Pathogen Reduction; Hazard Analysis and Critical Control Point (HACCP) Systems; Technical Corrections and Amendments

DEPARTMENT OF AGRICULTURE

Food Safety and Inspection Service

9 CFR Parts 308, 310, 381, and 416

[Docket No. 93-016T]

RIN 0583-AC28

Pathogen Reduction; Hazard Analysis and Critical Control Point

(HACCP) Systems; Technical Corrections and Amendments

AGENCY: Food Safety and Inspection Service, USDA.

ACTION: Final rule.

-----------------------------------------------------------------------

SUMMARY: FSIS is making technical corrections and amendments to the

final rule, ``Pathogen Reduction; Hazard Analysis and Critical Control

Point (HACCP) Systems,'' published on July 25, 1996. This document

responds to technical and scientific questions raised in the final rule

regarding E. coli testing and to issues discussed at the ``Technical

Conference Regarding E. coli Verification Testing,'' the ``Pathogen

Reduction/HACCP National Implementation Conference,'' and the

``Regional Implementation Conferences.'' Also, this document clarifies

ambiguities brought to FSIS' attention and provides guidance on various

technical issues. Additionally, this document corrects inadvertent

omissions and addresses minor editorial oversights.

EFFECTIVE DATE: June 12, 1997.

ADDRESSES: Reference materials cited in this docket will be available

for public inspection in the FSIS Docket Room, Room 3806, 1400

Independence Ave SW, Washington, DC 20250 from 8:30 a.m. to 1:00 p.m.

and from 2:00 p.m. to 4:30 p.m., Monday through Friday.

FOR FURTHER INFORMATION CONTACT: Ms. Patricia Stolfa, Assistant Deputy

Administrator, Office of Policy, Program Development and Evaluation,

(202) 205-0699.

SUPPLEMENTARY INFORMATION:

Background

On July 25, 1996, FSIS published a final rule, ``Pathogen

Reduction; Hazard Analysis and Critical Control Point (HACCP)

Systems,'' (61 FR 38806). The new regulations (1) require that each

establishment develop, implement, and maintain written sanitation

standard operating procedures (Sanitation SOP's); (2) require regular

microbial testing for generic E. coli by slaughter establishments to

verify the adequacy of the establishments' process controls for the

prevention and removal of fecal contamination and associated bacteria;

(3) establish pathogen reduction performance standards for Salmonella

that slaughter establishments and establishments producing raw ground

products must meet; and (4) require that all meat and poultry

establishments develop and implement a system of preventive controls

designed to improve the safety of their products, known as HACCP

(Hazard Analysis and Critical Control Points).

With respect to the generic E. coli testing requirement, a number

of questions were posed in the final rule, especially about how the

requirement would be applied and what testing results might indicate in

establishments that slaughter livestock.

Responses to those questions were received through written

comments; through presentations and discussions at a public meeting

convened by FSIS on September 12-13, 1996, specifically to discuss the

generic E. coli testing requirement; at a national implementation

conference in Washington, DC, September 30--October 3, 1996; and six

subsequent regional implementation conferences occurring on October 15,

17, 22, 24, November 7 and 13, 1996, and at numerous briefings

presented by FSIS representatives to a variety of audiences.

Additionally, FSIS held the conference, ``Sanitation Standard Operating

Procedures (Sanitation SOP's) and E. coli Testing Requirements,'' on

January 23, 1997.

Through these comments and meetings, a number of technical

questions have arisen which indicate the need for further

clarification. Some of these have required a change in the

[[Page 26212]]

regulation; others simply require further technical guidance.

Technical Amendments

Sanitation Standard Operating Procedures

Questions were raised at the public meetings about the corrective

actions an establishment might take in response to Sanitation SOP

failures. Commenters suggested that the Agency make it clear that, in

certain cases, improving the execution of the existing Sanitation

SOP's, instead of revising the Sanitation SOP's, would be appropriate

corrective action. In response to this concern, FSIS is amending

section 416.15(b) to clarify that satisfactory corrective actions can

include appropriate improvements in the execution of Sanitation SOP's.

Applicability of E coli Testing Requirement

Species Required to be Sampled and Tested for E. coli

At the E. coli meeting, implementation conferences, and other

briefings, numerous questions were raised about the applicability of

the generic E. coli testing requirement. There were questions about

whether generic E. coli testing was required for all types of

livestock, i.e., cattle, sheep, swine, goats, horses, mules and other

equines (9 CFR 301.2). There were questions about whether generic E.

coli testing was required for all types of poultry, i.e., chickens,

turkeys, ducks, geese and guineas. There were also questions about

whether generic E. coli testing was required for all market classes of

livestock and poultry.

Clarification is needed because of inconsistencies in terminology

used in the preamble and the regulatory text. For example, the preamble

states that ``establishments that slaughter livestock or poultry will

be required to begin sampling and testing for E. coli''(61 FR 38844).

This statement is inconsistent with section 310.25 of the regulations

which refers only to ``cattle and/or hogs'' and subsequently ``swine''

and ``market hogs.'' This inconsistency also makes it necessary to

amend the regulations to clarify that all market classes of cattle,

swine, chickens, and turkeys must sample and test for generic E. coli.

FSIS intends that all establishments slaughtering livestock and

poultry sample and test for generic E. coli. However, the regulatory

requirement codified in section 310.25(a)(1) is limited to cattle and

swine. FSIS will propose rules in the future to carry out its goal of

applying the generic E. coli testing requirement to other types of

livestock, such as sheep, goats, and equines. Until that rulemaking is

completed, only cattle and swine are required to be sampled and tested

for generic E. coli (9 CFR 310.25(a)(1)).

