severe consequences, at least with
regard to the Antiguan farmers, may be the impact on non-target
organisms and disruption of endemic disease stability. In addition
to the Tropical Bont Tick, Antigua also supports populations of
Boophilus microplus, a one-host cattle tick found extensively in
Africa, Australia, Asia, South and Central America, and the
Caribbean. Boophilus ticks transmit the parasites, Babesia bigemina
and B. bovis, responsible for "cattle tick fever" or babesiosiss". This
is an economically important disease of cattle in tropical and
subtropical areas of the world and is considered a major obstacle to
livestock herd and production improvement in these areas (Kuttler,
1988). It is estimated that one half billion cattle throughout the
world are at risk of infection with the disease (Haile et aL., 1992).
Endemic stability of babesiosis is a delicately balanced system
which depends on passive immunity, natural challenge and an
"optimal" Boophilus population (Young, 1988; Smith, 1983). Endemic
zones typically support stable populations of Boophilus with tick
numbers sufficient to ensure that all calves receive a natural
challenge inoculation with the parasite before about nine months of
age. Colostral antibodies and age resistance protect exposed calves
from developing severe reactions. If calves do not receive a natural
challenge inoculation by about nine months of age, passive immunity






wanes and these animals become susceptible to acute disease, and
possibly death, if infected later in life (Mahoney & Ross, 1972).
No published studies could be found that attempted to evaluate
the economic impact of a babesiosis epidemic resulting from
reintroduction to previously endemic areas nor are there
comprehensive data on the impact of an epidemic in a newly infected
region. However, the United States itself provides an example of the
potential devastation associated with an epidemic episode of
babesiosis. In 1868, infested and apparently infected cattle were
shipped from Texas to Illinois. The subsequent babesiosis outbreak
resulted in the death of 15,000 head of cattle, representing a
mortality rate approaching 90% (Kuttler, 1988). A general
estimation for the mortality losses associated with the introduction
of susceptible cattle into babesiosis-endemic areas is 50%
(McCosker, 1981). Eradication of the Boophilus tick has proven
difficult, even though it is a one-host tick. After the epidemic of
the 1800's, the US instituted a nationwide eradication campaign in
1906 including 965 counties in 15 southern and southwestern states
(Schnurrenberger et al., 1987). Eradication was completed in 1960
after a long and sometimes violent campaign that finally required
the depopulation of the native deer population from Florida (Graham
& Hourrigan, 1977; Drummond, 1993, pers. comm.). A quarantine
zone is now maintained between the US and Mexico to prevent the
return of the tick and babesiosis. Constant surveillance for spread
of the tick is conducted and periodic infestations still occur (Mount
et al., 1991; Kuttler, 1988).






Boophilus was eradicated from Puerto Rico in 1941 but the
campaign had to be reinstituted in 1947 after ticks were discovered
(Graham & Hourrigan, 1977). In 1952 the island was again
considered cleared but the tick has since returned (Mount et al.,
1991). In the 1978 joint USDA-Commonwealth of Puerto Rico
Tropical Bont Tick eradication proposal, there was stated concern
that the rapid spread of Boophilus ticks and the damage they cause
livestock makes them an even greater economic threat than the Bont
ticks. The report noted that the economic benefits of eradication of
the TBT would not be as great as indicated in the feasibility study if
there were not subsequent control or eradication of the cattle fever
ticks. Eventually, the cattle fever tick would spread into the herds
where the Bont tick had been eradicated and negate most of the
benefits of eradication.
Boophilus ticks and babesiosis appear to be endemic on
Antigua. Morrow et al. (1989), reported that approximately 60% of
the cattle entering the St. John's abattoir during the rainy season of
1987 had infestations of Boophilus microplus. The El study of 1989-
90 found that over the eight month course of the study 65-82% of
the cattle entering the abattoir had infestations of the tick. An
unpublished serological survey indicated that 97% of 274 cattle
tested on Antigua exhibited antibodies to Babesia bovis organisms.
Widespread presence of the tick, serological confirmation of Babesia
organisms, and the very low level of reported cases of acute
babesiosis lead to the conclusion that this tick-born disease is
endemically stable at the present time on Antigua. Amitraz, the






chosen acaricide, has limited efficacy against Boophilus ticks and
will, therefore, impact the now-stable populations of this tick.
The following section discusses the USDA/AID proposed
eradication project and develops alternatives to the eradication
objective. The risks and impacts determined by use of the risk
framework to be potentially significant are then compared across
the various program alternatives.


