This is the accessible text file for GAO report number GAO-06-157R 
entitled 'Mad Cow Disease: An Evaluation of a Small Feed Testing 
Program FDA Implemented in 2003 With Recommendations for Making the 
Program a Better Oversight Tool' which was released on November 9, 
2005. 

This text file was formatted by the U.S. Government Accountability 
Office (GAO) to be accessible to users with visual impairments, as part 
of a longer term project to improve GAO products' accessibility. Every 
attempt has been made to maintain the structural and data integrity of 
the original printed product. Accessibility features, such as text 
descriptions of tables, consecutively numbered footnotes placed at the 
end of the file, and the text of agency comment letters, are provided 
but may not exactly duplicate the presentation or format of the printed 
version. The portable document format (PDF) file is an exact electronic 
replica of the printed version. We welcome your feedback. Please E-mail 
your comments regarding the contents or accessibility features of this 
document to Webmaster@gao.gov. 

This is a work of the U.S. government and is not subject to copyright 
protection in the United States. It may be reproduced and distributed 
in its entirety without further permission from GAO. Because this work 
may contain copyrighted images or other material, permission from the 
copyright holder may be necessary if you wish to reproduce this 
material separately. 

Washington, DC 20548: 

United States Government Accountability Office: 

October 11, 2005: 

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

The Honorable Thad Cochran: 
United States Senate: 

The Honorable Richard J. Durbin: 
United States Senate: 

Subject: Mad Cow Disease: An Evaluation of a Small Feed Testing Program 
FDA Implemented in 2003 With Recommendations for Making the Program a 
Better Oversight Tool: 

In 1997, the Food and Drug Administration (FDA) banned the use of most 
proteins derived from mammals (referred to as prohibited material) in 
feed intended for cattle and other ruminants.[Footnote 1] The feed-ban 
rule is one of the primary actions taken by the federal government to 
protect U.S. cattle from bovine spongiform encephalopathy 
(BSE),[Footnote 2] commonly known as mad cow disease, which is believed 
to be spread through feed that contains malformed protein found in 
certain tissue--particularly brain and central nervous system tissue-- 
of BSE-infected animals.[Footnote 3] Earlier this year, mad cow disease 
was found for the first time in a 12-year old animal born and raised in 
the United States. 

In January 2002, we reported on the effectiveness of federal actions to 
prevent the introduction and spread of BSE in the United States and 
identified a number of areas where improvements were needed to 
strengthen FDA's oversight of firms in the feed industry.[Footnote 4] 
In February 2005, we issued a follow-up report that examined the 
effectiveness of FDA's actions since the 2002 report to ensure industry 
compliance with the feed-ban rule and protect U.S. cattle from 
BSE.[Footnote 5] Our report concluded that while FDA has taken a number 
of positive steps, its processes still have room for improvement. Our 
February 2005 report also noted that FDA had begun a small, discrete 
feed testing program in August 2003. We reported that we would provide 
information on this new feed testing program, which FDA described as a 
unique effort, once FDA provided us with data on the feed tests. FDA 
later gave us the information we required to examine those feed testing 
activities. Accordingly, this report assesses FDA's small feed testing 
program and examines the extent to which this feed testing program 
helps FDA better assure industry compliance with the feed-ban rule. 
This report is the final component of our follow-up work on FDA's BSE 
prevention efforts. 

FDA established the feed testing program in an assignment memorandum 
issued in August 2003, entitled Assignment Memorandum--Sample 
Assignment for Domestic Products, which contained instructions for 
implementing the program. The purpose of the feed testing program was 
to collect and analyze cattle and other types of animal feed and feed 
ingredients to determine whether feed that could be fed to cattle might 
contain material prohibited by FDA's feed-ban rule. Under the program, 
FDA collected 641 feed samples through the end of fiscal year 2004 and 
planned to collect 900 feed samples during fiscal year 2005. 

The 2003 guidance gave FDA's district offices responsibility for 
collecting samples and submitting them to an FDA laboratory where 
analysts test the samples using a procedure called feed microscopy--a 
visual (microscopic) examination for potentially prohibited material, 
such as particles of bone, hair, or muscle fiber from certain animals. 
If an analyst detects what appears to be prohibited material, the 
findings are confirmed by a second analyst. According to FDA officials, 
some samples were tested using a more specialized method called 
polymerase chain reaction (PCR), a test that FDA has been piloting, 
which can differentiate ruminant DNA from other animal DNA.[Footnote 6] 

The guidance noted that because FDA had designated a number of cattle- 
derived exemptions to the feed-ban rule, including blood, milk protein, 
and plate waste, the laboratory tests could not definitively determine 
violations but, rather, could identify potential violations. The 
guidance directs the districts to conduct follow-up reviews on each 
potential violation to determine whether the facility represented by 
the sample actually violates the feed ban. On the basis of the follow- 
up reviews, the districts assign final compliance determinations--that 
the facility where the sample was collected has complied with or has 
violated the feed-ban rule. 

In June 2005, FDA issued a directive that all feed sample analysis and 
follow-up actions be recorded in FDA's central data system--the Field 
Accomplishments and Compliance Tracking System (FACTS)--and that 
districts complete follow-up reviews of potential violations within 30 
working days. In July 2005, FDA issued a revised assignment memorandum 
that, among other things, enhances the testing protocol by adopting the 
PCR test for sample retesting and directs districts to provide 
sufficient narrative explanation in FACTS to explain their final 
determination on samples that laboratories identify as potential 
violations. 

For the purpose of this report, we use the term "feed testing program" 
to distinguish the samples FDA collected for the feed-testing 
assignments from samples FDA and states collected in conjunction with 
routine BSE inspections. We included only the samples that FDA 
collected for the assignments. To examine the extent to which FDA's 
feed testing program provides better assurance of industry compliance 
with the feed-ban rule, we reviewed FDA's data on 1,206 samples 
collected through June 2005. We identified 989 feed samples collected 
by FDA's district offices and analyzed by FDA laboratories between 
August 2003 and June 2005, under the feed testing assignment/program 
implemented under the August 2003 guidance document. We compared sample 
collection, analysis, and follow-up with the program instructions in 
the August 2003 assignment memorandum. In order to assess FDA's 
timeliness in analyzing feed samples and to determine results of these 
analyses, we analyzed data on feed sample collection and laboratory 
analysis maintained in FACTS on the 989 feed samples. In order to 
assess the types of follow-up activities carried out by the districts 
and the basis for their final determinations on potential violations, 
we obtained and analyzed additional electronic files from FDA districts 
and discussed those activities and determinations with officials in the 
19 FDA district offices. We also obtained detailed district-specific 
data and information on sample collection, follow-up, and enforcement 
activities in interviews with the officials in the 19 FDA district 
offices and discussed this information with FDA headquarters officials. 
To assess the reliability of the FACTS data, we analyzed the feed 
sample records in this database as of June 7, 2005. We analyzed the 
data to identify problems with completeness, accuracy, or timeliness of 
data entry, and reviewed system documentation on controls. We 
determined that the data were sufficiently reliable for the purposes of 
this report. The testing program data assessed for this report, 
including documentation in FACTS, spreadsheets maintained by individual 
district offices, documents describing district follow-up actions for 
individual samples, and all written guidance documents, were provided 
in response to our specific requests for all such documentation and 
data related to the feed testing program. Finally, we examined the feed 
testing program guidance that FDA provided in the June 2005 field 
management directive and the July 2005 assignment memorandum and 
compared it with the instructions and guidance FDA provided in the 
August 2003 memorandum. We performed our work from February through 
August 2005 in accordance with generally accepted government auditing 
standards. Our work included an assessment of FDA's feed testing 
program data reliability and internal controls. 

Results in Brief: 

The feed testing program is a small part of FDA's BSE oversight effort 
and is one of several methods FDA uses to monitor for compliance with 
the feed-ban rule. However, several weaknesses in the design and 
implementation of the feed testing program need to be addressed to 
improve its effectiveness. Specifically, under the program guidance, 

* FDA did not require districts to document their follow-up reviews or 
the basis for their final determinations on samples that the 
laboratories identified as potentially containing banned protein 
products. Although the districts may have conducted rigorous follow-up 
and exercised sound judgment, the basis for their decisions cannot be 
reviewed and confirmed. 

