The Centre for Science in the Public Interest (CSPI) is pleased to submit comments in
relation to Health Canadas Consultation Document Standards of Evidence for the Evaluation
of Foods with Health Claims: A Proposed Framework (hereinafter, the Consultation Document).
Given the vastly superior public health potential of mandatory nutrition labelling on all
foods and its broad public support, Health Canada will be doing consumers a great disservice if it
establishes a permissive health claims regime without a strong, comprehensive mandatory
nutrition labelling requirement for all foods. We are disappointed that the support for mandatory
nutrition labelling from experts and the public at large has not yet produced a commitment from
Health Canada, while the industry-driven proposal to authorize health claims on a few
manufacturer-selected products has diverted its attention and resources.
Although there is a theoretical potential for a voluntary health claims program to educate
consumers about the benefits of certain foods, this potential has not been demonstrated in either
Japan or the U.S., where regulatory programs permitting such claims have been in place for
several years. Indeed, a comprehensive report issued by CSPI and other members of the
International Association of Consumer Food Organizations (IACFO) indicated that the Japanese
and American programs have been beset with problems, many of which Health Canada appears
willing to replicate here.(2)
If Health Canada decides to proceed with permitting health claims, and takes steps to
rectify the serious and numerous problems we have identified in this submission, it should then
require all eligible foods to make applicable generic claims on their labels. Health claims on a
few products may simply mimic the problems of voluntary nutrition labelling. By contrast,
mandatory health claims for all eligible foods would ensure that consumers have complete health
information about all foods that qualify to make claims. Foods that are eligible to make more
than one claim on labels (e.g., most fruits and vegetables) could either make all claims on each
label or rotate the use of eligible claims. This approach is analogous to warning labels that appear
on cigarette packages except that, in this case, all of the messages would be positive and would
promote the sale of the food bearing the label rather than discourage it.
If health claims are to be permitted, standards for scientific substantiation of these claims
and the actual benefits of their use, must be set at a high level one that provides the optimum
additional health benefit to consumers. The Consultation Document fails to ensure this in several
fundamental ways that warrant corrective measures before the consultation proceeds further.
Mandatory comprehensive nutrition information on all foods is almost certain to produce
more healthful dietary practices than the voluntary use of marketing-oriented messages that some
manufacturers will make for a few selected foods based on profit-maximizing, not public health,
considerations. Mandatory nutrition labelling should be implemented first.
While health claims have a theoretical potential to educate consumers about diet-disease
relationships, Health Canadas plan for evaluating claims must be significantly changed to
increase the likelihood of that outcome. First, the approach for pre-market evaluation of the
safety, claim validity, and health impact of health and health-related claims that is described in the
Consultation Document should be revised in the manner described below. Second, if health
claims are permitted in Canada and an adequate regulatory program is established to control them,
they should be mandatory for all eligible foods.
1. Efficient use of public health resources
Health Canada must ensure that the net public health benefit of authorizing any generic or
other types of health claims on foods is commensurate with the costs of administering that
program, and that other more cost-effective options (such as mandatory nutrition labelling) are
pursued first if resources are insufficient to do both.
Properly reviewing numerous applications for health claims, particularly food-specific and
product-specific claims, could be a very resource-intensive task even if claims were permitted on
the basis of pre-determined, generic product monographs. In either case and especially the
latter significant public resources would be required to monitor compliance with these
standards in the retail food market. It is questionable whether these resources are effectively
allocated to measures that do not show clear promise of delivering public health benefits. Health
Canada and the Canadian Food Inspection Agency (CFIA) should not be in the position of
devoting considerable scientific, administrative and enforcement resources to administering a
program which merely re-distributes market share among food manufacturers without improving
the healthfulness of Canadian dietary intakes. Unlike mandatory nutrition labelling, a poorly
resourced health claims regulatory program could lead to considerable misleading advertising and
labelling that could both worsen the diets of consumers and injure the reputations of Health
Canada and CFIA.
The resources problem cannot be avoided by devolving duties to a third party private
system of review (a possibility suggested on page 39 of the Consultation Document) without
jeopardizing the public interest. A third party system in which reviewers are remunerated by
applicants would very likely introduce a host of problems related to conflict of interest and bias in
the approval/administration process. Also, in the end, consumers themselves will pay for the
costs of the review through the tax system or food prices regardless of who actually conducts the
review. As such, they are entitled to a program designed and operated by independent public
Without adequate resources to administer a health claims program, and without evidence
of net public health benefit of health claims on foods, Health Canada should adopt mandatory
comprehensive, easy-to-read nutrition labelling for all foods sold in Canada. This, we submit,
would be a more cost-effective means of enabling consumers to make healthful food choices on
the basis of product information on all foods.
2. Optimizing public health benefits and food safety should be paramount
Health Canada must pronounce a clear, unambiguous commitment to using public health
factors as overriding considerations in the claims evaluation process. As currently written, the
Consultation Document sets a very low standard for protection of public health. We recommend
that the standards require applicants to present compelling evidence that (a) proposed claims will
have significant positive expected health outcomes (and not merely avoid negative impacts), and
(b) proposed claims will advance national public health and nutrition policies (and not merely
refrain from conflicting with those policies).