With regard to poultry, the preamble of the final rule states that

minor species, such as ducks, geese, and guineas, would be addressed at

a later date. The rulemaking proposal to extend the E. coli testing

requirement to all types of livestock will also propose extending the

requirement to all types of poultry. However, until that rulemaking is

completed, only chickens and turkeys are required to be sampled and

tested for generic E. coli (9 CFR 381.94(a)).

At this time, FSIS is making technical amendments to ensure that

the terminology in sections 310.25 and 381.94 of the regulations

applicable to generic E. coli testing is consistent with other FSIS

regulations promulgated under the Federal Meat Inspection Act and the

Poultry Products Inspection Act.

Therefore, in section 325.10(a)(1), the phrase ``cattle and/or

hogs'' will be replaced with the phrase ``cattle and/or swine.'' In

section 310.25(a)(2)(iii), ``Sampling frequency,'' the word ``Cattle''

will replace the word ``Bovines.'' In section 310.25(a)(2)(v)(A) the

word ``cattle'' will replace the word ``bovines.'' In section

310.25(a)(5), Table 1, the phrase ``type of livestock'' will replace

the phrase ``slaughter class;'' the ``Steers/heifers'' and ``Cows/

bulls'' lines will become a single ``Cattle'' line having the lower

limit, upper limit, number of samples and maximum number of marginal

now permitted in both these slaughter classes; and ``Market hogs,''

will be redesignated as ``Swine.'' In section 381.94(a)(5), Table 1,

the phrase ``slaughter class'' will be replaced with ``type of

poultry,'' and the term ``broilers'' will be replaced by ``chickens.''

These terminology changes also will clarify that all market classes

of cattle or swine are categorized as ``cattle'' or ``swine,'' and that

all market classes of chickens and turkey are categorized as

``chickens'' or ``turkeys.''

Testing Requirements for Market Classes

Commenters and questioners also expressed confusion and sought

clarification about the applicability of the generic E. coli testing

requirement when no specific m/M criteria are available. They assumed

that if FSIS has not performed baseline studies and established m/M

criteria for evaluation of results, the requirement would not apply.

Commenters and questioners expressed their expectations that FSIS would

perform baseline studies for a large variety of market classes of

livestock and poultry, such as spent hens, sows and boars, calves, as

well as numerous types of livestock and poultry that are slaughtered,

dressed or chilled by non-traditional methods. At virtually every

public meeting where generic E. coli testing was discussed,

participants identified new livestock or poultry categories for

baseline data collection.

All market classes of cattle, swine, chickens and turkeys must be

sampled and tested for generic E. coli. FSIS's initial baseline studies

were conducted on separate market classes of cattle, swine, and

chickens. In future baseline studies, the Agency will sample from all

market classes of a type of livestock or poultry to develop m/M

criteria representative of that type of livestock or poultry. The

baseline study being developed for turkeys includes samples from all

market classes.

FSIS considered whether the m/M criteria for broilers could be

applied to all market classes of chickens, such as, fowl, heavy

broilers, and rock Cornish hens. FSIS determined that this would be

acceptable for three reasons:

1. The processing parameters likely to affect levels of generic E.

coli on carcasses, such as the use of automatic eviscerating equipment

and common bath chillers, the permitted levels of chlorine in poultry

processing waters, and the likely handling during processing were

essentially the same for all market classes of chickens.

2. The m/M levels of generic E. coli on chickens are expressed as

CFUs/ml, rather than total CFUs per carcass, and the actual values at

the 80th and 98th percentile have been rounded to the nearest whole log

10; both of these practices have the effect of minimizing

variability and normalizing values.

3. Broilers constitute the vast majority (94%) of chickens

slaughtered in the United States. An alternative to using the broiler

criteria for all chickens would be to conduct a baseline that includes

all market classes. However, the preponderance of broiler results will

mean that other market classes are highly unlikely to affect the

criteria.

These factors, taken together, mean that it would take very large

differences among market classes to necessitate a change in the

criteria found in the regulations. Accordingly, no amendment is being

made and the criteria published in the July final rule will be

applicable to all market classes of chickens.

FSIS expects that cattle and swine establishments will collect

samples by sponging carcasses. If so, they will evaluate tests by the

use of statistical process control, discussed below, and

[[Page 26213]]

the published m/M criteria in the regulations do not apply. FSIS will

sample all market classes of either cattle or swine in its baseline

studies to develop m/M criteria for samples collected by sponging

carcasses.

Cattle and swine establishments collecting samples by excising

tissue from carcasses will use the published m/M criteria. In the

regulations the m/M criteria for the market classes cows/bulls and

steers/heifers are the same. FSIS contends that these m/M criteria are

applicable for other market classes of cattle because of the similarity

in processing parameters and the methodology used to develop the m/M

criteria. FSIS also believes that the m/M criteria for market hogs are

applicable to other market classes of swine for the same general

reasons. Therefore, the published m/M criteria apply to all market

classes of cattle and swine.

While FSIS baseline surveys provide an appropriate national data

base for establishment of m/M criteria, microbiological data bases with

comparable accuracy and utility can be developed outside of FSIS. FSIS

encourages industry members, academia, and other groups to work with

the Agency to develop protocols for independent databases against which

the 80th and 98th percentile definitions can be applied. In

consultation with industry and consumer groups, FSIS may propose to

publish these m/M values as criteria for evaluating results.

