Mad Cow Disease: FDA's Management of the Feed Ban Has Improved,  
but Oversight Weaknesses Continue to Limit Program Effectiveness 
(25-FEB-05, GAO-05-101).					 
                                                                 
More than 5 million cattle across Europe have been killed to stop
the spread of bovine spongiform encephalopathy (BSE), commonly	 
called mad cow disease. Found in 26 countries, including Canada  
and the United States, BSE is believed to spread through animal  
feed that contains protein from BSE-infected animals. Consuming  
meat from infected cattle has also been linked to the deaths of  
about 150 people worldwide. In 1997, the Food and Drug		 
Administration (FDA) issued a feed-ban rule prohibiting certain  
animal protein (prohibited material) in feed for cattle and other
ruminant animals. FDA and 38 states inspect firms in the feed	 
industry to enforce this critical firewall against BSE. In 2002, 
GAO reported a number of weaknesses in FDA's enforcement of the  
feed ban and recommended corrective actions. This report looks at
FDA's efforts since 2002 to ensure industry compliance with the  
feed ban and protect U.S. cattle.				 
-------------------------Indexing Terms------------------------- 
REPORTNUM:   GAO-05-101 					        
    ACCNO:   A18236						        
  TITLE:     Mad Cow Disease: FDA's Management of the Feed Ban Has    
Improved, but Oversight Weaknesses Continue to Limit Program	 
Effectiveness							 
     DATE:   02/25/2005 
  SUBJECT:   Animal diseases					 
	     Cattle						 
	     Federal regulations				 
	     Feed additives					 
	     Feed industry					 
	     Infectious diseases				 
	     Inspection 					 
	     Interagency relations				 
	     Livestock products 				 
	     Materials handling 				 
	     Noncompliance					 
	     Performance measures				 
	     Regulatory agencies				 
	     Contamination					 
	     Disease surveillance				 
	     Livestock						 
	     Bovine Spongiform Encephalopathy			 
	     FDA Field Accomplishment and Compliance		 
	     Tracking System					 
                                                                 
	     Variant Creutzfeldt-Jacob Disease			 

******************************************************************
** This file contains an ASCII representation of the text of a  **
** GAO 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.                                         **
**                                                              **
******************************************************************
GAO-05-101

Report to Congressional Requesters

February 2005

MAD COW DISEASE

FDA's Management of the Feed Ban Has Improved, but Oversight Weaknesses
Continue to Limit Program Effectiveness

Contents

Tables

February 25, 2005Letter

The Honorable Saxby Chambliss Chairman The Honorable Tom Harkin Ranking
Member Committee on Agriculture, Nutrition, and Forestry United States
Senate

The Honorable Thad Cochran The Honorable Richard J. Durbin United States
Senate

Bovine spongiform encephalopathy (BSE), commonly known as mad cow disease,
is an always fatal neurodegenerative animal disease that has been found in
cattle in 26 countries since it was first identified in the United Kingdom
in 1986. In December 2003, the United States discovered its first case of
BSE in a cow in Washington State. The U.S. Department of Agriculture
(USDA) later determined that this cow was imported from Canada. The agent
believed to be responsible for BSE is a malformed type of protein called a
prion, found in certain tissue-particularly brain and central nervous
system tissue-of infected animals. Cattle contract BSE by eating feed
derived from the remains of BSE-infected animals.1 In Europe, more than 5
million head of cattle have been killed to thwart the spread of the
disease. Scientists also generally believe that a fatal disease in
humans-knows as variant Creutzfeldt-Jacob Disease (vCJD)-is linked to
eating beef contaminated with the malformed protein. Research suggests
that vCJD is difficult for humans to contract-about 150 people have died
worldwide from vCJD. Both diseases have long incubation periods during
which they are undetectable-2 to 8 years in cattle and possibly up to 30
years in humans.

USDA is primarily responsible for detecting the disease in cattle, and the
Department of Health and Human Services' Food and Drug Administration
(FDA) is primarily responsible for preventing its introduction and spread
through animal feed. Both agencies recognize the importance of preventing
BSE from becoming established in the United States-not only to protect the
safety of the U.S. food supply but also to protect the economic viability
of the $70 billion U.S. beef industry. With 95 million head of cattle, the
United States is the world's largest beef producer, exporting a record 2.6
billion pounds of beef, valued at over $3.1 billion, in 2003. In January
2002, we reported that the potential impact of even a small outbreak of
BSE in the United States could be economically devastating.2 Indeed,
between January and September 2004, the industry lost more than 80 percent
of its export trade, or an estimated $2 billion, following the discovery
of the one BSE-infected animal in December 2003. Although most countries
stopped importing U.S. beef for some period of time, domestic consumption
did not drop. In fact, changing dietary trends have led to increased U.S.
beef consumption in the last several years. The United States is in
discussions with its major trading partners about renewing U.S. beef
imports.

To protect U.S. cattle and consumers, USDA and FDA have put in place three
primary firewalls. These include the following:

o Controls over imports. Since 1989, USDA has prohibited the importation
of live cattle and certain cattle products from countries where BSE is
known to exist. In 1992, FDA began identifying medical products and other
FDA-regulated foods and products derived from cattle from countries with
BSE. USDA and FDA, in cooperation with the Department of Homeland
Security's Customs and Border Protection, screen shipments of such
products.

o Animal surveillance. Since 1990, to detect BSE, USDA has been testing
brain tissue, primarily from cattle that exhibit neurological symptoms and
adult cattle that die from unknown causes, as well as from cattle
slaughtered for meat.

o Feed ban.3 In 1997, FDA banned the use of most proteins derived from
mammals in feed intended for cattle and other ruminants to keep
potentially infectious tissue out of cattle feed.4

This report focuses on FDA's implementation and enforcement of the animal
feed-ban rule, which many industry and consumer groups consider the most
important firewall against the introduction and spread of BSE in the
United States.

Under the feed-ban rule, FDA requires firms to (1) label feed and feed
ingredients that contain or may contain most proteins from most mammals
(referred to hereafter as prohibited material) with a cautionary statement
that reads "Do not feed to cattle or other ruminants," (2) have procedures
to protect against commingling or cross-contamination if they handle both
prohibited and nonprohibited material for feed and feed ingredients, and
(3) maintain records so that feed and feed ingredients that contain or may
contain prohibited material can be tracked from receipt through
disposition.5 Firms that transport both types of materials also must have
procedures to prevent commingling.

FDA's feed-ban rule applies to feed for cattle and other ruminants, such
as sheep and goats. The material prohibited for use in cattle feed may
continue to be used in pet food and in feed for poultry, swine, horses,
and other nonruminant animals.

The feed-ban rule designates a number of cattle- and other animal-derived
items as exempt from the ban, and hence allowable in cattle feed. These
exempt items include blood and blood products, plate waste, gelatin, and
milk and milk proteins.6 In addition, poultry litter (a protein source
comprised of poultry waste material, bedding, and spilled feed) is allowed
in cattle feed. FDA has published, but not taken action on, several
advance notices of proposed rulemaking for revising the ban to, among
other things, end most of the exemptions and require that feed
manufacturers and other such firms use dedicated equipment for cattle
feed.

To oversee compliance with the feed ban, inspectors from FDA and the 38
states that have contracts or agreements with FDA periodically inspect
firms, using FDA guidance and an inspection form that FDA developed to
document inspection results.7 Since 1997, FDA and states have identified
and inspected about 14,800 firms that are subject to the feed-ban rule.
The types of firms inspected include renderers, protein blenders, feed
mills, farms that mix their own feed, feed transporters, pet food
manufacturers, and other firms subject to the feed-ban rule.8 In 2002, FDA
began using a risk-based priority approach to determine which firms to
inspect annually. Under this approach, FDA has designated firms that
manufacture, blend, and otherwise directly process with prohibited
material as the highest risk for potentially exposing U.S. cattle to BSE.
Firms that do not process with prohibited material are designated as lower
risk. FDA had designated about 570 firms as high risk in fiscal year 2004.
FDA assigns a list of high-risk firms for inspection to each of its
district offices and gives the district offices some discretion in
deciding which lower risk firms to inspect. FDA also negotiates with
states over the number of inspections that FDA will pay for under
contract. States conduct about 70 percent of feed-ban compliance
inspections. When FDA determines that firms are out of compliance, it can
issue warning letters, encourage firms to conduct voluntary recalls, or
seek court orders to seize feed or feed ingredients. FDA district offices
review inspection results for accuracy and enter inspection information
into FDA's central data system-the Field Accomplishments and Compliance
Tracking System (FACTS)-and periodically post inspection results on FDA's
Web site.

Our January 2002 report identified a number of weaknesses in federal BSE
prevention and detection efforts.9 Regarding the feed ban, we reported
that FDA had not promptly taken actions, such as issuing warning letters
or reinspecting firms that were out of compliance, to enforce the feed ban
and keep prohibited material out of cattle feed. We also reported that
FDA's data on inspections were so severely flawed that the agency could
not know the extent of industry compliance. We made a number of
recommendations to strengthen FDA's enforcement of the feed ban and its
oversight and monitoring of compliance inspections.

As you requested, this report examines the effectiveness of the actions
FDA has taken, since our 2002 report, to ensure industry compliance with
the animal feed ban and protect U.S. cattle from BSE. In addition,
appendix III provides a chronology of FDA's and USDA's actions in response
to the two cases of BSE discovered in North America in 2003.

In conducting our work, we examined in detail 404 inspection reports from
fiscal years 2003 and 2004, which we randomly selected from FDA's 18
district offices responsible for inspections in the 50 states. We
interviewed FDA district officials in the 18 districts and observed FDA
and state investigators conduct 19 inspections in 12 states. The sites
were selected to cover a range of firm types and sizes in various
geographic locations with concentrations of cattle feeding operations,
including dairy cattle. We met with FDA headquarters' officials
responsible for procedures and actions taken to (1) oversee and enforce
the feed ban, (2) maintain the inspection data system, and (3) propose and
analyze regulatory decisions. We also surveyed state agency officials in
the 38 states that had contracts or agreements with FDA in fiscal year
2004 regarding their inspection programs, testing of animal feed and feed
ingredients, and the training and guidance they received from FDA. The
state survey instrument and summary of responses appear in appendix IV.
Appendix I contains a detailed description of our scope and methodology.
We performed our work from October 2003 through January 2005, in
accordance with generally accepted government auditing standards, which
included an assessment of FDA's BSE program data reliability and internal
controls.

