# Preliminary Regulatory Impact Analysis and Initial Regulatory Flexibility Analysis of the Proposed Rules to Ensure the Safety of Juice and Juice Products

> Briefs, arguments, decisions, and more.

URL: https://www.frixlaw.com/law-library/documents/fr%3A98-11530

## Record

- **Collection:** Federal Register
- **Document type:** Proposed Rule
- **Published:** May 1, 1998
- **Citation:** 63 FR 24254

## Text

SUMMARY: The Food and Drug Administration (FDA) is publishing the
preliminary regulatory impact analysis (PRIA) that it has prepared
under Executive Order 12866 and initial regulatory flexibility analysis
(IRFA) that it has prepared under the Regulatory Flexibility Act (RFA),
as amended by the Small Business Regulatory Enforcement and Fairness
Act (SBREFA), on the costs and benefits of FDA's proposed regulations
regarding the Hazard Analysis Critical Control Points (HACCP) and
labeling for juice and juice products. FDA is issuing those proposals
because of recent outbreaks of foodborne illness and deaths caused by
consumption of juice products that were not pasteurized or otherwise
processed to control pathogenic microorganisms. Those proposals are
intended to ensure that juice and juice products are safe.

DATES: Submit written comments by May 26, 1998 on aspects of this
analysis related to labeling for juice and juice products and by July
8, 1998 on aspects of this analysis related to HACCP for juice and
juice products.

ADDRESSES: Submit written comments to the Dockets Management Branch
(HFA-305), Food and Drug Administration, 12420 Parklawn Dr., rm. 1-23,
Rockville, MD 20857. Comments should be identified with the docket
numbers found in brackets in the heading of this document.

FOR FURTHER INFORMATION CONTACT: David J. Zorn, Center for Food Safety
and Applied Nutrition (HFS-726), Food and Drug Administration, 200 C
St. SW., Washington, DC 20204, 202-205-4729.

SUPPLEMENTARY INFORMATION:

Table of Contents

I. Background
II. Introduction
III. Factors Considered in Developing This Analysis
IV. Regulatory Options
V. Benefits
VI. Costs
VII. Summary of Benefits and Costs
VIII. Initial Regulatory Flexibility Analysis
IX. References
X. Request for Comments

I. Background

This document constitutes FDA's PRIA and IRFA of the proposed rules
to amend the food labeling regulations and to require HACCP for juice
and juice products. Because the industries affected by both proposed
rules substantially overlap and because both proposals address the same
public health problem, the safety of juice and products containing
juice, the agency has chosen to analyze the economic impact of both
proposed rules in a single PRIA and IRFA. These documents analyze both
the costs and benefits of the proposed rules as well as the expected
impacts on the affected small entities. FDA has found that these rules
may constitute significant rules under Executive Order 12866 because
they could have a significant impact on one sector of the economy
(producers of minimally processed juice). In addition, FDA has
determined under the RFA that each proposal would present a significant
impact on a substantial number of small entities.

II. Introduction

FDA has examined the impacts of these proposed rules under
Executive Order 12866. Executive Order 12866 directs Federal agencies
to assess the benefits and costs of available regulatory alternatives
and, when regulation is necessary, to select regulatory approaches that
maximize net benefits (including potential economic, environmental,
public health and safety effects; distributive impacts; and equity).
Under the Executive Order, a regulatory action is ``significant'' if it
meets any one of a number of specified conditions, including having an
annual effect on the economy, competition, or jobs, or if it raises
novel legal or policy issues. FDA finds that each of these proposed
rules may constitute a significant regulatory action as defined by
Executive Order 12866, as discussed as follows.
In addition, FDA has determined that these rules are not
significant rules under the Unfunded Mandates Reform Act of 1995 (UMRA)
requiring benefit-cost and other analyses. Under UMRA significant rule
is defined as ``a Federal mandate that may result in the expenditure by
State, local and tribal governments in the aggregate, or by the private
sector, of $100,000,000 (adjusted annually for inflation) in any 1
year''.
Finally, in accordance with the SBREFA, the Administrator of the
Office of Information and Regulatory Affairs of the Office of
Management and Budget (the Administrator) has determined that these
proposed rules are major rules for the purpose of congressional review.
A major rule for this purpose is defined as one that the Administrator
has determined has resulted or is likely to result in an annual effect
on the economy of $100 million or more; a major increase in costs or
prices for consumers, individual industries, Federal, State, or local
government agencies, or geographic regions; or significant adverse
effects on competition, employment, investment, productivity,
innovation, or on the ability of U.S.-based enterprises to compete with
foreign-based enterprises in domestic or export markets.

III. Factors Considered in Developing This Analysis

This analysis estimates costs and benefits for two proposed
regulations, published in the Federal Register of April 24, 1998 (63 FR
20450 and 20486), that would affect the safety of juice products. The
first rule requires warning statements on minimally processed packaged
juice. That is, juice that has not been processed in a manner that will
produce, at a minimum, a 5-log reduction, for a period at least as long
as the shelf life of the product when stored under normal and moderate
abuse conditions, in the pertinent microorganism. The ``pertinent
microorganism'' is the most resistant microorganism of public health
significance that is likely to occur in the juice. In the remainder of
this analysis, this will be referred to as the ``5-log reduction.''\1\
The second rule requires manufacturers of most juice to implement a
HACCP program with the same 5-log reduction performance criteria.
However, FDA is proposing to exempt retailers who, for the purposes of
this rule, the agency has tentatively decided will include very small
businesses that make juice on their premises and whose total sales of
juice and juice products do not exceed 40,000 gallons per year and who
sell directly to consumers or directly to consumers and other
retailers.
---------------------------------------------------------------------------

\1\ That is, the total combined effect of all controls have the
effect of reducing the number of colony forming units (cfu's) by a
factor of 100,000. This implies that even if the product should
contain 1,000 cfu's per gallon (gal.) prior to processing, the final
product after processing would contain only .01 cfu's per gal.
---------------------------------------------------------------------------

The effective date for the labeling rule is proposed to be 60 days
following publication of the final rule with

[[Page 24255]]

warning statements required either on the labels or, in the case of
products which do not bear the warning statement on the label, on
labeling (e.g., on signs or placards at the point of sale) on juices
that have not been processed in a manner that will produce, at a
minimum, a 5-log reduction. Packaged juices produced by large firms are
required to bear warning labels beginning on January 1, 2000, and
packaged juices produced by small and very small firms\2\ are required
to bear warning labels beginning on January 1, 2001. The agency expects
that the HACCP rule, because of its complexity, will not be finalized
for at least 1 year following finalization of the juice labeling rule.
The HACCP rule is proposed to be effective for large firms, 12 months
following publication of the final HACCP rule; for small firms, 24
months following publication of the final HACCP rule; and for very
small firms, 36 months following publication of the final HACCP rule.
For purposes of this rule, the agency is proposing to define large
processors as those who have more than 500 employees, small processors
as those who have less than 500 employees and very small processors as
those who have either: (1) Total annual sales of less than $500,000, or
(2) that have total annual sales of greater than $500,000 but total
annual food sales of less than $50,000, or (3) that employ fewer than
100 full-time equivalent employees and annually sell less than 100,000
units of the juice in the United States.
---------------------------------------------------------------------------

\2\ The labeling rule does not define ``very small firms'' but
the HACCP rule does give a separate definition of ``very small
firms'' as a subset of ``small firms'' as defined in the labeling
and HACCP rules. Therefore, the term ``very small firms'' has been
used here in relationship to the labeling rule to make clear where
this subset fits in the context of both of these rules. The HACCP
rule defines small businesses as those with fewer than 500
employees. It defines very small businesses as those with total
annual sales of less than $500,000 or those with total annual food
sales of less than $50,000 or those with fewer than 100 employees
and less than 100,000 units of juice sold annually.
---------------------------------------------------------------------------

To a large extent, benefits and costs will depend on how processors
of juice who do not currently implement controls sufficient to achieve
a 5-log reduction respond to the warning label regulation. That is,
firms will choose whether to display the warning statement or to comply
early with the 5-log reduction. The agency has no information to
indicate the choices that specific processors will make.
The actual choice that each processor will make depends on several
factors: (1) The revenue that processors expect to lose because of
consumers' responses to the Government's announcement of the rules and
the warning label, (2) the costs of and length of time allowed to make
label changes, (3) the costs of achieving a 5-log reduction in
pathogens, and (4) the revenue that processors expect to lose if
consumers respond negatively to the changes in product characteristics
caused by processing the juice.
Processors will choose to discontinue juice production if they
perceive that either labeling or a change in processing practices will
lower profits below a ``normal'' return.\3\ In other words, processors
will go out of the juice business rather than comply with these
regulations only if one of the two following conditions is satisfied:
(1) The combination of the cost of displaying the warning labeling and
the reduction in revenue caused by the negative response of consumers
to the warning results in below normal profits; or (2) a combination of
increased costs from processing and a reduction in revenue caused by
the negative response of consumers to the changes in product quality
results in below normal profits.
---------------------------------------------------------------------------

\3\ A normal return on profits is the average market return on
capital that a processor could receive, for example, by investing in
the stock market.
---------------------------------------------------------------------------

For the purposes of this analysis, the agency has assumed that, in
order to avoid having their products associated with the warning to
consumers, all establishments that will eventually be covered by the
HACCP rule will implement controls sufficient to achieve a 5-log
reduction when the labeling rule takes effect. The agency has also
assumed for the purposes of this analysis that those establishments not
covered by the HACCP rule will display the warning statement for
packaged juice products. However, in order to avoid displaying the
warning statement, these establishments may choose to process their
juice in a manner sufficient to achieve a 5-log reduction in pathogens
or under an adequate voluntary HACCP plan.

IV. Regulatory Options

The preambles in the accompanying proposed regulations describe the
compelling public need for these regulations. For example, in recent
years, pathogens have been discovered in fresh juices after having
caused severe illness in humans. These products were previously not
known to be vehicles for such hazards, given their low pH. Because
these events have occurred, the agency tentatively finds that it is
prudent to require the adoption of preventative controls for hazards
now associated with juice where controls may not have been previously
thought to be necessary.
There are a number of regulatory options that FDA has preliminarily
considered to reduce the risks associated with consuming juice
products. FDA requests comments on benefits, costs, and any other
aspect of these options.

A. Take No New Regulatory Action

Choosing this option would imply either reliance on: (1) Existing
Federal regulation, (2) State and local regulatory activity, (3)
business interests, (4) consumer demands, and (5) product liability
pressures to reduce risks incurred by consumers of juice products or
acceptance that the risks that juice currently presents are risks that
consumers are unwilling to pay to reduce. In the first case, it is
unlikely that the market will adjust to eliminate the risks present in
juice because of the difficulty of establishing the link between the
various kinds of illnesses, whether acute or chronic, to consumption of
juice. Generally, this link may only be established when there are
large, geographically focused outbreaks of acute illness. However,
research indicates that most cases of foodborne illness are sporadic
and geographically dispersed and not associated with any identifiable
and focused outbreaks (Ref. 1). In the second case, it is presumed that
consumers are willing to pay to reduce these risks given the sizeable
estimated benefits of the proposed rules. Finally, while industry and
State governments have undertaken steps in many areas to reduce risks
associated with juice, FDA believes that the changes have been made
with the expectations of Federal regulation. It is unlikely that the
market would fully adjust to reduce the risk without additional Federal
action.

B. Regulate Only High-Risk Juice Products or High-Risk Hazards

FDA could choose to make these rules applicable only to juice
products that have been associated by epidemiology or by inspection
history with health hazards. This option is discussed in the appendix
supporting this analysis (Ref. 9). In the appendix, the agency
concluded that unpasteurized or otherwise nonheat treated juices
present the largest risk to consumers because pathogens pose the
highest risk of the several categories of hazards. FDA is proposing
that all chemical, physical, and biological hazards be included under
HACCP, despite the differences in relative risk posed by different
types of hazards. It is important to note that processors may, under
the umbrella of

[[Page 24256]]

HACCP, adjust for the probability and severity of hazards by adjusting
critical limits, the frequency of monitoring, intensity of corrective
action, or any number of other margins. FDA has not evaluated the
benefits and costs of structuring HACCP based on this option, and seeks
comments on it, especially on the option of covering only some types of
juice.