FSIS is still in the process of developing its long-term plan for

baseline data collection studies. The plan will identify the types of

livestock and poultry to be included in future baseline data collection

efforts. Tentatively, the Agency has determined that types of livestock

and poultry identified in the regulatory definitions are top priority

candidates for FSIS baseline studies. For livestock, FSIS is

considering developing baseline data collection studies for sheep,

goats, and equines. For poultry, FSIS is conducting a baseline study

for turkeys and is considering baseline data collection for ducks and

geese. Representatives of State inspection programs and others have

raised questions about FSIS intentions for baseline data collection on

the voluntarily-inspected species, such as rabbits and ratites. FSIS

will consider these requests for baseline data in developing its long-

term plan.

Use of Statistical Process Control

The current m/M criteria apply to all classes of chickens, and to

cattle and swine samples collected by excising tissue from carcasses.

The m/M criteria for turkeys are still being developed. At this time,

cattle and swine establishments collecting samples by sponging and

turkey establishments will use statistical process control techniques

to evaluate E. coli test results.

Statistical process control techniques are based on the principle

that every product is produced by a process. All processes are subject

to variation, which can be understood and controlled by statistical

methods. A process that is in control is stable in terms of average

level and degree of variation, i.e., it is predictable within limits

and is ``doing its best.'' Control is attained, often by degrees, by

detecting and eliminating special causes of variation, that is, causes

not present at all times or not affecting all product output.

Statistical process control initially involves evaluating data to

determine process capability (the typical process performance level),

then checking subsequent data to see whether they are consistent with

this baseline level to ensure the process is in control and variations

are within normal and acceptable limits. This is accomplished by

checking for unreasonably high results, trends, etc., and looking for

and correcting problems in the process when the signals occur.

Specific techniques of statistical process control include time

plots, which chart measurements over time. This is the first technique

to use with data collected over time and analyzed for patterns. Another

technique is the control chart, which plots data over time but also

displays an upper control limit for specific measurements, and a

centerline, above and below which is an equal number of sample results.

The centerline is in effect a median average. A sample result above the

upper control limit would indicate the likely presence of a special

cause of variation that needs to be addressed. Results within control

limits indicate that the process is in control. Control charts are used

for after-the-fact analysis of process performance and to assist in

gaining and maintaining control of a process. In most situations more

than one type of control chart is applicable. More detailed information

on time charts and control charts can be found in texts on statistical

process control, under the topic ``control charts.''

FSIS has concluded that statistical process control techniques will

provide experience in ``process thinking'' (a central tenet of HACCP),

develop an historical record of performance, and permit evaluation of

the long-term stability of a process and determination of process

capability (that is, how the process is actually working), and track

the effectiveness of process improvement actions.

FSIS emphasizes that the value of microbiological testing is not

negated by the lack of national m and M criteria against which to

evaluate results. E. coli testing is intended to provide verification

of process control for fecal contamination within individual

establishments. While there is utility in being able to compare

individual establishment data with national norms (i.e., national m and

M criteria), the intent of the rule is to have microbial testing

integrated into the overall process control procedure that

establishments are implementing. In this context, establishment-

specific databases, developed as establishments begin microbial

testing, are also of value to individual establishments as a means of

verifying their process control procedures.

FSIS is amending section 310.25 of the regulations to require

establishments slaughtering cattle or swine to use either a three-site

sponging or a three-site excision sample collection technique. This

amendment to the meat regulations is necessary because of the inability

to develop a conversion factor for results derived from two or three-

site sample collection by sponging which correlates to the m/M criteria

developed based on excision sampling methods used in conducting the

baseline studies. If sponging is chosen, results must be evaluated

using statistical process control techniques, because the m/M criteria

derived from the baseline studies have not been validated for sample

collection using sponging. If an establishment chooses to use the

excision sample collection technique, results will be evaluated against

national norms as expressed in the m/M criteria drawn from baseline

studies. FSIS intends to give high priority in its baseline plan to

collecting data that will support establishing m/M criteria using

sponge sample collection techniques.

FSIS also is amending section 381.94 of the regulations to require

turkey establishments to evaluate results using statistical process

control techniques. This amendment is necessary because FSIS has not

completed the development of m/M criteria for turkeys.

Establishments evaluating test results using statistical process

control techniques will be subject to the regulatory provisions for

failure to test and record (9 CFR 310.25(a)(7) and 381.94(a)(7)). Such

establishments will not be subject to the regulatory provisions for the

failure to meet criteria (9 CFR 310.25(a)(6) and 381.94(a)(6))

[[Page 26214]]

until such time as m/M criteria are developed and added to the

regulations. The Agency intends to establish m/M criteria for each type

of livestock and poultry based on national norms. Therefore, the

requirements to utilize statistical process control techniques is

temporary.

Sampling Frequencies

There are three amendments related to the following topics: (1) The

requirement that establishments sample at the greater of one sample per

week or the published frequency for each type of livestock or bird; (2)

the requirement that all establishments are required to sample only the

type of livestock or poultry which they slaughter in the largest

number; (3) adjustments to sampling directions for very low volume

establishments that do not operate each week or operate on a seasonal

basis. Each of these three amendments is discussed below.

Sampling Frequencies For Very Low Volume Establishments

The final rule states that very low volume establishments ``shall

collect one sample per week starting the first full week of June and

continuing through August of each year.'' FSIS is aware that some very

low volume establishments do not operate every week or operate only

seasonally. Therefore, this requirement is amended to provide

flexibility and accommodate all very low volume establishments. The

revised regulations require that very low volume establishments begin

sampling the first full week they operate after June 1 and continue

collecting one sample per week in each week they operate until they

have met their sampling requirement.

