IN THE UNITED STATES DISTRICT COURT
FOR THE NORTHERN DISTRICT OF TEXAS
SUPREME BEEF PROCESSORS, INC., §
v. § Civil No. 3-99CV2713-G
UNITED STATES DEPARTMENT OF §
BRIEF IN SUPPORT OF SUPREME BEEF
PROCESSORS, INC.’S MOTION FOR SUMMARY JUDGMENT
The amici curiae submitting this Brief in Support of Supreme Beef Processors, Inc.’s Motion for Summary Judgment are non-profit trade associations that represent meat packers and processors.
I.INTERESTS OF AMICI With one exception, the amici submitting this Brief are identified and their interests in this litigation are described in the Amicus Curiae Brief of the National Meat Association, the American Association of Meat Processors, the Southeastern Meat Association, and the Southwest Meat Association at 1-3 (Dec. 9, 1999). The exception is the North American Meat Processors Association (NAMP). NAMP is an international trade association representing the needs of meat, poultry, seafood, and game processors, many of whom are subject to federal inspection under the Federal Meat Inspection Act, 21 U.S.C. § 601 et seq, and to the pathogen reduction performance standard for Salmonella, 9 C.F.R. § 310.25(b). NAMP is headquartered at 1920 Association Drive, Suite 400, Reston, Virginia 20191-1547.
These associations and their members are committed to the safest possible products. Meat safety sells meat. Amici appear in this action because of their belief that the Defendant’s current ground beef Salmonella testing program is both an illegal and ineffective use of food safety resources, which could better be targeted to testing for Salmonella at points in the process when there are meaningful opportunities for pathogen reduction.
An agency of Defendant, the Food Safety and Inspection Service (FSIS), has asserted the authority to suspend inspection if a beef grinder fails to comply with the Salmonella performance standard regulation, 9 C.F.R. § 310.25(b). Amici respectfully submit that FSIS lacks the statutory authority to suspend inspection services for non-compliance with the salmonella performance standard. Moreover, as demonstrated below, the ground beef Salmonella performance standard is fatally flawed, both procedurally and substantively: procedurally because the agency did not provide a meaningful opportunity for comment before it promulgated the standard; substantively because the ground beef performance standard lacks any valid justification.
A. Events Leading to The Adoption of the Salmonella Performance Standard
1. The FSIS proposes to modernize meat inspection procedures .
In 1995, as a response to a variety of studies that established the "need for fundamental change in the FSIS [meat and poultry inspection] program" to improve food safety, 60 Fed. Reg. 6,774, 6,783, col. 3 (February 3, 1995), FSIS proposed to modernize its inspection system. This massive proposal, entitled "Pathogen Reduction: Hazard Analysis, Critical Control Points (HACCP) Systems," but commonly referred to as the "Mega-Reg," was designed to replace the antiquated organoleptic inspection system with a modern HACCP system that identifies food safety hazards and uses scientifically validated controls to eliminate or prevent the hazards. Id. at 6,774.
2. Salmonella and the proposed modernization
One of the many changes FSIS proposed was that those establishments producing raw meat and poultry products regularly conduct microbial testing for Salmonella " to verify the performance of an establishment’s system of process controls." 60 Fed. Reg. at 6,802, col. 1.
Under the proposal, FSIS would compare the establishment's Salmonella results to a certain "interim target" identified in the proposed regulation. Proposed 9 C.F.R. § 310.25 (60 Fed. Reg. at 6,837). To establish the "interim target," FSIS "used data from its Nationwide Microbiological Baseline Data Collection Program." Id. at 6,800, col. 2. In the case of ground beef, the agency relied on its July 1995 Draft Report, "Nationwide Federal Plant Raw Ground Beef Microbiological Survey, August 1993 to March 1994" (hereinafter "1995 Draft Survey"); [App. 1-25]) which reported a Salmonella prevalence of 4% based on the 563 samples collected in this survey. 60 Fed. Reg. at 6,800, col. 3.
Under the proposed regulation,, a grinding establishment would be deemed in compliance with this 4% "interim target" if it had no more than two positive Salmonella samples per thirty-eight samples using the statistical technique of a moving window. Proposed 9 C.F.R. § 310.25(e) (60 Fed. Reg. at 6,839). According to the agency, this target would ensure that if the establishment was actually in compliance (operating at or below the 4% interim target), it would successfully pass the tests 80% of the time. 60 Fed. Reg. at 6,804, col. 1. Stated another way, an establishment actually in compliance with the interim target would fail one out of five times. Id. at col. 2.