C. Do Either One of the Proposed Rules but Not Both

One option would be to eliminate the HACCP requirement for juices,
one of the two proposed actions, and only require that juices that are
not processed to achieve a 5-log reduction be labeled with a warning to
consumers. The purpose of this labeling is to alert consumers who are
at increased risk to avoid these products and to inform all consumers
of the risk of these products relative to other juices. However, it is
difficult to predict what products consumers would switch to once they
encounter the warnings. It is possible that some consumers may reduce
their health status by choosing less nutritious substitutes in order to
avoid the products with the warning labels. Although labeling may be
effective for changing both producer behavior (particularly to avoid
displaying the warning) and consumer behavior, the agency believes that
labeling alone is unlikely to be sufficient to address all health
hazards associated with consumption of juice products.
Another option would be to eliminate the labeling rule and only
require that juice processors implement HACCP. This option would reduce
the possibility that some consumers might overreact and avoid all
juice. This option would also allow fresh juice to be marketed without
warnings and would result in some cost savings for products that will
not need to pay for labeling costs. However, it would also result in
some reduction in benefits because the HACCP rule will take longer to
implement than the labeling rule and because the proposed labeling rule
covers juice made at the point of sale and the proposed HACCP rule does
not cover retailers.

D. Require New Current Good Manufacturing Practices

FDA could develop and require current good manufacturing practices
(CGMP's) or sanitation standards specific to juice products to improve
the safety of juices. The use of CGMP's would assist processors in
ensuring the safety of their juices by providing guidance on how to
reduce insanitary manufacturing practices and on how to protect against
food becoming contaminated. While FDA currently has general CGMP's that
provide guidance to all food processing industries, it does not have
specific CGMP's for the juice industry.
There are three reasons that this alternative alone may be
undesirable. First, CGMP's by themselves are unlikely to have a
sufficient impact on the safety of juice, particularly relative to
HACCP. That is, CGMP's do not provide: (1) A structure for each
processor to align specific hazards unique to the processor's
operations with specific control measures; (2) assurance that the
processor will establish specific performance standards appropriate to
the processor's unique operation; (3) records that document that the
performance standards are met; and (4) records of frequent audits to
verify that controls are being applied, all of which are associated
with HACCP. Identifying specific hazards, designing controls that are
specific and unique to each operation, and verifying that these
controls are being applied as specified are essential elements of a
control program that will provide an improved level of food safety.
Secondly, under the HACCP approach being proposed, the industry is
required to use FDA's general CGMP's in part 110 (21 CFR part 110) and
to develop and adopt sanitation standard operating procedures (SOP's)
as part of their prerequisite programs for their HACCP plan. Therefore,
the HACCP approach builds on the foundation of CGMP's at the same time
it avoids the limitations of this alternative.
HACCP is designed for use in all segments of the food industry from
growing, harvesting, processing, manufacturing, distributing, and
merchandising to preparing food for consumption. Prerequisite programs
such as current good manufacturing practices (CGMP's) are an essential
foundation for the development and implementation of successful HACCP
plans.
The production of safe food products requires that the HACCP system
be built upon a solid foundation of prerequisite programs. Each segment
of the food industry must provide the conditions necessary to protect
food while it is under their control. This has traditionally been
accomplished through the application of CGMP's. These conditions and
practices are now considered to be prerequisite to the development and
implementation of effective HACCP plans. Prerequisite programs provide
the basic environmental and operating conditions that are necessary for
the production of safe, wholesome food.

E. Require Pasteurization

FDA could require that all juice be pasteurized rather than
requiring HACCP with a specified 5-log reduction. Although FDA is not
currently aware of other practical methods to achieve this level of
control, solely requiring pasteurization would inhibit new
technological innovation and it would only address one type of hazard
(pathogens that are not heat resistant). In this analysis, the agency
has, in fact, evaluated the costs of pasteurization for those juices
not now pasteurized. It should be pointed out that, by volume, the vast
majority of juices are now pasteurized or otherwise equivalently
treated. Thus, the marginal costs and benefits of requiring
pasteurization only apply to the small fraction of juice that is not
heat treated.
The agency requests comment on the appropriateness of the 5-log
reduction performance standard and if other approaches, such as
establishing a minimal acceptable risk standard for juices, could be
used that would ensure the safety of the juice. The agency requests
comments on what such a minimal acceptable risk standard should be and
how it would be implemented. The agency also invites interested persons
to submit scientific data concerning the acceptability of a 5-log
reduction requirement or whether a more or less stringent performance
standard (e.g., 3- or 7-log reduction) for specific juices would be
more appropriate or whether different approaches consistent with a
minimal acceptable risk standard for juices might be appropriate for
specific juices based on their unique characteristics.

F. Set Different Performance Standards for Processing of Different
Products

One regulatory option would be to establish different performance
standards for processing different types of juice products to decrease
the number of pathogens. In the proposal, the agency has tentatively
proposed that any combination of processing steps which cumulatively
result in a 5-log (a 100,000-fold) reduction in pathogens should be
applied to the production of all types of juice. However, different
products may warrant different processing stringencies because of a
number of factors, including: (1) The initial microbial counts on raw
produce are likely to vary, (2) different types of produce are likely
to harbor different kinds of pathogens, and (3) different products
provide different environments for microbial growth. This

[[Page 24257]]

option could either be exercised as part of the final rule in response
to comments or the proposed standards could remain with the option to
further petition the agency for a different standard. The benefits and
costs of the standard will vary directly with the stringency of
different performance standards. However, FDA does not have data to
estimate preliminarily the costs and benefits of this option.

G. Expand HACCP Rule Coverage

FDA has tentatively concluded that the retail sector should not be
included in the HACCP rule and has asked for comments on the
appropriateness of this conclusion. The expansion of coverage of the
HACCP rule to include retailers that process juice at the point of sale
would add an estimated additional 14,300 restaurants and 1,300 grocery
stores and supermarkets for a total of approximately 16,000
establishments. If the cost for these establishments to implement HACCP
was equivalent to that of very small processors who would be required
to initiate pasteurization ($26,000 in the first year and $11,900 in
subsequent years), then the total additional cost of this option would
be approximately $416 million in the first year and approximately $190
million in subsequent years. However, the agency does not have direct
information about the cost of implementing HACCP in a retail setting
for juice and the actual costs may vary significantly from these
estimates.

H. Use of One of Various Alternatives

An alternative approach to mandating HACCP would be to provide a
more flexible array of options tailored to the microbial risk present
in the particular juice. Manufacturers of apple cider would be provided
a permanent option choosing between labeling or implementing a HACCP
program with a 5-log pathogen reduction. All juices other than
untreated apple cider would be provided a permanent option of choosing
between labeling, implementing a HACCP system, or achieving a 5-log
pathogen reduction. However, FDA believes that this option provides
only weak incentives for processors to implement a HACCP system.
Processors could label hazardous products without taking steps to
improve the safety of juice or choose to achieve a 5-log reduction for
microbial pathogens without addressing other hazards. The agency
believes that labeling would not achieve the same level of product
safety. Additionally, there would be less incentive for processors to
implement a HACCP system, which includes, among other things,
developing and implementing sanitation SOP's and recordkeeping at
critical control points in addition to achieving a 5-log reduction.
Other hazards that would not be addressed include chemical
contaminants, hazardous metals, including lead and tin, mycotoxins,
pesticides, and physical hazards, such as glass.
Another regulatory option would be to include labeling for
unpackaged juice products for all retail outlets, such as restaurants.
This option would also require any very small retailer (as defined for
the purposes of this rulemaking) who is manufacturing less than 40,000
gallons of juice per year and selling it directly to consumers and
other retailers to either label or achieve a 5-log kill until a
requirement for HACCP would become effective 36 months from the date of
publication of the final rule.
If this option is combined with both proposed rules, FDA has
estimated the benefits to be $383 to $478 million annually and
estimated the costs in the first year to be $54 million and the costs
in subsequent years to be $28 million.

V. Benefits

This analysis provides estimates of three additive, independent
benefits of these two proposed rules: (1) Reduced expenditures related
to regulatory enforcement, (2) reduced adverse health effects, and (3)
other benefits. To some extent, the benefits of the two rules are
intertwined. Because of the earlier compliance dates, the impact of the
labeling rule will be to achieve some of the benefits faster. That is,
if firms choose to achieve a 5-log reduction through their processing
practices to avoid labeling, then some of the future benefits that
would be otherwise achieved under HACCP will be achieved sooner because
of the incentive provided by the labeling rule. Also, if at-risk
consumers avoid unpasteurized juices as a result of the labeling, there
will be reduced adverse-health effects prior to the introduction of
HACCP. On average, the labeling rule will achieve some of the benefits
2 years faster than the HACCP rule.

A. Enforcement Benefits

To the extent that these proposed rules are effective at reducing
contaminated juice, they should reduce the number of safety-related
enforcement actions (for both domestic and imported products) taken by
the agency for juice products. The enforcement activities chosen as a
baseline for juice products fall between the period 1992 and 1996
(inclusive) and involve import detentions and domestic recalls.
In the final regulatory impact analysis for FDA's seafood HACCP
rule, FDA used an assumption that the rule would prevent 50 percent of
the current number of annual enforcement actions. The agency did not
receive comments on this assumption in that rule and does not yet have
data from implementation of the rule to validate it. However, this may
be a conservative assumption. If HACCP plans are properly conceived,
implemented and validated, it is likely that the vast majority of
problems will be caught and corrected in the plant, rather than result
in foodborne disease outbreaks or be caught through Federal sampling of
the final product. Thus, the agency will continue to make this
assumption but requests comment on it.
1. Import Enforcement
Over the period 1992 through 1996, there were a number of imported
juice products detained for various violations of the Federal Food,
Drug, and Cosmetic Act (the act). A detention is a procedure for
preventing violative products from entering the United States.
Following a determination that a sample of a product is violative,
three steps occur: (1) FDA sends a detention notice to the importer
providing an opportunity to introduce testimony as to the condition of
the product; (2) the importer may contact an attorney, submits a
response application, and introduces evidence regarding the product;
and (3) FDA makes a determination about what should be done with the
shipment. There are three actions that FDA can specify for a detained
shipment: (1) The product is allowed to be ``reshipped'' out of the
country, (2) the product is reconditioned so as to bring it into
compliance with U.S. law, or (3) the product is destroyed under Federal
supervision. Assume that the cost per shipment of the three steps to
all parties involved is $5,000. Then the remaining cost of detention is
the cost per shipment of the three actions which is related to the
value of the shipment.
Table 1 gives the number of shipments detained and the total dollar
value of juice products detained for violations of the act for the
entire period 1992 through 1996.
The average value per shipment of imported juice products refused
entry is approximately $10,000. The average number of imported juice
product shipments detained annually is 23.

[[Page 24258]]

Table 1.--Totals of Juice Import Detentions for 1992 through 1996 by Reason for Detention
--------------------------------------------------------------------------------------------------------------------------------------------------------
Poisonous or Violative
Reason for Detention Food Additive Deleterious Pesticide Chemical New Drug Microbial Total
Issues Substances Residues Contamination Residues Hazards
--------------------------------------------------------------------------------------------------------------------------------------------------------
Number of Shipments 44 17 53 1 1 1 117
Value of Shipments $122,000 $112,000 $802,000 $79,000 $20,000 $2,000 $1,137,000
--------------------------------------------------------------------------------------------------------------------------------------------------------

If, on an annual basis, 23 imported juice product shipments are
detained at an average Federal enforcement and industry negotiation
cost of $5,000 per shipment (60 FR 65189), and if all 23 shipments
(with an average value of $10,000 per shipment) are destroyed so that
the entire $10,000 value of the shipment is lost, then the total annual
cost of all juice detentions is approximately $345,000 (23 shipments x
($10,000 value of shipment + $5,000 enforcement and negotiation cost)).
If 50 percent of these enforcement costs are prevented, then the
benefits related to import enforcement are approximately $175,000.
2. Recalls
Recalls tracked by FDA for pathogens or pesticides in juice
products are infrequent. For the period 1992 through 1996 there was one
class 1 recall and there were seven class 2 recalls\4\ for such
hazards, or about two recalls per year. A class 1 recall may cost as
much as $3 to $5 million between expenditures by the manufacturer,
retailers and State, local, and Federal authorities. However, the
typical juice recall is smaller and less costly than this. If the
combination of industry and government costs per recall on average is
$1 million, then the total annual cost of juice recalls is
approximately $2 million (2 recalls per year at $1 million each). This
assumption is based on FDA conversations with industry for both large
and small recalls. FDA acknowledges that this may not be the true
average cost of a recall and requests comment on this assumption. If 50
percent of these enforcement costs are prevented, then the benefits
related to recalls tracked by FDA are $1 million. However, FDA may not
be aware of all recalls that take place, particularly for less
hazardous reasons. Assuming that the recalls that FDA is not aware of
are considerably smaller, perhaps costing $100,000, and that FDA may
only hear about 10 percent of such recalls, then the total annual cost
of such recalls could be $1 million. If 50 percent of these enforcement
costs are prevented, then the benefits related to recalls not tracked
by FDA would be $500,000. Thus, the total annual benefits of the HACCP
rule related to recalls is estimated to be $1.5 million.
---------------------------------------------------------------------------

\4\Class 1 recalls are for dangerous or defective products that
predictably could cause serious health problems or death. Class 2
recalls are for products that might cause a temporary health
problem, or pose only a slight threat of a serious nature.
---------------------------------------------------------------------------

In addition to those benefits, when firms have recalls that are
made public they will generally suffer a loss of sales, at least
temporarily, from lost ``goodwill.'' This alone does not result in a
social cost but rather a social transfer as other firms will step
forward to capture sales lost from the recalling firm. However, in
addition to the resources invested in recalling the product, the
recalling firm may invest real resources in advertising to recapture
lost goodwill, a social cost. FDA cannot quantify this cost.

