Food Safety: Procedures for Inspecting Canadian Meat Imports (Testimony,
04/02/97, GAO/T-RCED-97-121).

Pursuant to a congressional request, GAO discussed the U.S. system for
inspecting Canadian meat imports, focusing on the: (1) Food Safety and
Inspection Service's (FSIS) system for selecting and inspecting imported
meat products; (2) availability of equipment at import inspection
facilities along the Canadian border that would enable the full
inspection of meat shipped in trucks; and (3) effectiveness of FSIS' new
procedure to verify that carcass samples selected by Canadian inspectors
are representative of the entire shipment.

GAO noted that: (1) FSIS considers the eligible foreign countries'
inspection system, not its own reinspection at the port of entry, to be
the primary control for ensuring that imported meat products meet U.S.
standards; (2) as a check on the foreign countries' inspection
performance, FSIS requires that every shipment of imported meat,
including shipments from Canada, receive some level of U.S. inspector
review at the border; (3) all Canadian meat shipments receive a visual
check for container damage and inaccurate labeling or paperwork; (4)
further, a sample of Canadian meat shipments (about 11 percent of total
Canadian meat shipments in calendar year 1996) receive a more intensive
examination; (5) the sampling approach used to perform this more
intensive examination varies by the type of product being shipped; (5)
for meat shipped in containers, such as boxed ground beef or palletized
bins of canned corn beef, FSIS inspectors unload the entire shipment and
select and examine a random sample of the product; (6) for meat
carcasses, FSIS inspectors select samples from the back of the trailer
because the existing inspection facilities lack the equipment needed to
unload an entire shipment of meat carcasses; (7) to facilitate its
sample inspection of meat carcasses, starting February 16, 1997, FSIS
began requiring that Canadian inspectors located at Canadian slaughter
plants select a random sample of carcasses and place them at the back of
the trailers in which they are to be shipped; (8) the FSIS import
inspector at the border will then examine the Canadian-selected samples;
(9) to verify that the Canadian sample selection is unbiased, FSIS
inspectors also examine at U.S. destinations, about 15 carcass shipments
per month that were not examined at the border; (10) for these
verifications, FSIS inspectors randomly select a sample of carcasses
from the shipment, examine that sample and the Canadian-selected sample
and compare the results of the sample examinations; (11) these data will
be accumulated over time to verify that the Canadian samples are
representative of entire shipments; and (12) while little experience has
been gained since the new procedures took effect in February, an FSIS
pilot test of the procedures found no evidence of bias in Canadian
sampling.

--------------------------- Indexing Terms -----------------------------

 REPORTNUM:  T-RCED-97-121
     TITLE:  Food Safety: Procedures for Inspecting Canadian Meat Imports
      DATE:  04/02/97
   SUBJECT:  Meat inspection
             Product safety
             Import regulation
             International trade regulation
             Information systems
             Safety regulation
             Contaminated foods
             Foreign governments
             Meat packing industry
IDENTIFIER:  United States-Canada Free Trade Area Agreement
             FSIS Automated Import Information System
             Sweetgrass (MT)
             
******************************************************************
** This file contains an ASCII representation of the text of a  **
** GAO report.  Delineations within the text indicating chapter **
** titles, headings, and bullets are preserved.  Major          **
** divisions and subdivisions of the text, such as Chapters,    **
** Sections, and Appendixes, are identified by double and       **
** single lines.  The numbers on the right end of these lines   **
** indicate the position of each of the subsections in the      **
** document outline.  These numbers do NOT correspond with the  **
** page numbers of the printed product.                         **
**                                                              **
** No attempt has been made to display graphic images, although **
** figure captions are reproduced.  Tables are included, but    **
** may not resemble those in the printed version.               **
**                                                              **
** Please see the PDF (Portable Document Format) file, when     **
** available, for a complete electronic file of the printed     **
** document's contents.                                         **
**                                                              **
** A printed copy of this report may be obtained from the GAO   **
** Document Distribution Center.  For further details, please   **
** send an e-mail message to:                                   **
**                                                              **
**                                            **
**                                                              **
** with the message 'info' in the body.                         **
******************************************************************