3. Proposed enforcement provisions to achieve compliance
The 1995 proposal also contained enforcement provisions. Under the proposed Salmonella testing requirement, if an establishment was operating over the "interim target" level, it was expected to undertake corrective action until it came into compliance. Proposed 9 C.F.R. § 310.25(e) (60 Fed. Reg. at 6,837-6,839). There were no provisions in the proposed rule to suspend or withdraw inspection in the event of noncompliance, nor did the proposal indicate that non-compliance would have any affect on the establishment’s HACCP or sanitation programs.
4. Comments on the FSIS proposal
NMA and the other amici strongly supported the move towards a more scientific inspection system. See, e.g., National Meat Association Comments on Pathogen Reduction: HACCP Systems, dated July 5, 1995 (hereinafter "NMA Comments") (Mucklow Affidavit, para 3, Attachment 1); [App. 26, 30-46]. On the issue of the use of Salmonella as a process control indicator, amici voiced their opposition as did other commenters.
During the rulemaking, FSIS published a set of issue papers in the Federal Register on October 24, 1995. 60 Fed. Reg. 54,450-54,457. In the third issue paper, the agency noted it was considering changing the Salmonella target to generic E. coli. Id. at 54,453. With no antecedent whatsoever, the agency went on to note that it "could" adopt a performance standard for controlling and reducing pathogens, such as Salmonella. Id. If the establishment failed to meet such standard, it "would be required to take corrective action and could be subject to other regulatory action, as appropriate." Id. at col. 3.
5. The final regulation issued after consideration of comments, included a new test and new enforcement provisions .
On July 25, 1996, FSIS issued its final rule. 61 Fed. Reg. 38,805-39,029. On the issue of microtesting to assess process control, FSIS adopted the position asserted by NMA and others that Salmonella is not an appropriate indicator of process control. 61 Fed. Reg. at 38,864-865. Accordingly, FSIS modified the proposal so as to use generic E. coli as the indicator microorganism to assess process control. 9 C.F.R. § 310.25(a).
Notwithstanding this modification, FSIS retained Salmonella testing, re-characterizing the "interim target" as "Salmonella performance standards." 9 C.F.R. § 310.25(b). This new test, not previously part of the proposed regulations, required FSIS to sample raw products for Salmonella and measure the establishment’s results against the national prevalence, as determined by the agency’s nationwide data collection programs and surveys. Id. The agency retained the 80% confidence level that an establishment actually meeting the standard would pass the sample set. 61 Fed. Reg. at 38,847, col. 3. In other words, there was still a one in five chance that an establishment operating in compliance with the standard would nonetheless exceed the permissible positives in a single sample set. Id. If the establishment failed on three consecutive tests, FSIS would summarily suspend inspection. 9 C.F.R. § 310.25(b)(3).
6. Purported justification for the new performance standards is questionable at best .
The new Salmonella performance standards were intended to be a technology-forcing regulation to "encourage" establishments to reduce the incidence of this pathogen in raw products. 61 Fed. Reg. at 38,846, col. 1. As the benchmark, FSIS adopted the prevalence levels found in the "April 1996 Final Report, Nationwide Federal Plant Raw Ground Beef Microbiological Survey, August 1993 - March 1994" (hereinafter "1996 Final Report"); [App. 58-66]. According to FSIS, because this prevalence represented the level which could be achieved with current technology, it was a reasonable standard to impose on industry. Id. at 38,852, col. 3. See also, Id. at 38,836, col. 2 ("The targets currently are set at the national baseline prevalence of contamination and reflect what is achievable using current technology"). However, as shown below, there was no basis to conclude that the level adopted as the Salmonella performance standard for ground beef accurately represented what was and is technologically achievable.
a. The survey data on which the performance standards were based were flawed .
In the case of ground beef, the agency adopted 7.5% as the acceptable Salmonella prevalence rate, based on the findings from the 563 samples analyzed and summarized in the April 1996 Final Report. 61 Fed. Reg. at 38,846-38,847. The prevalence level for ground beef was different than the 4% level identified in the proposed rule, even though the same 563 sample analyses were used. To determine compliance with this raw ground beef prevalence, the standard was set at no more than five positives per sample set of fifty-three samples (not using the moving window as proposed) and with the FSIS laboratory, not a private laboratory, analyzing the samples. 9 C.F.R. § 310.25(b)(1)-(2).