Results in Brief

FDA has taken a number of important actions, as we recommended in our 2002
report, to improve its implementation of the feed ban. FDA developed a
uniform format for federal and state inspectors to document inspection
results, implemented a new data system that more reliably tracks
inspection results, and entered inspection results into the data system in
a more timely fashion. FDA also issued guidance and trained its inspectors
along with state inspectors on how to conduct BSE inspections. However, we
found the following weaknesses in FDA's oversight and enforcement of the
feed ban, which continue to limit the effectiveness of this critical BSE
firewall and could place U.S. cattle at risk for BSE:

o FDA acknowledges that more firms are subject to the feed ban than the
nearly 14,800 that have been inspected to date, but it does not have
uniform procedures for identifying additional firms. Because these firms
have never been inspected, FDA has no assurance that they are in
compliance with the feed ban. FDA officials told us the agency has asked
Congress for more resources, which it plans to use, in part, to support
states' efforts to identify and inspect additional firms. We observed one
possible approach to help FDA identify additional firms with existing
resources: some inspectors wrote down the names of the suppliers and
customers of firms during inspections to check against the inventory of
active firms. Inspectors do not routinely note such information, however,
because FDA's guidance does not instruct them to do so.

o We found that about 2,800 firms had not been reinspected since 1999 or
earlier. While those early inspections indicated that most did not process
with prohibited material at that time, the firms could have changed their
practices over the last 5 years. Our analysis showed that about two-thirds
of those firms were farms that fed cattle and did not feed other types of
animals; FDA believes such farms are unlikely to change their practices.
However, about 400 firms were feed mills, which FDA would consider at high
risk of potentially exposing cattle to BSE if they started to use
prohibited material. Because firms are not required to notify FDA if they
change their operations and begin to process feed using prohibited
material, FDA would not target them for annual inspection as high-risk
firms.

o FDA's inspection guidance does not include routinely sampling feed
intended for cattle, in cases where such tests would be useful, to augment
the visual examination of facilities and equipment and review of documents
carried out during inspections. According to FDA, the presence of exempt
items, such as cattle blood, which are allowed in cattle feed, would
negate the value of the tests because the tests cannot distinguish between
prohibited material and these exempt items. However, 18 of the 38 states
that conduct BSE inspections under agreement with FDA told us they take
samples of feed during inspections to test for animal material. State
officials told us that tests could confirm the presence of potentially
prohibited material in cattle feed at firms that assert they do not use
exempt items. Tests would also be useful to confirm the adequacy of
procedures for cleaning equipment and vehicles used for both cattle feed
and feed with prohibited material. However, FDA began testing bags of feed
sold at retail stores and bulk feed sold to cattle feedlots in August
2003. These samples were not taken as part of the compliance inspections
and were not selected systematically. According to FDA officials, tests on
some of these samples indicated the presence of animal material and the
agency was investigating those test results at the completion of this
report. We plan to provide our analysis of FDA's collection, testing, and
follow-up of these samples later this year.

o FDA's regulations do not require the cautionary statement-"Do not feed
to cattle or other ruminants"-on feed or feed ingredients that contain
prohibited material if they are intended for export, although that feed
could be intentionally or inadvertently redirected back into feed for U.S.
cattle. In addition, the exported feed containing prohibited material
could be fed to cattle in other countries and meat from those animals
could subsequently be imported into the United States. However, according
to FDA officials, FDA cannot require feed intended for export to carry the
cautionary statement without a change to the law that governs the export
of food and feed.

o Although FDA has procedures for alerting USDA and states when it
discovers that cattle may have consumed feed that contains prohibited
material, FDA officials told us that they had never given such
notification, even though they had identified instances when prohibited
material had been used in cattle feed in the past. FDA said that
notification was not needed because BSE had not been discovered in a cow
born in the United States. However, FDA's position is inconsistent with
the purpose of the feed ban-to be a firewall for safeguarding U.S. cattle
from the introduction and spread of BSE. On one inspection we observed, an
inspector discovered that a firm's process had been allowing prohibited
material into cattle feed for nearly a year. The firm voluntarily
conducted a recall, but FDA did not alert USDA and the state. FDA
maintained that the recall was sufficient; however, USDA officials told us
that the department would have tracked the animals that may have been fed
contaminated feed and tested them for BSE when slaughtered.

o FDA has not identified or inspected many transportation firms. In
addition, inspectors do not routinely review and document firms'
procedures for ensuring that the vehicles they use to haul cattle feed are
free of prohibited material. Routine review and documentation does not
occur in part because FDA's inspection form does not have specific
questions to capture that information. Eighty-two of the inspection
reports we examined were for renderers, protein blenders, feed mills, and
other firms that handled cattle feed and feed ingredients and also
processed with prohibited material. Inspectors documented vehicle
clean-cut procedures for only 11 of those 82 firms. Research suggests that
cattle can get BSE from ingesting even a small amount of infected
material-an amount that could be introduced in feed that was transported
in a poorly cleaned vehicle. FDA told us it has requested resources to
identify and inspect more transportation firms. However, because thousands
of trucks could transport cattle feed, we believe it would be more
effective to review and document cleanout procedures and inspect vehicles
as part of inspections at feed mills or other firms that use the vehicles
to haul their cattle feed or feed ingredients.

In addition to these weaknesses in the feed-ban firewall, we also
identified a related issue that needs to be addressed. FDA is reporting
information to Congress and the public on industry compliance without
providing a full and complete context for that information. That is, FDA
reported a 99 percent compliance rate in January 2004. While FDA noted the
rate was based on renderers, protein blenders, and feed mills that process
with prohibited material, it did not note that the rate was based on
inspections of only about 570 firms. Some industry officials have cited
that high rate of compliance as support for their position that FDA does
not need to strengthen the feed-ban rule. Furthermore, FDA does not
include all serious violations in its calculations of compliance on its
Web site because it reclassifies firms as "in compliance" once they
correct violations, regardless of how long the problem may have existed.
In addition, in 42 of the 404 inspection reports that we analyzed in
depth, FDA had counted firms as "in compliance" that lacked written
procedures to prevent commingling or cautionary statements on feed that
contained prohibited material-violations that can result in cattle being
fed prohibited material. Because of these concerns and the fact that FDA
is still identifying firms subject to the ban-as well as the fact that
inspections are largely paperwork reviews without tests to confirm
compliance, and some inspections are 5 or more years old-we do not believe
that FDA has enough information or enough current information to cite a
rate of compliance. Any compliance information FDA cites must be reported
in its complete context.

To further strengthen oversight and enforcement of the animal feed ban and
better protect U.S. cattle and American consumers, we are making nine
recommendations to the Commissioner of FDA, including that FDA develop
procedures for identifying additional firms subject to the ban; ensure
that it alerts USDA and states when inspectors discover that feed with
prohibited material may have been fed to cattle; and develop guidance for
inspectors to use tests to verify the safety of cattle feed and confirm
the adequacy of procedures for ridding equipment and vehicles of
prohibited material before using them for cattle feed.

In commenting on a draft of this report, FDA said we had conducted a
thorough and diligent study. However, FDA believes that the weaknesses we
identified are not sufficiently material to place U.S. cattle at risk for
BSE and that its risk-based inspection approach assures adequate oversight
of the feed-ban rule. We believe that the problems described in this
report are serious and that, given the fact that BSE has been discovered
in North American cattle, breaches in FDA's oversight of the feed-ban rule
place U.S. cattle at risk for BSE. FDA generally disagreed with four of
our nine recommendations. FDA did not agree that, among other things, it
should use tests as part of compliance inspections, as we recommend,
because current tests cannot detect the prions that cause BSE. That is
true. However, the existing test can detect animal tissue, and FDA is
using it to test samples of bagged feed and feed sold at mills. We believe
tests, in conjunction with document review and visual examination carried
out during compliance inspections, will give FDA greater assurance that
inspection results are accurate. FDA also disagreed with our
recommendation that it require firms that process with prohibited material
to notify FDA. FDA believes it would need significant additional resources
to implement a notification program and said that its current approach of
working collaboratively with states gives FDA a good opportunity to learn
when firms change to using prohibited material. If there are not
significantly more high-risk firms than the approximately 570 firms FDA
already knows about, then the cost of implementing this recommendation
would be minimal. However, if the number of firms that process with
prohibited material is significantly larger, FDA needs to know that.
Appendix VI contains FDA's written comments and our detailed response.

Background

BSE and vCJD belong to a family of diseases known as transmissible
spongiform encephalopathies (TSE). Other TSEs include scrapie in sheep and
goats, chronic wasting disease in deer and elk, feline spongiform
encephalopathy in domestic cats, and mink encephalopathy. Currently, no
therapies or vaccines exist to treat TSEs and a definitive diagnosis can
only be made from a post-mortem examination of the brain. The infective
agent that gives rise to TSEs is generally thought to be a malformed type
of protein, called a prion, which causes normal molecules of the same type
of

protein in the brain to become malformed and eventually results in
death.10 Prions are neither viruses nor bacteria and contain no genetic
material-no deoxyribonucleic acid (DNA). Prions cannot be readily
destroyed by conventional heat, irradiation, chemical disinfection, or
sterilization procedures.11 TSE prions have been found to accumulate in
central nervous system tissue-specifically the brain, spinal cord, and
eye-and have been found in other body tissues, such as the tonsils and
small intestines, of animals and humans. For BSE, the precise amount of
infective material needed to cause disease is unknown, but research
suggests that it is very small. According to scientific experts in the
European Commission, in careful feeding experiments, less than 1 gram of
infected brain tissue induced disease in all the recipient cattle.

The original source of BSE is not known with certainty. However, based on
available evidence, experts generally agree that the practice of recycling
the remains of diseased animals, specifically scrapie-infected sheep, into
feed for livestock, including cattle, was responsible for the emergence
and spread of BSE in the United Kingdom. In 1986, BSE was first identified
in the United Kingdom; and in 1988, that government banned the practice of
feeding ruminant-derived protein to ruminants to thwart its spread. The
number of new cases of BSE has declined from a high in 1992 of 37,316 to a
total of 764 new cases in 2004. BSE has been found in about 189,000
animals worldwide, most of which (about 184,000) were discovered in the
United Kingdom. The remaining cases were discovered in 26 countries,
including Canada and the United States. Three nations-the United States,
Oman, and the Falkland Islands-have only detected the disease in imported
animals. The following are the number of reported cases, by region and/or
country:12

o Europe. United Kingdom-184,045; the rest of Europe-5,107;

o North America. Canada-4; United States-1;

o Middle East. Oman-2; Israel-1;

o Asia-Pacific. Japan-14; and

o South America. Falkland Islands-1.