B. Health Benefits

This section presents quantitative estimates of health benefits
from this rule. This is accomplished by the following steps:
1. The most significant hazards in juice are described in terms of
severity and duration;
2. The hazards are described in terms of resulting health effects
and symptoms when they cause illness;
3. The health effects and symptoms are translated into consumer
utility losses;
4. The utility losses are translated into values in terms of lost
dollars (this gives the cost per case for every combination of level of
severity and for the specified duration for each hazard);
5. The average annual number of reported cases associated with
juice are distributed according to the percentages associated with each
level of severity;
6. The factors used to account for under reporting of foodborne
illness are estimated;
7. The reported cases are multiplied by the under reporting factors
to get the estimated average annual number of cases;
8. The percentages of each type of hazard expected to be prevented
by the proposal are listed; and
9. The total health benefits of the proposal are derived by
multiplying numbers 4, 7, and 8.
That is, TB = RC x CF x CR x V, where
TB = total health benefits in dollars,
RC = number of reported cases,
CF = under reporting correction factor,
CR = percent of cases reduced,
V = dollar value per case averted (medical costs + value of pain and
lost function).
1. Description of Microbial Hazards in Juice
Most of the significant health risks associated with juice products
are microbial. In the last 5 years the hazards associated with
commercially processed, packaged juice produced by nonretail
establishments include Bacillus cereus, Escherichia coli O157:H7, and
Salmonella non typhi.\5\ Table 2 lists these hazards with associated
severities and duration of severities. These hazards have been directly
linked to orange and apple juice products. However, all juices take
farm produce as an input; all use similar types of processing steps;
and all are distributed in similar ways. Therefore, although other
types of juices are less likely to be associated with foodborne disease
outbreaks primarily because consumption of orange and apple juice
greatly exceeds consumption of all other types of juice combined, all
juices are similarly vulnerable to microbial contamination. All juices
are sensitive to potential contamination by pathogenic microorganisms
due to the way fruits and vegetables are grown and harvested.
---------------------------------------------------------------------------

\5\ Most of the information in section V of this document
(Benefits) is taken from Ref. 9. It includes hazards other than
those for which benefits have been estimated in this analysis. The
hazards considered in section V of this document are those for which
the risk is highest. That is to say they are the most significant in
terms of probability of occurrence and severity.
---------------------------------------------------------------------------

Based on current scientific understanding, potential vehicles or
mechanisms for pathogenic cross contamination common to most fruit and
vegetable harvesting and juicing operations include water; manure
fertilizer; worker, field, and facility sanitation and transportation,
handling and processing. While most of the potential for contamination
would appear on the surface of the fruit or vegetable, the process of
juicing this

[[Page 24259]]

fruit or vegetable would potentially incorporate the pathogenic
microorganisms into the final juice product. Ref. 10, page 31, lists
the pH of some fruit and vegetable juices.

Table 2.--Description of Microbial Hazards in Juice
------------------------------------------------------------------------
Duration of
Hazard Severity Percent3 Illness (days)
------------------------------------------------------------------------
E. coli O157:H7
Mild 50 5
Moderate 32 9
Severe-acute 18 32
Severe- 2 26,645\1\
chronic
Death 1
Salmonella (non
typhi)
Mild 65 2
Moderate 30 5
Severe 5 17
Reactive 2 25
arthritis-
short term
Reactive 5 18,2502
arthritis-
long term
Death .1
B. cereus
Mild 99 .75
Moderate 1 1
Severe 0 NA
Death 0 NA
------------------------------------------------------------------------
\1\ Symptoms lasting 26,645 days, or 73 years, implies that it is
generally very young children who experience these severe chronic
effects (Ref. 2-3).
\2\Symptoms lasting 18,250 days, or 50 years. This estimate and other
information in section V of this document (Benefits) relating to
reactive arthritis are taken from Ref. 10.
\3\Percentages are taken from Ref. 10.

Symptoms of illness that results from exposure to each hazard may
be classified as mild, moderate, or severe. In general, mild cases are
not brought to the attention of a medical professional. Moderate cases
receive medical attention but do not require hospitalization. Severe
cases involve hospitalization and some of these result in death. The
``Percent'' column in Table 2 gives an estimate of the percentage of
the total number of cases that are classified in these four categories
of severity for each hazard. Note that the categories are not
necessarily mutually exclusive, for example, severe-chronic cases of E.
coli O157:H7 follow only after severe-acute cases of E. coli O157:H7,
and deaths follow only after severe cases. However, the ``Percent''
column reports each category of severity as a percentage of total cases
so that there is no double counting. Another factor that tends to
distinguish the categories of severity is the duration of time that
symptoms are experienced. The ``Duration'' column gives the general
duration of symptoms (in days) that are associated with the categories
of severity for each hazard.
2. Description of Health Effects and Symptoms of Microbial Hazards in
Juice
In order to quantify the loss (disutility) that individuals
experience from becoming ill, the pain, suffering, and mobility loss
must be scaled. Tables 3, 4, and 5 represent the outcome of one type of
scaling of these effects. Individuals who become ill experience
different levels of functional status in terms of mobility, ability to
do other physical activity, and ability to engage in social activities.
The ``Functional Status Code'' column in Table 3 represents the status
code which correlates with the categories of severity for each hazard.
Individuals who become ill also experience additional disutility due to
the symptoms of the illness. The ``Symptom/Problem Complex Code''
column represents the symptom/problem complex codes which correlate
with the categories of severity for each hazard. Descriptions of the
functional status and symptom/problem complex codes are given in Tables
4 and 5. FDA requests comment on this scaling model.

Table 3.--Description of Health Effects and Symptoms of Microbially Related Illnesses in Juice
----------------------------------------------------------------------------------------------------------------
Symptom/Problem Complex
Hazard Severity Functional Status Code1 Code2
----------------------------------------------------------------------------------------------------------------
E. coli O157:H7
Mild L20 8, 12, 13, 29
Moderate L19 8, 12, 13, 16, 19, 29,
32
Severe-acute (L1 x .2) + (L6 x 8, 12, 13, 16, 19, 29,
.8)\3\ 32
Severe-chronic L31 9
Salmonella (non typhi)
Mild L20 12, 13, 29
Moderate L20 12, 13, 29
Severe L6 12, 13, 16, 29
Reactive arthritis L35, L41, L42, L434 19
B. cereus
Mild L19 12, 13, 29
Moderate L19 12, 13, 29

[[Page 24260]]

Severe NA NA
----------------------------------------------------------------------------------------------------------------
\1\ Functional Status Codes are described in Table 4.
\2\ Symptom/Problem Complex Codes are described in Table 5.
\3\ The disutilities for two functional status codes were taken for severe cases of E. coli O157:H7 because
functional status varies among severe cases of this hazard.
\4\ Functional Status Code varies, Ref. 10.

In Table 4, the last column, ``Level of Disutility,'' represents
the degree of departure from perfect functionality. Thus, a person
would be functioning at about half capacity if the level was .5 and
would be even more diminished at .75. Code L42 is used whenever the
mobility, physical activity, and social activity conditions apply and a
person is experiencing a symptom described in Table 5. Code L43 is used
whenever the mobility, physical activity, and social activity
conditions apply and a person is experiencing no symptoms. In Table 5,
``Level of Disutility'' refers to the amount of pain and suffering such
that .03 would be minor pain and suffering relative to .3.

Table 4.--Description of Functional Status Codes1
------------------------------------------------------------------------
Function Status Physical Social Level of
Levels Mobility Activity Activity Disutility
------------------------------------------------------------------------
L1................ In special In bed or Had help .5626
care unit chair with
self-
care
L6................ In In bed or Had help .5301
hospital chair with
self-
care
L19............... In house Walked Performed .4176
with self-
physical care but
limitati not
ons work,
school,
or
housewor
k
L20............... In house Walked Limited .4448
with in work,
physical school,
limitati or
ons housewor
k
L23............... In house Walked Performed .3512
without self-
physical care,
limitati but not
ons work,
school,
or
housewor
k
L31............... Did not Walked Limited .4087
drive, without in work,
needed physical school,
help with limitati or
transport ons housewor
ation k
L35............... Drove car Walked Limited .3980
and used with in work,
transport physical school,
ation limitati or
without ons housewor
help k
L41............... Drove car Walked Did work, .3145
and used without school,
transport physical or
ation limitati housewor
without ons k, but
help other
activiti
es
limited
L42............... Drove car Walked Did work, .2567
and used without school,
transport physical or
ation limitati househol
without ons d, and
help other
activiti
es
L43............... Drove car Walked Did work, .0000
and used without school,
transport physical or
ation limitati househol
without ons d, and
help other
activiti
es
------------------------------------------------------------------------
\1\ Ref. 4.

Table 5.--Description of Symptom/Problem Complex Codes1
------------------------------------------------------------------------
Symptom/Problem Level of
Complex Description Disutility
------------------------------------------------------------------------
8................. Itching, bleeding or pain in
rectum .0379
9................. Pain in chest, stomach, side,
back, or hips .0382
12................ Sick or upset stomach, vomiting,
or diarrhea (watery bowel
movements) .0065
13................ Fever chills with aching all
over and vomiting or diarrhea .0722
16................ Headache, dizziness, or ringing
in ears .0131
19................ Pain, stiffness, numbness, or
discomfort of neck, hands,
feet, arms, legs ankles, or
several joints together .0344
29................ General tiredness, weakness, or
weight loss .0027
32................ Loss of consciousness such as
seizures (fits), fainting, or
coma (out cold or knocked out) .1507
------------------------------------------------------------------------
\1\ Ref. 4, p. D-14.

3. Utility Losses From Microbial Hazards in Juice
The ``Functional Status Code'' translates into values of disutility
given in the ``Functional Disutility'' column in Table 6. The symptom/
problem complex code translates into values of

disutility given in the ``Symptom/Problem Disutility'' column in Table
6. The ``Total Disutility'' column is the sum of the ``Functional
Disutility'' and the ``Symptom/Problem Disutility'' columns. The
``Utility Losses for Survivors'' column is derived by multiplying the
total disutility per day by the number of days that symptoms of the
illness persists. This gives the utility loss for survivors in terms of
the number of quality adjusted life days (QALD's) for each case of the
categories of severity for each hazard.\6\ FDA requests comment on this
estimation of utility loss.
---------------------------------------------------------------------------

\6\ A QALD is a day of perfect health.