* For nearly half the 989 samples, FDA took longer than 30 days from 
the date the sample was collected until the date the laboratory 
completed its analysis--including 21 samples that took longer than 100 
days. This extended period does not include the time FDA's districts 
would have spent following up on samples that indicated potential 
violations. FDA and industry agree that cattle feed is consumed very 
quickly. By the time FDA conducted its follow up to determine whether a 
violation had occurred, the feed may have been consumed. 

* FDA managers in headquarters did not adequately oversee the feed 
testing program. Specifically, FDA managers did not receive periodic 
reports or have other oversight controls in place to assure that the 
program was implemented correctly. Moreover, FDA did not identify 
intended program goals and, as a result, does not know whether or to 
what extent the feed testing program is contributing to the agency's 
BSE oversight efforts. 

FDA's June 2005 directive and July 2005 revised instructions--issued 
nearly 2 years into the program--includes (1) a requirement that follow-
up actions and compliance determinations be fully documented in FDA's 
centralized FACTS compliance tracking system with sufficient 
explanation to allow the reader to understand the basis for the 
decision and (2) a time limit for districts to complete follow-up 
reviews. 

To ensure that the feed testing program contributes to FDA's BSE 
oversight efforts, we are recommending that FDA (1) fully implement the 
June 2005 field management directive and July 2005 assignment 
memorandum, (2) assure that districts and laboratories adhere to time 
limits on collecting samples, completing sample analysis, and carrying 
out follow-up activities to minimize cattle's exposure to potentially 
contaminated feed, and (3) require sufficient oversight by headquarters 
managers to assure the program is achieving its intended goals. 

In commenting on a draft of this report, FDA expressed concern that GAO 
was issuing a report that focused on one small aspect of FDA's BSE 
oversight efforts. We agree that it is a small component of FDA's 
overall efforts, but it vies for FDA's limited BSE oversight resources. 
Furthermore, as we pointed out in our more comprehensive February 2005 
report, we looked at this small program separately because FDA did not 
provide program data in time for its inclusion in the broader report. 
FDA also disagreed with two of our recommendations in a draft of this 
report: that it set a time period for laboratories to complete sample 
analyses and that headquarters managers exercise sufficient oversight 
to assure the program operates as intended. FDA indicated that it had 
some target timeframes for laboratories. Because we could not pinpoint 
where delays were occurring, we revised our recommendation to address 
the need to minimize overall time--from sample collection through 
analysis and follow-up activities--in order to minimize cattle's 
exposure to potentially dangerous feed. With regard to our 
recommendation for better management oversight, FDA disagreed with our 
assertion that the program was not sufficiently monitored and noted the 
activities its managers have undertaken. We modified that 
recommendation to clarify what we believe is needed in terms of 
management oversight. 

Background: 

BSE is an always fatal neurodegenerative animal disease, first 
identified in 1986. The disease has been found in cattle in 26 
countries, including the United States, which discovered its first 
native-born case in a 12-year old cow in June 2005. The agent believed 
to be responsible for BSE is a malformed protein found in certain 
tissue--particularly brain and central nervous system tissue--of BSE- 
infected animals. Cattle contract BSE by eating feed derived from the 
remains of an infected animal. Scientists also generally believe that a 
rare but fatal disease in humans--known as variant Creutzfeldt-Jacob 
Disease--is linked to eating products containing cattle tissue 
contaminated with the malformed protein. 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. 

Under FDA's 1997 feed-ban rule, firms in the feed industry must (1) 
label feed and feed ingredients that contain or may contain most 
proteins from most mammals (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 firms handle cattle feed and feed ingredients (in the same facility) 
as well as material intended for other animal species that is 
prohibited in cattle feed, and (3) maintain records for 1 year so that 
feed and feed ingredients that contain or may contain prohibited 
material can be tracked from receipt through disposition.[Footnote 7] 
Firms that transport both types of materials also must have procedures 
to prevent commingling. 

The feed ban prohibits the use of certain mammalian proteins in the 
feed for cattle and other ruminants, such as sheep and goats; however, 
the material prohibited for use in cattle feed can be used in pet food 
and in feed for poultry, swine, horses, and other nonruminant animals. 
In addition, FDA designated a number of cattle-and other animal-derived 
items as exempt from the feed-ban rule and, hence, allowed in cattle 
feed. The exempt items include blood and blood products, plate waste, 
gelatin, and milk and milk proteins.[Footnote 8] In addition, poultry 
litter (composed of poultry waste material, bedding, and spilled feed 
that is used as a protein source) is allowed in cattle feed.[Footnote 
9] Consequently, the presence of animal protein in a feed sample may or 
may not indicate a violation of the feed-ban rule.[Footnote 10] 

Under the risk-based inspection approach that FDA adopted in 2002, FDA 
has designated firms that manufacture, blend, and otherwise directly 
process with prohibited material as posing the highest risk for 
potentially exposing U.S. cattle to BSE. Firms that do not process with 
prohibited material are designated as posing a lower risk. FDA 
documents the results of BSE inspections in the FACTS compliance data 
system and periodically posts inspection results on the FDA Web site. 

According to the August 2003 assignment memorandum implementing the 
feed testing program, the program objective was to "collect and analyze 
domestic feed, feed ingredients and other animal feed products for the 
presence of animal tissue using the feed microscopy method to monitor 
for compliance with [the feed-ban rule.]" The memorandum instructed 
districts to (1) select samples from animal feed, feed ingredients, and 
other animal feed products, such as medicated feed; (2) collect at 
least 50 percent of samples from products intended for ruminants; (3) 
select products that are labeled as containing animal protein but do 
not have a caution statement that they not be fed to cattle, which is 
required by the feed-ban rule; (4) include samples from feed that does 
not list mammalian protein in their name or ingredients; and (5) select 
each sample from a different source, processor, or manufacturer, if 
possible. In addition, FDA officials told us samples were being taken 
from "finished" feed--sold in bags or bulk--and that the testing 
program would give FDA an additional way to review products in the 
marketplace. 

The August 2003 assignment memorandum assigned the district offices 
responsibility for regulatory and administrative follow-up of 
laboratory findings. It directs districts to obtain additional 
information on samples that the laboratories classify as identifying 
potential violations through reviews of firms' records, trace-back 
inspections to suppliers, and interviews with individuals in the chain 
of receipt and use of materials in the sampled feed. When districts 
confirm a violation, the guidance directs districts to remove the feed 
or feed ingredients from distribution, either by voluntary recalls or 
by seizure. According to the guidance, decisions to take additional 
enforcement actions, such as issuing warning letters, depend on the 
history of the firms, the scope of the violations, and the source of 
the prohibited material. 

In May 2004, FDA headquarters conducted an internal evaluation of the 
feed testing program based on a review of sample collection and 
laboratory analysis information on samples collected and testing in the 
first 8 months of the program. That evaluation did not include 
information on follow-up reviews by the districts. In May 2005, FDA's 
Center for Veterinary Medicine reported that FDA follow-up reviews at 
feed mills and elsewhere in the feed chain revealed a high level of 
compliance with the feed-ban rule.[Footnote 11] 

The August 2003 assignment memorandum instructed the districts to 
collect a total of 600 samples through the end of fiscal year 2004; in 
fact, FDA collected 641 feed samples in that period. Enclosure I shows 
the number of feed samples assigned to each district and the number 
collected and analyzed through the end of fiscal year 2004 and for 
fiscal year 2005. 

The Feed Testing Program Has Not Provided FDA Additional Assurance of 
Compliance with the Feed-Ban Rule Because of Weaknesses in Its Design 
and Implementation: 

The effectiveness of FDA's feed testing program has been limited by 
three design and implementation weaknesses. First, in designing the 
program FDA did not require districts to document their follow-up 
activities on samples that potentially violated the feed-ban rule or 
the basis for their final compliance determinations of those samples. 
Second, it was designed and implemented without time frames for 
promptly collecting and analyzing samples and following up on test 
results. Finally, FDA headquarters managers did not maintain adequate 
oversight responsibility for ensuring the program met the intended 
goals. FDA's June 2005 directive and July 2005 revised guidance address 
some of these concerns but will be useful only when fully implemented. 

