Identifying Research Needs for Risk Assessment of U.S. Food Supply Security

1
Identifying Research Needs for Risk Assessment of
U.S. Food Supply Security
Bruce Hope Oregon Department of Environmental
Quality Portland, Oregon

• Society of Toxicology
• Risk Assessment Specialty Section (RASS)
• Monthly Teleconference
• April 13, 2005

2
Why risk?

• Much is possible, not all is probable.
• Hazard is about possibility
• Risk is about probability
• Risk is -
• Probability that exposure to a pathogen will
  result in a negative consequence of some
  magnitude.
• When probability is ignored
• High low priority scenarios may appear equal
• Scarce resources may go to low priority scenarios
• Confusion over priorities decision paralysis
• Caveat unconventional ? improbable

3
A complex human food supply chain
meat poultry dairy fish seafood eggs
R
ANIMAL PRODUCTION
grain elevators shippers feed lots
fresh products prepared products
R
R
T
T
R
R
R
FARM PRODUCTASSEMBLERS
FARM INPUTS
MULTIPLE FARM SOURCES
T
R
END-PRODUCT PROCESSORS
R
R
T
T
R
seed feed fertilizer pesticides
R
R
CROP PRODUCTION
R
R
T
INGREDIENT PROCESSORS
R
T
T
T
grains oil seeds fruits vegetables
milling slaughtering crushing starches flavorings
R
R
MULTIPLE FARM SOURCES
STORAGE
T
R
WHOLESALERS
supermarkets convenience stores co-op restaurants
hotels schools hospitals military
bases vending farmers markets community feeding
R
general home specialty products food service
T
R
RETAILERS
R
CONSUMERS
4
Is essentially

• Credibility and urgency of a threat is directly
  proportional to the overall probability of this
  specific chain of events
• Probability of whole process, not just one part,
  controls

EXPOSURE SCENARIO (plausible)
PATHOGEN (source)
TARGET (susceptible)
5
A context for research needs
Release Characterization
Dose-Response Assessment
EXPOSURE SCENARIO
PATHOGEN
TARGET
Exposure Assessment
Hazard Characterization
6
So we need to consider

• The probability of their being a source and a
  target, connected by a plausible chain of events.
• The probabilities associated with the actuality
  of source, target, and connecting events.
• The biological, environmental, and cultural
  factors which influence these probabilities.
• Research to identify, understand, and quantify,
  as well as help alter, these probabilities.

7
Risk assessment process
ACTION!
HAZARD CHARACTERIZATION
RISK MANAGEMENT
PROBLEM FORMULATION
RELEASE CHARACTERIZATION
EXPOSURE ASSESSMENT
EXPOSURE-RESPONSE ASSESSMENT
RISK CHARACTERIZATION
8
Hazard characterization

• Factors affecting probability of a pathogen being
  chosen as a bioweapon, then obtained, handled,
  deployed, and transported successfully.
• Virulence and pathogenicity
• Pathologic characteristics and diseases caused
• Survival and multiplication
• Resistance to environmental control measures
• Host specificity
• Infection mechanism and route portal of entry
• Ecology and natural history potential for 2
  spread
• More attention to non-weaponized agents.

9
Hazard characterization (RI 1984)

• Rajneeshee Incident The Dalles, Oregon 1984
• Motivation / objectives / needs
• Prevent vote for county court hostile to cult
• Sicken dead OK, but no mass casualties
• Difficult to detect
• Agent needs v. capabilities - Salmonella
• Not highly lethal, common, hard to detect as
  bioweapon
• Not robust, not persistent affects release
  mechanism
• High probability of obtaining, handling,
  deploying
• Had technically trained personnel laboratory
  facilities

10
Release characterization

• Factors affecting probability of release for
  differing exposure scenario.
• Logistical
• Deployment requirements (technical)
• Special equipment / handling requirements
• Access to desired exposure scenario(s)
• Biological (pathogen characteristics)
• Environmental requirements (hardiness)
• Compatibility with desired exposure scenario(s)

11
Release characterization (RI 1984)

• Bioagent characteristics dictated an exposure
  scenario with a release point close to target
• Using a technically and logistically undemanding
  release mechanism
• Cult members placed agent in creamer salad
  dressing at 3 restaurants in county
• Several bumbling, abortive attempts in other
  venues
• Easy access, low probability of detection /
  interdiction
• Plausible deniability its just food
  poisoning
• Low probability of release failure, yet...

12
Exposure assessment

• Factors affecting probability of exposure for
  differing exposure scenarios.
• Transport fate characteristics
• Routes of exposure and transmission potential
• Pathogen ecology
• Environmental reservoirs amplification, die-off,
  persistence
• Characteristics of exposed population
• Demographics (age, immune status, size, etc.)
• Feeding behavior (preferences cultural
  practices)
• Counter-measures
• Probability of detection in various media and
  biological matrices detection methods.

13
Exposure assessment (RI 1984)

• Transport fate characteristics
• Direct ingestion exposure, limited transport
  ability
• Not persist, no environmental reservoirs
• Exposed population assume food is clean, salad is
  not processed further
• Counter-measures
• Low probability of detection, high probability of
  interdiction if detected
• Actual exposure dose impossible to quantify

14
Fault Tree Analysis Estimates probability of dose
(D) of organisms reaching the target
Probability of bioagent being present
AND
Bioagent present in node
Bioagent not interdicted in node
OR
OR
Enters from intentional release in node
Enters from external environmental source
Present from internal source
Enters from prior node
Interdiction
AND
Detection
BT has resources to access node
BT has technical dissemination capabilities
Environmental conditions
AND
AND
Access resources
Access requirements
Dissemination capabilities
Dissemination requirements
15
Exposure response assessment

• Improved quantification of dose-response for
  known pathogens.
• Model fit to data
• Applicability of animal models
• Assessing multiple exposures
• Development of mechanistic (v. empirical) models
• Differential response by sensitive
  sub-populations
• Dose-response studies for new and emerging
  pathogens.
• Only a few of 150 enteric viruses studied for
  dose-response in human volunteers.

16
Exposure-response (RI 1984)
Risk Assessments of Salmonella in Eggs and
Broiler Chickens - 1 - Interpretative Summary.
World Health Organization, Food and Agriculture
Organization of the United Nations (2002)
17
Risk characterization (RI 1984)
Applied dose not quantified, but sufficient to
induce 751 reported cases of illness.
18
Suggestions

• Bioagent of choice need not be
• Weaponized
• Restricted to a particular pathway (e.g.,
  inhalation)
• Harmful to humans or difficult to handle
• Hard to obtain
• Food supply as a delivery system
• Good for localized impact, poor for mass
  casualties
• Possibly with a specialized bioagent?
• Food supply as the target
• Loss of system services
• Loss of confidence in system
• Significant 2 health and economic impacts