As discussed in the final rule, FSIS requires slaughter

establishments to record and evaluate E. coli results in a ``moving

window'' of 13 consecutive results, and the Agency is permitting very

low volume establishments to conduct as few as 13 tests per year, in

part because of their relatively simple and stable production

environments.

If there are published m/M criteria for the type of livestock or

poultry a very low volume establishment slaughters in the largest

number, the establishment must sample that type of livestock or poultry

at a minimum frequency of once per week until a series of 13 tests has

met those m/M criteria.

If there are no m/M criteria for the type of livestock or poultry

slaughtered in the largest number, a very low volume establishment must

sample a minimum of once per week until 13 samples are collected. If

the establishment does not slaughter their primary type of livestock or

poultry for 13 weeks per year, the establishment must still collect one

sample each week in which they conduct those slaughter operations. This

provision will be eliminated once m/M criteria are developed for the

type of livestock or poultry that is slaughtered in the greatest

number.

One Type per Establishment

The final rule states that if a very low volume establishment

slaughters multiple types of livestock or poultry, the establishment

shall collect samples from the type it slaughters in the largest

number. FSIS intended that this provision apply to all establishments.

However, because of an inadvertent omission, this language was not

incorporated into the regulatory text for all establishments.

Therefore, FSIS is amending the regulations so that each slaughter

establishment, regardless of size, conducts generic E. coli testing on

the type of livestock or poultry that it slaughters in the largest

number.

The purpose of the testing is not lot acceptance, but rather to

provide each establishment with a microbial indication of how effective

its sanitary dressing procedures are in preventing contamination of

carcasses by fecal material, ingesta, and associated bacteria. The

preamble stated that the required testing and criteria are intended to

provide an initial basis for slaughter establishments and FSIS to begin

using microbial testing to evaluate the adequacy of process control. To

meet this regulatory objective, it is not necessary that all slaughter

types be sampled. Whether the establishment slaughters one type or

multiple types, E. coli test results provide information that

establishments can use to verify their process controls over sanitary

dressing.

Minimum Sampling Frequencies

The preamble to the final rule stated that establishments, except

for very low volume establishments, must test at the frequencies

established in the regulations or at a minimum of at least once per

week. This weekly minimum requirement was inadvertently not

incorporated into the regulatory language for other than very low

volume establishments. These technical amendments add the once per week

minimum to the regulatory language. Under this amendment, an

establishment slaughtering 9,000 cattle and sampling at the once per

week minimum shall collect 52 samples, rather than 30, as required by 1

test per 300. Obviously, the minimum of 52 assumes the establishment

slaughters cattle each week during the year.

Sampling Sites

Two specific questions raised in the final rule with respect to the

technical specifications of the generic E. coli testing requirement for

cattle and swine carcasses addressed the issue of sample sites on

carcasses. The questions were: ``[a]re there more appropriate

anatomical sites for microbial testing than those adopted?'' and

``[a]re there worker safety concerns regarding sampling from difficult

to reach carcass sites, and how can they be mitigated?''

The final reports, ``Analysis of ARS Baseline and Sponge Data'' and

``FSIS Comparison of Baseline Excision and Two-Site Sponge Method,''

describe results of data collection efforts by ARS and FSIS in cattle

and swine establishments to compare sponge and baseline excision

sampling methods and to seek conversion factors that would make sponge

results comparable to baseline results. The baseline excision method

for each slaughter class was defined in the protocol for the baseline

study and specified the sites to sample, the area of tissue to analyze,

and the amount of buffer to add to the tissue.

The final rule specified sampling cattle and swine with a sponge

from the same three sites from which FSIS collected excision samples in

baseline studies. During the comment period, industry representatives

expressed concerns over inefficiencies and safety hazards associated

with sampling the rump of cattle and the ham of swine. During

preparation of the final rule, FSIS initiated a data collection effort

by ARS to evaluate sponge methods with one or three sites, and to seek

conversion factors that would make sponge results comparable to

baseline results and to the m/M values derived therefrom. In response

to the comments on the 3-site sponge method, the Agency conducted

further data collection to compare a 2-site sponge method with the

baseline method.

ARS compared the baseline method with the final rule's three-site

sponge method and with a one-site sponge method, the one site being

flank for cattle and belly for swine. They collected data on a total of

280 carcasses in one cattle establishment and one swine establishment

and presented summaries of their results at the September 12-13, 1996,

FSIS E. coli conference. FSIS later performed further statistical

analyses on the results in response to comments at the conference. The

results of these analyses are described in detail in the reports, and

summarized here.

[[Page 26215]]

Because the lowest detectable levels (LDLs) of the sponge methods

were well below the LDL of the baseline method, the sponge methods were

expected to find more E. coli positives than the baseline method. The

three-site sponge resulted in more E. coli positives than the baseline

method for both cattle and swine. However, whereas the one-site sponge

method found more E. coli positives for cattle, it gave less for swine

(i.e., the difference in sites appeared to affect the prevalences found

by the two sponge methods).

Since the two-site sponge method had not been included in the ARS

study, FSIS undertook comparison of this sampling method with the

baseline method. The Agency collected data on a total of 825 carcasses

in three cattle establishments and four swine establishments. Results

of this effort are presented in detail in the reports and are

summarized here.

Once again, the sponge method was expected to result in more E.

coli positives than the baseline method because of its lower LDL, and

it did for all three cattle establishments sampled. However, sponging

resulted in considerably fewer E. coli positives than the baseline

method in three of the four swine establishments. One establishment,

however, had 100 percent E. coli prevalence by both the two-site sponge

and the baseline methods. That establishment also had higher levels of

E. coli than the other swine establishments.

In addition to the qualitative comparison of sponge and baseline

methods in terms of prevalence, FSIS also evaluated sponge results

quantitatively in terms of recovery of bacteria relative to the

baseline method. It was evident from the results that the sponge

methods generally gave lower average microbial counts than the baseline

method.