In 1996, the United Kingdom reported the first case of the human disease,
vCJD. Scientists believe vCJD is linked to exposure to the BSE prion, most
likely through consuming beef and beef products infected with BSE.13 While
scientists and regulatory officials believe that millions of people in the
United Kingdom may have ingested BSE-infected tissue, many also believe
vCJD is difficult to contract. As of December 1, 2003, 153 cases of vCJD
had been reported worldwide, with 143 of these cases in the United
Kingdom. The Department of Health and Human Services' Centers for Disease
Control and Prevention, which is responsible for surveillance of vCJD,
reported that almost all of the vCJD victims had multiple-year exposures
in the United Kingdom during the height of the outbreak of BSE-infected
cattle-between 1980 and 1996. Most vCJD victims have been young-the
average age at death was 28-and half died within 13 months from the time
they first showed symptoms.

The first indigenous case of BSE in North America was discovered in Canada
in May 2003. (Canada's first infected cow, discovered in 1993, had been
imported from the United Kingdom.) A Canadian government investigation
concluded that the infected cow discovered in 2003 most likely contracted
the disease by consuming feed containing BSE-contaminated ruminant
material, probably before Canada imposed its feed ban in 1997.14 Canadian
authorities believe that BSE entered the feed chain through slaughtered
and rendered cattle imported from the United Kingdom. In December 2003, an
animal infected with BSE was discovered in the United States. According to
U.S. authorities, that animal-a dairy cow in Washington State-had been
part of a herd of 81 cattle imported from Canada in September 2001.
Appendix III describes FDA's and USDA's actions in response to the 2003
discoveries. In January 2005, Canada discovered two more cases of BSE.

Following the discovery of the infected cow in the United States, U.S.
beef exports dropped precipitously. The United States is currently engaged
in discussions with its major trade partners to reestablish beef exports.
In October 2004, Japan, previously the largest importer of U.S. beef,
agreed in principle to resume imports of certain beef products from cattle
slaughtered at 20 months or younger; as of February 11, 2005, the two
countries were working out the details of this agreement.

To detect potentially prohibited material in feed, FDA uses a test called
"feed microscopy," which is a visual examination of a sample under a
microscope for the presence of animal tissue, such as hair and bone
particles. According to FDA officials, when performed by an experienced
analyst, the species can sometimes be identified. FDA is evaluating a more
sensitive test called "polymerase chain reaction" (PCR), which detects
animal DNA and can distinguish ruminant DNA. However, feed containing
exempt items (e.g., milk and blood proteins) derived from ruminants would
test positive for ruminant DNA using PCR.

When inspectors find violations of the feed-ban rule, FDA can issue
warning letters, and firms may conduct voluntary feed recalls. FDA has the
authority to take immediate enforcement action, including seeking a court
order to seize feed products that violate the feed ban or obtaining a
court-ordered injunction ordering a firm to cease operations. Of the 38
states we surveyed, 37 told us they have authority to take action for
violations of the feed ban. FDA directs its districts to issue warning
letters within 30 workdays-approximately 45 calendar days after the
inspection. Warning letters give firms the opportunity to voluntarily take
corrective action before FDA initiates enforcement actions.

Under the risk-based priority inspection system that FDA adopted in 2002,
FDA and states have focused inspection resources on the following types of
firms, which FDA has designated as high-risk for potentially exposing
cattle to BSE:

o renderers that accept dead ruminant animals and/or the waste materials
from beef slaughter facilities;

o feed mills that use prohibited material, which can include FDA-licensed
mills that handle certain new animal drugs for use in animal feeds and
nonlicensed mills that do not handle such animal drugs; and

o protein blenders that use prohibited material.

Other firms subject to the feed ban include the following:

o firms that manufacture only pet food;

o firms that transport or distribute animal feed;

o firms that salvage animal feed or pet food; and

o other firms that handle animal feed, including retailers, grocery
warehouses, and specialty food companies.

In addition to inspections of high-risk firms, FDA asks states to perform
a number of inspections at the lower risk firms under their contracts or
agreements with FDA. FDA also performs inspections of some lower risk
firms. Table 1 shows the number of firms inspected during fiscal year
2004.

Table 1: Number of Firms Inspected by FDA and States for Compliance with
the Feed-Ban Rule, by Firm Type, Fiscal Year 2004

Firm type                                                  Number of firms 
Renderers                                                              195 
Protein blenders                                                       164 
FDA-licensed feed mills                                                747 
Nonlicensed feed mills                                               2,615 
Othersa                                                              2,285 
Total                                                                6,006 

Source: GAO's analysis of FDA's online database.

Note: Since many firms engage in more than one activity (for example, a
feed mill may also be a protein blender), this analysis makes assumptions
about firms' primary activities to avoid counting firms in more than one
category.

aOther firm types include cattle feeders, transporters, pet food
manufacturers, and retail establishments.

Our 2002 report found that

o FDA was not acting promptly to compel firms to keep prohibited materials
out of cattle feed and to label animal feed that cannot be fed to cattle;

o FDA's data on feed inspections was so severely flawed that FDA did not
know the full extent of industry compliance;

o FDA had no clear enforcement strategy for firms that do not obey the
feed ban and did not know what enforcement actions states had taken; and

o FDA had been using inaccurate, incomplete, and unreliable data to track
and oversee feed-ban compliance.

A 2001 study by the Harvard Center for Risk Analysis noted that the
greatest risk of BSE exposure to cattle in the United States is through
mishandling, mislabeling, or contaminating cattle feed.15 The study
developed a simulation model for predicting the number of infected animals
that would result from the introduction of BSE into the United States.
Using this model, the Harvard study concluded that, if 10 cattle infected
with BSE were imported into the United States, only three new cases of BSE
would likely occur, on average, and that BSE is virtually certain to be
eliminated from the United States within 20 years following its
introduction. According to the study, any new cases of BSE would come
primarily from industry's failure to comply with the feed ban. A
subsequent 2003 Harvard reassessment-following the discovery of the
BSE-infected cow in Canada that year-arrived at a similar conclusion.16

FDA Has Taken Important Steps to Improve Implementation of the Feed Ban

Since our January 2002 report, FDA has changed the way it collects,
tracks, and reports inspection data. In April 2002, FDA implemented a
uniform inspection form for federal and state inspectors to document
inspection results. Although FDA had an inspection form earlier,
inspectors were not always completing the required information, and
several states did not use FDA's form.

FDA has also issued feed-ban inspection guidance and appointed BSE
coordinators in each of its district offices to review inspection forms
for completeness.17 The district BSE coordinators told us that FDA has
trained inspectors on using the inspection form and carrying out
inspections. Although most states reported that this training was
sufficient, a few told us that they had not received training since the
late 1990s or were not able to attend training because of state budget
constraints. However, in commenting on a draft of the report, FDA
officials said that the agency always offers to provide training to
states, when requested.

Regarding the data deficiencies we reported in 2002, FDA implemented a
newly designed feed-ban database and data entry procedures in its Field
Accomplishment and Compliance Tracking System (FACTS) in April 2002.
According to our analysis, this new approach and data system are designed
to more reliably track feed-ban inspection results. As a result, FDA has a
better management tool for overseeing compliance with the feed-ban rule
and a data system that better conforms to standard database management
practices. Specifically, FDA's new approach makes the following
improvements:

o All firms have unique identifiers. Inspection records in FDA's data
system-including those that were previously missing unique identifiers-now
have them, according to our data reliability analysis. Before the new
approach, about 45 percent of FDA's feed inspection records lacked
information to identify individual firms. As a result, the earlier data
could not be used to reliably determine the number of firms inspected,
compliance trends over time, or the inspection history of an individual
firm. These problems should not occur with FDA's new system.

o Information is substantially complete and accurate. FDA has corrected
information problems we had identified in our 2002 report, according to
our data reliability analysis of the inspections conducted since April 15,
2002. The new FACTS database contains edit checks to detect any incomplete
or inaccurate data. Furthermore, FDA's current feed-ban inspection
guidance directs district BSE coordinators or their designees to review
BSE inspection forms for completeness and accuracy. Previously,
headquarters staff had entered the data received from district offices and
did not have sufficient knowledge to detect irregularities in the data
they were entering. In addition, states that have contracts or agreements
with FDA are now using the same inspection forms as FDA. Previously,
several states used state-developed forms, which did not always provide
comparable information.

o Data are more timely. Since April 15, 2002, about 95 percent of
inspections with serious violations have been entered into the FACTS
database within 45 days of the inspection date, according to our analysis.
This rate of entry is a significant improvement over the timeliness of
entry rates we reported in 2002. At that time, we found that some
inspections were entered into FDA's database 2 or more years after the
date of inspection. For such inspections, FDA could not accurately report
on firms' compliance with the feed ban and could not clarify inconsistent
or conflicting information, or obtain answers to missing
information-situations that FDA's new approach should help avoid.

As a result of these improvements, FDA is able to present more reliable
feed ban inspection information on its Web site for the approximately
10,000 firms inspected since April 15, 2002, or about two-thirds of the
approximately 14,800 firms inspected since 1997. Appendix II provides a
detailed description of actions FDA has taken on the recommendations in
our 2002 report.

Program Weaknesses Continue to Limit the Effectiveness of FDA's Animal
Feed Ban

While FDA has made many improvements to its oversight and enforcement of
the feed ban in response to our 2002 report recommendations, we found a
number of oversight weaknesses that limit the effectiveness of the ban and
could place U.S. cattle at risk for BSE. Specifically, we found that FDA
does not

o have a uniform procedure to identify all firms subject to the feed ban,

o require firms to notify FDA if they process with prohibited material,

o routinely use tests to verify compliance with the feed ban,

o alert USDA or states when cattle may have been fed with feed containing
prohibited material, and

o adequately overseeing the procedures for cleaning vehicles that haul
cattle feed.