[[Page 24261]]

Table 6.--Utility Losses From Microbial Hazards in Juice
----------------------------------------------------------------------------------------------------------------
Symptom/
Functional Problem Total Utility Losses
Hazard Severity Disutility Disutility Disutility for Survivors
(per day) (per day) (per day) (QALD's)
----------------------------------------------------------------------------------------------------------------
E. coli O157:H7
Mild .4448 .1193 .5641 2.8
Moderate .4176 .1668 .5844 5.3
Severe-acute .5464 .3175 .8639 27.8
Severe-chronic .4087 .0382 .4469 11,907.7
Salmonella (non typhi)
Mild .4448 .0814 .5262 1.1
Moderate .4448 .0814 .5262 2.6
Severe .5301 .0945 .6246 10.6
Reactive arthritis- .3980 .0344 .4324 10.8
short term
Reactive arthritis- .2582 .0280 .2862 5,223.2
long term
B. cereus
Mild .4176 .0814 .4990 .4
Moderate .4176 .0814 .4990 .5
Severe 0 0 0 0
----------------------------------------------------------------------------------------------------------------

4. Value of Losses From Microbial Hazards in Juice
FDA values a QALD at $630. This value derives from the statistical
estimate of a unit-risk reduction (commonly referred to as the value of
a statistical life (VSL)) which the Department of Health and Human
Services assigns the value of $5 million. Using $5 million for a full
lifetime yields a value for a quality adjusted life year (QALY) of
approximately $230,000, when discounted at 7 percent. (A QALY is the
estimated value of a year spent in perfect health. These values are
discounted to reflect time preferences for investments in health. That
is, as with any other commodity, people have a stronger preference for
good health now than they have for good health in the future. Costs or
benefits realized in the future are ``discounted'' to make them
comparable to today. Essentially, discounting is the inverse of the
interest rate. Thus, if a benefit of $1.10 were to be realized 1 year
in the future, this would be equivalent, at approximately a 10 percent
discount rate, to a benefit of $1 realized today. This is the reverse
of saying that $1 invested today at a 10 percent annual interest rate
is worth $1.10 1 year from now.) Dividing this value by 365 days per
year yields a value for a QALD of approximately $630. The ``Value of
Utility Losses for Survivors'' column in Table 7 comes from multiplying
the number of QALD's lost due to the illness (see ``Utility Losses for
Survivors'' in Table 6) by the value of a QALD, $630. This represents
the value of pain and mobility losses that individuals experience.
Additionally, there are the societal costs of medical treatment. These
costs are shared generally between insurance companies and individuals.
They include all aspects of medical expenses (e.g., physician visits,
laboratory tests, prescriptions and therapies, hospital stays). These
are estimated in the ``Medical Costs'' column in Table 7 (Ref. 2-3, pp.
19 and 40 and Ref. 10). The ``Value of Losses per Case'' column in
Table 7 is the sum of the ``Value of Utility Losses for Survivors''
column and the ``Medical Costs'' column for the categories of severity
for each hazard. FDA requests comment on these valuations.

Table 7.--Value of Losses From Microbial Hazards in Juice
----------------------------------------------------------------------------------------------------------------
Value of Utility
Losses for Value of Losses per
Hazard Severity Survivors Medical Costs Case (VSL=$5,000,000)
(QALD=$630) (QALD=$630)
----------------------------------------------------------------------------------------------------------------
E. coli O157:H7
Mild $1,800 $0\1\ $2,000
Moderate $3,300 $200\1\ $4,000
Severe-acute $17,200 $16,000\2\ $33,000
Severe-chronic $995,700 $225,000\3\ $1,221,000
Death NA NA $5,000,000
Salmonella (non typhi)
Mild $700 $200\4\ $1,000
Moderate $1,600 $800\4\ $2,000
Severe $6,700 $9,100\4\ $16,000
Reactive arthritis- $6,800 $1005 $7,000
short term
Reactive arthritis- $970,0005 $5,8605 $976,000
long term
Death NA NA $5,000,000
B. cereus
Mild $300 $0\6\ $300
Moderate $300 $100\6\ $400
Severe $0 $0 $0

[[Page 24262]]

Death NA NA $5,000,000
----------------------------------------------------------------------------------------------------------------
\1\ Ref. 2-3, p. 40.
\2\ Explained in Table 8.
\3\ Recalculated from data in Buzby et al., pp. 41-45 in order to arrive at the present value of the cost per
case using a 7 percent discount rate.
\4\ Buzby et al., pp. 18-19. Mild Salmonella medical costs are recalculated from data in Cohen, M. L. et al. so
as not to include productivity in medical costs.
\5\ Ref. 10.
\6\ The medical cost estimates for B. cereus were made by FDA for this analysis. The extremely brief duration of
mild cases suggests that there would be no medical costs for this level of severity. For moderate cases one
visit to a doctor with medical tests are estimated to cost approximately $100.

Table 8.--Medical Costs for Severe-Acute Cases Associated With E. coli O157:H71
----------------------------------------------------------------------------------------------------------------
Acute Hemorrhagic Average Severe-
Factors Colitis Acute HUS Acute Case
----------------------------------------------------------------------------------------------------------------
Percent of Severe Cases 80% 20%
Present Value per Case $11,000 $36,000
Weighted Present Value per Case $8,800 $7,200 $16,000
----------------------------------------------------------------------------------------------------------------
\1\ Ref. 2-3, p. 40.

5. Distribution of the Reported Cases per Year for Microbial Hazards in
Juice
Table 9 estimates the number of cases associated with each hazard
by severity. The ``Average Total No. of Cases Reported per Year''
column represents the average number of reported cases for each hazard
from 1992 through 1996. Cases for each hazard are divided among the
four categories of severity according to the percentages described in
Table 8. Only those reported cases associated with commercially-
produced juices sold in interstate commerce as beverages or used as
ingredients in beverages are included in the averages presented.

Table 9.--Distribution of the Reported Cases per Year for Microbial
Hazards in Juice
------------------------------------------------------------------------
Average No. of
Hazard Severity Percent Cases Reported per
Year
------------------------------------------------------------------------
Mild 50 8
Moderate 32 5
Severe-acute 18 3
Severe- 2 .3
chronic
Death 1 .2
E. coli O157:H7 Total cases 16\1\

Mild 65 8
Moderate 30 4
Severe 5 1
Reactive 2 .2
arthritis-
short term
Reactive 5 1
arthritis-
long term
Death .1 .01
Salmonella (non Total cases 12
typhi)

Mild 99 17
Moderate 1 .2
Severe 0 0
Death 0 0
B. cereus Total cases 17
------------------------------------------------------------------------
\1\ Total cases per pathogen are accurate. The sum of the number of
cases for all levels of severity per pathogen may not equal the total
number of cases per pathogen due to rounding.

6. Estimates of Factors Needed to Offset Underreporting of Foodborne
Illness
The cases reported in column 4 in Table 10 are the lower bound of
the likely total number of these cases. The total number of foodborne
illness is much greater than those numbers reported to the Centers for
Disease Control and Prevention (CDC) for several reasons. First,
individuals who become ill do not always go to doctors. This is
particularly true for milder cases of foodborne disease. Obviously, if
people do not go to health care professionals, the illnesses will not
be captured in any data base and will not be picked up by CDC. Second,
even when people go to health care professionals, they are not
necessarily diagnosed as having foodborne disease as the symptoms for
many types of

[[Page 24263]]

foodborne disease are common to influenza and other diseases. There is
often little incentive to culture stools to definitively identify a
pathogen if the disease is thought to be of short duration and not
requiring treatment. Even where a pathogen is identified, there is even
less incentive to identify the food or other vehicle which carried it.
Third, even when a correct diagnosis is made, State and local health
professionals do not always report these cases upwards, particularly
going as far as CDC. Again, milder cases are less likely to be reported
than more severe cases.\7\ To complicate matters, the rate of under
reporting is not observable, and, even if it were known in any 1 year,
it may fluctuate dramatically from year to year. Nevertheless, in order
to compensate for the rate of under reporting, the number of known
cases associated with a hazard (i.e., reported to CDC) is multiplied by
factors which are estimated to account for underreporting.
---------------------------------------------------------------------------

\7\ The CAST Report expands these three categories of reasons
that a case of illness may not be recognized as foodborne into six
reasons (Ref. 6).
---------------------------------------------------------------------------

In Foodborne Pathogens: Risks and Consequences (the CAST Report)
there are two estimates given of the actual number of foodborne
illnesses: One estimate made by Bennett et al., and one made by Todd
(Ref. 6, p. 46). Both Bennett et al. and Todd estimate the total number
of cases and the total number of deaths for each hazard. By dividing
Bennett's et al. and Todd's estimates of the actual number of cases and
deaths by the number of reported cases and deaths (Ref. 6, p. 42), the
respective implicit factors needed to correct for underreporting of
these categories for each hazard are derived. Based on these correction
factors, FDA has estimated correction factors for each category of
severity. The agency has taken the correction factor for the number of
cases as the correction factor for mild cases and the correction factor
for the number of deaths as the correction factor for severe cases. For
moderate cases, the agency has interpolated between the factors for
mild and severe cases. E. coli O157:H7 was not a recognized food-safety
hazard at the time that Bennett's et al. work was done. For a more
complete description of how these estimates were derived see the
Appendix attached to this document (Ref. 9).
In Table 10, the third column, ``Estimate of Underreporting
Correction Factor (Bennett),'' and the fifth column, ``Estimate of
Underreporting Correction Factor (FDA based on Todd),'' give the exact
implicit correction factors that can be derived from the work of
Bennett and Todd et al. The fourth column, ``Estimate of Underreporting
Correction Factor (FDA based on Bennett),'' and the sixth column,
``Estimate of Underreporting Correction Factor (FDA based on Todd),''
give FDA's interpolations of the work of Bennett and Todd et al. for
each of the identified categories of severity. In general, each
researcher's estimate of the underreporting correction factor for total
cases was used as the estimate for mild cases, and each researcher's
estimate of the underreporting correction factor for deaths was used as
the estimate for deaths and severe cases. FDA interpolated between each
researcher's estimates of underreporting for total cases and deaths to
derive under reporting rates for moderate cases. FDA requests comment
on these estimates of underreporting.

Table 10.--Estimates of Factors Needed to Offset Underreporting of Foodborne Illness
----------------------------------------------------------------------------------------------------------------
Estimate of
Estimate of Underreporting Estimate of Estimate of
Underreporting Correction Underreporting Underreporting
Hazard Severity Correction Factor (FDA Correction Correction
Factor based on Factor (Todd) Factor (FDA
(Bennett) Bennett) based on Todd)
----------------------------------------------------------------------------------------------------------------
Mild 195
Moderate 20
Severe 7
Death 7 7
E. coli O157:H7 Total cases ND\1\ 195

Mild 307 474
Moderate 307 45
Severe 246 4
Reactive arthritis- 307 474
short term
Reactive arthritis-long 307 474
term
Death 246 246 4 4
Salmonella (non typhi) Total cases 307 474

Mild 96 1,615
Moderate 96 1,615
Severe NA NA NA NA
Death NA NA NA NA
B. cereus Total cases 96 1,615
----------------------------------------------------------------------------------------------------------------

7. Estimates of Juice-Associated Cases per Year
In Table 11, FDA has estimated ranges of the likely annual number
of cases that occur for each of the four pathogens studied. The column
``Estimate of Actual No. of Juice Associated Cases per Year (FDA based
on Bennett)'' in Table 11 is derived by multiplying the ``Average Total
No. of Reported Cases per Year'' column in Table 9 by the ``Estimate of
Underreporting Correction Factor (FDA based on Bennett)'' column in
Table 11. The column ``Estimate of Actual No. of Juice Associated Cases
per Year (FDA based on Todd)'' in Table 11 is calculated in a similar
manner.

[[Page 24264]]

Table 11.--Estimates of Juice-Associated Cases per Year
----------------------------------------------------------------------------------------------------------------
Estimate of Estimate of Estimate of Actual Estimate of Actual
Underreporting Underreporting No. of Juice- No. of Juice-
Hazard Severity Correction Factor Correction Factor Associated Cases Associated Cases
(FDA based on (FDA based on per Year (FDA per Year (FDA
Bennett) Todd) based on Bennett) based on Todd)
----------------------------------------------------------------------------------------------------------------
Mild ND 195 ND 1,560
Moderate ND 20 ND 100
Severe-acute ND 7 ND 20
Severe-chronic ND 7 ND 2
Death ND 7 ND 1
E. coli O157:H7 Total cases ND 1,700

Mild 307 474 2,460 3,790
Moderate 307 45 1,230 180
Severe 246 4 150 2
Reactive 307 474 60 100
arthritis-
short term
Reactive 307 474 180 280
arthritis-long
term
Death 246 4 2 .04
Salmonella Total cases 3,800 4,000
(non typhi)

Mild 96 1,615 160 2,750
Moderate 96 1,615 2 30
Severe 0 0 0 0
Death 0 0 0 0
B. cereus Total cases 200 2,800
----------------------------------------------------------------------------------------------------------------

8. Percent of Cases Preventable by HACCP Proposal
In general, most pathogens will be eliminated when juice is heat-
treated. For example, E. coli O157:H7, and Salmonella should all be
completely eliminated from juice by standard methods of flash
pasteurization (absent extraordinarily high counts, detrimental human
intervention, or equipment failure). However, hazards associated with
B. cereus will not necessarily be eliminated by heat treatment. This
bacterium forms spores which are more difficult to kill by heat. After
heat treatment, if the spores survive, they may grow out and produce a
toxin which causes illness. Ideally, the best way to reduce illness
associated with B. cereus is by killing the bacterium in its nonspore
state before any toxin has been produced. For most types of heat-
treated juice, there is a small probability that the heat treatment
will take place when B. cereus is in its nonspore state. To the extent
that processors adopt controls for these hazards other than flash
pasteurization which are less effective, the percentage of cases
prevented may be smaller than those estimated here. FDA requests
comment on these estimates. Based on information from USAA, FDA
estimates that the exemption from the HACCP rule for retailers and
small retail processors will affect 14 percent of the volume of
unpasteurized juice. Therefore, the agency estimates that though
pathogen controls may be 100 percent effective in controlling some
hazards, such controls will only prevent 86 percent of the cases of
illness from these hazards.