FDA's Districts Have Not Documented Follow-up Activities or the Basis 
for Their Determinations on Feed Samples: 

FDA's districts may have conducted rigorous follow-up and exercised 
sound judgment. However, they did not document their follow-up actions 
and the basis for their compliance determinations on whether firms 
violated the feed-ban rule because FDA did not require districts to 
clearly document those activities and decisions. As a result, the basis 
for their decisions cannot be reviewed and confirmed. Without this 
documentation, FDA has no assurance that the districts' actions were 
thorough and correct. FDA laboratories identified 215 of the 989 
samples we examined as identifying potential violations. Based on their 
follow-up reviews, however, the districts determined that 214 samples 
did not show violations--that only one of the firms chosen for 
obtaining a surveillance sample violated the feed-ban rule (see table 
1). 

Table 1: Classification of 989 Feed Samples by FDA Laboratories and 
Districts, by Feed Type, August 2003 through June 2005: 

[See PDF for image] 

Source: GAO analysis of FDA data. 

[End of table] 

The one sample FDA determined demonstrated a violation of the feed-ban 
rule was from cattle feed collected at a feed mill. The laboratory 
classified the sample as identifying a potential violation because it 
contained cattle hair. The label indicated that the feed contained 
poultry meal. The district's follow-up review determined that the 
renderer that supplied the poultry meal to the feed mill had previously 
processed prohibited material and failed to use adequate clean-out 
procedures to prevent commingling or cross-contamination with the 
ingredients intended for cattle feed. FDA issued a warning letter to 
the renderer for not maintaining adequate procedures or labeling the 
product with the required cautionary statement that the ingredients not 
be fed to cattle or other ruminants. 

We were unable to independently verify the follow-up reviews on other 
potential violations or confirm the districts' final compliance 
determinations of samples, because the documentation supporting the 
districts' determinations was lacking or incomplete. When we asked FDA 
for this information, FDA acknowledged that it did not require the 
districts to document their follow-up activities. FDA headquarters 
contacted its districts and told them to reconstruct an accounting of 
their follow-up actions and final compliance determinations. Thus, FDA 
compiled this information several months after the fact for most 
samples. The information we received was unclear and did not contain 
sufficient sample-specific information. Table 2 summarizes the type of 
district follow-up activities compiled by FDA. 

Table 2: District Follow-up Action on Samples Identifying Potential 
Violations by the Laboratories: 

[See PDF for image] 

Source: GAO analysis of FDA data. 

[End of table] 

Likewise, the narrative information that FDA compiled from the 
districts on their final compliance determinations, which we summarize 
in table 3, does not give sufficient information to verify the basis 
for those determinations. 

Table 3: District Compliance Determinations on Samples Identifying 
Potential Violations by the Laboratories: 

[See PDF for image] 

Source: GAO analysis of FDA data. 

[End of table] 

In order to verify the basis for their determinations, districts must 
be able to provide clear and sufficient information for a reviewer to 
understand the decisions made and the reason for making those 
decisions. That is, when a district follows up on a sample that a 
laboratory has classified as evidencing a potential violation, the 
district would describe the specific evidence it uses to reach a 
determination that a firm has not violated the feed-ban rule. FDA's 
July 2005 guidance recognizes the importance of this critical step and 
directs the districts to provide sufficient narrative explanation in 
FACTS to allow an FDA manager to understand the basis for those 
decisions. If, for example, the analyst observes particles of bone, 
tissue, or hair in cattle feed, and the district is relying on records 
from a recent BSE inspection, we would expect the district to provide a 
detailed description of the animal material the firm used and the date 
it used that material to manufacture the feed. This description would 
have to fully explain what the laboratory observed. If information from 
a recent inspection is not available, we would expect FDA to conduct a 
follow-up inspection at the firm and describe the documents, such as 
dated invoices, that verify the type of animal material used that fully 
explains what the laboratory observed. 

When we met with FDA officials in September 2005, they acknowledged 
that headquarters and field managers did not have an easily accessible, 
uniform method for tracking districts' follow-up actions and compliance 
determinations that would enable them to perform thorough oversight and 
analyze trends in the program. Officials stated that the new June 2005 
directive should alleviate these shortcomings and that FDA will make 
further changes if managers determine that the directive does not 
address all of the weaknesses. 

FDA Did Not Ensure That Samples Were Promptly Sent to Laboratories and 
Analyzed and That Potential Violations Were Quickly Followed Up: 

FDA's program guidance did not include timeframes for ensuring that 
laboratories analyzed samples and districts follow up on test results 
promptly; as a result, FDA had no assurance that these activities were 
carried out expeditiously to minimize the risk that cattle would be fed 
potentially contaminated feed. After FDA received the draft report for 
comment, it informed us that laboratories are to complete their 
analysis of samples taken under the feed testing program within 20 
working days, although this timeframe is not in the August 2003 program 
guidance for laboratories. FDA could not provide data we requested to 
determine the amount of time that samples were undergoing analysis or 
the time districts spent in following up on potential violations and 
reaching a final determination because it does not track this 
information. 

However, FDA did provide the date each sample was collected and the 
date the laboratory reported the results of its analysis to the 
district because the districts and laboratories were entering that 
information into the FACTS compliance tracking system. In analyzing 
these data, we found that for nearly half of the samples we examined 
(473 of 989), more than 30 days elapsed before the laboratories 
reported sample findings to the districts. That included 38 samples for 
which more than 60 days--and in some cases more than 100 days--elapsed 
before the laboratory findings were reported to the districts (see fig. 
1). 

Figure 1: Number of Days for FDA Laboratories to Analyze Feed Samples 
and Report to Districts from Date of Collection, August 2003 through 
June 2005: 

[See PDF for image] 

Source: GAO Analysis of FDA Data. 

[End of figure] 

The districts initiated follow-up activities on potential violations of 
the feed-ban rule after the laboratories reported their analyses. 
However, FDA did not provide consistent information on the timeliness 
of district follow up actions because they were not tracking this 
information. Therefore, we could not determine how much more time 
passed before districts took follow-up actions on the 215 samples that 
the laboratories identified as potentially demonstrating violations of 
the feed-ban rule. 

According to FDA and industry officials, however, cattle feed is 
consumed very quickly. Because FDA did not include timeframes in the 
August 2003 guidance for laboratories to analyze samples and for 
districts to follow up on samples identifying potential violations, by 
the time inspectors determined that cattle feed was contaminated, all 
the feed in question could have been consumed by cattle. In commenting 
on a draft of this report, FDA indicated that it plans to evaluate the 
number of days that laboratories are spending on analyzing feed samples 
as those data are compiled. 

FDA Headquarters Managers Did Not Exercise Adequate Oversight of the 
Feed Testing Program: 

FDA's managers in headquarters designed the feed testing program and 
issued the August 2003 assignment memorandum. However, those managers 
did not exercise oversight once the program was implemented. 
Specifically, FDA managers had no controls in place to ensure that the 
August 2003 guidance was consistently followed, that results were 
carefully tracked, and that the program was operating as intended and 
achieving its intended goals. FDA did not identify program goals and, 
as a result, does not know whether or to what extent the feed testing 
program is contributing to the agency's BSE oversight efforts. In past 
reports, we have stressed the importance of performance measures as 
critical internal control standards that enable federal agencies to 
compare and analyze actual performance data against expected or planned 
goals for their activities and programs.[Footnote 12] However, FDA had 
no such controls in place to compare and analyze feed testing 
activities carried out by its laboratories and districts. Under the 
Government Performance and Results Act of 1993, agencies must use 
outcome-oriented goals and performance measures that assess results, 
effects, or impacts of a program or activity compared with its intended 
purpose.[Footnote 13] Without such measures, FDA cannot assess whether 
its feed testing efforts achieved the intended results or how well 
districts and laboratories collected and analyzed samples and followed 
up on samples that potentially violated the feed-ban rule. 