Furthermore, we found that cautionary statements are not required on feed
or feed ingredients intended for export that contain prohibited materials.
In addition, FDA has not been reporting BSE inspection results to Congress
and the public in a full and complete context.

FDA Does Not Have Uniform Procedures to Identify Additional Firms Subject
to the Feed- Ban Rule

When the feed ban took effect in 1997, FDA first focused on identifying as
many firms as possible that were subject to the ban. As of September 30,
2004, FDA officials had identified approximately 14,800 firms that are
subject to the feed ban (see table 2). That is about 4,200 more firms than
the 10,576 firms FDA had identified approximately 3 years earlier. FDA
officials acknowledge that the agency has not identified all firms subject
to the feed-ban rule.

Table 2: Number of Firms FDA Has Identified that Are Subject to the Feed
Ban, by Firm Type, as of the End of Fiscal Year 2004

Firm type                                                  Number of firms 
Renderers                                                              249 
Protein blenders                                                       281 
FDA-licensed feed mills                                              1,061 
Nonlicensed feed mills                                               4,922 
Othersa                                                              8,252 
Total                                                               14,765 

Source: GAO analysis of FDA's online database.

Note: Since many firms engage in more than one activity (for example, a
feed mill may also be a protein blender), this analysis makes assumptions
about firms' primary activities to avoid counting firms in more than one
category.

aOther firm types include cattle feeders, transporters, pet food
manufacturers and retailers.

FDA has identified firms by reviewing

o its list of firms that manufacture feed that contains certain new animal
drugs; FDA knew about these firms because it requires them to be licensed
and because it has certain regulatory responsibility over these firms.

o a list of the firms or individuals that USDA has identified as violating
USDA's and FDA's requirements with respect to chemical and drug residues
in animals slaughtered for human consumption.18

o lists of firms that states identified. For example, 27 of the 38 states
we surveyed register renderers, 28 register protein blenders, and 34
register feed mills that FDA has not licensed, and provide this
information to FDA during meetings to set up annual inspection plans.

o membership lists of industry associations, such as the National
Renderers Association.

In addition, FDA officials told us that FDA districts have used multiple
approaches, including looking through telephone books to identify the
names of additional firms. However, FDA has not developed a systematic
approach for identifying additional firms subject to the feed ban. For
example, FDA does not have an approach for identifying additional
nonlicensed feed mills in states that do not provide that information. FDA
also acknowledged that it has identified only a small percentage of the
thousands of transportation firms that may haul cattle feed. Moreover, in
commenting on a draft of this report, FDA told us that there are an
estimated 1 million businesses (e.g., dairy farms feedlots, and other
facilities) that feed cattle and other animals. FDA also told us that it
does not consider farms that mix their own feed or feed cattle as well as
other animals as low risk. However, FDA does not have a strategy for
ensuring that this industry sector is in compliance with the feed-ban
rule.

We observed one approach for expanding the number of firms subject to the
feed ban: some FDA and state inspectors we accompanied on firm inspections
wrote down the names of the firm's suppliers and customers during the
inspection and checked these names against FDA's inventory of firms to
help identify additional firms. According to officials in one district
where we observed this practice, they inspect these additional firms as
resources allow. However, FDA does not have guidance for inspectors to do
this routinely, and we observed other inspectors who did not record the
names of firms' suppliers and customers. The approach we observed was one
that may be largely applied with existing resources. Congress provided FDA
with an additional $8.3 million in the fiscal year 2005 budget, which

FDA officials told us would be used, in part, to funds states' efforts to
identify and inspect additional firms.19

FDA Does Not Require Firms to Notify the Agency if They Process with
Prohibited Material

Under FDA's risk-based inspection system, FDA's goal is to annually
inspect all renderers, feed mills, and protein blenders that process with
prohibited material-about 570 firms-and to inspect a number of other firms
that FDA considers lower risk. The number of other firms varies according
to the inspection resources available. As previously stated, in total, FDA
and states inspected 6,006 firms in fiscal year 2004. However, once FDA
has inspected a firm and determined that it does not process with
prohibited materials, FDA may not reinspect that firm for many years. In
the interim, FDA does not know whether the firm has changed operations and
now processes prohibited materials because it does not require firms that
do so to notify the agency. FDA and state agencies only learn of a change
in operations if they inspect the firms. Without a requirement to notify
FDA, these firms are not annually inspected to monitor for compliance with
the feed ban, as are other high-risk firms.

We found that 2,833 or about 19 percent, of these firms FDA has identified
as subject to the feed-ban rule have not been reinspected in 5 or more
years. These firms include

o 1,224 farms that fed ruminant animals;

o 846 farms that mixed their own feed;

o 377 feed mills; and

o 386 other types of firms, such as distributors and retailers.

According to FDA officials, of these four types of firms that have not
been reinspected, about 2,100 or two-thirds are farms, which FDA believes
are not likely to change their practices. However, feed mills, which
account for about 400 of the firms, would be classified as high risk if
they process with prohibited material.

FDA officials also believe that the number of firms processing with
prohibited material is declining and that in all likelihood firms that
have not been inspected for a number of years would not change their
practices and start doing so. As FDA pointed out, firms may decrease their
use of prohibited material because of the requirement that they maintain
records sufficient to track all receipt, processing, and distribution of
that material. Nonetheless, some firms that did not use prohibited
material when they were last inspected may begin to use that material in
processing their feed.

FDA officials told us that they have considered options for identifying
firms that process feed with prohibited material, including requiring
those firms to be licensed. The officials noted, however, that some firms
may not comply with a notification requirement; thus, FDA would still not
know about all high-risk firms, and it would incur the additional costs of
overseeing the notification requirement.

FDA Does Not Routinely Sample Feed and Feed Ingredients during Inspections
for Analysis to Verify Compliance with the Feed Ban

While FDA inspection procedures include guidance for reviewing firm
documents and procedures, examining their invoices, and inspecting
facilities and equipment, they do not include guidance on when samples
should be taken and tested. For example, the feed-ban inspection guidance
does not instruct inspectors to routinely sample cattle feed to verify
firms' claims that they do not use prohibited materials or exempt
ingredients, or to ensure that firms' cleanout and flushing procedures to
prevent commingling are followed and are effective.

We recognize that the usefulness of testing is limited at firms that use
exempt items-cattle and other ruminant blood, milk proteins, poultry
litter, and plate waste-as ingredients in cattle feed. FDA officials told
us that they did not want to routinely test samples at firms during
inspections because the tests would likely have many false positives as a
result of the exemptions. Consequently, officials believed testing would
not use resources wisely.

However, in 9 of the 19 inspections we observed, inspectors could have
used tests to verify feed-ban compliance because the firms claimed they
did not use any animal-derived exempt items. Even in these instances,
where tests would be beneficial, inspectors did not sample the feed. For
instance, inspectors did not take samples to confirm the adequacy of
cleanout procedures at firms that use nondedicated production facilities
to manufacture cattle feed but do not use any exempt materials. FDA's
feed-ban inspection guidance allows inspectors to draw samples at their
discretion, but FDA officials told us that inspectors rely on their
judgment of whether the cleanout procedures appear to be adequate and
rarely use testing to verify their assessment. FDA officials did not give
us a clear reason why they would not advise testing in situations where
tests would be useful to help confirm compliance.

Some states have also done significant testing that FDA could use to
verify compliance with the feed ban but do not provide their test results
to FDA, although that information could give FDA a more complete picture
of feed ban compliance. In response to our survey, 18 of the 38 states
that have agreements with FDA to conduct feed-ban inspections told us they
had collected and tested over 1,500 feed samples during 2003.20 For
example, according to a North Carolina Department of Agriculture official,
the state collected and tested 738 samples; and, according to a Kansas
Department of Agriculture official, the state collected and tested 94
samples. In these states, if the tests find what appears to be prohibited
material, the states followed up with the firms to determine what
ingredients they used. According to the officials, no contaminated cattle
feed was found. In California, which collected and tested about 100
samples, officials found tests to be useful for demonstrating to cattle
feed manufacturers the difficulties of cleaning equipment that has been
used for prohibited material. FDA and state agency officials told us that
most California feed firms have switched to using dedicated equipment for
cattle feed. Eleven of the 18 states share test results with FDA, but FDA
does not use these results to verify industry compliance with the feed
ban.

In August 2003, FDA instructed its districts to begin testing finished
feed and feed ingredients, such as bags of feed sold at retail stores and
bulk feed sold to cattle feedlots. These tests were not taken in
conjunction with feed-ban compliance inspections. FDA inspectors took 660
samples nationwide. The samples were submitted to FDA regional
laboratories for analysis, where analysts used feed microscopy. Although
in its instructions to districts for the collection effort, FDA called the
tests "a method to monitor for compliance with" the feed ban, FDA
officials told us that the test results could not be the sole basis for
enforcement action at individual firms because microscopic analysis cannot
distinguish prohibited bone and tissue from exempted material.
Nonetheless, the officials also told us the testing gives FDA further
assurance of industry's compliance with feed ban.

Because FDA did not use an approach that allows it to generalize the
results, the test results cannot be used as assurance of industry
compliance. In fact, because FDA did not provide instructions on how to
randomly select firms for sampling and how to take a random sample of feed
at the firms, the results cannot even help confirm compliance by the
stores, feedlots, and other firms where the samples were taken. In
initiating this effort without a sampling plan, FDA wasted its already
limited inspection resources. FDA has committed resources to collect and
analyze 900 additional samples in fiscal year 2005. With the same
resources, FDA could have developed a sample design that would have
allowed it to generalize the test results to industry.

FDA officials told us the agency would have to conduct an investigation to
determine whether an enforcement action was warranted. FDA provided us
some information on test results for the 660 samples that were taken and
analyzed. The data showed 145 potential violations, including 8 that FDA's
laboratories originally classified as serious. About one-third of the 145
samples with potential violations were of cattle feed. Several of those
samples had evidence of mammalian matter. Without more information, we
could not determine whether the cattle feed contained exempt items or
prohibited material. As of February 2005, FDA was in the process of
gathering the information we requested from its district offices on the
results of its investigation of the 145 potential violations and what, if
any, enforcement actions were taken based on the tests and follow-up
investigations. We plan to provide our analysis of FDA's collection,
testing, and follow-up of these samples later this year.