Where possible given the available data, FSIS evaluated recovery by

two alternate methods suggested at the E. coli conference. However,

there were several difficulties with getting reasonable estimates of

recovery. First, numerous negative baseline results left recovery

undefined for many carcasses. Second, the two recommended methods of

defining recovery gave seriously different recovery values. Third, the

sponge method gave appreciably more negative results for swine than the

baseline method. All of these difficulties caused FSIS to abandon the

effort to find a conversion factor.

In view of these findings, FSIS has determined that, at the present

time, the third sampling site is necessary. If data can be developed

that support a change to fewer, more accessible sampling sites, the

Agency is very willing to consider them. In addition, as described

above, livestock slaughtering establishments that want to relate their

results to national norms may use the excision technique and the m/M

criteria associated with the baseline studies.

Sampling Locations

Sampling location in the process is a factor for comparability of

an establishment's results with the criteria derived from baseline

studies. Establishments that slaughter, dress or chill types of

livestock or poultry by using non-traditional methods, such as hot

boning of swine and poultry and chilling of split turkey carcasses, may

not be able to collect samples at the exact location in the slaughter

process as was used in the baseline studies. FSIS is amending section

310.25(a)(2)(ii) to provide for sample collection after final wash, if

sampling chilled carcasses is not possible. Similarly, FSIS is amending

section 381.194(a)(2)(ii) to provide for sample collection after the

final wash, if sampling at the end of the drip line is not possible.

Additionally, questions have arisen about whether random carcass

sampling can only occur when carcasses are in the cooler. It is not

FSIS's intention to limit random carcass sample selection in the

cooler. The random sampling can be carried out before carcasses enter

the cooler so that carcasses selected for sampling can be placed in a

separate and convenient location in the cooler. The regulations require

establishments to include in their written procedures how sampling

randomness will be achieved (section 310.25(a)(2)(i)).

Technical Guidance

This section provides technical guidance for the following areas:

(1) definition of very low volume slaughter establishment; (2) counting

employees to determine establishment size for HACCP implementation; (3)

FSIS intentions on rules of practice.

Very Low Volume Slaughter Establishments

The regulations define very low volume establishments for cattle,

swine, chickens, and turkeys. These definitions are expressed in terms

of the number of animals or birds slaughtered annually. Establishments

should use 1996 slaughter data to determine whether they meet the

definition. Livestock and poultry slaughtered under the custom

exemption need not be counted.

Size Categories For HACCP Implementation

For purposes of determining whether an establishment is large,

small, or very small, FSIS has established the following guidelines for

counting employees. These guidelines combine the Small Business

Administration procedures for counting employees to determine

establishment size and the FSIS definition of ``official

establishment.'' All paid employees who work within the official

establishment are to be counted, whether full time, part time, or

temporary. Employees should be counted whether or not they perform

duties related to inspected products. Employee numbers should be

averaged over a year.

One exception to the above guidance covers situations where

headquarter's employees for firms with multiple establishments are

located at one official establishment and their assigned duties are

related to the company and not specifically to the official

establishment where they are located. Such employees need not be

counted. In addition, administrative staff, for example, billing and

bookkeeping staff, working outside the official premises need not be

counted. Unpaid family members of the owner or operator also need not

be counted. Large firms that have employees engaged full-time in buying

or selling products should count such staff even though they usually

work outside the establishment.

Establishments are very small if they have fewer than 10 employees

or annual sales of less than $2.5 million. In calculating annual sales,

establishments should count all sales of inspected meat and poultry

products produced at the establishment. Inspected product excludes

product produced under a retail or custom exemption provision.

Furthermore, ``Pass Through'' product that is produced in another

establishment and resold without any further processing need not be

counted. ``Pass Through'' includes the operation referred to as

``breaking bulk,'' if this operation involves only separating and

resorting ``intact'' packages prepared at another establishment.

FSIS Intention Regarding Rules of Practice

The final rule stated that, upon an establishment's failure to test

and record, inspection would be suspended in accordance with rules of

practice that ``will be adopted for such proceedings upon a finding by

FSIS that one or more provisions of subparagraphs (a) (1)-(4) of this

section have not been complied

[[Page 26216]]

with and written notice has been provided to the establishment.'' FSIS

has determined that a separate set of rules of practice for generic E.

coli testing is not necessary. The Agency does, however, intend to

review and propose revisions to its rules of practice (9 CFR 335.1 and

381.230). It plans to complete this process before the first HACCP

implementation date, January 26, 1998. In the meantime, the Agency will

use existing rules of practice for enforcement of Sanitation SOP's

requirements and for enforcement actions when establishments fail to

test and record results of generic E. coli analysis.

Technical Corrections

FSIS is making three technical corrections to the final rule. The

first corrects the inadvertent requirement that custom and retail

exempt establishments, as defined in section 303.1 of the Federal meat

inspection regulations, comply with the requirements for Sanitation

SOP's. These establishments are required to meet general sanitation

requirements, including those in section 308.3. When FSIS drafted the

final rule, it amended section 308.3 to ensure that meat and poultry

establishments not only meet the general sanitation requirements but

also comply with the regulations in Part 416, which require Sanitation

SOP's. However, FSIS never intended to require custom and retail exempt

establishments to comply with Sanitation SOP's. To clarify that point,

FSIS is amending section 308.3 to include language that explicitly

exempts these establishments from the part 416 requirements.