However, the 1996 Final Report was boggy ground upon which to build. FSIS should have anticipated difficulties with the Salmonella performance standard for ground beef given the limitations expressly mentioned in the FSIS-generated 1996 Final Report. First, in discussing the data limitations, the Report noted that "[t]he survey was designed to provide estimates of national prevalence." 1996 Final Report at 4 (emphasis added); [App. 62]. Second, the Report warned that the sampling "occurred over a relatively short period of time (August 1993 - March 1994) and not over a yearly period. As a result, the estimates may not reflect possible seasonal differences." Id. at 5; [App. 63]. Indeed, a review of the raw data, which identify the sample collection dates, show that approximately 80% of the samples were drawn in the first three months of the eight month survey. Third, the standard error of the prevalence summarized in the Report is quite large (3.1) given the average of 7.5%. Id. at Table 1; [App. 64]. Fourth, "plants with higher production volumes will have a greater impact" on the prevalence standard. Id. at 4; [App. 62]. Therefore, the prevalence unlikely reflects what smaller plants are capable of achieving.
The failure of FSIS to understand how the 1996 Final Report calculated the prevalence continues. In a January 10, 2000, letter from the FSIS Administrator to NMA, the Administrator discussed whether the 563 samples analyzed in the 1996 Final Report represented an estimated 94 million pounds of weekly ground beef production. January 10, 2000, Letter from FSIS Administrator Thomas J. Billy to Rosemary Mucklow, NMA (Mucklow affidavit, para 9 (Attachment 4); [App. 27, 53-57]. In that letter, Administrator Billy stated that the estimate of production did not enter into the development of the ground beef Salmonella standard. Id. at 3; [App. 55]. But, the summary of how the ground beef standard was calculated, attached to his letter, notes that: "FSIS took the results from the 563 samples, weighed the results using the production data and adjusted the results for non-responses." Id. at 5 (emphasis added); [App. 57].
Beyond the limitations of the data, any analysis of the data would have revealed significant variation due to the geographic location of the grinders. NMA requested and finally received the actual raw sample data underlying the 1996 Final Report and the FSIS Salmonella compliance testing data pursuant to Freedom of Information Act requests filed in September and December 1999. Mucklow affidavit, para 5-6; [App. 26-27]. NMA staff reviewed the actual raw sample data that formed the basis for the Salmonella prevalence level. From this data, NMA determined the geographic locations where the samples were drawn. Id. para 5 and attachment 2; [App. 26, 47]. When the raw data used in the 1996 Final Report is analyzed by comparing establishments in the north with establishments in the south, approximately 73% of the 563 samples analyzed were from the northern states. Id. Yet, only 55% of the positive findings were from the north. The southern states accounted for only 27% of the total samples but 45% of the positives. Id. This phenomena is more pronounced in Texas. Although ground beef produced in Texas only constituted 5.2% of the 563 samples analyzed, ground beef from Texas constituted 13.8% of the positive findings. Id.
b. The FSIS failed to identify technology to meet the performance standards for grinders .
FSIS justified the Salmonella performance standard by asserting that "the targets currently are set at the national baseline prevalence of contamination and reflect what is achievable using current technology." 61 Fed. Reg. at 38,836, col. 2. That generalization is without support in the administrative record.
On how industry could use current technology to achieve compliance with the Salmonella performance standards, FSIS provided detailed guidance only for slaughtering establishments. 61 Fed. Reg. at 38,846, col. 3. According to the agency: "Slaughter establishments . . . have available a wide range of technologies shown to reduce the level of pathogens." Id. The agency then went on in the preamble to list such technologies, including steam and hot water pasteurization, steam vacuuming, and whole carcass rinsing with anti-microbial substances. Id.
However, the listing of technologies available to establishments manufacturing ground beef was not as detailed:
Establishments producing raw ground product from raw meat or poultry supplied by other establishments cannot use technologies for reducing pathogens that are designed for use on the surfaces of whole carcasses at the time of slaughter. Such establishments may require more control over incoming raw product, including contractual specifications to ensure that they begin their process with product that meets the standard, as well as careful adherence to their Sanitation SOP’s and HACCP plan.