Program Alternatives

This section includes three alternatives. The first,
eradication, was the only alternative fully developed and evaluated
in the PEP EA. Integrated pest management was the only alternative
to eradication proposed by the EA team at that time. A third
alternative, control, is developed here for the first time. A control
alternative could and should have been considered at the time of PEP
EA preparation. Even though the scope of specific tasks required
only nonchemical alternatives be developed, a stated purpose of the
EA report was that the team outline means by which any Caribbean-
wide effort, including alternatives, could be improved. The
information to be gained from the development and evaluation of a
control alternative does have bearing on the greater Caribbean-wide
management of the TBT.

Eradication

To review, the operational design of the PEP calls for three-
member teams to spray 150 head of livestock/day, five days/week,






52 weeks/year for two years. This scheduling of treatments is
based on the targeted adult parasitic stage of the tick life cycle.
The average engorgement period for female ticks was found by Barre
& Garris (unpublished) to be 13.5 days. Based on this information,
the Puerto Rico tick eradication program instituted a 14-day
treatment interval and this was also applied to the proposed PEP. It
cannot be stressed too heavily that in order to effect eradication all
identified final host species must be included in the program. There
are no identified wild animal species present on Antigua known to
serve as final hosts which indicates a better chance for success
than might be the case on other islands. However, there are many
feral dogs on the island and they are not presently included in the
spray campaign. The PEP plan includes the spraying of all cattle,
sheep, goats, and equine. Farmers are required to present their
animals for treatment on predesignated days at predetermined sites.
The water-soluble acaricide is to be administered by spray route
from 100 and 200 gallon truck-mounted sprayers. Several small
sprayers are included for spot treatments. The animals must each
be soaked with from one to two gallons of diluted acaricide,
depending on the animal specie.
The operational design of the PEP was based on the eradication
effort conducted on Puerto Rico. In that program, it was possible
for a three-member spray team to treat 150 head/day, five
days/week. However, the ticks were confined to isolated, identified
farms and animals were easily collected and controlled for
treatment. On Antigua, upwards of 70% of livestock belong to
landless producers who would be required to collect and drive their






animals to a predesignated spray area once every two weeks for two
years. These animals are typically tethered by chain to a stake or
bush and are unaccustomed to crowding or excessive handling. The
small El cohort study incorporated a treatment regime identical to
the one designed for the proposed PEP. It was found during the
course of the study that the animals quite often became entangled
with each other and extremely stressed as a result of the crowding
and handling. Frequently producers did not present their animals and
the team had to track down the missing animals and drive them in
for treatment; often they were not successful and much time was
lost as a result.
Although the logistical problems associated with an
eradication objective are many, so too are the benefits of a
successful eradication campaign. Elimination of TBT from Antigua
and the Caribbean would serve two goals. The threat to continental
areas and currently noninfested islands would be eliminated. On
infested islands, production which is currently limited by the
presence of the tick and its associated diseases would be improved.
The entire program would have a prescribed time frame of
approximately five years. All costs to the US government and its
agencies would end upon completion of the eradication effort,
although the islands would have recurrent costs associated with
surveillance and quarantine measures. If the tick was successfully
eradicated from all islands the chance of reintroduction from Africa
is essentially negligible. Eradication is theoretically possible based
on the biology of the tick and the availability of efficacious






pesticides and therefore, is considered to be the preferred program
objective by some of the authors of the various proposals.


Integrated Pest Management

As previously noted in Chapter 4, integrated pest management
(IPM) can be defined as two or more techniques integrated into a
management strategy to reduce the losses to one or more pest
species to an acceptable economic threshold. This may be
considered the most conservative approach to tick control,
requiring, in most cases, an intensification of management
practices. The EA team examined various options for an IPM
approach to TBT control but concluded that given the nature of the
people, their cultural background and current priorities, IPM
techniques requiring major changes in forage production and range
management were unrealistic unless there were substantial outside
assistance. The most successful and relevant IPM methods for
Antigua would include the use of resistant cattle and pasture
management. Currently, a lack of land ownership and inability of
farmers to exclude others from improved pasture areas preclude any
incentive they might have for improving pastures as a means to
control ticks and increase production. IPM approaches appear
impractical for TBT management on Antigua at the current time.