Secondly, FSIS is updating the footnotes in the ``Salmonella

Performance Standards'' table (Table 2) in section 381.94(b)(1) of the

poultry products inspection regulations. Footnote ``b'' states that the

``Broiler'' data was based on partial analysis and was subject to

confirmation upon publication of the baseline survey. The baseline

survey is complete and published. There are no changes to the numbers

related to broilers. FSIS is, therefore, removing footnote ``b.'' Also,

with the deletion of footnote ``b,'' the footnote designated as ``d''

(an editorial oversight omitted a footnote ``c'') will be redesignated

as footnote ``b.''

Finally, FSIS is correcting the references in sections 325.10(a)(3)

and 381.94(a)(3) to the AOAC International by updating the regulatory

text and a footnote in the regulatory text to reflect the

organization's new name and the current edition of its publication.

Also, FSIS is clarifying what establishments must do if they intend to

have samples analyzed by a method approved by a scientific body other

than the AOAC International.

Executive Order 12866 and Regulatory Flexibility Act

This final rule has been determined to be significant and,

therefore, has been reviewed by the Office of Management and Budget.

The Administrator has determined that this final rule will not have

a significant economic impact on a substantial number of small

entities, as defined by the Regulatory Flexibility Act (5 U.S.C. 601).

The Pathogen Reduction/HACCP final rule included a Final Regulatory

Impact Assessment (FRIA) (61 FR 38945). The technical corrections and

amendments do not change the cost and benefit estimates and impact

assessments presented in the FRIA.

The technical amendments regarding Sanitation SOP's clarify the

regulatory language to more accurately explain what FSIS intended

corrective actions to encompass. There is no change in regulatory

impact or cost of Sanitation SOP's. Similarly, the regulatory

amendments that change terminology in sections 310.25(a) and 381.94(a)

do not affect any regulatory requirements.

The technical amendments regarding statistical process control

clarify how turkey establishments and livestock establishments

collecting samples by sponging will analyze test results until m/M

criteria are developed. This change will not affect the cost estimates.

In the Preliminary Regulatory Impact Analysis (PRIA) for the

Pathogen Reduction/HACCP proposed rule, FSIS concluded that for each

microbiological sample it would take 5 minutes ``* * * to prepare the

paperwork and review the results of the sample analysis and plot the

results on a statistical process control chart.'' In the FRIA, the

Agency used this 5 minute estimate as the time it takes to record a

window of E. coli test results and compare such results with m/M

criteria. The Agency still believes that it takes approximately the

same amount of time to conduct either of these processes.

FSIS has amended the regulations to clarify how sampling and

testing must be conducted on hot-boned or hide-on product. These are

not new requirements.

The FRIA estimated generic E. coli testing costs using an upper

bound estimate of 24 dollars per sample. To develop this upper bound

estimate for E. coli sampling, FSIS examined cost estimates reported in

the PRIA and current cost estimates for FSIS testing programs. The

proposed rule required establishments to collect Salmonella samples by

excising tissue from carcasses, and therefore, the cost estimate

factored in the time it takes to sample in such a manner. Similarly,

FSIS samples are collected by excising tissue, and FSIS cost analyses

of its testing program reflect this fact. Because sponging carcasses

presumably takes less time to perform than excising tissue from

carcasses, FSIS is confident that the cost estimates reported in the

FRIA are upper bound estimates. FSIS expects all establishments to use

the sponging method because excising tissue takes more time and

devalues the carcasses. However, because the cost estimates were based

on excision, establishments choosing to excise tissue should not incur

costs greater than 24 dollars a sample.

The three technical amendments relating to sampling frequencies do

not change the regulatory impact and cost to establishments. In the

FRIA the Agency based its cost estimates on the assumption that

establishments would sample at a minimum of 52 times a year. Also, the

cost estimates assumed that establishments would only sample and test

the type of livestock or poultry slaughtered in the largest number.

Lastly, FSIS's analysis assumed that very low volume establishments

sample and test once per week until the results show that they meet the

published criteria.

Executive Order 12988

This final rule has been reviewed under Executive Order 12988,

Civil Justice Reform. In this final rule: (1) all state and local laws

and regulations that are inconsistent with this rule will be preempted;

(2) no retroactive effect will be given to this rule; and (3)

administrative proceedings will not be required before parties may file

suit in court challenging this rule.

Paperwork Requirements

The Pathogen Reduction/HACCP final rule included a paperwork

analysis (61 FR 38862) prepared in accordance with the Paperwork

Reduction Act. FSIS has determined that the technical corrections and

amendments in this rule do not change any information collection burden

hours. The paperwork and recordkeeping burden hours were developed

using the assumptions in the FRIA, discussed above.

Final Rules

List of Subjects

9 CFR Part 308

Meat inspection.

[[Page 26217]]

9 CFR Part 310

Meat inspection, Microbial testing.

9 CFR Part 381

Poultry and poultry products, Microbial testing.

9 CFR Part 416

Meat inspection, Poultry and poultry products.

For reasons set forth in this preamble, 9 CFR chapter III is

amended as follows:

PART 308--SANITATION

1. The authority citation for part 308 continues to read as

follows:

Authority: 21 U.S.C. 601-695; 7 CFR 2.18, 2.53

2. Section 308.3 is amended by revising the last sentence of

paragraph (a) to read as follows:

Sec. 308.3 Establishments; sanitary conditions; requirements.

(a) * * * The provisions of part 416 of this chapter apply to all

establishments, except establishments that are exempt in accordance

with Sec. 303.1 of this chapter.