Table 12.--Percent of Cases Preventable by HACCP Proposal
------------------------------------------------------------------------
Percent of Cases
Hazard Preventable by
HACCP Proposal
------------------------------------------------------------------------
E. coli O157:H7 86
Salmonella (non typhi) 86
B. cereus 9
------------------------------------------------------------------------

9. Estimates of Annual Benefits for HACCP Proposal
The total benefits for the categories of severity for each hazard
are derived by multiplying the percentage of cases preventable by the
HACCP proposal by the estimates of the number of actual cases. The sum
of those benefits for each hazard is the total benefits of the HACCP
proposal for pathogen control. Table 13 gives the estimate of benefits
for each hazard using each source of information on the appropriate
correction factor for underreporting.

Table 13.--Estimates of Annual Benefits for HACCP Proposal
------------------------------------------------------------------------
FDA Estimate of FDA Estimate of
Hazard Severity Annual Benefits Based Annual Benefits Based
on Bennett on Todd
------------------------------------------------------------------------
Mild $2,680,000
Moderate $360,000
Severe- $660,000
acute
Severe- $2,442,000
chronic
Death $5,000,000
E. coli O157:H7 Total $11,142,000

Mild $2,120,000 $3,260,000

[[Page 24265]]

Moderate $2,120,000 $300,000
Severe $2,080,000 $32,000
Reactive $350,000 $630,000
arthriti
s-short
term
Reactive $146,400,000 $234,240,000
arthriti
s-long
term
Death $10,000,000 $200,000
Salmonella (non Total $163,070,000 $238,662,000
typhi)

Mild $42,000 $711,000
Moderate $1,000 $12,000
Severe 0 0
Death 0 0
B. cereus Total $43,000 $725,000
------------------------------------------------------------------------

Table 14 presents a range of estimates of annual benefits based on
the estimates in Table 13. The low and high estimates do not represent
lower and upper bounds of benefits, but only a range of potentially
likely estimates.

Table 14.--Range Estimates of Annual Microbially Related Benefits for
HACCP Proposal
------------------------------------------------------------------------
Low Estimate of Annual High Estimate of Annual
Hazard Benefits Benefits
------------------------------------------------------------------------
E. coli O157:H7 $11,142,000 $11,142,000
Salmonella (non
typhi)\1\ $163,070,000 $238,662,000
B. cereus\1\ $43,000 $725,000
Totals $174,000,000 $251,000,000
------------------------------------------------------------------------
\1\ Ranges for these two pathogens are taken from two different
estimates that exist in the public health literature. The estimates
for the other pathogen was made by FDA, alone.

10. Percent of Cases Preventable by Labeling Proposal
FDA does not have direct estimates of the effects of a warning
label on the incidence of illness from juice consumption. FDA
indirectly estimates the effects by estimating how warning labels will
change consumption, assuming that changes in the number of illnesses
are proportional to changes in consumption. FDA believes that the
labeling rule will cause a reduction in the consumption of
unpasteurized juice, but the size of the reduction is uncertain. As a
likely value, FDA estimates that consumption and illnesses will decline
by 5 percent in response to the warning label. The 5 percent reduction
is the estimated effect on cooking practices of the USDA meat safe
handling label, as found in a recent survey (Ref. 11). However, there
are some dissimilarities between the meat and juice labels, most
particularly that the juice label is targeted at sensitive consumers.
If, for example, parents redirect children away from nonheat-treated
juice, then consumption and illness will decline by 16 percent, which
is the proportion of apple cider consumed by children under the age of
6 (Ref. 12). This estimate embodies the assumptions that cider
consumption is a good proxy for unpasteurized juice consumption, and
that parents will not let their children consume unpasteurized juices.
11. Estimates of Annual Benefits for Labeling Proposal
Table 11 shows FDA's estimate that there are approximately 5,600
cases of foodborne illness associated with commercially processed,
package juice produced by nonretail establishments. In addition to
these cases, an average of 6 cases annually of Cryptosporidium parvum
have been associated with commercially processed, packaged juice
produced by retail establishments exempted from the HACCP rule. Table
15 shows the agency's estimate of the actual number of cases per year
by severity.

[[Page 24266]]

Table 15.--Estimates of Juice-Associated C. parvum Cases per Year
------------------------------------------------------------------------
Average No.
of Cases FDA Estimate of FDA Estimate of
Severity Reported Underreporting Actual No. of Juice-
per Year Correction Factor\1\ Associated Cases per
(1992-1996) Year
------------------------------------------------------------------------
Mild 5 100 500
Moderate 1 10 10
Severe .06 5 .3
Death .001 5 .005
Total 6 500
------------------------------------------------------------------------
\1\Because C. parvum was not a recognized food safety hazard at the time
that Bennett et al. and Todd's work was done, FDA has made its own
estimates of the factors needed to correct for underreporting of this
hazard.

Table 16 gives the agency's estimate of the value of the loss per
case of C. parvum.

[[Page 24267]]

Table 16.--Estimate of Value of Losses Associated with Case of C. parvum
--------------------------------------------------------------------------------------------------------------------------------------------------------
Symptom/ Value of
Duration of Function Problem Total Utility Losses Utility Losses Value of Losses per
Severity Percent Illness (in Status Code\1\ Complex Disutility for Survivors for Survivors Medical Costs Case (VSL=$5,000,000)
days) Code\2\ (per day) (QALD's) (QALD=$630) (QALD=$630)
--------------------------------------------------------------------------------------------------------------------------------------------------------
Mild 90 9 L41 12, 13, 29 .3959 3.6 2,300 $0\3\ $2,000
Moderate 9 17 L41 12, 13, 29 .3959 6.7 $4,200 $400\3\ $5,000
Severe 1 24 L6 12, 13, 29 .6115 14.7 $9,300 $8,300\4\ $18,000
Death .02 NA NA $5,000,000
--------------------------------------------------------------------------------------------------------------------------------------------------------
\1\Functional Status Codes are described in Table 4.
\2\Symptom/Problem Complex Codes are described in Table 5.
\3\Medical Costs for mild and moderate cases of C. parvum were calculated by multiplying the per day medical costs for E. coli 0157:H7 for these levels
of severity by the duration of illness of C. parvum. The symptoms of C. parvum for these levels of severity are similar to those of E. coli 0157:H7.
\4\Medical Costs for severe cases of C. parvum were calculated by multiplying the per day medical costs for severe cases of acute hemorrhagic colitis by
the duration of illness of C. parvum. The implicit assumption is that the medical costs for acute hemorrhagic colitis (bloody diarrhea) are equivalent
to the medical costs for watery diarrhea associated with C. parvum.

[[Page 24268]]

The labeling rule is expected to prevent some cases of foodborne
illness as people avoid juice that is labeled. Because B. cereus is, in
general, not disproportionately associated with minimally processed
juice, cases of B. cereus are not expected to be prevented by the
labeling. However, to the extent that the label is effective and to the
extent of the volume of juice that is labeled, the labeling rule will
reduce the number of cases associated with E. coli 0157:H7, Salmonella
and C. parvum.
Combining the estimates of the number of illnesses in Tables 11
and 15, the total number of estimated cases associated with minimally
processed juice for these 3 hazards is 6,100 per year associated with
consumption of the 70 million gallons of minimally processed juice
produced annually. FDA has estimated that 14 percent of minimally
processed juice (10 million gallons) will be exempt from the HACCP rule
but will be covered by the labeling rule. Therefore, the number of
illnesses that may be associated with this volume of juice (10 million
gallons) will be exempt from the HACCP rule but will be covered by the
labeling rule. Therefore, the number of illnesses which may be
associated with this volume of juice (10 million gallons) is
approximately 900 and 5,200 illnesses are associated with minimally
processed juice covered by the HACCP rule.
As stated earlier, FDA estimates that consumption of labeled,
minimally processed juice will decline by 5 percent in response to the
warning label. This leads to the conclusion that the labeling rule is
expected to prevent approximately 50 illnesses annually (900 x .05). If
juice consumption decreases by as much as 16 percent in response to the
warning label, then the labeling rule may prevent as many as 140
illnesses per year.
The value of this reduction in illness depends on the type of cases
prevented. FDA assumes that these cases will be distributed according
to the share of illnesses associated with each of these hazards. Table
17 shows the expected distribution of cases prevented by labeling
across the hazards and severities.

[[Page 24269]]

Table 17.--Distribution of Cases Prevented by Labeling Proposal
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Low Estimate of High Estimate of Low Estimate of No. High Estimate of No. Low Estimate of No. High Estimate of No.
Actual No. of Juice- Actual No. of Juice- of Cases Prevented by of Cases Prevented by of Cases Prevented by of Cases Prevented by
Hazard Severity Associated Cases per Associated Cases per a 5% Consumer a 5% Consumer a 16% Consumer a 16% Consumer
Year Year Response to Labeling Response to Labeling Response to Labeling Response to Labeling
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Mild 1,560 1,560 13 13 36 35
Moderate 100 100 1 1 2 2
Severe-acute 20 20 .2 .2 .5 .5
Severe-chronic 2 2 .02 .02 .05 .05
Death 1 1 .008 .008 .02 .02
E. coli 0157:H7 Total 1,700 1,700 14 14 39 38

Mild 2,460 3,790 20 31 58 87
Moderate 1,230 180 10 1 29 4
Severe 150 2 1 .02 4 .05
Reactive arthritis-short term 60 100 .5 .8 1 2
Reactive arthritis-long term 180 280 1 2 4 6
Death 2 .04 .02 .0003 .05 .0009
Salmonella (non typhi) Total 3,800 4,000 32 32 90 91

Mild 500 500 4 4 11 11
Moderate 10 10 .08 .08 .2 .2
Severe .3 .3 .002 .002 .006 .006
Death .005 .005 .00004 .00004 .0001 .0001
C. parvum Total 500 500 4 4 11 11

Total 6,000 6,200 50 50 140 140
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

[[Page 24270]]

Table 18.--Value of Losses Prevented by the Labeling Proposal
--------------------------------------------------------------------------------------------------------------------------------------------------------
High Estimate of High Estimate of
Low Estimate of Value Value of Losses Low Estimate of Value Value of Losses
Hazard Severity of Losses Prevented Prevented by a 5% of Losses Prevented Prevented by a 16%
by a 5% Consumer Consumer Response to by a 16% Consumer Consumer Response to
Response to Labeling Labeling Response to Labeling Labeling
--------------------------------------------------------------------------------------------------------------------------------------------------------
Mild 26,000 26,000 72,000 70,000
Moderate 4,000 4,000 8,000 8,000
Severe-acute 7,000 7,000 17,000 17,000
Severe-chronic 24,000 24,000 61,000 61,000
Death 40,000 40,000 100,000 100,000
E. coli 0157:H7 Total 101,000 101,000 258,000 258,000

Mild 20,000 31,000 58,000 87,000
Moderate 20,000 2,000 58,000 8,000
Severe 16,000 300 64,000 1,000
Reactive arthritis-short term 4,000 6,000 7,000 14,000
Reactive arthritis-long term 976,000 1,952,000 3,904,000 5,856,000
Death 100,000 2,000 250,000 5,000
Salmonella (non typhi) Total 1,136,000 1,993,000 4,341,000 5,971,000

Mild 8,000 8,000 22,000 22,000
Moderate 400 400 1,000 1,000
Severe 0 0 100 100
Death 200 200 500 500
C. parvum Total 9,000 9,000 24,000 24,000

Total 1,000,000 2,000,000 5,000,000 6,000,000
--------------------------------------------------------------------------------------------------------------------------------------------------------

12. Pesticide Residues
Tolerances for pesticides in foods are established by the
Environmental Protection Agency (EPA) and enforced by FDA. FDA collects
samples for both surveillance and compliance purposes. Since the
incidence of violative pesticide residues in fruit and vegetable juices
is relatively low, few compliance samples are taken.
This discussion pertains to surveillance samples of fruit and
vegetable juices from 1991 through 1997 (see Table 15). The lab
classification scheme used for pesticide residues is:
1 = in compliance;
2 = not in compliance, but not of regulatory concern; and
3 = not in compliance, and of regulatory concern.
The class 2 and 3 violative sample data are summarized in Table 15.
Of the 1,196 surveillance samples of juice taken and analyzed during
this period, only three (approximately one quarter of one percent) were
class 3 violative. One was apple cider and the other two were apple
juice, and the violative pesticide residue was acephate in each case.
There were also five class 2 violations, in which trace quantities of a
pesticide with no tolerance (i.e., the pesticide was not approved for
use in the commodity) were found. The products with class 2 violations
were grape juice, watermelon juice concentrate, strawberry/nectarine
juice (2 samples), and apple juice concentrate; the pesticides were
chlorpyrifos, acephate, and methamidophos.
Pesticides present some potential chronic risks to humans at very
low levels of exposure. There is a small background risk associated
even with nonviolative pesticide residues and, in the case of products
with violative levels, an added risk from the violative residues.
(Violative residues are residues above tolerance or residues of
pesticides with no tolerance.)