The Cautionary Statement Is Not Required on Feed Intended for Export

Animal feed and feed ingredients containing prohibited material (including
material from rendered cattle) are not required to be labeled with the
cautionary statement, "Do not feed to cattle or other ruminants," when
that material is intended for export. Shipping containers for such
material, however, must be labeled that they are for export only; and, if
prohibited material is put back into domestic commerce, the containers
must be relabeled with the cautionary statement.

Not placing the warning label on exported feed poses a potential risk to
U.S. and foreign cattle and consumers from two perspectives. First, feed
with prohibited materials could be intentionally or inadvertently
redirected into feed for U.S. cattle if firms fail to add the cautionary
label to the product that they had initially intended to export. Second,
exported feed containing prohibited material could mistakenly be fed to
cattle that are subsequently imported into the United States or whose meat
and other products are imported into the United States.

We observed one situation where a problem could occur because a cautionary
statement was not on an exported product. One firm we visited processed
fishmeal, which is normally considered a safe ingredient for cattle feed.
However, this plant processed the fishmeal on the same equipment it used
for prohibited materials. If it were sold domestically, the fishmeal would
have to be labeled with the cautionary statement because it is potentially
contaminated with prohibited materials. However, the product was shipped
to overseas customers without the cautionary statement. Because the
fishmeal was not labeled, and fishmeal would not be expected to contain
prohibited material, customers could unwittingly mix the fishmeal with
other ingredients for their cattle. The FDA inspector did not document in
the inspection report which countries were sent the fishmeal. When we
asked FDA officials about this situation, they were concerned only about
whether feed intended for export was actually being diverted to domestic
cattle, a situation that they believed was unlikely to occur because FDA
rules prohibit it. However, according to the report by the international
panel of experts on BSE convened by USDA, the United States has an
obligation to act responsibly toward its global neighbors when exporting
feed and feed ingredients.

FDA officials told us that FDA cannot require the cautionary statement on
feed intended for export without a change to the Federal Food, Drug, and
Cosmetic Act.21 Under that act, animal feed intended for export only
cannot be deemed to be adulterated or misbranded if it (1) meets the
foreign purchasers specifications, (2) is not in conflict with laws of the
country to which it is intended for export, (3) is labeled on the outside
of the shipping package that it is intended for export, and (4) is not
sold or offered for sale in domestic commerce.

FDA Did Not Alert USDA or State Regulatory Authorities When It Learned
That Cattle Feed Containing Prohibited Material Was Marketed

When an FDA district office learns that ruminant animals may have been fed
contaminated feed, the feed-ban inspection guidance directs the district
office to oversee efforts to appropriately dispose of the contaminated
feed and to ensure that the animals that had consumed this feed are not
slaughtered for human food or other animal feed. The guidance also advises
FDA to consider coordination with USDA and the affected states.

While FDA districts have monitored voluntary recalls of feed that did not
comply with the feed ban, they had not been alerting USDA or state
departments of agriculture when they learned that such feed had been given
to cattle and other ruminants-in some cases for an extensive period of
time. FDA district and headquarters officials responsible for the feed-ban
program were not aware that the guidance instructed FDA to alert USDA and
states.

In our observations at inspections and our review of inspection records,
we found the following instances in which FDA did not alert USDA or state
authorities or take further action.22

o A producer of cattle, hogs, and goats had inadvertently fed salvaged pet
food containing prohibited materials to goats, which are ruminants.23 We
observed the mislabeled feed in a March 2004 inspection. The feed mill
that manufactured and sold the feed had not labeled the salvaged pet food
with the required cautionary statement "Do not feed to cattle or other
ruminants." Shortly after this discovery, the firm recalled the misbranded
feed. In April 2004, a state feed inspector found out about the misfed
animals from the feed mill, not from FDA, and alerted his state program
managers. The state contacted FDA, and after determining that FDA did not
intend to take action beyond issuing a warning letter, the state seized
and destroyed the animals in May 2004 under state authority to prevent the
meat from entering the food supply. FDA did not alert the state or USDA
and did not issue the warning letter to the feed mill until June 2004.

o A feed mill had inadvertently contaminated cattle feed with prohibited
material. The firm had made a mistake in designing and placing equipment
in the manufacturing process, which allowed spilled feed containing
prohibited material to become commingled with ingredients used to make
cattle feed. We observed this problem during an April 2004 inspection. FDA
issued a warning letter in June 2004 demanding that the firm correct the
violations; the firm also conducted a voluntary recall of the feed in
June. Because the mill operated with this flawed system for about 1 year
before the discovery, potentially contaminated feed was marketed and sold
for cattle feed for that period of time. FDA did not contact USDA or state
authorities to alert them that cattle had consumed the feed.

o A feed mill did not clean mixing equipment and transportation vehicles
used for processing and transporting feed containing prohibited and
nonprohibited materials. The firm also failed to properly label feed
containing prohibited materials with the required cautionary statement and
did not maintain sufficient records for tracking the sale of cattle feed
to its customers, as FDA requires. We identified these problems during our
review of inspection reports. The inspection occurred in March 2003. The
firm corrected the violations and recalled all cattle feed that had not
yet been consumed in March 2003. FDA issued a warning letter to the firm
in May 2003 and took no further action.

When we discussed these findings with FDA headquarters officials, they
told us they were not familiar with the guidance recommending this
communication. As a result, FDA, USDA, and state authorities had not
assessed the health risk to humans and the animals that may have ingested
that feed and may not have taken sufficient action to prevent those cattle
and other ruminants from entering the human food or animal feed supply.
The FDA officials said they had not considered coordinating with USDA and
state officials but that USDA and the states were notified of the recalls
because the recalls are posted on the FDA Web site. However, we found that
the posted recall notices do not include information on whether, or for
how long, cattle or other ruminants had been given the contaminated feed.
Furthermore, FDA officials asserted that no action was needed beyond a
recall in these incidents because BSE has not been discovered in a cow
born in the United States. According to the officials, the meat would not
make people ill and the feed would not make cattle ill. Before this report
was issued, these same FDA officials told us that in the future, FDA will
alert USDA and states when cattle may have consumed prohibited feed. USDA
officials told us that they were not aware of these three incidents. They
said that, had they known, USDA would have tracked the animals and tested
them for BSE when they were slaughtered.

FDA Has Limited Assurance That Vehicles Used to Haul Cattle Feed Comply
with the Feed Ban

According to FDA's feed-ban rule, transportation firms that haul
prohibited material and use the vehicles to haul feed or feed ingredients
for cattle must have and use procedures to prevent commingling or
cross-contamination. The procedures must provide for cleaning out the
vehicles or other adequate preventative measures. Research suggests that
cattle can get BSE from ingesting even a small amount of infected
material-an amount that could be introduced in feed that was transported
in a poorly cleaned vehicle. As part of an inspection of transportation
firms, inspectors review the adequacy of these procedures, but the
inspection form does not prompt them to do so during inspections of other
types of firms. The following two problems impede the effectiveness of
FDA's current procedures:

o FDA has not identified and does not inspect many transportation firms.
According to FDA officials and transportation data, thousands of
independent truckers, large and small trucking companies, and rail
companies may carry cattle feed and feed ingredients. FDA officials told
us that it would be virtually impossible to identify and inspect all of
these firms, given its limited resources. However, FDA agrees that
transportation compliance is important. In commenting on a draft of this
report, the agency noted that it is planning to increase oversight of
transportation firms based on FDA's assessment of compliance and risk in
this industry sector.

o Inspecting transportation firms at their home base would not ensure that
the required procedures are being used and that the nearly 200,000 large
trucks that haul animal feed would be clean at the time they picked up
cattle feed, in part, because vehicles that carry prohibited material may
also carry cattle feed and other loads in succession before returning to
their home base. For example, at an inspection of one high-risk protein
blender, we observed an FDA inspector talking with an independent trucker
who had dropped off a load of cattle feed ingredients, was picking up
prohibited materials at the protein blender, and was scheduled later to
pick up a load of corn, which could be used in cattle feed. The trucker
explained that if he saw anything in the truck between loads, he would
climb in and sweep the material out with a broom; if he did not see
anything, he did not sweep out the truck between loads. The trucker also
said it would be extremely difficult to find washout facilities to clean
the truck between loads while on the road.

Consequently, we believe that it would be more effective to require FDA
and state inspectors to review and document procedures that feed mills and
other firms use to ensure that the vehicles they use to haul cattle feed
and feed ingredients are free of prohibited material as part of their
inspections at feed mills and other firms. During our observations of
inspections, we found that some FDA and state inspectors were already
doing so. However, our observations and analysis of inspection reports
showed that the inspectors did not routinely do so and did not uniformly
report on the adequacy of the firms' procedures for preventing the
introduction of prohibited material. We believe that inspectors were
overlooking the adequacy of firm's procedures to ensure the safe transport
of cattle feed because the BSE inspection form does not have any questions
to capture that information. Specifically, 82 of the 404 inspection
reports we reviewed were for renderers, protein blenders, feed mills, and
other firms that processed with prohibited material and handled cattle
feed and feed ingredients. We found that inspectors had documented the
required cleanout procedures for transportation equipment at only 11 of
these 82 firms. Without requiring inspectors to uniformly review and
document vehicle cleaning procedures, FDA has insufficient assurance that
the vehicles are safe to carry cattle feed and feed ingredients.24

FDA Does Not Fully Report BSE Inspection Results

In January 2004, FDA's Deputy Commissioner testified that inspectors "at
least annually, targeted BSE inspections of 100 percent of known
renderers, protein blenders, and feed mills processing" with prohibited
material. He testified that compliance by those firms was "estimated to be
better than 99 percent." Subsequently, some industry officials claimed
that overall compliance with the feed ban is nearly 100 percent and used
that figure to support their claim that the feed ban does not need to be
strengthened. However, as noted earlier, those groups are comprised of
about 570 firms-approximately 4 percent of the firms in FDA's inventory.

In addition, FDA periodically publishes compliance information on its Web
site for all industry segments. This information has also been used to
cite high industry compliance. However, FDA and industry do not have a
basis for citing a compliance rate for a segment of firms subject to the
feed ban or industrywide because there are too many unknowns.
Specifically, FDA does not know the status of compliance for firms that

o have never been inspected,

o have not been reinspected in 5 or more years, and

o may have started to process with prohibited materials since their last
inspection.