Following are some examples where laboratories and districts did not 
implement the 2003 assignment consistently and FDA headquarters 
managers did not have oversight in place to discover these 
inconsistencies: 

* FDA laboratories classified 29 samples as "in compliance" that 
analysts described as containing mammalian protein from an 
unidentifiable source. Based on the August 2003 guidance, however, 
analysts should have classified these samples as identifying potential 
violations, thus flagging them for district follow-up. 

* One of the six FDA laboratories continued to misclassify samples as 
demonstrating definite violations--a classification that current 
testing technology does not support--after the May 2004 evaluation 
revealed that this type of misclassification was occurring. 

* Eighteen districts collected nearly all samples from firms that had 
previously undergone a BSE inspection, while one district collected 
samples at retail stores that did not manufacture feed and typically 
had not undergone a BSE inspection. FDA's risk-based inspections target 
firms that manufacture, blend, and otherwise directly process with 
prohibited material; however, the 2003 assignment instructions appear 
to focus on feed samples collected at the retail level and FDA 
officials told us that samples collected for the feed testing program 
were to be taken from finished feed sold in bags or in bulk, which 
would give FDA an additional way to review products in the marketplace. 

* Laboratories reported that labels and ingredient lists were missing 
for 28 of the 215 samples with potential violations, although the 
August 2003 assignment instructed districts to submit these items with 
samples. The July 2005 assignment continues to instruct districts to 
submit labels with samples. 

FDA headquarters did not have oversight mechanisms in place to monitor 
the feed testing program nor performance indicators to compare program 
results across laboratories and districts to detect these 
implementation differences. 

In addition, headquarters had no controls in place to discern that 
districts were not recording or tracking sample follow-up actions and 
compliance determinations. Although its FACTS compliance tracking 
system contains data fields for documenting a narrative explanation of 
what action was taken, the rationale for the action, the final district 
classification, and the date of the decision, the 2003 assignment did 
not direct the districts to use FACTS and FDA's oversight did not 
detect and correct this until the June 2005 directive and July 2005 
revised assignment. The FACTS compliance tracking system is the 
centralized database that FDA implemented agency wide for the expressed 
purpose of capturing this information. 

Furthermore, after FDA headquarters conducted the internal evaluation 
in May 2004, it did not act to implement internal controls to ensure 
that the testing program would achieve its intended goals and correct 
the problems identified in that review. The internal review looked at 
370 samples taken during the first 8 months of the program. The report 
identified 70 samples classified by the laboratories as potentially in 
violation of the feed ban, including 42 samples of feed intended for 
cattle. FDA based its evaluation on the laboratories' descriptions and 
any label or ingredient information submitted with the samples, but did 
not consider any district follow up on laboratory findings. The 
evaluation "encouraged" the districts to follow up on 14 of the 42 
cattle feed samples and 26 samples from feed intended for other species 
that could include prohibited material. FDA headquarters did not 
question how the districts addressed the review findings. 

The feed testing program cannot provide FDA with additional assurance 
of compliance with the feed ban unless headquarters exercises adequate 
oversight and implements internal controls to address these program 
weaknesses. 

New Procedures Address Some Feed Testing Program Weaknesses: 

FDA officials have acknowledged weaknesses in the August 2003 
memorandum and told us that the June 2005 directive and July 2005 
revised assignment memorandum are intended to address those problems. 

FDA's June 2005 directive requires, among other things, that: 

* all feed sample analysis and follow-up actions are documented 
accurately and in a timely fashion in the agencywide FACTS compliance 
tracking system; 

* program managers at headquarters, regions, districts, and 
laboratories implement internal audit procedures and controls to verify 
that sample analysis and follow-up actions are timely and accurately 
documented in FACTS; and: 

* districts complete and document follow-up actions within 30 working 
days following receipt of sample results from the laboratory. 

The July 2005 revised program instructions clarify sample selection 
criteria and require, among other things, that: 

* laboratories use PCR to verify samples that indicate the possible 
presence of mammalian bone or hair, and: 

* districts document their assessments of samples found to be in 
potential violation of the feed ban and their final determinations with 
sufficient narrative explanation to allow a reviewer to understand the 
basis for their decisions. 

The new directive and instructions went into effect immediately. FDA 
officials told us that the districts are entering the required 
information in FACTS for all samples followed up in fiscal year 2005. 
However, if districts enter the same type of information that they 
provided to us without, for example, citing the specific documentation 
used and actions conducted to reconcile laboratory findings, then these 
additions to FACTS may not be useful for oversight. 

Conclusions: 

FDA's June 2005 directive and the July 2005 revised assignment include 
important new controls that address many of the weaknesses we found in 
the feed testing program. However, the new directive and guidance will 
be useful only when FDA ensures their full implementation. One 
important requirement in the directive and guidance--for districts to 
document their follow-up activities and compliance decisions--will 
allow FDA to use the program results to supplement the agency's other 
BSE oversight activities. FDA's districts and laboratories believe they 
have implemented the feed testing program diligently and thoroughly, 
using their best professional judgment. That notwithstanding, until the 
districts' actions are documented in a fashion that fully explains the 
basis for their compliance determinations, FDA cannot verify and hence 
cannot confidently rely on the testing program results. 

Another important requirement in the new directive is the addition of a 
30-day time limit for districts to complete their follow-up actions and 
make final compliance determinations for feed samples that identify 
potential violations of the feed-ban rule. That new guidance 
notwithstanding, we remain concerned about the overall time frame. We 
found that more than 30 days elapsed between the date samples were 
collected and the date laboratories completed their analysis for nearly 
half the samples, and that these two steps took more than 100 days in 
some instances. Only then would districts have begun their follow-up 
activities. However, both FDA and industry agree that cattle feed is 
consumed very quickly. Consequently, by the time FDA completes its 
follow up activities and determines that a violation has occurred, the 
feed may have been consumed. We believe that both the districts and the 
laboratories need to carry out their feed testing program 
responsibilities promptly to minimize cattle's exposure to potentially 
contaminated feed. 

While FDA's new assignment instructions recognize the importance of 
management accountability, they do not include specific oversight 
requirements that will address the deficiencies we identified. Even 
though the feed testing program is small, adequate management oversight 
of the program is critical, because the resources FDA spent on the 
program since August 2003 came directly from the agency's limited BSE 
oversight funding. If they exercise appropriate oversight, FDA 
headquarters managers can help ensure that future results of the feed 
testing program will be reliable, and that BSE resources will be 
carefully spent. In this regard, we believe that periodic reports using 
FACTS data would be useful. Other internal controls may also provide 
useful management oversight, and FDA would benefit if it developed 
performance indicators and set goals for its managers to use to 
determine whether and to what extent the feed testing program is 
contributing to the agency's BSE oversight efforts. 

Finally, feed testing has the potential to be an important tool in 
FDA's feed-ban oversight arsenal as technology improves, and we believe 
FDA would benefit by encouraging the development, testing, and 
implementation of new feed testing technologies. PCR is a better tool 
than feed microscopy, and the capabilities of PCR are being refined and 
improved. As more accurate and effective PCR and other feed testing 
technologies emerge, the value of feed testing to FDA's BSE oversight 
will increase. 

Recommendations for Executive Action: 

To ensure that the feed testing program is a useful tool for helping 
FDA oversee industry compliance with the feed-ban rule, we are 
recommending that the Commissioner of FDA take the following three 
actions: 

* Fully implement the June 2005 field management directive and July 
2005 assignment memorandum revising the feed testing program. 

* Assure that districts and laboratories adhere to time limits on 
collecting samples and completing sample analysis and follow-up 
activities to minimize cattle's exposure to potentially contaminated 
feed. 

* Require FDA headquarters managers to exercise sufficient oversight, 
with periodic reports from FACTS or other management controls, and 
identify appropriate performance indicators for the feed testing 
program, to assure that the program operates as intended and achieves 
its intended goals. 