Table 19.--Violative Pesticide Residues in Fruit and Vegetable Juices, 1991 through 1997
----------------------------------------------------------------------------------------------------------------
Commodity Fiscal Year Pesticide Amount Found, ppm Tolerance, ppm Class Violation
----------------------------------------------------------------------------------------------------------------
Grape juice 1993 Chlorpyrifos Trace None 2
Apple cider 1995 Acephate 0.075 None 3
Apple juice 1995 Acephate 0.052 None 3
Apple juice 1995 Acephate 0.040 None 3
Watermelon juice, 1995 Acephate Trace None 2
concentrate
Strawberry/ 1996 Methamidophos Trace None 2
nectarine juice
Strawberry/ 1996 Methamidophos Trace None 2
nectarine juice
Apple juice, 1997 Methamidophos Trace None 2
concentrate
----------------------------------------------------------------------------------------------------------------

There are two potential benefits associated with the regulation of
pesticides: (1) Decreases in cancer and other illness caused by chronic
consumption of pesticide residues and, (2) social benefits associated
with reductions in the costs of recapturing firm goodwill. The U.S. EPA
is responsible for determining the benefits of reducing exposure to
pesticide residues and, it is assumed, that the health benefits of the
enforcement actions proposed here are already accounted for when
regulatory tolerances are established. As to the

[[Page 24271]]

latter benefit, when firms have products with violative residues either
over tolerance for legal pesticides or any residue of an illegal
pesticide and a recall of the violative product becomes publicly known,
the sales of those firms are reduced, at least temporarily. Because
other firms will step in to supply the product, that loss of sales
alone does not constitute a social cost. However, it is likely that
real resources will be expended to recapture the lost ``goodwill'' that
would be in addition to the real expenditures made to actually recall
the product. FDA cannot quantify the cost savings that will occur
because of more vigilant monitoring of pesticide residues by firms
under a HACCP rule.

C. Other, Nonquantified Benefits

1. Firm Efficiency
The principle benefits from HACCP reported by the pilot firms are
more effective and efficient operations, a higher level of confidence
in the safety of the product, and greater customer satisfaction. The
pilot firms attributed these benefits to HACCP because of the following
results.
(1) Training makes the employees more aware of safety and needed
control measures, and empowers employees to prevent problems and
respond properly when deviations occur. Improvement in employee
performance was perhaps the most significant benefit from HACCP
expressed to FDA by the pilot firms. One firm reported that ``due to
increased HACCP awareness, employees have been instrumental in
designing new processes/procedures for monitoring and control.'' The
firm gave an example of a processing step that was changed to reduce
the likelihood of occurrence of a physical hazard. FDA is unable to
estimate the societal cost savings in terms of reduced product costs
which will, ultimately, affect the cost of implementing HACCP.
(2) SOP's and other documented procedures enable employees to
implement their tasks more consistently and effectively, and result in
smoother operations.
(3) Prerequisite programs and incoming ingredient controls prevent
hazards from being introduced into the process; continuous monitoring
reveals problems quickly and enables prompt correction and continuation
of production with less waste.
(4) Recordkeeping and review makes employees more accountable and
conscientious about safety.
(5) Validation and verification activities provide management with
greater control over their operations and documentation of the safety
of their product.
Perhaps the most significant benefit in terms of firm efficiency
will be cost savings from greater awareness by firms of violative
product runs, and the resulting increase in response to such violative
runs. Although the benefits of formal recalls have already been
accounted for, many pilot plant managers suggested that the continuous
monitoring required by HACCP enabled them to decrease the amount of
waste associated with production-line problems. For example, one
manufacturer noted that glass breakage was a constant problem on the
line and that, prior to HACCP, almost an entire lot would have to be
discarded because the manager could not be sure exactly when a problem
had started. With continuous HACCP monitoring, problems were caught
more quickly and the problem corrected more promptly, thereby
minimizing the amount of lost product.
The cost savings may be substantial from this source of benefits
but FDA is unable to quantify them. FDA requests comments on these and
other potential benefits.
2. Increased Shelf Life
Nonheat-treated juices have a limited shelf life. Heat-treated
juices have longer shelf lives. Depending upon temperature used,
increases of 7 days or more have been reported. Longer shelf life
allows more flexibility in the conditions of distribution and sale of
products. The agency requests comments on how this potential benefit
may be quantified.

D. Summary of Benefits

Table 20 summarizes the benefits of these two rules.

Table 20.--Benefits of Juice Proposals
------------------------------------------------------------------------
Type of Benefit Description Annual Value
------------------------------------------------------------------------
Enforcement: Reduced waste and Federal activity $175,000
Import from detaining violative juice
Detentions imports
Enforcement:Prod Reduced numbers of domestic recalls $1,500,000
uct Recalls of violative juice products
Health Benefits: Reduced illness and death from $174 to 251
HACCP controlling pathogens in juice million
Health Benefits: Reduced illness and death from $1 to $6 million
Labeling avoidance of minimally processed
juice
Health Benefits: Reduction of consumption of Not quantified
Pesticides violative pesticide residues in but small
juice and social losses from lost
goodwill
Other Benefits: Some offsetting reductions in Not quantified
Firm Efficiency manufacturing costs due to but potentially
increased worker productivity and large
less product waste
Other Benefits: Product Shelf life may be increased Not quantified
Increased Shelf for products achieving a 5-log but potentially
Life reduction of pathogens large
Total Quantified .................................... $180 to 260
Benefits million
------------------------------------------------------------------------

VI. Costs

A. General Industry Information Used Throughout This Analysis

The costs of these rules have been estimated by analyzing the costs
for each proposed requirement on a per-plant basis and multiplying
these costs by the number of plants affected by each requirement. Cost
per plant will vary by current practice, product, and size. In order to
determine the number of plants covered, the analysis will first analyze
coverage qualitatively.
1. Types of Plants Covered
The labeling rule and the HACCP rule do not equally affect an
identical subset of the food industry.
2. HACCP Rule Coverage
For the purpose of this rule, FDA has tentatively decided that
retailers will include processors who are very small businesses and who
make juice on their premises and directly sell juice or juice products
to consumers and other retailers provided that retail sales of juice
and juice products do not exceed 40,000 gallons per year. The HACCP
rule covers all processors of juice except those who are retailers.
Retailers may include grocery stores, supermarkets, farms, roadside
stands, restaurants and eating places.

[[Page 24272]]

3. Labeling Rule Coverage
The labeling rule covers processors and retailers of packaged
minimally processed juice. The labeling rule is also applicable to
packaged beverages that have not received further processing to control
microbial hazards and that contain minimally processed juice. Such
beverages include diluted juice beverages, ``smoothies,'' sports
drinks, flavored bottled waters, and carbonated beverages that contain
juice that was not processed to control pathogens.
Table 21 provides examples of the types of products and processors
covered and not covered by the two rules.

Table 21.--Coverage of Juice Proposals
------------------------------------------------------------------------
Covered by Covered by
Processor Type Labeling Rule HACCP Rule\3\
------------------------------------------------------------------------
Processors of packaged beverages sold as Yes Yes
juice\1\
Processors of packaged purees sold as Yes Yes
juice
Processors of juice used as an Yes Yes
ingredient in a beverage (e.g., the
cranberry juice in cranberry juice
cocktail)
Processors of juice which retail the Yes Yes
juice at a different location from
which it is produced
Processors of beverage concentrates sold Yes Yes
as juice
Processors of beverage bases of a fruit Yes Yes
origin or other beverage bases
including dried or powdered juice
mixes\2\
Processors of packaged baby (infant and Yes Yes
junior) fruit juices and drinks
Processors of juice that ship to a Yes Yes
different location (e.g., the juice
processing plant owned by a supermarket
chain that then ships the juice to the
chain's stores or very small processors
that sell juice from their own roadside
stand and to other retailers)
Retailers of packaged juice processed by Yes No
other establishments (e.g.,
supermarkets, restaurants and roadside
stands that sell juice produced by
another processor) Note: the juice sold
by these retailers is covered by the
HACCP rule but the retailer is not
covered by the HACCP rule.
Processors of packaged juice that do not Yes No
ship juice to different locations but
retail the entire production on the
premises (e.g., supermarkets, and
roadside stands that produce juice at
the point of sale)
Processors of beverages that include Yes No
juice as an ingredient but which do not
produce the juice itself
Retailers of juice processed for No No
immediate consumption
Processors of non-beverage products that No No
include juice as an ingredient
Processors of hard cider or other No No
alcoholic beverages
Processors of oils No No
Processors of purees not sold as No No
beverages (e.g., tomato puree)
Processors of juices not sold as No No
beverages (e.g., vinegar or borscht)
Processors of imitation juice flavorings No No
Processors of coffees, teas, or cocoa No No
products
------------------------------------------------------------------------
\1\ Juice types are berry; citrus; core fruit; mixed fruit; pit fruit;
subtropical and tropical fruit; vine fruit; other fruit; beans, peas
and corn; fruits used as vegetables; leaf and stem vegetables; mixed
vegetables; root and tuber vegetables; and other vegetables.
\2\ Beverage bases of fruit origin are berry, citrus, core fruit, mixed
fruit, pit fruit, subtropical and tropical fruit, vine fruit, and
other fruit.
\3\ A ``yes'' in this column applies only to processors producing in
excess of 40,000 gallons of packaged juice per year. Very small
businesses processing packaged juice, producing 40,000 gallons of
juice or less annually are classified as retailers for the purpose of
the HACCP rule and are therefore exempt from it.