PART 310--POST MORTEM INSPECTION

3. The authority citation for part 310 continues to read as

follows:

Authority: 21 U.S.C. 601-695; 7 CFR 2.18, 2.53.

4. Section 310.25 is amended by revising paragraphs (a)(1),

introductory test (a)(2)(ii),(a)(2)(iii), (a)(2)(v)(A), (a)(3), (a)(4),

and (a)(5) to read as follows:

Sec. 310.25 Contamination with microorganisms; pathogen reduction

performance standards for Salmonella.

(a) * * *

(1) Each official establishment that slaughters cattle and/or swine

shall test for Escherichia coli Biotype 1 (E. coli). Establishments

that slaughter more than one type of livestock or both livestock and

poultry, shall test the type of livestock or poultry slaughtered in the

greatest number. The establishment shall:

* * * * *

(2) Sampling requirements.

(i) * * *

(ii) Sample collection. The establishment shall collect samples

from all chilled swine or cattle carcasses, except those boned before

chilling (hot-boned), which must be sampled after the final wash.

Samples shall be collected by either sponging or excising tissue from

three sites on the selected carcass. On cattle carcasses,

establishments shall sponge or excise tissue from the flank, brisket

and rump, except for hide-on calves, in which case establishments shall

take samples by sponging from inside the flank, inside the brisket, and

inside the rump; on swine carcasses, establishments shall sponge or

excise tissue from the ham, belly and jowl areas. 1

---------------------------------------------------------------------------

\1\ A copy of FSIS's ``Guidelines for E. coli Testing for

Process Control verification in Cattle and Swine Slaughter

Establishments'' is available for inspection in the FSIS Docket

Room.

---------------------------------------------------------------------------

(iii) Sampling frequency. Slaughter establishments, except very low

volume establishments as defined in paragraph (a)(2)(v) of this

section, shall take samples at a frequency proportional to the volume

of production at the following rates:

Cattle: 1 test per 300 carcasses, but at a minimum one sample each

week of operation.

Swine: 1 test per 1000 carcasses, but at a minimum one sample each

week of operation.

* * * * *

(v) Sampling in very low volume establishments.

(A) Very low volume establishments annually slaughter no more than

6,000 cattle, 20,000 swine, or a combination of cattle and swine not

exceeding 6,000 cattle and 20,000 total of both types. Very low volume

establishments that collect samples by sponging shall collect at least

one sample per week, starting the first full week of operation after

June 1 of each year, and continue sampling at a minimum of once each

week the establishment operates until June 1 of the following year or

until 13 samples have been collected, whichever comes first. Very low

volume establishments collecting samples by excising tissue from

carcasses shall collect one sample per week, starting the first full

week of operation after June 1 of each year, and continue sampling at a

minimum of once each week the establishment operates until one series

of 13 tests meets the criteria set forth in paragraph (a)(5)(i) of this

section.

* * * * *

(3) Analysis of samples. Laboratories may use any quantitative

method for analysis of E. coli that is approved as an AOAC Official

Method of the AOAC International (formerly the Association of Official

Analytical Chemists) 2 or approved and published by a

scientific body and based on the results of a collaborative trial

conducted in accordance with an internationally recognized protocol on

collaborative trials and compared against the three tube Most Probable

Number (MPN) method and agreeing with the 95 percent upper and lower

confidence limit of the appropriate MPN index.

---------------------------------------------------------------------------

\2\ A copy of the current edition/revision of the ``Official

Methods of AOAC International,'' 16th edition, 3rd revision, 1997,

is on file with the Director, Office of the Federal Register, and

may be purchased from the Association of Official Analytical

Chemists International, Inc., 481 North Frederick Ave., Suite 500,

Gaithersburg, MD 20877-2417.

---------------------------------------------------------------------------

(4) Recording of test results. The establishment shall maintain

accurate records of all test results, in terms of CFU/cm\2\ of surface

area sponged or excised. Results shall be recorded onto a process

control chart or table showing at least the most recent 13 test

results, by type of livestock slaughtered. Records shall be retained at

the establishment for a period of 12 months and shall be made available

to FSIS upon request.

(5) Criteria for evaluation of test results.

(i) An establishment excising samples from carcasses is operating

within the criteria when the most recent E. coli test result does not

exceed the upper limit (M), and the number of samples, if any, testing

positive at levels above (m) is three or fewer out of the most recent

13 samples (n) taken, as follows:

Table 1.--Evaluation of E. Coli Test Results

Maximum

Number of number

Type of livestock Lower limit of marginal Upper limit of marginal sample permitted

range range tested in marginal

range

(m).................... (M).................... (n) (c)

----------------------------------------------------------------------------------------------------------------

Cattle.............................. Negative a............. 100 CFU/cm \2\......... 13 3

[[Page 26218]]

Swine............................... 10 CFU/cm \2\.......... 10,000 CFU/cm \2\...... 13 3

----------------------------------------------------------------------------------------------------------------

a Negative is defined by the sensitivity of the method used in the baseline study with a limit of sensitivity of

at least 5 cfu/cm\2\ carcass surface area.

(ii) Establishments sponging carcasses shall evaluate E. coli test

results using statistical process control techniques.

* * * * *

PART 381--POULTRY PRODUCTS INSPECTION REGULATIONS

5. The authority citation for part 381 continues to read as

follows:

Authority: 7 U.S.C. 138f, 450; 21 U.S.C. 451-470, 7 CFR 2.18,

2.53

Subpart K--Post Mortem Inspection; Disposition of Carcasses and

Parts

6. Section 381.94 is amended by revising paragraphs (a)(1)

introductory text, (a)(2)(ii), (a)(2)(iii), (a)(2)(v)(A), (a)(3),

(a)(4); and (a)(5) Table 1; by redesignating paragraph (a)(5) as

(a)(5)(i); by adding a new paragraph (a)(5)(ii); and by removing the

footnote b in Table 2 of paragraph (b)(1) and removing the symbol ``b''

as it appears after the term ``Broiler'' and redesignating footnote d

as footnote b to read as follows:

Sec. 381.94 Contamination with microorganisms; process control

verification criteria and testing; pathogen reduction standards.