Control

In practice, a strategic control objective treatment regime
would be designed such that some level of tick infestation remains






but the economic damage resulting from the presence of the tick is
reduced to an acceptable level. For a control program on Antigua to
be appropriate and desirable, it would have to be designed to reduce
overall TBT burdens, reduce or eliminate the incidence of acute
dermatophilosis and fit into the current management practices of
producers. Based on accrued information and current control
programs on other islands, it is possible to reduce the incidence of
clinical cases of dermatophilosis in TBT infested herds from 10% to
3% with control measures (Alderink & McCauley, unpublished).
Morrow et al. (1989), found 31-35% of cattle belonging to landless
farmers and 40% of the animals seen at the abattoir exhibiting
dermatophilosis skin lesions upon inspection. It was estimated in
the risk/benefit analysis of the EA, that approximately 50% of the
Antiguan cattle have some degree of dermatophilosis but only 10%
exhibit measurable production losses as a result. The projected
reduction in dermatophilosis incidence to 3% with control would
occur in this latter 10% group of infected animals.
The TBT on Antigua has been found to exhibit strong
seasonality with regard to host infestation levels. The adult tick
burden per host is reported to increase late in May, remain high
during the summer rainy season, and drop-off late in the fall at the
beginning of the dry season (Dicks et al., 1992; Morrow, 1989; Ford,
1919). Effective tick control could be gained by treating all cattle
every two weeks beginning May 1 and continuing through August 31
(Kaiser et al., 1988; Sutherst et al., 1979). A control program of
this type does not require that producers refrain from their own tick
control practices. In fact, the 10% or so of cattle raised under






improved conditions with more intensive management could still be
treated regularly. The difference is this program would eliminate
the need for the buying, storing and using of pesticides for tick
control by the majority of those farmers currently employing some
form of tick control. Sheep and goats are affected by
dermatophilosis to a much lesser degree than are cattle and account
for only a small percentage of the livestock industry on Antigua. In
addition, adult TBT's parasitize small ruminants at half the rate
they do cattle, both in terms of number of animals infested and in
tick load/host. As a result, small ruminants need not be included in
a control program.
It is often stated that a severe drawback of control programs
is the indeterminate period of time for which control practices must
continue. This includes what is often perceived as an indefinite
stream of program costs and because of this, control programs are
considered second best to eradication. However, researchers in the
human health field are beginning to realize that the assumption of
static conditions with regard to health practices and treatment
possibilities is unrealistic. In fact, a 50 year time frame for
estimating discounted cost and benefit streams has been suggested
as being too long given the current rate of medical advances. A
control program can be costed and implemented with a fixed time
frame. During the course of the program, assessment of program
success may indicate extending the control objective beyond the
scheduled conclusion date based on new information, availability of
new treatments, or improved technology. There is currently
available another acaricide, flumethrin, that, based on routes of






application, residual effect, and other parameters, would be more
efficacious in controlling or eradicating the TBT (Consortium for
International Crop Protection, 1989). However, this compound is not
FDA approved and therefore, can't at the present time be used in the
PEP. Means other than topical acaricide treatment are currently
being researched, such as the use of pheromone attractants, which in
the future might allow for species specific control or eradication,
leaving other tick species unaffected (Meltzer, 1994, pers. comm.).


Comparison of Risks across Control and Eradication Alternatives

The risk assessment framework described in Chapter 3 is
again useful for examining the pathways which might have an effect
on or be affected by the outcome of a control program. Many of the
concerns highlighted when the framework was applied to the
eradication objective are still pertinent in the case of a control
objective. However, the significance of potential risks and impacts
changes with the less restrictive program objective of control.
Some factors are unchanged over the control versus eradication
objective. These are the current availability of efficacious
acaricides, and adequate understanding of the biology of the tick to
allow predictive evaluation of treatments on TBT populations and
dermatophilosis incidence. Social and environmental factors still
impact on the possibility of program success, however, a less than
100% treatment schedule compliance does not have the magnitude of
impact on the feasibility of the control objective as it does on the
feasibility of an eradication objective. Risks and impacts vary






dramatically across alternatives, however, when the impacts of a
failed eradication program, a successful eradication program and a
control program are evaluated.