Agency Comments and Our Evaluation: 

We provided FDA with a draft of this report for review and comment. In 
its comments on the draft report, FDA included an overview of its BSE 
oversight program to put the feed testing effort in context. FDA 
expressed concern that we were issuing a report that focused on one 
small aspect of that effort. As we explained in our more comprehensive 
February 2005 report, we analyzed and are reporting separately on this 
small program because FDA did not provide program data in time for its 
inclusion in the broader report. 

With respect to our first recommendation, FDA indicated that it plans 
to fully implement the June directive and July guidance. We have 
included this as a recommendation to help FDA maintain its momentum and 
attention to a program that commands a portion of its limited BSE 
oversight resources. With respect to our second recommendation, FDA 
told us that its laboratories have a time limit of 20 working days to 
analyze feed samples from this program. However, FDA could not provide 
data to document whether laboratories were meeting this time limit and 
our analysis of the elapsed time for the two steps of sample collection 
and data analysis often showed the time spent to be excessively long-- 
from 60 to 100 days and longer in some instances. Because the overall 
time frame is the period of concern, we revised our recommendation to 
address overall timeliness to minimize cattle's exposure to potentially 
contaminated feed. We believe that when FDA implements better tracking 
under the 2005 directive and guidance, it will have data to help 
determine specifically where timeliness can be improved. This will give 
FDA data to assess laboratory timeframes, which it indicated that it 
plans to do. Regarding our third recommendation for better management 
oversight, FDA disagreed with our assertion that the program has not 
been adequately monitored. However, FDA did not provide evidence that 
its managers received periodic reports assessing program performance or 
that they had other adequate management oversight controls in place. We 
believe that our revised recommendation, if implemented, will put FDA 
in a position to determine whether and to what extent the feed testing 
program is contributing to its BSE oversight efforts. 

As agreed with your offices, unless you publicly announce the contents 
of this report earlier, we plan no further distribution until 30 days 
from the date of this letter. At that time, we will send copies of this 
report to the congressional committees with jurisdiction over FDA and 
its activities; the Secretary of Health and Human Services; the 
Secretary of Agriculture; and the Director, Office of Management and 
Budget. In addition, this report will be available at no charge on the 
GAO Web site at http://www.gao.gov. 

If you or your staff have any questions about this report, please 
contact me at (202) 512-3841 or robinsonr@gao.gov. Contact points for 
our Office of Congressional Relations and Public Affairs may be found 
on the last page of this report. Key contributions to this report were 
made by Erin Lansburgh, Assistant Director; Jeremy Manion; Lynn Musser; 
George Quinn; Carol Herrnstadt Shulman; John C. Smith; and Amy Webbink. 

Signed by: 

Robert A. Robinson: 

Managing Director, Natural Resources and Environment: 

Enclosures: 

Number of Feed Samples Assigned and Collected and Analyzed, as of June 
7, 2005, by FDA District: 

[See PDF for image] 

Source: GAO analysis of FDA data. 

[A] We excluded samples that were collected by the districts when the 
analysis was not also included in the files provided by FDA. 

[End of table] 

Enclosure II: Comments from the Food and Drug Administration: 

Note: GAO comments supplementing those in the report text appear at the 
end of this enclosure. 

DEPARTMENT OF HEALTH & HUMAN SERVICES: 
Public Health Service: 
Food and Drug Administration:
Rockville MD 20857: 

Robert A. Robinson: 
Managing Director, Natural Resources and Environment: 
Natural Resources and Environment Team: 
United States Government Accountability Office: 
441 G Street, NW: 
Washington, DC 20548: 

Dear Mr. Robinson: 

Please find enclosed the general comments in response to the General 
Accountability Office's correspondence entitled, "Mad Cow Disease: FDA 
Needs to Fully Implement New Procedures and Correct Other Design Flaws 
for the Feed Testing Program to Help Assure Industry Compliance with 
the Feed Ban." 

We appreciate the opportunity to review and comment on this draft 
correspondence before it is published, as well as the opportunity to 
work with your staff in its development. 

Sincerely, 

Signed by: 

Lester M. Crawford, D.V.M., Ph. D.,
Commissioner of Food and Drugs: 

Enclosure: 

FDA General Comments on the Government Accountability Office's Draft 
Letter on Mad Cow Disease Follow-up: Feed Testing (GAO-05-904R): 

FDA values the opportunity to review and comment on the Government 
Accountability Office's (GAO) draft document. Even though we understand 
based on our exit conference discussion that some changes will be made 
in the draft document, our comments are based on our review of the 
draft provided on August 17, 2005, including the draft recommendations 
for executive action. We are concerned that GAO's issuance of a 
document that focuses on a small component of the overall BSE control 
program places an undue emphasis on that component's impact on the 
program. For that reason, FDA will present an overview of the BSE 
program operating in the United States, and attempt to place into 
perspective the role feed testing contributes to the overall program. 
FDA will also offer comments on GAO's three recommendations. 
Additionally, it is important to note that FDA spent nearly 600 (ORA 
and CVM) hours providing information to GAO on this supplemental study. 
This includes time accounted for when GAO conducted interviews with 
every FDA District office, as well as time spent by Headquarters' staff 
compiling information and responding to the 33 questions that GAO asked 
in order to assess if feed testing helped FDA better assure industry 
compliance with the feed ban. 

GAO provided three draft recommendations for executive action. For the 
first recommendation, FDA appreciates GAO's recognition that the June 
2005 field management directive and July 2005 BSE sampling assignment 
memorandum provide controls that enhance the agency's ability to 
evaluate the utility of this program in the future. Regarding the 
second recommendation, FDA disagrees that laboratory testing timeframes 
to complete sample analysis for domestic surveillance samples are a 
critical element to minimize cattle's exposure to potentially 
contaminated feed. Lastly, with regard to the third recommendation, FDA 
also disagrees with GAO's assertion that the sampling assignment was 
poorly implemented, and that FDA did not adequately oversee the 
assignment. 

While GAO identifies in its report that feed testing is a small 
component of FDA's BSE oversight effort, we believe it is critical to 
clarify the current perspective on the utility that feed testing plays 
in relation to the overall U.S. BSE control program. The U.S. cattle 
BSE control program consists of a multifaceted system to keep the 
disease from entering and spreading. The components consist of import 
controls on ruminant animals and animal feed, a BSE surveillance cattle 
testing program, an effective response to the finding of BSE positive 
animals, and a feed ban. This program remains a top priority for FDA, 
USDA, and other government agencies. While we focus on the feed ban in 
this discussion, we will touch on the other facets to make certain the 
overall control program is understood. 

Import control is the critical safeguard preventing the BSE agent from 
entering the United States. FDA and the USDA's Animal and Plant Health 
Inspection Service (APHIS) work in close cooperation with Customs and 
Border Protection on controls related to imports of animals and animal- 
derived products. Imports were restricted in 1989 from the United 
Kingdom, and those restrictions were expanded to imports from countries 
where BSE has been reported, and to countries identified to be at high 
risk for BSE, as needed. It is important to note that testing imported 
feed is a valuable component of our import controls for BSE, yet the 
GAO did not include the import testing program in its audit or report. 

USDA's active surveillance program to test cattle for BSE is another 
element of the prevention efforts to date. USDA has tested more than 
460,000 cattle from June 1, 2004 to mid-September 2005 and has found 
one BSE-positive cow. Prior to June 1, 2004, USDA's surveillance 
program identified an additional BSE-positive cow from Canada. In 
addition to its continued testing of the higher risk cattle population, 
USDA is planning to test a population of apparently healthy cattle over 
30 months of age at slaughter. 