Furthermore, as we previously discussed, because FDA does not routinely
sample feed to confirm compliance, inspection results are largely based on
a review of paper documents and a visual inspection. All these concerns
apply to compliance information FDA reports to Congress and the public on
its Web site.

Additionally, our analysis of inspection reports also disclosed that FDA
was not including all serious violations in its calculation of the
compliance rate because it reclassified firms as "in compliance" once they
correct violations, regardless of how long the problem may have existed.

Finally, we found that FDA has classified 42 firms as having less serious
violations that it counted as "in compliance" with the feed ban.
Inspectors reported that 18 of these firms failed to include a cautionary
statement on feed containing prohibited materials. Although FDA's feed-ban
inspection guidance designates the lack of a cautionary statement as a
serious violation, and lack of such a statement should result in the feed
being deemed misbranded under the Federal Food, Drug, and Cosmetic Act,
FDA excluded the violations at these firms from its calculation of the
compliance rate. Inspectors also reported that the remaining 24 firms had
procedures for preventing commingling but did not have these procedures in
writing. FDA's guidance designates the lack of written procedures as a
less serious violation, but we believe these violations should be
classified as serious. Without written procedures, FDA has no assurance
that the firms consistently take the necessary steps to prevent
commingling. FDA officials told us that the guidance is advisory and
therefore gives the agency the discretion to reclassify the violations
based on its review.

Conclusions

Diligent FDA oversight and enforcement of the feed ban is essential, not
only because of the potential threat to public health but also because of
the economic impact on the cattle and beef industry; this impact was
clearly demonstrated by the sharp drop in U.S. beef exports after one
infected cow was discovered in 2003. The ongoing discussions and
agreements to reopen beef export markets could be derailed if more cattle
were discovered with BSE.

FDA has taken positive steps since our 2002 report. Today FDA can say with
greater confidence that it has more timely and reliable inspection data.
Also, the risk-based system FDA has adopted to target inspection resources
on high-risk firms will increase the likelihood that firms inspected
annually will remain in compliance with the feed ban.

FDA's processes, however, still have considerable room for improvement.
FDA does not have uniform procedures for identifying additional firms that
are subject to the ban but have never been inspected or for learning about
firms that change their practices and begin to handle prohibited material.
Furthermore, because inspectors are not using tests optimally-to help
confirm, when appropriate, that cattle feed, production equipment, and
transportation vehicles are free of prohibited material-FDA is limiting
its ability to assure that firms are in compliance with the feed ban and
that cattle feed is safe. Additionally, FDA is not taking advantage of
state test results to provide greater assurance that industry is adhering
to the feed ban and is not using its own program for sampling finished
feed and feed ingredients in a manner that will allow it to project test
results.

Moreover, the lack of a requirement for warning labels on feed and feed
ingredients intended for export that contain prohibited material, creates
opportunities for having the material fed to domestic or foreign cattle,
either intentionally or inadvertently. As the international group of BSE
experts convened by USDA pointed out, the United States has an obligation
to act responsibly toward its global neighbors when exporting feed and
feed ingredients.

Especially troubling was our discovery that FDA did not alert USDA and
state authorities when it became aware that cattle had been given feed
that contained prohibited material. FDA, and its key partner, USDA,
together provide critical firewalls that the federal government has in
place to protect U.S. cattle and consumers. In addition, the lack of
notification was contrary to FDA's own guidance and FDA's inaction
prevented USDA and states from being able to make an informed decision on
how to respond to the discovery that cattle had consumed prohibited
material.

Given these weaknesses and the fact that FDA does not include all
violations in its estimates, we believe FDA is overstating industry's
compliance with the animal feed ban and understating the potential risk of
BSE for U.S. cattle in its reports to Congress and the American people.
Despite the problems in FDA's calculation, some in the feed industry claim
that overall compliance with the feed ban is nearly 100 percent-a claim
that FDA's compliance information does not support.

Recommendations for Executive Action

To further strengthen oversight and enforcement of the animal feed ban and
better protect U.S. cattle and American consumers, we recommend that the
Commissioner of FDA take the following nine actions:

o Develop uniform procedures for identifying additional firms subject to
the feed ban.

o Require firms that process with prohibited material to notify FDA. If
FDA believes it does not have the necessary statutory authority, it should
seek that authority from Congress.

o Develop guidance for inspectors to systematically use tests to verify
the safety of cattle feed and to confirm the adequacy of firms' procedures
for ridding equipment and vehicles of prohibited material before they are
used for processing or transporting cattle feed or feed ingredients.

o Collect feed test results from states that sample feed to help verify
compliance with the feed ban.

o Develop a sample design for FDA's inspectors to use for sampling
finished feed and feed ingredients that will allow FDA to more accurately
generalize about compliance with the feed ban from the test results.

o Seek authority from Congress to require the cautionary statement on feed
and feed ingredients that are intended for export and that contain
prohibited material.

o Ensure that USDA and states are alerted when inspectors discover that
feed or feed ingredients with prohibited material may have been fed to
cattle.

o Modify the BSE inspection form to include questions inspectors can use
to document whether firms that process or handle cattle feed or feed
ingredients have procedures to ensure the cleanliness of vehicles they use
to transport cattle feed and feed ingredients.

o Ensure that inspection results are reported in a complete and accurate
context.

Agency Comments and Our Evaluation

We provided FDA with a draft of this report for review and comment. FDA
stated that our report was thorough and that it recognized the
enhancements FDA has put in place in its feed-ban program. However, FDA
said the report did not identify material weaknesses to support our
position that oversight weaknesses limit FDA's program effectiveness and
place U.S. cattle at risk of spreading BSE. FDA believes that its current
risk-based inspection approach is adequate to protect U.S. cattle.
According to FDA, given the wide variety of firms subject to the feed ban
and its resource limitations, it "is obligated to set priorities for
inspecting a meaningful subpopulation of these regulated firms." We
recognize that FDA has made many improvements, including adopting a
risk-based approach for inspections, that have substantially improved its
oversight of the feed-ban rule. However, our report identifies significant
problems in FDA's oversight that continue to place cattle at risk for BSE.
The importance of a strictly enforced feed ban is heightened now that BSE
has been found in North American cattle. As Harvard and the international
panel of experts pointed out, the feed ban is the most important fire wall
against the spread of BSE. Given the problems we identified and the
significance of a well enforced feed ban, it is important that FDA
improves its feed ban oversight and optimizes its use of resources.

In addition, FDA does not agree with our criticism of its compliance
reporting. FDA believes that it provides the inspection results in a
transparent, complete, and accurate context. FDA notes that the BSE
inspection data posted on its Web site "allows the user to analyze the
data, in a multitude of ways, to provide their own contextual reference."
Our concern is precisely that the data are being analyzed and interpreted
in an erroneous context. Specifically, when FDA and industry used those
data to assert a 99 percent compliance rate with the feed ban, they took
that information out of context. While FDA's calculation of compliance by
a subset of regulated industries may in fact be quite high, FDA's data are
not sufficient to make that projection to all regulated industries. In
addition, FDA does not know the status of compliance for firms that have
never been inspected or have not been reinspected in years. Nor does it
know if previously inspected firms have started using prohibited material.
Furthermore, because FDA reclassifies firms from "out-of-compliance" to
"in-compliance" on its Web site when the firms correct violations, the
information posted on that Web site does not tell the user when serious
and/or long-standing violations have occurred. Lastly, inspection results
are largely based on a review of paper documents and a visual inspection,
with little or no feed testing. Given these data concerns and compliance
unknowns, FDA's data should not be used to project industry compliance;
and, anytime those data are cited, they should be reported in a complete
and accurate context.

Regarding the nine recommendations we make in the report, FDA did not take
issue with the need for five and generally disagreed with four. Although
FDA noted implementation concerns, it did not take issue on the need for
(1) developing uniform procedures for identifying firms subject to the
feed ban, (2) collecting test results from the states that sample feed,
(3) including a cautionary statement on feed and feed ingredients intended
for export, (4) notifying USDA and states when feed or feed ingredients
containing prohibited material may have been fed to cattle, and (5)
modifying the inspection form to include questions to better oversee the
cleanliness of vehicles used to transport cattle feed or feed ingredients.

FDA disagreed with our recommendation that it require firms that process
with prohibited material to notify the agency. FDA believes that it is
already getting information on changes to firms' practices from states and
that requiring an additional notification process would be costly to
implement. However, FDA acknowledged that it has generally not identified
high-risk feed salvagers and farms that mix their own feed or those that
feed cattle as well as other animals. The cost of the notification program
will depend on the requirements FDA puts in place. In developing the
program, FDA could target the notification to firms that pose a
potentially high risk for exposing cattle feed to prohibited material. We
believe that FDA should know which firms are high risk and that industry
self-reporting is a mechanism that would help the agency identify those
firms and help it ensure compliance with the feed ban.

FDA also disagreed with our recommendation to systematically use tests in
conjunction with compliance inspections. While we recognize the
limitations of current test methodologies, we believe that tests are
useful. In fact, states and FDA are currently using these tests on feed.
Our recommendation speaks to systematically using these tests where
appropriate, to augment inspections, which are largely observation and
paperwork reviews. We expanded the recommendation to recognize that FDA
may validate other tests in the future.

With respect to our recommendation that FDA develop a sample design for
testing finished feed and feed ingredients, FDA disagreed with the need
for a sample design that will allow it to more accurately generalize about
compliance. FDA stated that tests alone cannot serve as a basis to
generalize compliance. We agree that tests that indicate potential
violations need to be confirmed, because of the limitations of the current
tests. However, FDA is using the test results to identify potential
problems, and it tested 660 samples in 2003/2004 and plans to test 900
samples this year. The point of our recommendation is that any testing
activity of this magnitude should have a sampling plan.

Finally, FDA believes that it already reports inspection results in a
complete and accurate context, as we recommend. We disagree. As noted
above, given the data concerns and compliance unknowns raised in this
report, FDA's data should not be used to project industry compliance.
Anytime those data are cited, they should be reported in a complete and
accurate context. FDA also provided technical comments, which we have
incorporated into this report, as appropriate. FDA's written comments and
our responses are in appendix VI.

We also provided USDA with a draft of appendix III, which summarizes FDA's
and USDA's actions in response to the 2003 discovery of BSE in North
America, for review and comment. USDA had no comments on the draft
appendix.