Id. In other words, FSIS could identify no controls a grinder can employ to prevent the introduction of Salmonella into the grinder’s facility. As FSIS recognized:
[grinding] establishments do not have the same opportunities to reduce Salmonella levels as do slaughter establishments. They basically must depend on the Salmonella levels of their incoming product in order to meet the performance standards."
61 Fed. Reg. at 38,977, col. 3. This "technology" is not viable in practice. First, the slaughter performance standard for steers/heifers is lower than that for cows/bulls, but there is a single performance standard for ground beef from both sources. Thus, grinders using the latter are more challenged in meeting the standard. Mucklow Affidavit at para 8; [App. 27 Second, grinders lack any effective means to control Salmonella in their operations. The agency’s assertions that grinders could eliminate suppliers whose products test positive for Salmonella or could impose purchase specifications are not practical given the realities of the beef market. Id. para 7; [App. 27]. In addition, as the FSIS Deputy Administrator testified at the Preliminary Injunction Hearing, given the heterogeneous nature of pathogens in ground beef, a negative result from one sample does not mean a pathogen is absent from the product. Testimony of Dr. Kaye Wachsmuth, page 133, lines 6-11; [App. 126]. Thus, testing cannot ensure that pathogens are absent from the grinders’ raw materials.
Although FSIS did not identify any technology that can be used to control Salmonella in grinding operations, the agency did recognize that as a result of the grinding operation, "Salmonella is more likely present on raw, ground, or comminuted products than on the carcasses from which they are derived." 61 Fed. Reg. at 38,848, col. 1. However, this recognition only meant that agency compliance sampling for ground beef would take priority over compliance sampling of carcasses. Id.
In sum, the flawed baseline data and the illusory technology do not support the ground beef Salmonella performance standard. It should not be surprising that in summarizing the agency’s appraisal of the effect of the performance standard on the industry, the Administrator confessed: "the Agency did not know what proportion of establishments would have trouble meeting the standards (see Final Regulatory Impact Analysis)." Mucklow Affidavit, para 9, Attachment 4 at 3; [App. 27, 55]
c. In the final rule the FSIS admitted to many limitations in the performance standards .
The final preamble also eliminates certain factors as a justification for the performance standard. First, the standard does not use Salmonella as an indicator of a process control. FSIS admitted that E. coli was a more effective indicator of process control. 61 Fed. Reg. at 38,838, col. 3. Second, the performance standard is not based on a legal adulteration standard. The presence of Salmonella on raw products does not render such products adulterated. Id. at col. 2. Third, the standard is not a public health standard since it is not based on a quantitative assessment of the risk posed by any particular incidence of Salmonella. Id. at 38,847, col. 1. Otherwise, the Salmonella performance standards would be identical for all raw products: beef, pork, and poultry. Id. (the standards are not based on a "determination of a ‘safe’ incidence or level").
7. The new enforcement provisions came as a surprise .
Importantly, under the new rule, FSIS would take much more aggressive enforcement action than simply requiring an establishment to undertake corrective actions in the event of non-compliance with the standard. If an establishment failed the performance standard three times in a row, FSIS would suspend inspection services. 9 C.F.R. § 310.25(b)(3). Because inspection is required to process meat products, suspension of inspection effectively closes a company. Not even in the FSIS issue papers published in October of 1995 (60 Fed. Reg. at 54,453) did the agency indicate it would take this draconian enforcement action.
8. Flawed implementation of the new regulations
FSIS did promise it would administer and apply the performance standards "in a fair, equitable, and common sense manner." 61 Fed. Reg. at 38,849, col. 1. However, given the deficiencies in the development and adoption of those standards, it was inevitable that an establishment would fail its third sample set and be subject to FSIS enforcement action. Not surprisingly, in light of the data developed by FSIS in setting the standard, that failure occurred in Texas. In October of 1999, one of NMA's members, Supreme Beef Processors, Inc. was apprised by FSIS that it had failed the third consecutive Salmonella sample set. See Supreme’s Complaint at 6. Following unproductive discussions with agency officials, FSIS notified Supreme that inspection would be suspended at the facility on November 30, 1999. Id. at 7. The instant lawsuit grew from that notice.