An additional control measure is an effective and coordinated response 
to the finding of a BSE-positive animal. In response to the December 
2003 BSE-positive animal in Washington State, FDA, USDA, and the 
Washington State Department of Agriculture coordinated closely in their 
epidemiological investigation tracing the positive cow to its farm of 
origin in Alberta, Canada, and in the traceforward investigation to 
control potentially contaminated product derived from the positive 
animal. A similarly coordinated response was seen in the investigation 
of the Texas cow that was declared positive for BSE in June 2005. 
Although the positive cow posed no risk to the human food or animal 
feed supply, FDA, APHIS, the Texas Animal Health Commission, and the 
Texas Feed and Fertilizer Control Service conducted an epidemiological 
investigation of the herd of origin, as well as a feed investigation. 
The feed investigation found that no feed or feed supplements used on 
the farm since 1997 were formulated to contain mammalian protein 
prohibited under the FDA Ruminant Feed Ban regulation, and that all the 
investigated rendering plants were operating in compliance with the 
regulation. This high level of coordination came about through 
extensive planning, including the sharing of BSE response plans and the 
conduct of three joint exercises with our federal and state partners to 
test our capabilities to respond. 

Another extremely important component in the BSE control program is 
FDA's ruminant feed ban, which was put in place in 1997 to prevent the 
amplification of BSE through feed in the event that the BSE agent had 
been introduced into the United States. The regulation prohibits the 
use of most mammalian protein in feeds for ruminant animals. This rule, 
found in 21 CFR 589.2000, became effective on August 4, 1997. The 
regulation reflected the best epidemiological knowledge at the time. 
FDA continues to evaluate the science related to BSE and recognizes the 
presence of BSE in the United States. Accordingly, FDA published 
jointly with USDA an Advance Notice of Proposed Rulemaking (ANPRM) in 
July 2004 to obtain information and comments on additional control 
measures being considered to further strengthen the feed ban. 

To implement its feed ban regulation, FDA has put into place a Ruminant 
Feed Ban compliance program. The key elements of the compliance program 
include education of industry on the requirements of the regulation, 
inspection of facilities subject to the ban, and appropriate 
enforcement of the regulation's requirements. Since a significant 
portion of the animal feed industry was impacted by the new regulation, 
the initial focus of the compliance program was to educate both the 
industry and regulators about the feed ban. FDA sponsored numerous 
workshops attended by state veterinarians and feed control officials 
from all 50 States and Puerto Rico. In addition, FDA held briefing 
sessions with trade associations and consumer groups and has developed 
guidance documents to assist industry in complying with the regulation. 
The major implementation tool to assure compliance by regulated parties 
is a rigorous program of establishment inspections by FDA and state 
regulatory counterparts. In collaboration with state feed control 
officials, FDA has conducted over 37,000 inspections of renderers, feed 
mills, protein blenders, as well as other firms subject to this 
regulation such as ruminant feeders, on-farm mixers, pet food 
manufacturers, animal feed salvagers, distributors, retailers, and 
animal feed transporters. FDA routinely provides reports on its feed 
ban enforcement activities. Additionally, FDA posts the results of 
every firm inspected under the Ruminant Feed Ban program on the FDA web 
site. Compliance with this regulation by renderers, feed mills, and 
protein blenders remains very high. Furthermore, FDA has monitored 
recalls, issued warning letters, and taken other enforcement actions 
when significant violations have been documented during inspections. 

The assignment to test domestic feed for the presence of prohibited 
material is a relatively small component of FDA's overall domestic 
Ruminant Feed Ban efforts. Since no test currently exists for the 
detection of the agent that causes BSE in feed, analysis of feed is not 
a means of verifying the safety of cattle feed. The substitute for the 
infectious agent used in all current tests is the mammalian protein 
prohibited under the Ruminant Feed Ban. However, the current tests 
used, feed microscopy and/or PCR, are not adequate methods to make 
compliance decisions about the presence of prohibited material since 
the methods have limitations and the rule has exemptions. Feed 
microscopy generally can only detect the presence of mammalian tissue, 
either through the identification of bone or hair. The present Ruminant 
Feed Ban allows for certain exemptions to the mammalian protein 
prohibition. Exempted materials include pure porcine and equine meat 
and bone meal, blood (from any animal species, including ruminants), 
gelatin, and milk protein. There is no prohibition on the use of non- 
mammalian proteins (e.g., poultry meal). Certain tissues, such as bone 
or muscle, may be present as a result of the use of exempt ingredients, 
such as pure porcine meat and bone meal or poultry meal. While the PCR 
method can detect ruminant mitochondrial DNA, it cannot differentiate 
between prohibited material and ingredients exempted by the feed ban, 
such as ruminant blood and inspected meat products which have been 
cooked and offered for human food and further heat processed for feed. 
Since neither method can differentiate prohibited material from other 
acceptable materials, the analytical results by themselves cannot be 
used to verify the presence or absence of prohibited material. The only 
way to determine compliance with the Ruminant Feed Ban rule is to 
conduct an inspection of the firm. 

The FDA assessment of the current weakness of the feed microscopy 
method, as a compliance tool, is similar to findings by the Canadian 
Food Inspection Agency (CFIA). CFIA conducted a recent trial on the 
usefulness of microscopy method for analyzing the composition of feed. 
CFIA found that the limitations of the microscopy method outweigh its 
usefulness. Further, CFIA found that physical inspection of facilities 
and records was necessary to determine compliance with Canada's BSE 
feed controls, which are nearly identical to ours. The report is 
available at www.inspection.gc.ca. 

While CFIA came to its position after a small pilot test, FDA is still 
assessing the potential usefulness of feed testing of domestically 
produced products in FDA's long-term enforcement of the feed ban. 
Further, the sampling assignment is designed to be an additional way to 
review products in the marketplace by providing a supplemental piece of 
intelligence to help in selecting firms for inspection coverage under 
the compliance program. A positive finding from analytical testing 
provides information for us to conduct targeted follow-up inspections, 
but does not, by itself, prove the presence of prohibited material and 
a violation of the Ruminant Feed Ban rule. 

On August 18, 2003, FDA/CVM issued the first sampling assignment to the 
FDA field staff for the collection of 600 domestic samples, which was 
subsequently increased to 900 samples for the current fiscal year. The 
fairly recent implementation of this unique sampling assignment has 
involved efforts to train laboratory personnel in the techniques of 
feed microscopy, to provide reference samples for confirmatory 
comparison, and to assess proficiency in the technique by all 
laboratories conducting analytical testing. Based on the agency's 
experience derived from the initial sampling assignment, FDA issued a 
second assignment in July 2005 that refines directions for sample 
collection and classification and incorporates PCR for additional 
analysis of samples found by the feed microscopy method to contain 
animal tissue. 

It is important to note that the scientific enhancements in this latest 
assignment are based on research the agency has conducted on feed 
testing methodology. In 2001, FDA published the results of its first 
method validation trial using a PCR-based approach to detect bovine 
materials in animal feed. This method was successfully validated at a 
detection level of 0.125 percent bovine meat and bone meal, on a weight-
to-weight (w/w) basis. While a robust method, it was labor- intensive 
to perform, requiring the analyst to prepare all the necessary 
reagents. The 24 hours needed to analyze a sample further limited the 
number of samples an analyst could process. 

Subsequently, efforts were initiated to develop a second-generation PCR-
based assay using a DNA extraction method that was easier to perform 
than the one used in the validated, first-generation PCR-based method. 
Using a commercially available DNA Forensic Kit, a suitable method was 
developed to permit extraction of DNA from animal feed and feed 
ingredients. This method significantly shortened the time required to 
analyze a single sample to under 8 hours, start-to-finish. In addition, 
because the DNA extraction portion is significantly easier to perform, 
the total number of samples that a single analyst can examine was 
doubled. The second-generation method has been successfully validated 
to permit detection of bovine, ovine, and porcine materials in animal 
feed and feed ingredients at a level of 0.1 percent (w/w basis). FDA 
has implemented this second-generation PCR-based assay in the enhanced 
sampling assignment issued in July 2005 following validation work in 
our field laboratories. 