Cover
================================================================ COVER


Before the Committee on Small Business,
U.S.  Senate

For Release
on Delivery
Expected at
9:00 a.m.  MST
Wednesday
April 2, 1997

FOOD SAFETY - PROCEDURES FOR
INSPECTING CANADIAN MEAT IMPORTS

Statement of Robert A.  Robinson, Director,
Food and Agriculture Issues
Resources, Community, and Economic
Development Division

GAO/T-RCED-97-121

GAO/RCED-97-121T


(150644)


Abbreviations
=============================================================== ABBREV

  FSIS -
  GAO -

============================================================ Chapter 0

Mr.  Chairman and Members of the Committee: 

We are pleased to be here today to respond to your inquiries about
the U.S.  system for inspecting Canadian meat imports.  You asked us
to (1) describe FSIS' system for selecting and inspecting imported
meat products, (2) determine the availability of equipment at import
inspection facilities along the Canadian border that would enable the
full inspection of meat shipped in trucks, and (3) assess the
effectiveness of FSIS' new procedure to verify that carcass samples
selected by Canadian inspectors are representative of the entire
shipment. 

To respond to these questions, we visited 5 of the 21 FSIS-approved
import inspection facilities located along the U.S./Canada border and
a Canadian slaughterhouse, where we observed the inspection and
sampling processes.  In addition, we met with FSIS officials and
current and former import inspectors to discuss the import inspection
process.  We also reviewed the availability of equipment required to
unload carcasses at import inspection stations and FSIS' data on the
quality of Canadian sampling. 

In summary, FSIS considers the eligible foreign countries' inspection
system--not its own reinspection at the port of entry--to be the
primary control for ensuring that imported meat products meet U.S. 
standards.  As a check on the foreign countries' inspection
performance, FSIS requires that every shipment of imported meat,
including shipments from Canada, receive some level of U.S. 
inspector review at the border.  All Canadian meat shipments receive
a visual check for container damage and inaccurate labeling or
paperwork.  Further, a sample of Canadian meat shipments (about 11
percent of total Canadian meat shipments in calendar year 1996)
receive a more intensive examination. 

The sampling approach used to perform this more intensive examination
varies by the type of product being shipped.  For meat shipped in
containers, such as boxed ground beef or palletized bins of canned
corn beef, FSIS inspectors unload the entire shipment and select and
examine a random sample of the product.  For meat carcasses, FSIS
inspectors select samples from the back of the trailer because the
existing inspection facilities lack the equipment needed to unload an
entire shipment of meat carcasses.  To facilitate its sample
inspection of meat carcasses, starting February 16, 1997, FSIS began
requiring that Canadian inspectors located at Canadian slaughter
plants select a random sample of carcasses and place them at the back
of the trailers in which they are to be shipped.  The FSIS import
inspector at the border will then examine the Canadian-selected
samples.  To verify that the Canadian sample selection is unbiased,
FSIS inspectors also examine at U.S.  destinations, about 15 carcass
shipments per month that were not examined at the border.  For these
verifications, FSIS inspectors randomly select a sample of carcasses
from the shipment, examine that sample and the Canadian selected
sample and compare the results of the sample examinations.  These
data will be accumulated over time to verify that the Canadian
samples are representative of entire shipments.  While little
experience has been gained since the new procedures took effect in
February, an FSIS pilot test of the procedures found no evidence of
bias in Canadian sampling. 


   BACKGROUND
---------------------------------------------------------- Chapter 0:1

The Federal Meat Inspection Act requires that meat imports meet U.S. 
standards for wholesomeness and be produced under inspection systems
that are equivalent to the U.S.  system.  Under the act, FSIS reviews
the inspection systems of exporting countries for equivalency with
the U.S.  system and reinspects imported meat at the port of entry as
a spot-check of the countries' inspection performance.  FSIS conducts
two types of spot checks which are commonly referred to as "skip"
assignments and "inspect" assignments.  Skip assignments require a
visual check for container damage and inaccurate labeling or
paperwork.  Inspect assignments require an FSIS import inspector to
conduct the labeling, paperwork and container review associated with
a skip assignment and to (1) randomly select and examine product
samples and (2) sometimes randomly select and send product samples to
an FSIS laboratory for testing.  In a product examination, the
inspector feels, smells, and visually examines exposed product
samples to discover defects such as blood clots, bone fragments,
extraneous materials, feces, hair, and lesions. 