Subsequent to this Court's order enjoining FSIS from enforcing the Salmonella performance standard at Supreme, FSIS officials publicly stated that, in their opinion, the order only applied to establishments operating in the Northern District of Texas. Mucklow affidavit para 10; [App. 27-28]. Accordingly, the agency has continued its Salmonella testing and, in at least one instance, has initiated enforcement action against another NMA member, also a Texas grinder, that failed its third sample set.
IV.ARGUMENT AND AUTHORITIES
The Defendant Lacks the Statutory Authority To Suspend Inspection For Failure to Comply with a Salmonella Performance Standard
A government agency may only exercise the powers granted by Congress. Bowen v. Georgetown University Hospital, 488 U.S. 204, 208 (1988); Federal Trade Commission v. National Lead Co., 352 U.S. 419, 428 (1957). "[T]he determinative question is not what [an agency] thinks it should do but what Congress has said it can do." Civil Aeronautics Board v. Delta Airlines, 367 U.S. 316, 322 (1961). Hence, regardless of whether an agency strongly believes it needs certain powers or whether an agency has promulgated a regulation asserting certain powers, the Administrative Procedure Act (APA) forbids an agency from acting "in excess of statutory jurisdiction, authority, or limitations, or short of statutory right." 5 U.S.C. § 706(2)(C).
In the regulation adopting the Salmonella performance standard, FSIS provides that if an establishment fails three consecutive Salmonella sets, FSIS will suspend inspection. 9 C.F.R. § 310.25(b). The statute does not permit this sanction for such a reason. The plain language of the statute, the 1986 Amendments, and the agency’s repeatedly unsuccessful attempts to obtain this authority far outweigh any unspecific or strained interpretation of existing authority. Because the enforcement provision is in excess of the agency’s authority, it is unlawful under the APA.
1. The FMIA does not provide express authority to suspend inspections here .
There are only two instances where the FMIA expressly authorizes FSIS to remove its inspectors from a processing establishment. First, FSIS can unilaterally suspend inspection services if the official establishment refuses to destroy product that the inspector has condemned. 21 U.S.C. §§ 604, 606. Second, FSIS can withdraw inspection if the establishment or a responsibly connected individual has been convicted of certain crimes, but only after a formal administrative proceeding, during the pendency of which the establishment remains in operation. 21 U.S.C. § 671.
Other than in these two discrete instances, there is no express authority in the FMIA to suspend inspection, unilaterally or otherwise. Nor, in light of subsequent legislative activities, can FSIS assert that it has the implied power.
2. Legislative activity is inconsistent with implied authority .
The FMIA was amended by the Processed Products Inspection Improvement Act of 1986, 100 Stat. 3556 (P.L. 99-641)(hereinafter 1986 Amendments). One of the provisions of the 1986 Amendments authorized FSIS to suspend inspection services if the continued operation of the official establishment posed a direct and substantial danger to the public health. Section 403(b)(6)(amending 21 U.S.C. § 671 by adding new subsection (d)); [App. 128-131]. However, FSIS could not impose this extreme sanction unilaterally. The sanction could only be imposed by a Federal District Court upon application from the agency. Id. Before the agency could file suit, it had to give the official establishment notice of the perceived deficiencies and the opportunity to bring its operations into compliance. Id. Before the District Court would order suspension, FSIS had to demonstrate, "on the basis of clear and convincing evidence, that [the establishment] repeatedly failed to comply [with the FMIA and regulations], in a manner that poses a direct and substantial threat to the public health or safety." Id.
The 1986 Amendments, including the suspension provisions, sunsetted on November 11, 1992 (§ 408(b)), Id., without ever having been used by FSIS.
If FSIS had the implicit authority to suspend inspection unilaterally, the 1986 Amendments would not merely be redundant, they would have actually restricted the agency’s authority by mandating that the sanction be imposed by a District Court and then only after a hearing. Moreover, the Conference Report accompanying the legislation would not have characterized this sanction as an extreme measure to be used in rare occasions if it was a restriction of existing implied authority. H. Conf. Report 99-995, 99th Cong. 2d Sess. (1986), reprinted in 1986 U.S.C.C.A.N. 6083 ("The conferees note that this provision constitutes extraordinary authority") (emphasis added).
On several occasions in the 1990's, both before and after the FSIS Mega-Reg was finalized, Defendant USDA sought additional legislative authority from Congress in the 103rd Congress; the 104th Congress; the 105th Congress and the 106th Congress. See affidavit of Stephen L. Lacey; [App. 132, 133-284]. In each and every one of these Bills, there were provisions to authorize the unilateral suspension of inspection services. None of these Bills passed either the House or the Senate. Repeatedly seeking express legislative authority is inconsistent with the agency’s assertion that it has implicit authority to suspend inspection services unilaterally.