Both the first and second-generation PCR-based methods are conventional 
PCR-based methods that rely on photo documentation followed by visual 
interpretation of the results. These methods are at best semi- 
quantitative. Currently, FDA scientists are working on a third- 
generation PCR-based method that will use an approach called real-time 
PCR. Real-time PCR assays measure the amount of product being formed 
while the PCR process is still ongoing. There is no post-PCR processing 
of the test samples as is required for the first and second-generation 
PCR-based methods. Therefore, a real-time PCR based method will result 
in a further reduction in time needed to analyze the feed samples. The 
process involved in assessing PCR-product formation yields data that 
are proportional to the amount of starting DNA in the test sample, and 
therefore, these results are proportional to the amount of prohibited 
proteins in the original sample. The chemistry involved with real-time 
PCR permits direct addition of internal controls to ensure assay 
functionality. Lastly, PCR products formed during real-time PCR permit 
a determination as to whether or not the final PCR product was derived 
from the expected species (e.g., bovine) by performing a melt-curve 
analysis. Each PCR product will have a specific temperature at which 
the double-stranded DNA molecules will separate. 

While feed microscopy and PCR are used by FDA and some state 
laboratories, these methods cannot be performed by all facilities and 
organizations that need to test for the absence of prohibited proteins. 
There are four companies that market various immunochemical test kits 
(antibody-based) that are simpler to use than either PCR or feed 
microscopy. These kits might also seem useful for FDA's purposes. 
However, these test kits are marketed without FDA's review and 
approval, as FDA does not have premarket approval authority over 
veterinary devices. FDA therefore initiated a research program to 
evaluate their performance characteristics. An essential aspect of this 
evaluation was the development of acceptance criteria and performance 
guidelines against which these tests would be assessed. An important 
component of these guidelines was the criteria that these tests be able 
to detect prohibited proteins at the same level as PCR and microscopy. 
To date, we have completed our initial evaluation of two of these test 
kits and are close to completing our evaluation of a third kit. The two 
test kits for which we have completed our evaluation did not meet our 
acceptance criteria and performance guidelines; one test had an 
unacceptable rate of false positives (true negatives that the test 
deemed to be positive) and the second test kit did not have the 
sensitivity set out in our performance criteria. 

FDA remains firmly committed to fostering the development of new 
technologies to better understand BSE. Future scientific. research may 
uncover a way to definitively test for the presence of prohibited 
materials in animal feed or to detect the presence of the agent that 
causes BSE in feed. FDA will continue to evaluate and make changes to 
the sampling assignment to reflect current validated scientific 
methods. 

In conclusion, FDA's Ruminant Feed Ban domestic sampling program is 
unlike other sampling programs where analytical methods are available 
to detect specific pathogens, toxins, or residues that are considered 
adulterants in FDA regulated products. For the Ruminant Feed Ban 
domestic sampling and analysis assignment evaluated by GAO in this 
study, the analytical contribution to assurances of compliance is more 
limited due to the constraints of the currently available test 
methodologies and existing exemptions for mammalian proteins in the 
regulation. For this reason, FDA conducts only limited testing under 
this surveillance assignment (in FY'04 this assignment only accounted 
for approximately 600 of the over 46,000 samples collected and analyzed 
by FDA regulatory laboratories). The domestic sampling and analysis 
component of the Ruminant Feed Ban may become a more valuable tool in 
the future, or may be discontinued, depending on development of new 
technologies or methods, enhancements to strengthen the feed ban that 
may be instituted, and/or a better scientific understanding of the 
epidemiology and pathogenesis of the BSE agent in animal populations. 

GAO Recommendations for Executive Action: 

To ensure that the feed testing program is a useful tool for helping 
FDA oversee industry compliance with the Feed Ban rule, GAO recommends 
that the Commissioner of FDA take the following three actions: 

* Fully implement the June 2005 field management directive and July 
2005 assignment memorandum revising the feed testing program. 

FDA Response: 

FDA appreciates GAO's recognition that the June 2005 field management 
directive and July 2005 BSE sampling assignment memorandum provide 
controls that enhance the agency's ability to evaluate the utility of 
this program in the future and intends to implement the current 
versions or subsequent versions of these documents that are issued. 

* Establish timeframes for laboratories to complete sample analysis to 
minimize cattle's exposure to potentially contaminated feed. 

FDA Response: 

FDA field laboratories currently have timeframes for completing sample 
analyses. The timeframes are 10 business days for compliance samples 
and 20 business days for surveillance samples. The samples taken under 
the BSE feed sampling assignment are surveillance samples, as opposed 
to compliance samples. Because import and "for- cause" samples are 
given priority, laboratories are sometimes unable to complete analysis 
of surveillance samples within 20 days. 

FDA will evaluate the number of days it has taken laboratories to 
complete their analyses under this program and will set new timeframes, 
if needed. This evaluation will be completed as part of management's 
oversight of the implementation of the June 2005 field management 
directive and July 2005 sampling assignment memorandum. 

GAO asserts that expedited laboratory testing will minimize cattle's 
exposure to potentially contaminated feed. FDA disagrees that 
laboratory testing timeframes to complete sample analysis for domestic 
surveillance samples are a critical element to minimize cattle's 
exposure to potentially contaminated feed. FDA relies on its overall 
inspection program, rather than analytical methods alone, to enforce 
the feed ban. 

* Require FDA headquarters' managers to exercise sufficient oversight, 
with periodic reports from FACTS or other management controls and 
identify appropriate performance indicators for the feed testing 
program to assure that the program operates as intended and achieves 
its intended goals. 

FDA Response: 

FDA disagrees with GAO's assertion that the sampling assignment was 
poorly implemented, and that FDA did not adequately oversee the 
assignment. The FACTS database was utilized in fully describing feed 
collection and laboratory activities. Feed collection information 
includes feed type, a feed description, a feed label summary, and the 
identity of the associated feed manufacturer and distributor. 
Laboratory information includes a full description of the test 
utilized, the observations of the analyst, and the laboratory 
classification of the sample results. Spreadsheets containing both 
types of information were generated on a weekly basis and distributed 
throughout FDA headquarters. These data spreadsheets were constantly 
examined and utilized by FDA headquarters' managers in providing 
feedback and education to District and laboratory personnel. 

Although the spreadsheet did not contain information regarding follow- 
up activities, FDA headquarters was in constant communication with 
Districts concerning questions and issues related to sample collection 
and interpretation of analytical results. This oversight and 
communication were reflected in an improved understanding of the 
sampling assignment's goals by Districts and laboratories as time 
progressed. 

The experiences gained through the August 2003 sampling assignment 
allowed for significant improvements as represented in the revised 
assignment issued in July 2005. The issuance of the revised assignment 
was somewhat delayed until all laboratories were fully prepared to 
utilize the new PCR method. The recording of follow-up summaries in the 
FACTS database and the development of additional database reports have 
enhanced the ability of FDA headquarters to more effectively evaluate 
all aspects of activities related to feed testing efforts. 

FDA notes that the assignment may again be revised based on experiences 
involving the incorporation of the new PCR method. FDA will continue to 
monitor and evaluate feed testing to determine whether it is a 
worthwhile component of FDA's long-term BSE prevention efforts. 

The following are GAO's comments on the Food and Drug Administration's 
letter received on Monday, September 19, 2005. 

GAO comments: 

1. At our September 7, 2005, exit meeting with FDA, FDA raised concern 
that the draft title could be taken out of context by U.S. trading 
partners who would not read the report and could construe that we were 
talking about weaknesses in FDA's bovine spongiform encephalopathy 
(BSE) oversight efforts in general. We told FDA officials that we would 
look at the title in that light. We revised the title to better ensure 
that readers would realize by the title alone that the report focused 
on the small feed testing program that FDA started in August 2003. At 
the exit meeting, FDA also provided us with an untitled and undated 
document that FDA officials identified as a list of time frames for 
laboratories to complete analysis on various testing programs, 
including the feed testing program. The targeted time frame for the 
feed testing program--from receipt of sample to classifying the sample 
in FACTS--was 20 working days. As our report states, the time frames 
from sample collection to documenting the laboratory result in FACTS 
exceeded 30 days for 473 of the 989 samples we assessed. These included 
17 samples that took from 60 to 100 days and 21 that took more than 100 
days. FDA officials agreed that these time frames were unacceptable and 
did not challenge our analysis. FDA did not give us data on whether 
laboratories are meeting the 20-working day target. Also, FDA could not 
provide information on the time it took for districts to follow up and 
make a final determination on the 215 potential violations we report 
because it did not track those time frames. The timeliness of the 
entire process from sample collection to final determination is a 
factor that directly affects cattle's exposure. The second 
recommendation in our draft report initially recommended that FDA 
establish time frames for laboratories to complete sample analysis to 
minimize cattle's exposure to potentially contaminated feed. Because 
FDA did not provide data to assess whether the delays are occurring 
during sample collection, laboratory analysis, or follow-up, we revised 
our recommendation to address the need to minimize the overall time 
frame to protect cattle. 