Canada, generally speaking, is subject to the same requirements as
other exporting countries, but is granted certain exceptions.  One
such exception is that for meat carcasses shipped from Canada,
Canadian inspectors, not FSIS inspectors, typically select the
carcass samples to be examined by FSIS.  In accordance with one goal
of the 1988 U.S./Canada Free Trade Agreement--facilitate commerce by
reducing trade barriers--both countries sought agreement to minimize
inspection procedures applicable to each other's meat imports. 


   AUTOMATED SYSTEM DETERMINES
   TYPE OF INSPECTION FOR IMPORTS
---------------------------------------------------------- Chapter 0:2

For meat being imported from all countries, FSIS relies on its
Automated Import Information System to automatically assign the type
of reinspection that an FSIS import inspector will perform on an
incoming shipment.  The automated assignment system works in two
ways.  When non-Canadian shipments arrive at a U.S.  port of entry, a
description of the shipment is entered into the automated system and
the type of reinspection is automatically assigned.  The proportion
of skip assignments and inspect assignments will vary with the
product type, country of origin and meat producer.  In calendar year
1996, approximately 70 percent of meat shipments from countries other
than Canada were given a skip and 30 percent were given the more
intensive inspect assignment. 

For Canadian shipments, the process is modified.  Rather than
assigning inspections on the basis of the product type, the automated
system is programmed to (1) randomly generate about 3,000 inspect
assignments annually, at a rate of about 250 per month, and (2)
automatically generate inspect assignments for those meat producers
that are placed in intensified inspection status because of recent
compliance problems.\1 For calendar year 1996, approximately 89
percent of Canadian meat shipments received skip assignments and 11
percent received inspect assignments (7 percent triggered by random
assignments and 4 percent triggered by previous compliance problems). 
Skip assignments for Canadian shipments also differ from those for
other countries in that rather than unloading the shipment,
inspectors check only what is visible when the rear doors of the
trailer are opened. 

As with imports from other countries, an inspect assignment for
Canadian meat shipments (for other than meat carcasses) requires that
the entire shipment be unloaded from the truck so that the inspector
can randomly select samples and perform specified product
examinations.  As discussed below, however, entire Canadian shipments
of meat carcasses are not unloaded because of the limited equipment
at the border inspection facilities. 


--------------------
\1 If a shipment of imported meat is refused entry, the foreign meat
producer is placed on "intensified" inspection status.  Non-Canadian
meat producers, depending on the reason for refusal and the type of
product refused, will automatically receive an inspect assignment
until the next 10 to 15 shipments of that product pass inspection. 
For Canadian meat producers, 15 consecutive shipments of the refused
product--equaling at least 15 times the weight of the refused
shipment--must pass inspection regardless of product type or reason
for refusal. 


   BORDER INSPECTION FACILITIES
   LACK EQUIPMENT NEEDED TO UNLOAD
   CANADIAN CARCASS SHIPMENTS
---------------------------------------------------------- Chapter 0:3

Only 1 of the 21 FSIS-approved inspection facilities along the
U.S./Canadian border has the necessary equipment to unload an entire
shipment of meat carcasses.  As a result, unlike non-carcass
shipments, Canadian meat carcass shipments are not entirely unloaded
at any of the border inspection facilities. 

Our analysis of FSIS data and our observations at the inspection
facilities we visited confirmed that only one inspection facility
located on the U.S./Canada border (facility number I-47 in
Sweetgrass, Montana) had sufficient overhead rail capacity to hang an
entire truckload of meat carcasses.  Nine of the border inspection
facilities had no rail capacity, and the remaining 11 facilities had
rail capacity ranging from 16 to 100 feet in length.  Because meat
carcasses are generally shipped in trailers having a total rail
length of about 160 feet, the rail capacity at these facilities is
not adequate to handle a full shipment of carcasses. 