3. Existing authority does not authorize suspension of inspections .
In the preamble to the Mega-Reg, FSIS discusses its authority to promulgate the regulations generally and the enforcement provisions specifically. 61 Fed. Reg. at 38,851-852. However, the agency does not, and indeed cannot, cite to any provision of the FMIA which authorizes suspension in the event the official establishment fails the Salmonella performance standard.
Admittedly, there is some discussion in the preamble regarding suspension because of insanitary conditions or the agency’s obligation to inspect and mark "inspected and passed" products that are not adulterated. Id, However, this authority does not and cannot justify the suspension of inspection in the context of the Salmonella performance standard.
The section of the FMIA dealing with sanitary conditions provides, in relevant part: "where the sanitary conditions of any such establishment are such that the meat or meat food products are rendered adulterated, [FSIS] shall refuse to allow said meat or meat food products to be labeled, marked, stamped, or tagged as ‘inspected and passed.’" 21 U.S.C. § 608. This section would not justify suspension because: (1) Salmonella on raw product does not render the product adulterated, 61 Fed. Reg. at 38,838, col. 2; and (2) FSIS has conceded that grinders "basically must depend on the Salmonella levels of their incoming product in order to meet the performance standards." 61 Fed. Reg. at 38,977, col. 3. Thus, any failure of the Salmonella performance standard is a function of incoming materials, not sanitation at the grinding facility.
The section of the FMIA dealing with the refusal to mark product "inspected and passed" also does not support suspension. That section provides that the "inspector shall mark . . . ‘inspected and passed’ all such products found to be not adulterated." 21 U.S.C. § 606. Salmonella is not an adulterant on raw ground beef. 61 Fed. Reg. at 38,838, col. 2. Therefore, the finding of Salmonella does not foreclose the application of the "inspected and passed" mark on the product.
Most importantly, neither of these authorities speak to or authorize suspension. Indeed, the authorities only permit FSIS to refuse to mark products "inspected and passed." The inspector must remain at the establishment.
B. The Defendant Failed to Follow the APA in Promulgating the Salmonella Performance Standard
Under the APA, an agency action is invalid, in part, if it is "without observance of procedure required by law." 5 U.S.C. § 706(2)(D). The provisions of the APA governing informal rulemaking, 5 U.S.C. § 553 (b) & (c), require that the public be given notice of a proposed regulation and a meaningful opportunity to comment. In the case of the Salmonella performance standard regulation, 9 C.F.R. § 310.25(b), the regulation is procedurally defective because FSIS did not provide a meaningful opportunity for comment before it promulgated the standard. Because the procedures were not followed here, the rule is flawed, unfair, and impracticable.
In determining whether the public received a meaningful opportunity to comment, courts will look, in part, to whether the regulation, as finalized, was significantly the same as the proposed rule. When an administrative agency adopts a provision that is substantially different from what was proposed and is not a logical outgrowth of the proposal, it is invalid. See Shell Oil v. E.P.A., 950 F.2d 741 (D.C. Cir. 1991); American Fed’n of Labor v. Donovan, 757 F.2d 330 (D.C. Cir. 1985). In other words, if the final rule bears little resemblance to the proposed rule, the agency must re-propose the final rule to provide parties the opportunity to review and comment on the change. Until the agency does so, there is no enforceable rule.
1. The differences between the proposal and final rule are significant .
In the Mega-Reg, as proposed, the Salmonella testing requirement had four components: (1) it was to be conducted by official establishments; (2) it was to assess whether the establishment’s process was in control (by means of the statistical device of a moving window); (3) compliance was measured against "interim targets;" and (4) failure to comply with the "interim targets" would require the establishment to take corrective action until it achieved compliance with the target. Proposed 9 C.F.R. § 310.25 (60 Fed. Reg. at 6,837-839).
When the Salmonella testing regulation was finalized, each and every major component of the proposal itemized above was changed. Now: (1) FSIS conducts testing at official establishments; (2) the agency testing is not to assess process control; (3) but to determine (by means of a sample set) whether the establishment is operating at a regulatory Salmonella prevalence performance level "achievable" with current technology; and (4) failure to comply with these regulatory performance standards would result in suspension of inspection upon the third failure. 9 C.F.R. § 310.25(b). These are significant differences.