4. Number of Establishments Covered
FDA's own Official Establishment Inventory (OEI, FDA's list of food
establishments under its jurisdication) lists approximately 900 juice
manufacturers. However, recent information from the U.S. Apple
Association (USAA) indicates that there are about 1,800 apple juice
plants, most of which are very small processors. A typical description
of these very small processors is an apple grower who operates a small
apple press and bottling operation on the same property. In general
these processors market their products in more than one way. The
channels of distribution include: Roadside stands owned by the
processors and stands owned by others, farmers' markets, grocery
stores, and restaurants. FDA has proposed to exempt retail
establishments from the HACCP rule. For the purposes of this rule, the
agency has tentatively decided that retailers will include very small
businesses that make juice on their premises and whose total sales of
juice and juice products do not exceed 40,000 gallons per year and who
sell directly to consumers or directly to consumers and other
retailers. Based on data supplied by the USAA, this exemption would
exempt from the HACCP rule 80 percent of apple juice processors. (Ref.
13). Such an exemption would leave approximately 360 apple juice
processors covered by both of these regulations, and all 1,800 would be
covered by the labeling rule.
The OEI lists about 200 plants in the United States that produce
core fruit (apple, crab apple, pear, quince, etc.) juice. If all of the
200 core fruit plants in the OEI are included in the USAA list and are
not exempt, then there would still be an excess of 160 apple juice
processing plants in the USAA list not exempt from the HACCP rule and
an excess of 1,600 (1,800-2000) plants in the USAA list not exempt from
the labeling rule. (Information from FDA's field inspections indicates
that very few of these 160 plants will be exempted from the HACCP rule
under the exemption for retailers of juice for immediate consumption.
Almost none of the very small apple juice processing plants recently
inspected by FDA retailed all of the juice that they produced at the
same location that it was processed. See Table 21 for a description of
the types of products and processors not covered.)
The agency is aware that there are also many very small orange
juice processors who grow oranges and who also operate a juicing and
bottling operation on the same property. However, the agency has no
direct information on the number of such orange juice processors. The
OEI lists about 300 plants in the United States that produce citrus
fruit juice. In this analysis, the agency has assumed that there is an
equivalent number (300) of very small processors who are not listed in
the OEI. It is likely that the proportion of very small orange juice

[[Page 24273]]

processors to OEI citrus juice makers is lower than the proportion of
very small apple juice processors to OEI apple juice makers because the
growing region for oranges in the United States is far smaller than the
region for growing apples.
FDA assumes for the purpose of this analysis, that 80 percent of
these very small orange juice processors will be exempt from the HACCP
rule based on their classification as retail establishments. This would
leave 60 very small orange juice processors covered by both of these
regulations, and all 300 covered by the labeling rule. FDA has assumed
that there are no vegetable juice processors which are not in the OEI
or which are not also very small processors of apple or orange juice as
estimated above. FDA requests comments on these assumptions.
FDA has assumed that 5 percent (about 50 plants (900 x .05)) of all
juice plants in the OEI would have implemented HACCP substantially in
the form required by this regulation by the time that this proposed
HACCP rule is finalized regardless of this regulatory action.
Therefore, approximately a total of 1,070 plants (850 plants in the OEI
plus 60 very small orange and 160 apple juice retailers) will be
affected by the HACCP rule.
The labeling rule will cover retailers (roadside stands and grocery
stores) of packaged minimally processed juice.
The agency does not have direct information on the number of
supermarkets and grocery stores that produce and package at the point
of sale and sell minimally processed juice. The agency believes that
only a portion of chain supermarkets and grocery stores do so. Duns
Market Identifier (DMI) lists approximately 9,400 chain supermarkets
(SIC 54110101) and approximately 3,800 chain grocery stores (SIC
54119904) making a total of approximately 13,000 chain supermarkets and
grocery stores. If 10 percent of these stores produce at the point of
sale and sell packaged minimally processed juice, then approximately
1,300 chain grocery stores and supermarkets will be affected by the
labeling rule. (In addition to these processors, there are other
retailers that do not process juice but which offer for sale the juice
produced by other processors, which should be labeled by the
manufacturer.)
Due to publicity about the hazards associated with minimally
processed juice, the agency believes that relatively few retailers are
offering such products for sale. DMI lists approximately 3,100
independent supermarkets (SIC 54110103) and approximately 31,000
independent grocery stores (SIC 54119905) making a total of
approximately 34,100 chain supermarkets and grocery stores. If 5
percent of these stores sell minimally processed packaged juice, then
approximately 1,700 independent grocery stores and supermarkets will be
affected by the labeling rule. The labeling rule will also affect
roadside markets and stands that retail packaged minimally processed
juice. For the purpose of this analysis, the agency assumes that there
are 1,000 such roadside markets and stands. However, the assumptions
that go into these calculations may be incorrect, and the agency
specifically requests comments on them.
Table 22 shows the estimated number of establishments affected by
each rule.

Table 22.--Number of Plants Affected by the HACCP and Labeling Rules
------------------------------------------------------------------------
No. of No. of
Establishments Establishments
Plant Type Affected by HACCP Affected by
Rule Labeling Rule
------------------------------------------------------------------------
Juice manufacturers in the OEI 850 201
Very small apple juice makers 160 1,600
Very small orange juice makers 60 300
Roadside retailers 1,000
Grocery stores and supermarkets
processing and packaging at the
point of sale 1,300
Total 1,070 4,220
------------------------------------------------------------------------
\1\ The number of juice manufacturers listed in the OEI affected by the
labeling rule is small (20) because most of these manufacturers are
already achieving a 5-log reduction. See Table 24.

5. Hourly Price of Labor
Throughout this analysis the hourly price of labor is taken to be
approximately $13. This is estimated by taking the 1996 average hourly
rural wage of $9.20 (Ref. 7) and increasing it by 40 percent (the
average amount for benefit costs paid by employers) (Ref. 8), or $3.70
to account for such costs in addition to wages, such as Social
Security, workers' compensation, unemployment insurance, paid leave,
retirement and savings, health insurance, and supplemental pay.
6. Length of Production Period
The agency is aware that many juice processors operate on a
seasonal basis. Information supplied by USAA indicates that 94 percent
of the apple cider producers process only seasonally. The season for
apple cider production runs primarily from September through December.
The other 6 percent operate year round. Many other processors covered
by the proposed HACCP rule (e.g., makers of beverage bases) may process
year round. The agency has assumed that 50 percent of the 850 plants in
the OEI plus all of the 220 very small juice makers affected by the
HACCP rule produce seasonally. Table 23 shows the length of the
production period for plants producing seasonally and year round.

Table 23.--Plants' Production Period
----------------------------------------------------------------------------------------------------------------
Weeks of Operation Hours of Operation
Production per Year per Day No. of Plants
----------------------------------------------------------------------------------------------------------------
Seasonal 16 12 645
Year Round 52 24 425
Total 1,070
----------------------------------------------------------------------------------------------------------------

[[Page 24274]]

B. Cost Estimates by Requirement

1. Costs have been estimated for the following sections of the
labeling regulation:
(1) Signs or Placards (Sec. 101.17(f)(3)(i) (part 101 (21 CFR part
101))
(2) Container Labels (Sec. 101.17(f)(3)(ii))
2. Costs have been estimated for the following sections of the
HACCP regulation:
(1) CGMP's (Sec. 120.5 (part 120 (21 CFR part 120))
(2) Prerequisite Program SOP's (Sec. 120.6)
(3) Hazard Analysis and HACCP Plan (Secs. 120.7 and 120.8)
(4) Corrective Actions (Sec. 120.10)
(5) Validation and Verification (Sec. 120.11)
(6) Records (Sec. 120.12)
(7) Training (Sec. 120.13)
(8) Imports and Foreign Processors (Sec. 120.14)
1. Labeling Costs
This cost depends strongly upon producers' responses to the
labeling requirements. Some producers may elect to comply early with
the HACCP rule and avoid the warning labels or labeling. Others may
choose to label until they are required to implement HACCP. Finally,
some firms may choose not to produce juice products because they
believe that either the cost of HACCP implementation or the negative
effect on revenue generated by consumer response to labels may depress
profits below a normal return for a substantial time period. Such
producers will be better served by reinvesting their capital into more
profitable ventures.
a. Signs or placards (Sec. 101.17(f)(3)(i)). The costs of signs and
placards may be estimated by multiplying the number of establishments
that must post placards by the cost per placard. As shown in Table 22
the agency estimates that the labeling rule covers approximately 4,220
plants. However, for the purpose of this analysis, the agency has
assumed that all those processors that will at some point be required
to implement HACCP will do so at the earliest possible date to avoid
the warning labeling, or delay operation until they implement a 5-log
pathogen reduction process.
The following analysis underlies this assumption. If displaying the
warning can be avoided by beginning pasteurization (or an equivalent 5-
log pathogen reduction process) sooner, some firms may marshal the
resources to do so. FDA does not have data, however, that will allow it
to predict how many firms will respond to this labeling regulation in
this fashion. However, one way to examine this choice is examine the
additional discounted costs of pasteurizing sooner. For example, if a
small firm's cost of initiating pasteurization is about $18,000, with
recurring costs of about $8,000, and the firm has an annual juice
revenue of $200,000, then a total sales decline caused by the warning
of 8 percent (a loss of approximately $16,000 discounted at a rate of 7
percent) or more spread over the course of 2 years (or approximately 4
percent for 2 years) would cause the firm to attempt to borrow the
funds needed to initiate pasteurization 2 years early or to delay
operation until it implements a 5-log pathogen reduction process. FDA's
predictions of consumer reactions to the labeling (for the purposes of
benefit estimations) are an expected loss of revenue of about 5
percent. Thus, there is a tentative conclusion that most firms that are
not exempt from the HACCP rule will choose to implement a 5 log
reduction in pathogens immediately rather than label and to delay
operation until such processes have been implemented.
However, there are many uncertainties contained in this simple
example. Because of the short time frame for labeling to begin, 60 days
from publication of the final rule, many firms may not be able to
purchase and install pasteurization equipment or find other means of
validating a 5 log reduction in the target organism. It is unclear how
manufacturers think that consumers will react to the warning signs,
they may believe that their customers will not reduce their purchases
of juice. Also, firms with larger sales or smaller pathogen reduction
costs will need a smaller percentage sales decline from labeling in
order to be induced to initiate 5 log pathogen controls early. Finally,
it is unclear how many firms will have immediate access to the capital
requirements imposed by this rule.
If, therefore, all processors which will eventually be covered by
the HACCP rule do not label, then they have no direct labeling cost.
The cost of the labeling rule to these processors is the extra expense
that results from implementing HACCP 2 years earlier than would be
required by the HACCP rule alone. This cost, as stated above, is
$16,000 (discounted for 2 years at 7 percent). Of the 1,070
establishments covered by the HACCP rule, all of the 20 firms in the
OEI which are also affected by the labeling rule (those estimated to be
producing minimally processed juice) plus all of the 220 very small
orange and apple juice processors covered by the HACCP rule are
affected in this way (240 plants in all). The agency assumes, based on
information from industry sources, that 30 percent of this set of
processors (72 plants) have already initiated or are in the process of
initiating pasteurization. Therefore, the total cost of the labeling
rule for this set of processors is $2,688,000 ($16,000 x 168 plants).
The establishments that will need to display warning labeling are
those 3,980 establishments covered by the labeling rule but not by the
HACCP rule. Based on information learned from FDA's nutrition labeling
rules, the average cost per placard (and periodic replacement) is
estimated to be $100. This estimate will encompass the possibility that
some firms may have to supply multiple signs to meet the requirement
that it will be available at the point of purchase. Therefore, the
total one-time cost for this set of processors is $398,000.
b. Container labels (Sec. 101.17(f)(3)(ii)). The cost of labeling
is estimated by multiplying the number of affected separable labels on
packaged products, normally referred to as stock keeping units (SKU's),
by the cost of changing the label to add the warning. Table 24 shows
FDA's estimate of the cost per SKU of placing a warning label on the
information panel for different lengths of the compliance period. These
costs decrease over time for several reasons. The primary reason is
that manufacturers change labels or, at least, reorder them at regular
intervals and a larger length of compliance period allows manufacturers
to incorporate regulatory changes into planned changes.

Table 24.--Label Change Costs per SKU for Different Lengths of the Compliance Period
--------------------------------------------------------------------------------------------------------------------------------------------------------
2 months 6 Months 1 Year 2 Years 3 Years
--------------------------------------------------------------------------------------------------------------------------------------------------------
Administrative costs $6,000 $1,800 $900 $450 $350
Redesign costs $1,500 $450 $450 $50 $50
Inventory loss $800 $250 $0 $0 $0

[[Page 24275]]

Totals $8,300 $2,500 $1,350 $500 $400
--------------------------------------------------------------------------------------------------------------------------------------------------------

