Title: Environmental Health Investigations: Conducting Traceback Investigations
1Environmental Health InvestigationsConducting
Traceback Investigations
2Goals
- Describe the basic steps of conducting a
traceback investigation - Identify when it is appropriate to conduct a
traceback investigation - Provide examples of recent outbreak
investigations that carried out traceback
investigations
3Quick reviewTraceback Investigations
- Process used to determine the production and
distribution chain of a vehicle implicated in an
outbreak. - Helps determine if (and where) you can conduct an
environmental health assessment - Used to clarify the point at which the implicated
vehicle may have become contaminated
4Traceback Investigations
- Identify
- Places where the implicated vehicle has been AND
- If the vehicles to which different cases have
been exposed had places in common - Identification of a common place strongly
suggests contamination at/before that point in
production or distribution of the vehicle
5Traceback Investigations
- Do not identify the source of the
problem/contamination - Tell investigators where to look
- Further investigation/inspection is necessary to
identify contributing factors or implement
control measures
6Traceback Investigation Example 1
- 1994 Nationwide outbreak of Salmonella
Enteritidis associated with a particular brand of
ice cream. (1) - Ice cream obtained by cases in 41 states from
multiple warehouses across the country - BUT, all ice cream eaten by cases prepared in one
plant in Minnesota
7Traceback Investigation Example 1 (cont.)
- Implicated plant obtained ice cream pre-mix
shipped via truck from two suppliers - Identification of a common ice cream plant
suggested contamination with S. Enteritidis
occurred at or before that point in the
production such as - Plant
- Tanker trucks
- Pre-mix suppliers
- Follow-up assessments found problems with
inadequate cleaning of tanker trucks that were
also used to carry non-pasteurized eggs
8Traceback Process
- Begins with information from cases about the
implicated vehicle - Extends backwards in time
- All points in the production and use of the
vehicle are considered - Retailers
- Point of service establishments (e.g.
restaurants) - Distributors
- Importers
- Producers
9Vehicle Distribution Patterns
- Distribution patterns may be complex
- Retailers may obtain vehicle from more than one
distributor, distributors may change over time - Distributors may have multiple sources, may
supply to other distributors - Producers may be domestic or foreign
10Traceback Investigation Example 2
- 1997 E. coli 0157H7 outbreak in Michigan (2)
- Case-control study revealed strong association
with eating alfafa sprouts - Traceback investigation initiated by collecting
data from cases - Places sprouts were purchased/eaten
- Locations of said establishments
- Dates of purchase/consumption
11Traceback Investigation Example 2 (cont.)
- Investigators used store and restaurant records
to identify likely sprouting facilities for
implicated sprouts - Source of sprouts traced for all 16 patients
- Source Facility A for 15 patients
- Source Facility B for 1 patient
- Seed sources identified
- Facility A 2 seed lots (Idaho Australia)
- Facility B multiple seed lots
12Traceback Investigation Example 2 (cont.)
- Concurrent outbreak of E.coli O157H7 in
Virginia, also tied to alfafa sprouts - Source of sprouts for 13 VA patients traced to a
single sprouting facility in VA using seed lot
from Idaho, same as in Facility A in Michigan - Facility B determined to have used Idaho seeds
also, but only for 2 days - Conclusion problem resided in seeds from Idaho
13Identification of Products
- Clear identification of implicated products is
essential to a traceback - Branded (commercial) products do not always
provide clear identification - If package is not available, consumers may not
remember exact name - One product may be repackaged and distributed
under different names
14Traceback Investigation Example 3
- 1998 49 cases of Salmonella Agona reported in
Illinois (3) - Nine other states also had increases in S. Agona
- PFGE subtyping suggested cases had common origin
- Matched case-control study linked S. Agona
infections with consumption of cereal but no
common brands were noted
15Traceback Investigation Example 3 (cont.)