Simulation of Eradication and Control Program Risks and Impacts

A model developed by Haile, Mount, and Cooksey (1992) for the
simulation of Boophilus population growth and Babesia transmission
was used in this dissertation to evaluate the potential non-target
impact of a TBT eradication attempt on Boophilus and babesiosis on
Antigua. This model was not available at the time the environmental
assessment for the PEP was prepared; however, much of the
qualitative information gathered from the model could have been
derived using expert opinion techniques such as a Delphi survey
approach. This tool has been used with success and is widely
accepted in the human health and veterinary health fields (Carmody
et al., 1984; Schoenbaum et al., 1976; Milholland et al., 1973). Of
particular interest in this case is a past application of the Delphi
survey tool to evaluate the likelihood and risks of an epidemic of
swine-like influenza for use in an economic analysis of vaccination
campaign strategies (Schoenbaum et aL, 1976). A Delphi survey for
eliciting expert opinion about the potential impacts of the
disruption of endemic babesiosis would have been used in this
research had it not been for the development of the computer
simulation model.
Haile's model of Babesia transmission simulates the
interactions between infective and susceptible individuals in tick
and cattle populations. It incorporates as a sub-model a life history






model for population dynamics of Boophilus ticks developed by Mount
et al. (1991). The structure of the Babesia model is designed to
represent the movement of the parasite between and within ticks
and cattle. Site-specific data necessary to run the model include;
monthly rainfall, average temperatures, relative humidity, pasture
density and type, cattle breed, and host density. Reliable
information on these parameters is available for Antigua, so the
model could be used to evaluate the impacts of tick and parasite
disruption. Unfortunately, a similar model has not yet been
developed for the simulation of the Tropical Bont Tick and its
associated diseases. Indeed, much less is known about the biology
of the TBT and about its impact on production parameters.
As a preliminary step to evaluate how well the Babesia
transmission model reflects the epidemiological situation on
Antigua, island-specific data were incorporated in an initial
evaluation to determine if the simulated tick-parasite population
equilibrium of the model is consistent with the observed situation
on the island. The results of the initial simulation indicated that in
the absence of any tick control practices, the stable endemic
situation is characterized by 1) an average daily standard female
tick load per cattle host of 190, 2) 100% of calves infected within
the critical nine-month period, and 3) an annual average Babesia-
related mortality loss of 13 cattle. In comparison, the observed
Boophilus situation may be evidenced by data collected for the El
study on livestock entering the St. John's abattoir from September,
1989 through April, 1990. Boophilus counts on over 800 cattle
yielded tick burdens/host ranging from zero to over 700. Tick






control measures taken by farmers vary in their frequency and
efficacy but some control is practiced and these measures can be
expected to affect the observed tick counts. As a result, it is
impossible to determine the equilibrium tick load that might be
observed under a situation of no tick control. The Boophilus counts
observed in the El study indicate that, given the current tick control
practices, "background" Boophilus tick loads average 34 adult female
ticks/host/day. According to the model, this count is consistent
with and supports the results of the unpublished serological survey
on Babesia infectivity rates. With regard to Babesia-related deaths,
at least one animal was presented to the slaughterhouse during the
El study exhibiting symptoms indicative of the final stages of acute
babesiosis. The 13 yearly Babesia-related deaths predicted by the
model may well be occurring as over 2,000 cattle are "lost" per year
on the island. There is no requirement for reporting the death of
animals occurring in the field and these carcasses cannot be
marketed in any way. There appears to be no evidence to suggest
that values predicted by the simulation do not reflect the general
Boophilus-Babesia situation on Antigua.
The Babesia transmission model developed by Haile et al, was
used here to examine the impact of the various program objectives
on expected Boophilus populations, Boophilus burdens and Babesia
deaths. Only the eradication and control objective alternatives were
modeled because the IPM approach was considered to be impractical
for TBT management on Antigua. Although program impacts on TBT
(Amblyomma) populations cannot be directly examined through use






of this model, some inferences may be drawn based on the results
seen for Boophilus populations.
Input options for simulation by the Babesia model include the
area control of ticks through the application of acaricides. Specific
variables for program options are; acaricide treatment interval in
weeks, number of treatments per year, starting week of treatments,
length of residual acaricide effect, and level of acaricide efficacy or
effectiveness (0-100%) each week for up to nine week's post-
treatment. An acaricide treatment interval of two weeks was used
for all simulation runs done in this work based on the primary PEP
program objective of TBT management, not Boophilus management.
The initial simulation run which established the background
epidemiological parameters of a stable, endemic Boophilus/Babesia
situation (see Appendix B) was used as the starting point for each
simulation conducted. If acaricide treatment is instituted, the
program of treatment must be inputted at the beginning of each
simulated year of the run. If a treatment schedule is not reentered
as an input to the simulation of the next year of the program the
results for that year reflect how the Boophilus population and
Babesia transmission respond to the cessation of management
measures. The simulation can be continued for any number of years,
each year incorporating the final epidemiological parameters of the
year before as a starting point. A summary of the vector and host
parameters can be obtained for each year of the run and includes;
percent of calves infected, cumulative Babesia deaths, and average
daily standard female Boophilus per host.