(a) * * *

(1) Each official establishment that slaughters poultry shall test

for Escherichia coli Biotype I (E. coli). Establishments that slaughter

more than one type of poultry and/or poultry and livestock, shall test

the type of poultry or livestock slaughtered in the greatest number.

The establishment shall:

* * * * *

(2) Sampling requirements.

(i) * * *

(ii) Sample collection. Samples shall be collected by taking a

whole bird from the end of the chilling process, after the drip line,

and rinsing it in an amount of buffer appropriate to the type of bird

being tested. If the bird is boned before chilling (hot boned poultry),

the sample shall be taken from the end of the slaughter line instead of

the end of the drip line.1

---------------------------------------------------------------------------

\1\ A copy of FSIS's ``Sampling Technique for E. coli in Raw

Meat and Poultry for Process Control Verification'' is available for

inspection in the FSIS Docket Room.

---------------------------------------------------------------------------

(iii) Sampling frequency. Slaughter establishments, except very low

volume establishments as defined in paragraph (a)(2)(v) of this

section, shall take samples at a frequency proportional to the

establishment's volume of production at the following rates:

Chickens: 1 sample per 22,000 carcasses, but at a minimum one

sample per each week of operation.

Turkeys: 1 sample per 3,000 carcasses, but at a minimum one sample

each week of operation.

* * * * *

(v) Sampling in very low volume establishments

(A) Very low volume establishments annually slaughter no more than

440,000 chickens or 60,000 turkeys or a combination of chickens and

turkeys not exceeding 60,000 turkeys and 440,000 birds total. Very low

volume establishments slaughtering turkeys in the largest number shall

collect at least one sample per week, starting the first full week of

operation after June 1 of each year, and continue sampling at a minimum

of once each week the establishment operates until June 1 of the

following year or until 13 samples have been collected, whichever comes

first. Very low volume establishments slaughtering chickens in the

largest number shall collect one sample per week, starting the first

full week of operation after June 1 of each year, and continue sampling

at a minimum of once each week the establishment operates until one

series of 13 tests meets the criteria set forth in paragraph (a)(5)(i)

of this section.

* * * * *

(3) Analysis of samples. Laboratories may use any quantitative

method for analysis of E. coli that is approved as an AOAC Official

Method of the AOAC International (formerly the Association of Official

Analytical Chemists) 2 or approved and published by a

scientific body and based on the results of a collaborative trial

conducted in accordance with an internationally recognized protocol on

collaborative trials and compared against the three tube Most Probable

Number (MPN) method and agreeing with the 95 percent upper and lower

confidence limit of the appropriate MPN index.

---------------------------------------------------------------------------

\2\ A copy of the current edition/revision of the ``Official

Methods of AOAC International,'' 16th edition, 3rd revision, 1997,

is on file with the Director, Office of the Federal Register, and

may be purchased from the Association of Official Analytical

Chemists International, Inc., 481 North Frederick Ave., Suite 500,

Gaithersburg, MD 20877-2417.

---------------------------------------------------------------------------

(4) Recording of test results. The establishment shall maintain

accurate records of all test results, in terms of CFU/ml of rinse

fluid. Results shall be recorded onto a process control chart or table

showing at least the most recent 13 test results, by type of poultry

slaughtered. Records shall be retained at the establishment for a

period of 12 months and shall be made available to FSIS upon request.

(5) Criteria for evaluation of test results.

(i) * * *

[[Page 26219]]

Table 1.--Evaluation of E. coli Test Results

Maximum number

Types of poultry Lower limit of Upper limit of Number of sample permitted in

marginal range marginal range tested marginal range

(m)............... (M)............... (n)............... (c)

----------------------------------------------------------------------------------------------------------------

Chickens....................... 100 CFU/ml........ 1,000 CFU/ml...... 13................ 3

Turkeys........................ N.A.a............. N.A............... N.A............... N.A.

----------------------------------------------------------------------------------------------------------------

a Not available; values for turkeys will be added upon completion of data collection program for turkeys.

(ii) For types of poultry appearing in paragraph (a)(5)(i) Table 1

of this section that do not have m/M criteria, establishments shall

evaluate E. coli test results using statistical process control

techniques.

* * * * *

PART 416--SANITATION

7. The authority citation for part 416 continues to reads as

follows:

Authority: 21 U.S.C. 451-470, 601-695; 7 U.S.C. 450, 1901-1906;

7 CFR 2.18, 2.53.

8. Section 416.15 is amended by revising paragraph (b) to read as

follows:

Sec. 416.15 Corrective Actions.

* * * * *

(b) Corrective actions include procedures to ensure appropriate

disposition of product(s) that may be contaminated, restore sanitary

conditions, and prevent the recurrence of direct contamination or

adulteration of product(s), including appropriate reevaluation and

modification of the Sanitation SOP's and the procedures specified

therein or appropriate improvements in the execution of the Sanitation

SOP's or the procedures specified therein.

Done at Washington, DC, on May 7, 1997.

Thomas J. Billy,

Administrator.

[FR Doc. 97-12397 Filed 5-7-97; 3:21 pm]

BILLING CODE 3410-DM-P