Processors of minimally processed packaged juice which are not
covered by HACCP will need to add the warning to their package labels
at the end of the 2-year compliance period. FDA estimates that 2,980
processors will be subject to this provision (1,440 very small apple
juice retailers and 240 very small orange juice retailers exempted from
the HACCP rule plus 1,300 grocery stores producing packaged juice). The
total cost for this provision is $1,490,000 (2,980 x $500) at the end
of the 2-year compliance period. For simplicity of reporting and
calculation with the other labeling costs, this cost will be added as
$1,301,000 (the present value of $1,490,000 discounted 2 years at 7
percent).
c. Summary of likely labeling costs. The agency estimates that the
likely total cost of the labeling rule is a one-time cost of $4,387,000
($2,688,000 + $398,000 + $1,301,000).
2. HACCP Costs
a. CGMP's (Sec. 120.5). This section of the proposal reaffirms the
applicability of the CGMP's in part 110 in determining whether facility
design, materials, personnel practices, and cleaning and sanitation
procedures are safe.
No costs are attributed to this section for this rulemaking. The
overwhelming majority of juice plants are in compliance with the
CGMP's. In 1996 only 6 percent of the plants inspected were cited for
official action. Therefore it is assumed that these rules will not have
any effect on the enforcement of the CGMP's for juice products.
b. Prerequisite program SOP's (Sec. 120.6). FDA is proposing to
require that processors control and document specific SOP's that
provide a foundation for the HACCP system and to have and implement
SOP's for prerequisite programs. In general, there are three activities
that are part of prerequisite program SOP's: (1) Developing SOP's, (2)
implementing sanitation controls with corrections of deviations from
SOP's, and (3) monitoring and documenting for SOP's.
i. Developing SOP's. Each processor must have a sanitation SOP. FDA
estimates that SOP's for juice plants could be developed with 20 hours
of labor. At the rural hourly cost of labor ($13), the cost per plant
of developing SOP's is approximately $260. If one half of the 900
domestic plants in the OEI and all of the 220 very small juice
processors do not currently have SOP's, then they will have to develop
them to comply with this regulation, if it is adopted. Under those
assumptions, the total cost for the industry to develop SOP's would be
approximately $174,200 ($260 x 670 plants).
ii. Implementing sanitation controls with corrections of deviations
from SOP's. Each processor must implement a sanitation SOP and correct
deviations from the prerequisite program SOP's in a timely fashion.
In 1996, 39 percent of the juice plants inspected were cited as VAI
(voluntary action indicated). This citation usually indicates that an
investigator noted deficiencies that were not significant enough to
warrant an administrative or regulatory action but which should be
corrected on a voluntary basis. Information from the inspection reports
indicates that approximately 30 percent of the juice plants inspected
had sanitation and food safety related deficiencies, 4 percent had
deficiencies which were related to low-acid canned food regulations,
and 4 percent had deficiencies for misbranding or mislabeling. Also in
1996, 6 percent of the juice plants inspected were cited as OAI
(official action indicated). This citation indicates that an
investigator noted deficiencies significant enough to recommend
regulatory or administrative sanctions. Information from the inspection
reports indicates that 3 percent of the juice plants had significant
deficiencies that could be related to food safety or low-acid canned
food regulations, 2 percent had significant deficiencies for
misbranding or mislabeling.
On a few of the VAI inspection reports, FDA investigators indicated
an estimate of the cost of correcting sanitation and food safety
related deficiencies indicated. Two-thirds of the reports estimated
costs of corrections at $0 to $99, and one-third of the reports
estimated costs of corrections at $1,000 to $4,999.\8\ Taking the
middle of these ranges gives an average estimated cost of corrections
of approximately $1,000 (($50 x 67 percent) + ($3,000 x 33 percent))
per plant for correcting sanitation and food safety related
deficiencies.
---------------------------------------------------------------------------

\8\ No reports estimated costs of $100 to $999.
---------------------------------------------------------------------------

The HACCP rule will mandate the implementation of daily monitoring
of sanitation controls. This should make the correction of sanitation
and food safety related deficiencies happen on the day that they occur
rather than months later. Regulatory inspections of juice plants are
made approximately once every 5 years. If food safety and sanitation
related deficiencies occur on average approximately once every 5 years
midway between inspections (to facilitate calculation), then the HACCP
rule should cause corrections to be taken an average of 2.5 years
earlier than would be the case without the rule. The cost of the rule,
then, is not the full cost of taking the corrections. Those corrections
would be taken even without the HACCP rule after the plant was
inspected and the deficiencies noted. The cost of the HACCP rule is the
present value of making the expenditures to correct the deficiencies at
an earlier date than would take place otherwise. The present value of
making an infinite series of $1,000 expenditures once every 5 years and
2.5 years earlier than they would otherwise occur is $500 when
discounted at 7 percent.
Based on information from inspection reports, FDA assumes that
about 30 percent of all 1,070 covered juice plants (about 320 plants)
are not likely to have sanitation controls that are sufficiently
implemented, but which do not warrant administrative or regulatory
action. If it costs each of these 320 plants $500 to implement
sanitation controls and to correct deviations from SOP's, then the
total cost borne by the industry for this requirement is $160,000,
which, because it is discounted, will be added as a one-time
expenditure in the total costs.
iii. Monitoring and documenting of SOP's. All procedures in the
prerequisite program SOP's are required to be conducted at the
frequencies specified and implementation of these procedures will have
to be monitored and documented.
FDA estimates that monitoring and documenting of SOP's will require
one-half hour of labor per operating week. The cost per plant of SOP
monitoring and documenting is given in Table 25.

[[Page 24276]]

Table 25.--Annual Per Plant Cost of SOP Monitoring and Documenting
----------------------------------------------------------------------------------------------------------------
Estimate Hrs. per Estimate Annual
Weeks of Operation Week for SOP SOP Monitoring and
Production per Year Monitoring and Wage ($/hour) Documenting Cost
Documenting per Plant
----------------------------------------------------------------------------------------------------------------
Seasonal 16 .5 $13 $100
Year round 52 .5 $13 $340
----------------------------------------------------------------------------------------------------------------

Table 26 shows the distribution of per plant and total industry
costs based on the estimate in Table 25 for SOP monitoring and
documenting needed to comply with this rule, if it is adopted. These
estimates assume that no plants are currently in compliance with these
particular requirements.

Table 26.--Total Annual Cost of SOP Monitoring and Documenting
----------------------------------------------------------------------------------------------------------------
Estimate Annual
SOP Monitoring and Estimate Annual
Production Documenting Cost No. of Plants SOP Monitoring and
per Plant Documenting
----------------------------------------------------------------------------------------------------------------
Seasonal $100 645 $64,500
Year round $340 450 $153,000
Totals 1,095 $218,000
----------------------------------------------------------------------------------------------------------------

c. Hazard Analysis and HACCP Plan (Secs. 120.7 and 120.8). Under
the proposal, processors are required to have a written hazard analysis
and to have and implement a written HACCP plan whenever a hazard
analysis reveals a food hazard that is reasonably likely to occur.
Requirements are set forth for the minimum contents of the plan and for
the signing and dating of the HACCP plan by specified personnel.
Failure of a processor to have and implement a HACCP system in
compliance with this rule, if adopted, will render the food products of
that processor adulterated.
i. Hazard analysis and HACCP plan development. Under the proposal,
each plant is responsible for developing a written hazard analysis of
hazards that are reasonably likely to occur in the product that a
processor can control. The hazards to be considered are any chemical,
physical, and biological hazards that may cause illness, injury, or
death in humans. Plant management must determine the likelihood of
occurrence of these hazards, either due to their introduction through
material inputs or processing or a possible failure to eliminate them
or to reduce them to acceptable levels in processing. Some Federal
Government sampling and illness outbreak data are available to provide
firms with a set of possible hazards that may affect a particular
product and process. In addition, section V of this document, the
accompanying appendix, and the preambles to these proposed rules
contain information on most of the hazards that have caused problems in
juice products in the past. Additional information may be forthcoming
in the HACCP final rule (after FDA evaluates the comments). Experience
from the HACCP pilot suggests that the hazard analysis for products
similar to juice took 16 to 24 hours. FDA's preliminary estimate is
that it will take approximately four individuals, including a plant
manager; 5 hours each to complete the hazard analysis; and another 15
hours each to formulate the HACCP plan. The HACCP plan requires that
the plant manager, quality control official and others establish
critical control points (CCP's) for every hazard identified in the
hazard analysis and critical limits at each CCP; establish a plan to
monitor those CCP's; determine how deviations from critical limits will
be handled; and establish procedures for verification and validation
that the plan is being followed and that it is properly controlling the
identified hazards. FDA assumes that part of this process will be to
determine the most cost-effective means to comply with this regulation
when developing the plan. Thus, the total number of person hours per
plant to develop both documents is 80 hours. At $13 per hour the total
cost per plant is about $1,000 per plant.
FDA has assumed that about 5 percent (50 plants) of all juice
plants in the OEI will have implemented HACCP substantially in the form
required by this regulation by the time that this regulation is
finalized regardless of this regulatory action. This assumption is
based on conversations with pilot plant firms who have indicated to FDA
that many large firms have begun both to do HACCP and require HACCP of
their suppliers. It is estimated that approximately 1,070 plants will
need to do hazard analyses and develop HACCP plans to comply with this
rule, if it is adopted. Therefore, the total cost of 1,070 plants at
$1,000 each to develop a hazard analysis and a HACCP plan is
approximately $1,070,000 million.
ii. Pesticide HACCP controls. Pesticides may be a component of
material inputs that must be controlled. If a processor has direct
knowledge of the amount of pesticide applied, either because the
produce is from the processor's own farm or because records showing the
application of pesticides accompanies the incoming produce, then the
processor may control pesticide hazards by means of a supplier
certificate. Under such an arrangement a supplier would only need to
provide the processor with a certification that any pesticides had been
properly applied to the produce so as not to exceed applicable
tolerances. As each arrives at the processing plant, a worker will need
to verify that the supplier for that shipment has supplied the
processor with a proper and up-to-date certification. FDA assumes that
verification of supplier certification requires 1 minute per shipment
which, at $13 per hour, represents a cost per shipment of approximately
$0.25.
FDA has estimated the number of shipments that will be verified in
this manner by working backward from the amount of juice consumed.
Annual juice consumption in the United States is 2.3 billion gallons
(gal). The agency assumes that 80 percent of this total (1.84 billion
gal) is produced by approximately 75 large firms (operating 225
plants). FDA believes that all large firms are currently doing a
sufficient

[[Page 24277]]

amount of sampling and monitoring (or receiving supplier certificates)
for pesticides. Therefore it is assumed that there are no costs for
large firms to comply with this requirement. That leaves 20 percent of
the total (460 million gal) produced by approximately 2,575 small and
very small firms. FDA assumes that all small and very small firms use
domestic produce only. If 15 pounds (lb) of produce are required to
make 1 gal of juice, then small firms use 6.9 billion lb of domestic
produce (460 million gal x 15 lb/gal). If 45,000 lb of produce (the
amount carried by a typical tractor trailer) constitutes 1 shipment of
produce, then small and very small firms use 153,000 shipments of
produce (6.9 billion lb 45,000 lb/shipment).
However, for the purposes of this proposed regulation FDA is
including as retailers very small businesses that make juice on their
premises, whose total sales of juice and juice products do not exceed
40,000 gallons per year and who sell directly to consumers or directly
to consumers and other retailers. This exemption decreases the
percentage of juice processed under pesticide controls by approximately
14 percent thereby reducing the number of shipments of produce to
132,000 (153,000 x 86 percent).
FDA assumes that 80 percent of small and very small firms covered
by the rule (676) will process shipments of produce that will be
accompanied by supplier certifications of pesticide application after
the HACCP rule is in place. Therefore, the number of shipments to be
handled under prerequisite program controls is 106,000 (132,000
shipments x 80 percent) per year. Thus, this analysis assumes that the
average small and very small plant receives approximately 160 (106,000
shipments 676 small plants) shipments per year. The total per
plant cost is about $40 (60 shipments x $0.25/shipment) for the 676
small and very small plants that can control this issue in this way.
Based on these calculations, the total marginal cost of this type of
control for pesticides is approximately $27,000 ($40 x 676 plants).
If such records cannot be obtained, different types of controls
need to be implemented. In this case, the processor must run pesticide
residue tests to ensure that there are no pesticides either over
tolerance or used on products for which there is no tolerance. To
determine the frequency of such testing, processors may avail
themselves of Government test results which indicate the likely
variance of illegal residues over a particular crop or region.
Current records indicate that, for domestic crops, only about .25
percent (one-quarter of 1 percent) are out of compliance. Furthermore,
as HACCP is adopted by more of the food industry, it is expected that
records, for some types of produce, will routinely accompany produce
intended for interstate commerce. However, many types of produce are
currently commingled at different stages in the distribution network.
This creates a problem for backtracking when there are either pesticide
or pathogen problems.
There are two potential costs associated with ensuring that
pesticide residues are legal: (1) Matching and shipping pesticide spray
records with crops and (2) costs of multiresidue testing. If records
are to accompany produce, fruits and vegetables may only be commingled
if all of the commingled produce has records showing it is under
tolerance. Otherwise, produce with paperwork must be kept separate from
produce without such paperwork. In the latter case, if it is to be used
to produce juice, multiresidue tests must be performed costing about
$150 per test. Just as was calculated for supplier certificates, FDA
calculates that there are 132,000 shipments which use 5,865 million
pounds of produce that must be covered by pesticide controls. As 80
percent has been considered to b

[Text truncated at 120,000 characters. The full text is on the page linked above.]

---

Source: Frix Law Library, https://www.frixlaw.com/law-library/documents/fr%3A98-11530. Public record. Not legal advice.