The number of treatments per head per year was set at 26
(treatments every two weeks) beginning with week one of the
calendar year to simulate an eradication objective. Simulation of a
control objective was carried out by setting the number of
treatments per year at eight, given every two weeks over a four
month period beginning week 17 (May 1). This starting date was
chosen because it corresponds to the seasonal increase in adult TBT
numbers seen on cattle. Acaricide efficacy rates were set after
consultation with Dr. Haile, developer of the Babesia transmission
model, and based on product information, and a previous study
(Sutherst et al., 1979). The expected level of effectiveness, or
efficacy, of amitraz on Boophilus was determined to be a maximum
of 95-97% during the first week post-treatment, and approximately
50% during the second week post-treatment. This relationship of
first and second week post-treatment efficacy rates in all
simulations is represented as a ratio, i.e. 97%/50%.
By varying the values inputted for acaricide efficacy in the
first and second week post-treatment it was possible to simulate
the impact of missed treatments. The cultural and environmental
risks to program success identified previously are characterized by
missed treatments, whether through owner noncompliance or
weather conditions. Incremental reductions of the acaricide
efficacy rate in the first week post-treatment were used to reflect
a respectively increasing percentage of missed treatments. Little
to no effect on simulation results was seen when the second week
post-treatment efficacy rates were varied. The impacts of reduced
efficacy rates and missed treatments are the same; some ticks






survive amitraz exposure or miss exposure altogether and complete
their life-cycle. Again, it must be noted that this model simulates
only Boophilus populations and Babesia transmission and how
program actions may impact the vector/agent cycle these species
represent. Only inferences can be drawn on how TBT populations
might respond under the conditions simulated in the model. The
empirical results of the simulations, i.e., daily tick burdens and
Babesia-related deaths, were valued and incorporated into the
risk/benefit analysis in the next chapter.
The simulation results of varying the efficacy rates were
grouped into six scenarios, three each for the eradication and
control objectives. Efficacy rates, scheduled number of treatments
per year, and the impacts on Boophilus and Babesia are presented in
Table 5 and described below.


Eradication Objective
Eradication Scenario 1. This scenario represents a best-case
outcome with respect to fulfillment of a TBT eradication directive.
All cattle (the model cannot simulate sheep and goats at the present
time) are scheduled to be sprayed once every two weeks for two
years and it is assumed that none (or a negligible number) of the
scheduled treatments are missed over the course of the program.
The efficacy of amitraz for Boophilus ticks in this scenario was set
at the maximum 97% for the first week post-treatment and 50% for
the second week post-treatment (97%/50%).
The results show that the tick would not be eradicated, but
rather would return as a result of surviving individuals. The model











(D D a) a) Q) 42)
.0 m- 4 4-m 4-.

c EE E E E
00
E O0 0 00 u






EB
c c c 0c
C LUJ '4
0 0E










*- e .c3
0 0 W 4 W 4 4- 4 L)
0 m 2m
(n VVD-0

0> 00 00
4JLJ* 4*4 414 4-1 41 "
u^ 0 0 0 0 0 0 .
w) cc c ccc
S4 -e 0














ou a
c So
) wE
E 0

















cc w
)u 4- 4






















U 4- 4- 4-4- 4-.4
noc W 00( w( 0 0 a)0V
.o E E E E E E
Q) 7 (n W (A A (u)n L MU
4- 4- 44-4 0 4- .
E 4E c c cc























j~ CC CC C C C D
0 a)0) a 0W
E E E E E E 3
.0 ) 4- 4- 4- 4- 4- 4-

-"' 0) 4 4-* 4-' 4-' 4~ .-' 4-* -.. 4-.
wa (D )

(' (Q 4- 4- fl )

EE
000

m 0
(I) CD ~U 4
> a) (A


> 4-'*^
4D C4
c C: -S.z


) -o 5

o I (D w c
0)0


C _C 4- 0
00 0 U) -)
fl Ln U :3
-o U0):=^ ^ -o
C 0 CL
CC 000 000 CQL0
> tn Lf L7fl Ln L L 0o
ou c
r- L0 C> u- 0



2n a) 0 )E

*i >,'4- M
m M *

LU m w
0)0

z 0)E E
.00 0) :

CUc c ccc *. (
I-w 0-0 (000

~0 00 Ocuc 0 0 0
U (> u O u u u ----
Li ni ((/)t> V)2A V ( V