As agreed with your offices, unless you publicly announce the contents of
this report earlier, we plan no further distribution of it until 30 days
from the date of this letter. We will then send copies to interested
congressional committees; the Secretary of Health and Human Services; the
Secretary of Agriculture; the Director, Office of Management and Budget;
and other interested parties. We will make copies available to others on
request. In addition, the report will be available at no charge on GAO's
Web site at http://www.gao.gov.

If you or your staffs have any questions about this report, please call me
at (202) 512-3841. Key contributors to this report are listed in appendix
VII.

Robert A. Robinson Managing Director, Natural Resources and Environment

Scope and MethodologyAppendix I

As discussed below, to assess the effectiveness of the Food and Drug
Administration's (FDA) actions to ensure industry compliance with the feed
ban and protect U.S. cattle from bovine spongiform encephalopathy (BSE),
we (1) analyzed 404 inspection reports for BSE inspections performed
during fiscal year 2003 and 2004; (2) observed 19 inspections in 12 states
that were conducted by either FDA or state inspectors; (3) assessed the
reliability of FDA's feed-ban inspection database; (4) interviewed
officials at FDA headquarters and district offices, state agencies, and
industry associations, as well as reviewed documents provided by these
officials concerning oversight of the animal feed ban; and (5) surveyed
state agency officials in 38 states.

To assess FDA's oversight, we analyzed BSE inspection records to identify
types of firms inspected; types of material processed (prohibited,
nonprohibited, or both); oversight of transportation equipment; violations
identified during inspections (if applicable); and final inspection
classifications. We randomly selected 413 inspection reports from the
universe of BSE feed inspections conducted during fiscal year 2003 and
fiscal year 2004 (up to February 7, 2004). For each of the 18 FDA
districts, responsible for inspections in the 50 states, we randomly
selected inspection reports from one state (most FDA district offices
cover more than one state). We included all of the 314 high-risk firms
that process prohibited materials for the 18 selected states. In addition,
we randomly selected 12 other firms that process with prohibited
materials; 68 firms that distribute prohibited materials; and 19 firms
that do not process or distribute prohibited materials. We examined only
404 of the 413 inspection reports because 9 of the report files that we
requested were still open-case files at the time of our review.

To evaluate the inspection process, we accompanied inspectors on 19 BSE
inspections of firms in 12 states covered by the feed ban. The sites were
selected to cover a range of firm types and sizes in various geographic
locations with concentrations of cattle feeding operations, including
dairy cattle. The 19 inspections included renderers, protein blenders,
feed mills, farms with ruminants and other animals, and pet food
manufacturers. Seven of these firms processed or handled only prohibited
material, and the remaining 12 processed or handled both types of
material. On 12 of the inspections, we accompanied FDA inspectors, and on
7 we accompanied state inspectors.

To assess the reliability of the data FDA uses when reporting industry
compliance, we analyzed the agency's database for inspections conducted on
or after April 15, 2002, when FDA implemented its newly designed feed-ban
database.1 Specifically, we analyzed the 9,230 inspection records in this
database, as of February 7, 2004. To complete the reliability assessment,
we (1) reviewed existing documentation related to the data sources; (2)
electronically tested the data to identify obvious problems with
completeness, accuracy, or timeliness of data entry; and (3) interviewed
knowledgeable agency officials about the data. We determined that the data
were sufficiently reliable for purposes of this report.

We interviewed officials or reviewed documents at FDA headquarters and at
the 18 FDA district offices that are responsible for overseeing and
enforcing the feed ban in the 50 states, maintaining the inspection
database system, and proposing and analyzing regulatory decisions. In the
18 district offices, we used a structured interview to uniformly gather
information on various issues, such as methods used to identify the
universe of firms subject to the feed ban; the process for selecting firms
for inspection; training programs for FDA and state inspectors; feed-ban
inspection guidance and procedures; the processes for reviewing inspection
results, classifying findings, and determining what, if any, enforcement
action should be taken; and oversight of contracts and agreements with
state agencies that perform BSE inspections. We received information and
documentation on FDA's oversight and enforcement of the feed ban from the
following specific FDA units: Center for Veterinary Medicine, Office of
Management, Office of Surveillance and Compliance; Office of Regulatory
Affair's Office of Regional Operations; Center for Food Safety and Applied
Nutrition's Office of the Director; and Office of the Chief Counsel. We
reviewed various FDA program documents, including the BSE/Ruminant Feed
Ban Inspections Compliance Program Guidance; BSE feed inspection form;
advance notices of proposed rulemakings to strengthen the feed ban,
including public comments; and the reports on the feed samples collected
and tested. We also interviewed state agency officials and reviewed
documents from the California Department of Food and Agriculture; the
Departments of Agriculture of Georgia, Illinois, Kansas, Missouri, North
Carolina, and Pennsylvania; and the Texas Feed and Fertilizer Control
Service. Lastly, we interviewed officials and reviewed documents from the
American Feed Industry Association, the Association of American Feed
Control Officials, the National Renderers Association, the Association of
Analytical Communities, and the Harvard Center for Risk Analysis.

To understand the role that states play in the feed inspection program, we
surveyed state officials in the 38 states that have contracts or other
agreements with FDA to perform feed-ban compliance inspections and report
the inspection results to FDA.2 The survey included questions about the
states' inspection programs, testing of animal feed ingredients, and FDA's
training and guidance for feed-ban inspections and enforcement. Before
implementing our survey, we pretested the questionnaire with state
agriculture officials in five states. During these pretests, we
interviewed the respondents to ensure that (1) questions were clear and
unambiguous, (2) terms were precise, and (3) the survey did not place an
undue burden on the staff completing it. We received completed
questionnaires from all 38 states surveyed. The state information
presented in this report is based on information obtained from this survey
and interviews with state officials.

We performed our work from October 2003 through January 2005, in
accordance with generally accepted government auditing standards, which
included an assessment of data reliability and internal controls.

GAO's Analysis of the Status of Actions on Recommendations to FDA in Our
January 2002 ReportAppendix II

Source: GAO analysis of FDA documents.

GAO's Summary of FDA's and USDA's Actions in Response to the Two Cases of
BSE Discovered in North America in 2003 Appendix III

Source: GAO analysis of FDA and USDA documents.

Survey of State AgenciesAppendix IV

Chronology of FDA's Feed Ban and Proposed RulemakingsAppendix V

August 4, 1997

FDA feed ban took effect, prohibiting certain materials in ruminant feed
to prevent the establishment and spread of BSE if it were to appear in
U.S. cattle herds. FDA took this action because it had been an industry
practice to feed proteins to ruminant animals that could transmit the
infective agent that causes BSE. Additionally, research in the United
Kingdom suggested that variant Creutzfeldt-Jacob Disease (vCJD) in humans
is linked to eating cattle infected with BSE. The feed ban requires that
firms, with some exceptions, take the following actions:

o label feed and feed ingredients that contain most proteins from mammals
(prohibited material) with a cautionary statement "Do not feed to cattle
or other ruminants,"

o have procedures to protect against commingling or cross-contamination if
they handle both prohibited and nonprohibited feed and feed ingredients by
using either equipment dedicated exclusively to feed or ingredients
intended for cattle or using cleanout procedures or other adequate means
to prevent carryover, and

o maintain records so that feed and feed ingredients that contain or may
contain prohibited material can be tracked from receipt through
disposition.

According to FDA's rules, firms that transport both types of materials
must also follow these procedures. Additionally, prohibited materials may
be used in pet food and in feed for poultry, swine, horses, and other
nonruminant animals. Lastly, FDA designated a number of cattle- and other
animal-derived items as exempt from the ban-and hence, allowable in cattle
feed. These items include blood and blood products, plate waste, gelatin,
milk and milk protein, and any product whose only mammalian protein
consists entirely of protein from pigs and horses. FDA has also not
regulated the use of poultry litter in feed.

October 30, 2001

FDA held a public hearing to solicit information and views regarding ways
in which the current feed ban and its enforcement might be improved or to
determine if any new objectives should be considered. FDA took this action
because BSE had spread beyond the United Kingdom to most countries in
western and central Europe and Japan. FDA asked for responses to 17
questions, including the following:

o Should FDA require dedicated facilities for the production of animal
feed containing mammalian protein?

o Should FDA require dedicated transportation of animal feed containing
mammalian protein?

o Should FDA license renderers and other firms engaged in the production
of animal feed containing mammalian proteins?

o Should FDA revoke or change any of the current exemptions in the current
rule?

o Should FDA require pet food to contain the cautionary statement?

o Should FDA extend the recordkeeping requirement beyond 1 year?

o Should FDA request authority to assess civil monetary penalties?

November 6, 2002

FDA published an advanced notice of proposed rulemaking announcing that it
was considering revising the feed ban and asking the public to comment on
certain possible modifications. FDA explained that shortly after its
October 2001 public hearing, USDA released a report by the Harvard Center
for Risk Analysis on the findings of a major, 3-year initiative to develop
a risk assessment model and assess the risk of BSE in the United States.
The model concluded that the risk to U.S. cattle and to consumers from BSE
is very low, but certain new control measures could reduce that small risk
even further. Therefore, based on comments received at the public hearing
and the findings of the Harvard Study, FDA asked for public comment on
various ways that the BSE feed ban could be strengthened, including the
following questions:

o Should tissues that are known to be at higher risk for harboring the
infective agent for BSE, such as brain and spinal cord from ruminants 2
years of age or older be excluded from all rendered products?

o How extensive is the use of poultry litter in cattle feed, what is the
level of feed spillage in poultry litter, and what would be the impacts
resulting from banning poultry litter in ruminant feed?

o Should pet food for retail sale carry the cautionary statement "Do not
feed to cattle or other ruminants?"

o Are there practical ways, other than dedicated facilities, for firms to
demonstrate that the level of carryover of prohibited material in a feed
mill could not transmit BSE to cattle or other ruminants? If so, what is
the safe level of carryover of prohibited material and what is the
scientific rationale for establishing this safe level?

o To what extent is plate waste used in ruminant feed and what would be
the impacts from excluding this material from ruminant feed?

January 26, 2004

FDA announced that it would be issuing interim final rules to strengthen
existing BSE firewalls, including banning a wide range of cattle material
from human food, dietary supplements, and cosmetics, and strengthening the
1997 feed ban through an extended list of banned feeding and manufacturing
practices.