FSIS-approved import inspection stations have never been required to
have the overhead rail capacity needed to unload a full truckload of
carcasses.  In January 1989 the United States and Canada, responding
to the goals of the U.S./Canada Free Trade Agreement, implemented new
meat inspection procedures for meat products traded between the two
countries.  According to the FSIS Import Inspection Director, prior
to 1989, Canadian meat carcasses were inspected at their U.S. 
destination, rather than at the border.\2 In January 1989, FSIS moved
its inspection activities for imported Canadian meat to approved
privately built and operated inspection facilities located along the
U.S./Canada border.  However, the privately owned facilities were
only required to have enough overhead rail capacity to unload the
Canadian-selected carcass samples for FSIS inspectors' examination. 
In July 1992 FSIS discontinued Canadian sample selection in response
to issues we raised during our 1992 report on Canadian meat
inspection.\3 In that report we expressed concern about giving
advance notice of inspection to Canadian meat plants and having
Canadian inspectors select samples for U.S.  inspection.  Because the
import facilities lacked the equipment to unload full carcass
shipments, FSIS import inspectors started selecting their own samples
from the carcasses accessible at the rear of a truck. 

After discontinuing Canadian sample selection and eliminating the
advance notice of inspection to Canadian meat plants, FSIS import
inspection officials told us that they had made a number of
unsuccessful attempts to provide for random sampling of carcass
shipments at the border.  For example, around August 1993, FSIS
proposed to Canadian inspection and meat industry officials that
carcasses be unloaded for random selection and that inspection
facilities be upgraded for this purpose.  The proposal was dropped
because (1) industry and Canadian government officials were concerned
about shipment delays and potential adverse impacts on meat hygiene,
and (2) import inspection facility owners objected to the increased
costs associated with providing the additional equipment and staff
that would be needed to unload an entire truckload of meat carcasses. 

A second FSIS proposal, made in December 1993, to sample carcasses
from randomly selected sections of truck trailers was also abandoned
when it was determined that the trailers might become unbalanced,
creating a safety hazard.  FSIS officials concluded that unloading
entire carcass shipments at the border inspection facilities was (1)
unnecessary, given the U.S.-equivalent Canadian inspection process
before shipment and (2) more risky, given the increased potential
contamination and spoilage associated with additional handling. 


--------------------
\2 Currently, two U.S.  destination facilities are authorized to
conduct import inspections.  Canadian importers may request
inspection at these facilities, rather than at the port of entry. 

\3 Food Safety and Quality:  USDA Improves Inspection Program for
Canadian Meat, But Some Concerns Remain (GAO/RCED-92-250, Aug.  26,
1992). 


   NEW VERIFICATION PROCEDURE
   CHECKS CANADIAN CARCASS SAMPLE
   SELECTION
---------------------------------------------------------- Chapter 0:4

FSIS recently revised its procedures for Canadian carcasses imports
to again allow for Canadian inspectors to select the samples that
will be examined by FSIS inspectors.  The revised procedures also
establish a system for FSIS to verify that the Canadian-selected
carcass samples are representative of the entire shipment.  Effective
February 16, 1997, Canadian meat producers exporting meat to the U.S. 
are required to place, at the rear of every truckload, marked
carcasses that were randomly selected by a Canadian government
inspector.  These are the carcasses that will be examined by the FSIS
import inspector if an inspect assignment is drawn from the automated
system at a U.S.  border inspection facility. 

However, unlike the inspection procedures that were in place prior to
July 1992, Canadian plants are not provided advance notice of
inspections and FSIS has implemented a process to verify that the
Canadian selected samples are unbiased.  To verify that Canadian
samples are representative of shipments, 15 shipments a month that
receive skip assignments at the border will be resealed and sent to
their U.S.  destination, where they will be met by an FSIS import
inspector.  The inspector will unseal the truck, remove the Canadian
selected sample, randomly select another sample from the remaining
carcasses, and compare the inspection results from the two samples. 
The results of the verifications will be accumulated over time to
statistically gauge the reliability of the Canadian sampling process. 
If the cumulative verification results exceed the pre-set limit of
variation established by FSIS, there is evidence of possible bias in
the Canadian sample selection process.  Results will be compiled for
individual Canadian slaughterhouses to ensure that the Canadian
sample selection process is unbiased. 

FSIS and Canadian inspection officials conducted a 60-shipment pilot
test of the new verification procedure from July 10, 1995, to October
16, 1995.  The test involved one of the largest Canadian meat
producers and two U.S.  destination plants.  The results of the pilot
test indicated that Canadian sampling was unbiased and that the
verification procedure is an effective means for ensuring that
Canadian-selected samples are representative. 


-------------------------------------------------------- Chapter 0:4.1

This concludes our prepared statement.  We would be happy to respond
to any questions you or members of the committee may have. 


*** End of document. ***