On the issue of whether the FSIS laboratory or the establishment’s own laboratory/an outside laboratory conducts the test, it is noteworthy that tests conducted by FSIS laboratories are secret. There is no opportunity for a company to even review the procedures, let alone challenge the results of the agency’s testing. See Testimony of Dr. Kaye Wachsmuth at Preliminary Injunction Hearing, page 137, lines 7-8 ("We have a lot of checks and balances so [a bad result] doesn’t happen"); [App. 127]. This change is significant in that it forecloses any issue of laboratory error in determining compliance.
On the issue of process control, NMA and the other commenters successfully persuaded the agency that indicator organisms, such as generic E. coli, are superior to the use of pathogens in evaluating process control. Having been successful in this regard, it was not anticipated FSIS would retain Salmonella testing for another purpose. Without knowing the purpose in advance, there was no way the public could comment and perhaps persuade FSIS that another approach would be superior, as was the case with the change from Salmonella to E. coli for assessing process control.
There is a great difference between "interim targets" and "performance standards." According to FSIS, failure to meet an "interim target" or "performance criterion," the term used in the final rule, will trigger greater inspection oversight. 61 Fed. Reg. at 38,843-844. However, it will not form the basis for any enforcement action. Id. A "performance standard," on the other hand, will be used for a suspension action if there are three failures. Id. at 38,849, col. 1. Obviously, whether a test is used as the basis for a suspension action constitutes a significant difference. The requirement to take corrective actions versus the specter of suspension is as dramatic a difference as can be imagined. Indeed, it goes from the least dramatic tool in the agency’s enforcement arsenal to its most powerful.
2. The final rule was not a logical outgrowth of the proposed rule .
It cannot be said that the Salmonella performance standards were a logical outgrowth of the proposal. By parsing the proposed Salmonella testing requirement, it is clear that the E. coli final regulation, not the Salmonella performance standard, grew out of the mandatory testing proposal. The E. coli testing, as it appeared in the final regulation and as codified at 9 C.F.R. § 310.25(a), tracks the Salmonella testing proposal. The E. coli testing: (1) was to be conducted by the official establishments; (2) to assess the establishment’s process control (by means of the statistical device of a moving window); (3) compliance was measured against national baseline performance criteria; and (4) failure to operate in accordance with performance criterion would require the establishment to take corrective action until it achieved compliance.
The Salmonella performance standards were made out of whole cloth, with no antecedent in the proposed Mega-Reg. The publication of the issue papers on October 24, 1995, (60 Fed. Reg. at 54,453) did not amount to notice because it did not specify the agency’s intended implementation or enforcement. In light of the agency’s failure to provide the public with a meaningful opportunity to comment on the Salmonella performance standards, the procedures followed by FSIS were not in compliance with the provisions of the APA governing informal rulemaking, 5 U.S.C. 553 § (b) & (c), and thus were "without observance of procedure required by law." 5 U.S.C.§ 706 (2)(D).
The Salmonella Performance Standards are Arbitrary and Capricious Because They Lack a Rational Basis
Under the APA, an agency action shall be held unlawful and set aside if it is "arbitrary, [or] capricious." 5 U.S.C. § 706(2)(A). The Supreme Court has stated that an agency rule is arbitrary and capricious "if the agency . . . entirely failed to consider an important aspect of the problem, offered an explanation for its decision that runs counter to the evidence before the agency, or is so implausible that it could not be ascribed to a difference in view or the product of agency expertise." Motor Vehicle Mnfrs. Ass’n of U.S., Inc. v. State Farm Mutual Auto Ins. Co., 463 U.S. 29, 43 (1983). See also National Treasury Employees Union v. Helfer, 53 F.3d 1289, 1292 (D.C. Cir. 1995) ("[The] court must determine whether the agency’s decision making was reasoned, i.e., whether it considered relevant facts and explained the facts and policy concerns on which it relied"); Olenhouse v. Commodity Credit Corp., 42 F.3d 1560, 1574 (10th Cir. 1994) (holding that the duty of court reviewing agency action under "arbitrary and capricious" standard is to ascertain whether agency "examined the relevant data and articulated rational connection between facts found and decisions made"); Organized Fishermen of Florida v. Hordel, 775 F.2d 1544, (11th Cir. 1985), cert. denied, 476 U.S. 1169 (1986 ).