2. FDA expressed concern that we were issuing a report that focused on 
one small aspect of its BSE oversight efforts. We agree that the feed 
testing program is a small component of FDA's overall efforts, but it 
vies for FDA's limited BSE oversight resources. As we pointed out in 
our more comprehensive February 2005 report--Mad Cow Disease: FDA's 
Management of the Feed Ban Has Improved, but Oversight Weaknesses 
Continue to Limit Program Effectiveness (GAO-05-101; Feb. 25, 2005)--we 
looked at this small program separately because FDA did not provide 
program data in time for its inclusion in the broader report. 

3. FDA stated that its field staff spent nearly 600 hours to provide 
information to us, and headquarters also spent time to compile 
information and to respond to questions about the program. We had to 
collect follow-up information directly from staff because it was not 
readily available in FDA's FACTS data system or other electronic data 
systems. We specifically asked FDA not to create documents or compile 
data after the fact. While routine interviews are always involved to 
clarify our understanding of agency documents and data, this study was 
designed and intended to be primarily an analysis of FDA data on the 
program. 

4. As we note in comment 1, we revised our second recommendation to 
address the need to minimize the overall time frame to protect cattle. 
Comments 11 and 13 discuss FDA's concerns with the other two 
recommendations. 

5. FDA points out that it also has a feed testing program for imported 
feed and feed ingredients. Our report focused on the domestic feed 
testing program that FDA identified as a component of its BSE oversight 
during our earlier study, which resulted in the February 2005 report. 
We did not assess the import feed testing program. 

6. We last examined USDA and other federal BSE detection and prevention 
efforts--other than FDA--in 2002 in a report entitled Mad Cow Disease: 
Improvements in the Animal Feed Ban and Other Regulatory Areas Would 
Strengthen U.S. Prevention Efforts (GAO-02-183; Jan. 25, 2002): 

7. We agree with FDA that feed testing alone may not be able to verify 
the presence of prohibited material and that follow-up is necessary to 
determine whether the feed ban has been violated. Thoroughly 
documenting follow-up actions and the rationale for compliance 
determinations is critical to FDA's effective oversight of the feed 
ban. Our report recommends that FDA fully implement the 2005 directive 
and revised assignment that require its districts to thoroughly 
document the basis for their decisions. 

8. FDA maintains that it is assessing the potential usefulness of feed 
testing. Because feed testing is using FDA's limited BSE oversight 
resources, it is imperative that FDA properly exercise oversight of the 
program by evaluating the costs and benefits, developing measurable 
goals, and periodically assessing trends to optimize the use of these 
resources. We believe that implementing our recommendations will help 
FDA in its assessment. 

9. FDA describes the refinement of PCR technology in the context of an 
ongoing technology evaluation. FDA officials made similar comments 
during the course of our work. However, FDA did not provide any 
information on the evaluation criteria it is using to measure 
performance or on the cost of developing and refining PCR technology. 

10. The draft and the final report clearly state that the feed testing 
program is a small part of BSE's oversight effort and provides FDA with 
additional information about some sample feed. 

11. We are continuing to include a recommendation that FDA fully 
implement the June 2005 directive and the July 2005 revised assignment 
to help ensure that FDA maintains its momentum and commitment to the 
new procedures. Because the feed testing program draws resources from 
FDA's BSE oversight activities, it is important that FDA avoid 
implementation weaknesses that limited the potential usefulness of 
testing under the 2003 assignment. In conjunction with our other 
recommendations, fully implementing the directive and the revised 
assignment will help FDA better assure the usefulness of the feed 
testing program as a tool in its BSE oversight efforts. 

12. See discussion of the second recommendation in comment 1. 

13. FDA disagreed with our assertion that the sampling assignment was 
poorly implemented and that it did not adequately oversee the program. 
According to FDA, the FACTS database was used to fully describe feed 
collection and laboratory activities and spreadsheets containing 
collection and laboratory information were distributed weekly and 
reviewed by FDA headquarters managers. FDA provided a copy of this 
spreadsheet that contained counts of the number of samples taken and 
the laboratory classification. However, each week's spreadsheet 
overrode the week before, and FDA's managers did not maintain previous 
versions. Furthermore, they could not provide any report that 
summarized their weekly review. Thus, FDA's managers could not do any 
comparative analysis, such as examining the type of feed sampled across 
districts. FDA also did not track or have any reports on follow-up 
activities or determinations to assess whether, for example, districts 
were using the same criteria. We envision a more substantive and 
meaningful oversight that might include comparisons of follow-up 
findings across districts, analyses of the number and types of new 
firms identified, assessments on how frequently follow-up involved only 
a file review or an on-site inspection, and decisions about what 
documents are consistently proving the most useful in expediting follow-
up. These or other types of analyses give managers better information 
to assess program performance. 

[End of section] 

(360563): 

FOOTNOTES 

[1] Ruminants are animals with four-chambered stomachs, including, but 
not limited to, cattle, buffalo, sheep, goats, deer, elk, and antelope. 
For the purpose of this report, "cattle" refers to cattle and all other 
ruminant animals and "cattle feed" refers to feed for cattle and other 
ruminant animals. 

[2] 21 C.F.R. 589.2000. 

[3] Adding protein (derived from animals or plants) to feed is a common 
nutritional practice used to speed animal growth. 

[4] GAO, Mad Cow Disease: Improvements in the Animal Feed Ban and Other 
Regulatory Areas Would Strengthen U.S. Prevention Efforts, GAO-02-183 
(Washington, D.C. Jan. 25, 2002). 

[5] GAO, Mad Cow Disease: FDA's Management of the Feed Ban Has 
Improved, but Oversight Weaknesses Continue to Limit Program 
Effectiveness, GAO-05-101, (Washington, D.C. Feb. 25, 2005). 

[6] The PCR test works by aiding in the differentiation of 
mitochondrial DNA between animal species. 

[7] The feed-ban rule is based on FDA's authority to regulate food 
additives, 21 U.S.C.  321(s), 348, as well as other authorities. 

[8] Plate waste is discarded meat and other food offered for human 
consumption from institutions, restaurants, and other dining 
facilities, which are collected by processors, recooked to eliminate 
bacteria, and used in animal feed as a protein source. Gelatin is made 
from boiling animal bones, cartilage, tendons, and skin. 

[9] FDA has published two advance notices of proposed rulemaking 
requesting comments and information revising the ban to, among other 
things, end most of the exemptions. 

[10] In September 2005, FDA announced that it would propose regulations 
that parallel regulations that Canada recently announced, banning at- 
risk tissue--brains, spinal cords, and other parts that may carry mad 
cow disease--from feed for all animals including chicken, pigs, and 
pets. 

[11] FDA Center for Veterinary Medicine Using the Science and Law to 
Protect Public and Animal Health, Annual Report Fiscal Year 2004, 
October 1, 2003 - September 30, 2004. Rockville, MD: May 2005. 

[12] See GAO, Results-Oriented Government: GPRA Has Established a Solid 
Foundation for 

Achieving Greater Results, GAO-04-38, (Washington, D.C. March 10, 
2004); Managing for Results: Strengthening Regulatory Agencies' 
Performance Management Practices, GAO/GGD-00-10 (Washington, D.C. Oct. 
28, 1999); Standards for Internal Controls in the Federal Government, 
GAO/AIMD-00-21.3.1, (Washington, D.C. Nov. 1999). 

[13] Pub. L. No. 103-62, 107 Stat. 285 (1993).