July 14, 2004

FDA, with USDA, announced that the agencies are considering additional
measures to protect the public from the health risk associated with BSE
and to prevent the spread of the disease in U.S. cattle and are asking for
public comment. The agencies are considering additional safeguards based
on the recommendations of a panel of international experts convened by the
Secretary of Agriculture to review the U.S. regulatory response following
the finding of a BSE-positive cow in Washington State in December 2003. In
addition to some of the measures FDA had planned to take in an interim
final rule, the international panel recommended broader measures, such as
banning all mammalian and poultry protein from ruminant feed. Since these
recommendations would require significant changes in current feed
manufacturing practices and could make some previously announced proposals
unnecessary, FDA requested additional information and public comment on
the panel recommendations and other measures, including the following:

o What information is available to support or refute the assertion that
removing tissues that are known to be at higher risk for harboring the BSE
infective agent, such as brain and spinal cord tissue, from all animal
feed is necessary to effectively reduce the risks of cross-contamination
of ruminant feed or of misfeeding on the farm?

o If FDA prohibits high-risk tissues from all animal feed, would there be
a need to require dedicated facilities, equipment, storage, and
transportation?

o What information is available to support banning all mammalian and
poultry meat and bone meal from ruminant feed?

o If FDA prohibits high-risk tissues from all animal feed, what
information is available to support banning all mammalian and poultry meat
and bone meal from ruminant feed?

o Can high-risk tissues be effectively removed from dead stock and
nonambulatory cattle so that the remaining material can be used in animal
feed, or is it necessary to prohibit the entire carcass from use in all
animal feed?

o Do FDA's existing authorities under the Federal Food, Drug, and Cosmetic
Act and under the Public Health Service Act provide a legal basis to ban
the use of high-risk cattle tissues and other cattle material in
nonruminant animal feed, given that such materials have not been shown to
pose a direct risk to these animals?

FDA also issued an interim final rule on July 14, 2004, to prohibit
certain cattle materials in FDA-regulated food, including dietary
supplements, and cosmetics, to minimize potential human exposure to the
BSE infective agent. Specifically, FDA prohibited use of the brain, skull,
spinal cord, and other specified tissues of cattle that are 30 months or
older; small intestine and tonsils of all cattle; material from
nonambulatory disabled cattle or cattle not inspected and passed for human
consumption; and beef that is mechanically separated from bones. FDA took
this action in response to the finding of a BSE-positive cow in Washington
State in December 2003 and to conform with an interim final rule issued by
USDA in January 2004 declaring these materials unfit for human
consumption.

Comments from the Food and Drug AdministrationAppendix VI

The following are GAO's comments on the Food and Drug Administration's
letter dated January 13, 2005.

GAO Comments

1.We believe the report identifies numerous oversight weaknesses that
continue to limit program effectiveness and place cattle at risk. The
purpose of the feed ban firewall is to prevent the exposure and spread of
BSE. A well enforced feed ban is even more critical now that BSE has been
discovered in cattle in North America. As shown in our report, FDA does
not know the compliance status or risks posed by firms it has not
identified, inspected or reinspected for many years. FDA acknowledged that
many more firms are subject to the feed ban than have been inspected to
date but said the agency must set priorities for the number and types of
firms it can identify and inspect with limited inspection resources. We
agree with FDA's use of a risk-based inspection approach; however, FDA
acknowledges the need to increase inspections of certain industry
segments, such as transporters and animal feed salvagers. Moreover, for
firms that FDA inspects, it does not routinely sample feed to verify
whether the operating procedures observed by its inspectors are actually
preventing prohibited materials from contaminating cattle feed. Our
recommendations are aimed at ensuring that FDA has a strategy for
maximizing the effectiveness of its limited inspection resources,
targeting inspections, and using feed tests to minimize the risk of cattle
being fed prohibited material.

2.Our concern is precisely that the data are being analyzed and
interpreted in an erroneous context. Specifically, when FDA and industry
used those data to assert a 99 percent compliance rate with the feed ban,
they took that information out of context. While industry compliance may
in fact be quite high for firms FDA has inspected recently, FDA's data are
not sufficient to project compliance industrywide. FDA does not know the
status of compliance for firms that have never been inspected or have not
been reinspected in years. In addition, compliance history is lost-firms
that had serious and long-standing violations are classified as
"in-compliance" once FDA determines that the problems are corrected. FDA
is not reporting that the firms were ever out of compliance or the length
of time that the feed ban was violated. Lastly, inspection results are
largely based on a review of paper documents and a visual inspection, with
little or no feed testing. Given these data concerns and compliance
unknowns, we believe that FDA's data should not be used to project
industry compliance and, anytime those data are cited, they should be
reported in a complete and accurate context.

3.FDA agrees that there are industry sectors (such as transporter and
animal feed salvagers) that need to be assessed to determine their
potential risk to U.S. cattle. In fact, FDA acknowledges that there are
millions of firms potentially subject to the feed-ban rule. At the same
time, FDA implies that it has identified all high-risk firms. FDA has no
basis for that assertion. The example we suggest in this report is one way
of identifying additional firms that we observed during our review. FDA
identified other approaches that its districts used to identify other
firms. We believe that any approaches FDA identifies as useful should be
applied uniformly across all FDA districts. We included information in the
report on how FDA plans to use the $8.3 million it received in the 2005
budget. We also revised the report to include FDA's estimate of the number
of firms that feed cattle and other ruminants and revised the
recommendation in recognition that it may be impossible for FDA to
identify all firms subject to the feed-ban rule.

4.FDA suggests that requiring notification would take significant
resources. The cost of the notification program will depend on the
requirements FDA puts in place. In developing the program, FDA could
target the notification to firms that pose potentially high-risk for
exposing cattle feed to prohibited material. According to FDA, of the
14,800 firms it has inspected, about 570 renderers, protein blenders, and
feed mills comprise the high-risk firms subject to notification because
they manufacture or process prohibited material. While we believe there
may be more firms that fall into this group, it should not be a
significantly larger number. If it is significantly larger, that is
something FDA needs to know. Furthermore, requiring industry to
self-report is another mechanism that would help FDA identify firms and
oversee compliance. Finally, FDA has registration requirements in place
for medicated feed firms and for food facilities, and could draw on its
experience with those programs for developing a notification program for
firms subject to the feed-ban rule.1 Because firms can change their
practices over time, we believe it is important that firms notify FDA
whenever such changes occur.

5.While we agree that the current test methods have certain limitations,
we believe that testing can be a valuable tool for helping FDA oversee
compliance with the feed ban. FDA maintains that, because the current test
methods cannot differentiate prohibited material from exempt material,
they cannot be used to verify the presence or absence of prohibited
material or to confirm the adequacy of cleanout measures. However, states
told us that they are using tests for these purposes. Moreover, FDA is
currently testing finished feed and using the test results, together with
follow-up inspections, to determine whether the feed ban had been
violated. We believe tests would help inspectors who now rely on only
paperwork review and visual examination to determine the adequacy of
cleanout procedures. Tests would also be useful for vegetable-based cattle
feed, where detecting the presence of animal protein would indicate a
violation. We revised the recommendation to recognize that FDA may elect
to use other test methods in addition to feed microscopy and polymerase
chain reaction (PCR). With respect to FDA's sampling of finished feed, the
660 samples FDA tested were not collected during feed-ban compliance
inspections. We plan to report later this year on FDA's sampling of
finished feed.

6.We agree that FDA's current test methodology will not allow it to use
test results alone to verify feed-ban violations. However, testing
combined with follow-up inspections would allow FDA to be in a better
position to generalize about compliance with the feed-ban rule if FDA
developed a random sample methodology for inspectors to use for sampling
finished feed and feed ingredients. (Also see comment 5.)

7.After clarifying FDA's comment with an attorney in FDA's Office of the
Chief Counsel, we revised the report and the recommendation to delete
references that FDA should encourage firms to include a cautionary
statement on feed exports that may contain prohibited material. We believe
that it would be more prudent for FDA to focus its efforts on obtaining
statutory authority to require that the cautionary statement be used on
such exports.

8.We revised the recommendation to clarify that FDA should be alerting
USDA and the affected states whenever inspectors discover that cattle may
have consumed feed with prohibited material.

9.Based on the inspections we observed and the 404 inspection reports that
we reviewed in detail, we believe that inspector activities during
feed-ban compliance inspections are driven by the checklist
items/questions on the BSE inspection form. Therefore, we believe the
checklist should include specific questions to prompt inspectors to
examine vehicles and firms' cleanout procedures on every inspection.

10.As noted in the report, FDA believes that it provides the inspection
results in a transparent, compete, and accurate context. FDA notes that
the BSE inspection data posted on its Web site "allows the user to analyze
the data, in a multitude of ways, to provide their own contextual
reference." Our concern is precisely that the data are being analyzed and
interpreted in an erroneous context. Specifically, when FDA and industry
used those data to assert a 99 percent compliance rate with the feed ban,
they took that information out of context. While FDA's calculation of
compliance by a subset of regulated industries may in fact be quite high,
FDA's data are not sufficient to make that projection for all regulated
industries because of the many problems we cite in the report.
Specifically, FDA does not know the status of compliance for firms that
have never been inspected or those that have not been reinspected in
years. FDA also does not know if a firm that it previously inspected and
classified as low-risk has started using prohibited material; and FDA
reclassifies a firm in the database from "out-of-compliance" to
"in-compliance" when it corrects a violation-even when the violation was
serious and long-standing. Lastly, inspection results are largely based on
a review of paper documents and a visual inspection, with little or no
feed testing. Given these data concerns and compliance unknowns, FDA's
data should not be used to project industry compliance and, anytime those
data are cited, they should be reported in a complete and accurate
context.

GAO Contacts and Staff AcknowledgmentsAppendix VII

GAO Contacts

Robert A. Robinson, (202) 512-3841 Erin Lansburgh, (202) 512-3017

Staff Acknowledgments

In addition to the individuals named above, Vincent Balloon, Jim Dishmon,
Natalie Herzog, Lynn Musser and John C. Smith made key contributions.
Other contributors included George Quinn, Carol Herrnstadt Shulman, Joan
Vogel and Amy Webbink.

(360391)
*** End of document. ***