According to FSIS, the performance standards "currently are set at the national baseline prevalence of contamination and reflect what is achievable using available technology." 61 Fed. Reg. at 38,836, col. 2. As shown below, all the evidence in the agency’s possession runs counter to this justification, at least with regard to ground beef.
1. The ground beef prevalence level is not justified by the data .
The 1996 Final Report did not provide sufficient evidence to support the conclusion that the 7.5% level accurately represented what was achievable on a national level. The Report identified significant limitations with the extrapolation of the results. First, "[t]he survey was designed to provide estimates of national prevalence." 1996 Final Report at 4 (emphasis added); [App. 62]. Second, the Report warned that the sampling "occurred over a relatively short period of time" (August 1993 - March 1994) and not over a yearly period. Id. at 5; [App. 63]. (80% of the samples were drawn in the first three months.) Third, the standard error of the prevalence summarized in the Report is quite large (3.1) given the 7.5% average. Id. at Table 1; [App. 64]. Fourth, large plants may have skewed the average. Id. at 4; [App. 62].
2. FSIS failed to assess relevant factors in establishing the standard .
Even though the data underlying the report shows significant geographic differences, with southern states accounting for approximately 30% of the total samples analyzed, but 50% of the positive samples, see Mucklow affidavit para 5 (Attachment 2); [App. 26, 47]. FSIS did not mention this variable in any document. Yet, according to the NMA data analysis, these geographic differences may be the single most determinative factor on whether a grinder can comply with the Salmonella performance standard.
There were other areas that are simply not discussed at all by the agency, including: (1) the possibility of prevalence variation due to environmental temperature variation, a corollary to the possibility of geographic variation; (2) the effect of the geographic location of the raw meat supplier; (3) the effect of the higher Salmonella incidence level in cows/bulls on those grinders using such raw materials; and (4) the possibility of laboratory error.
3. FSIS failed to recognize the limitations of available technology .
The performance standards allegedly constitute the level that could be achieved with current technology. Id. at 38,852, col. 3. However, nowhere in the proposed rule, final rule, or in any document issued since, has FSIS identified a single technological device or process to prevent or eliminate Salmonella in raw ground beef. As the preamble to the final rule notes, grinders lack the opportunities to eliminate Salmonella; they "basically must depend on their suppliers". Id. at 38,977, col. 3.
Supplier control is not an effective technology, let alone practical. It should be noted that grinders by law must buy their raw materials from FSIS inspected slaughter establishments. Yet, neither FSIS nor a slaughter establishment can guarantee the raw materials are free of Salmonella. Indeed, the performance standard for beef carcasses expressly recognizes a prevalence of between 1 and 2.7 %. 9 C.F.R. § 310.25 (Table 2). Likewise, testing does not guarantee the absence of Salmonella. Wachsmuth Testimony at 133, lines 6-11; [App. 126]. Moreover, certifications or testing assumes a supplier will be amenable to such an option or that a grinder can cease doing business with a particular supplier. That solution is simply not practical given the realities of the beef market. Mucklow Affidavit, para 7; [App. 27].
4. There are no other justifications FSIS can advance for the standards .
FSIS cannot claim some other justification for the performance standards. FSIS cannot claim the standards evidence a grinder’s process control -- the E. coli is the better indicator of that. 61 Fed. Reg. at 38,838, col. 3. FSIS cannot claim that the standard measures adulteration or reflects on the grinder -- Salmonella is not an adulterant in raw ground beef (Id. at col. 2), and non-compliance with the standard is "basically" caused by the incoming raw material from a FSIS-inspected facility (Id. at 38,977, col. 3); implicitly then, noncompliance is not primarily the fault of the grinder. FSIS cannot claim the standard is a food safety standard because FSIS admitted that the standards are not based on a "determination of a ‘safe’ incidence or level." Id. at 38,847, col. 1.
In short, there are no facts or data that support the agency’s assertion that the Salmonella performance standard for ground beef is achievable through current technology. Because this was the justification for the standard, and it is not correct, the regulation is arbitrary and capricious under the APA, 5 U.S.C. § 706(2)(A), and is therefore unlawful.
For the reasons discussed above, the Plaintiff’s Motion for Summary Judgment should be granted.