- Traceback investigation revealed implicated
cereal was produced by a Minnesota company,
Malt-O-Meal - Cereal sold under brand name Toasty-Os
- BUT also sold as Toasted Oats under in-house
labels in 39 grocery stores - All cases had consumed a common cereal
16Product Identification Is Not Enough
- Most outbreaks are due to a temporary problem in
production chain - Knowing a brand name is not enough, you also
need - Item description
- Place, date of purchase
- Manufacturer, supplier and lot number
- Locations of farms, production facilities,
supplier information, delivery schedules
17Getting Started
- For a traceback investigation information is
collected from - Cases
- Retailers or point of service establishments
- Distributors/importers
- Producers/processors
18Traceback First Steps
- Begin with information available from interviews
of the cases about the time and place of purchase
of implicated item - Original packaging and labels with identifying
features (lot numbers, names of manufacturers)
are helpful
19Traceback Next Steps
- Continue investigation at place(s) where cases
obtained items such as retail stores or
restaurants - Expand investigation to distributors
- Includes brokers and importers and may have
multiple levels - Business documentation (invoices, inventory
records, air bills, bills of lading) and customs
forms can be helpful
20Traceback Final Steps
- A traceback investigation is completed when firms
that supplied, processed and produced the
implicated product are identified - Investigation may include a visit to the source
to verify information collected (production
dates, locations of farms/facilities) - Validity of a traceback investigation is highly
dependent on proper documentation - Missing documentation (such as receipts or
labels) makes the investigators job difficult
21To Trace or Not to Trace?
- Tracebacks can be time consuming
- While important, such investigations should be
undertaken selectively - Criteria for undertaking a traceback
- Certainty that the vehicle in question is truly
implicated in the outbreak AND that - The source of the contamination is within the
production chain
22To Trace or Not to Trace?
- Other key questions to consider
- Is there solid epidemiologic evidence linking the
outbreak and the implicated products? - Is there historical precedent for the product
being contaminated with the organism in question? - Is there microbiological evidence linking the
outbreak and the implicated products?
23To Trace or Not to Trace?
- Key questions, continued
- Does the vehicle have chemical and physical
characteristics conducive to the survival and
growth of the causative agent? - Has mishandling or environmental contamination of
the product by the consumer, retailer or point of
service establishment been ruled out? - Could the product be commercially distributed in
a way that is consistent with the outbreak?
24To Trace or Not to Trace?
- Another factor to consider is the likelihood of
success - Tracebacks are most successful with
- Commercially packaged products (identifying
label) - Unusual products (salami)
- Products with a long shelf-life (frozen
hamburger) - Traceback investigations are unproductive with
- Products with a short shelf-life (fresh fruits
and vegetables) - Products derived from multiple sources (blood
clotting factor from multiple donors)
25To Trace or Not to Trace?
- Weigh benefits of the investigation
- How serious is the disease?
- Death, hospitalization or permanent disability?
- Is exposure likely to be ongoing?
- Are vulnerable populations at risk?
- Infants, elderly, immunocompromised persons?
- A yes to any/all of these questions might
provide justification to undertake a traceback
investigation
26Success Factors
- Coordination among many players
- Local, state or federal agencies
- Good solid epidemiological data
- Poor epidemiologic studies might implicate the
wrong vehicle and misdirect the investigation - Information obtained from epidemiological
investigation about the vehicle is often the
first step in the traceback investigation
27Conclusion
- To trace the source of an outbreak you need to
look back in time - Use a traceback investigation as a first step in
determining where problems may have occurred in
the chain of production of an implicated vehicle - Traceback investigations provide data for the
next step the environmental health assessment
28References
- 1. Hennessy TW, Hedberg CW, Slutsker L, et al. A
national outbreak of Salmonella Enteritidis
infections from ice cream. N Engl J Med.
19963341281-1286. - 2. Breuer T, Benkel DH, Shapiro RL, et al. A
multistate outbreak of Escherichia coli O157H57
linked to alfalfa sprouts grown from contaminated
seeds. Emerg Infect Dis. 20017977-982. - 3. CDC. Multistate outbreak of Salmonella
Serotype Agona infections linked to Toasted Oats
cerealUnited States, April-May, 1998. MMWR Morb
Mortal Wkly Rep. 199847462-464.