The Consultation Document provides little uncontradicted assurance that high standards
of consumer protection and health improvement will be observed. For instance, section 1.3 of the
proposal requires that applicants for health claims provide only "reasonable assurance that foods
with health claims do not have negative nutritional and health impacts at recommended levels of
intake [emphasis added]." This is not an appropriate basis upon which to seek a mandate from
stakeholders, especially consumers. Consumers are entitled to a higher safety standard than
3. Proof of proposed claims impact on dietary practices
Health Canada must ensure that, before any health claim is authorized, its proponent
provides compelling empirical evidence showing that the asserted claim will actually motivate
consumers to eat more healthful foods. We are not aware of any studies conducted by Health
Canada to evaluate the positive or negative effects of health claims on consumption patterns. By
contrast, Denmark, Norway, Australia, and New Zealand are testing these effects as a precursor to
making a final decision on whether to permit health claims in their countries.
Some food industry proponents of health claims argue that their public health benefits are
demonstrated by citing (a) increased product sales for foods bearing such claims in the U.S., and
(b) improvements to the diets of Americans since regulations authorizing health claims were
finalized in the U.S. in 1993. Both are faulty inferences.
Increased product sales of foods bearing health claims may be at the expense of declining
sales and consumption of equally or more nutritious foods. Similarly, the same regulations
promulgated pursuant to the Nutrition Labelling and Education Act required nutrition information
on virtually all (about 97.5%) packaged foods. By contrast, the proportion of stock keeping units
bearing health claims has never exceeded 1% in the U.S. Thus, mandatory nutrition labelling, not
optional health claim labelling, is likely responsible for the vast majority of improvements in the
dietary intakes of American consumers.
The only public health rationale for permitting health claims on foods is that such claims
will motivate consumers to select more healthful foods over less healthful foods when making
purchasing and consumption choices. At a minimum, applicants should be required to provide
empirical evidence showing that the claim proposed will likely produce more healthful eating
practices. Yet the Consultation Document proposes (at section 5.4, part III) that providing this
evidence in applications be optional.
4. Pre-market government approval
Health Canada must require government pre-market approval of all health claims. This
requirement is necessary to prohibit poorly substantiated and misleading claims before they
appear on food labels. We urge Health Canada to confirm that pre-market approval by the
department will be required for all health claims. Reviews should not be conducted by third party
systems (as suggested in section 8.5 of the Consultation Document) because they will raise the
spectre of undue industry influence in the application process.
5. Significant scientific agreement
Health Canada must set the standard for scientific justification for claims high at the
level of "significant scientific agreement among qualified experts based on the totality of
available evidence" and explicitly prohibit claims based on preliminary or otherwise
inconclusive evidence. The Consultation Document is imprecise about the quantity of evidence
required to establish health claims.
The last paragraph of table 2 in the Consultation Document states that moderate to strong
evidence indicating that the agent has the opposite effect to what is asserted in the claim
constitutes "inadequate substantiating evidence." We recommend that this point be revised to
state: "At a minimum, moderate or strong evidence that the agent has an effect that is neutral or
opposite to what is asserted in the claim always constitutes inadequate substantiating evidence."
Similarly, the essential causality criteria (see pages 12-13) states only that "There should
be no equally strong opposing evidence." We expect affirmative evidence of a claims validity to
be compelling and specific, and to satisfy the "significant scientific agreement" standard.
6. Evidence of claim validity
a. Causality: When seeking to make inferences about human populations, generally, human
studies are better than animal studies and experiments are better than observational studies.
However, ethical, financial and practical factors limit the feasibility of long-term, high dose,
experimental designs involving humans. As such, the Consultation Document should have
devoted more attention to setting rigorous standards for evaluating the merits of observational
studies. For instance, observational studies that properly assess causal relationships employ
multivariate statistical techniques that: are sensitive to non-linear dose-response effects (e.g., by
using categorical independent variables), include appropriate control variables (but are sensitive
to multi collinearity issues), and include assessment of interaction effects (including, but not
limited to, interactions with other components of the food matrix).
b. Statistical techniques: Similarly, in section 5.4 (part 1, step 1), no acknowledgement is given
to the preferability of using longitudinal studies (with adequate privacy protections) or employing
statistical techniques that are sensitive to analysing effects characterized by time lags.
c. Magnitude of effect: The magnitude of effect should be significant from a clinical
perspective and from a public health perspective. Of course, independent variables should all be
statistically standardized to avoid the result that units of measurement (which differ among
independent variables) do not disproportionately affect the dependent variables. See: section 5.4,
part I, step 2, B. Essential Causal Criteria.
d. Probability: The proposed standards of evidence require that there be a "statistical
relationship." This is needlessly imprecise. The word "statistical" should be replaced with
"statistically significant." Although small effects can be statistically significant when large
samples are used, the previous criterion addresses magnitude of effect. See: section 5.4, part I,
step 2, Essential Causal Criteria, C.
In addition, it is essential, in experimental and observational studies, that a sound sampling
strategy be used in order to properly analyse interaction effects and to confidently make
inferences about the population parameters to which the proposed claim relates.
e. Temporal Relationship: The proposed standards of evidence should require that the time lag
between the introduction of the agent and the causal effect be consistent with the most plausible
biological mechanism. See: section 5.4, part I, step 2, Essential Causal Criteria, D.
f. Reversal or Cessation of Effects: The meaning is unclear in section 5.4, part I, step 2, G.
Supporting Causality Criteria.
g. Optional information: Evidence about the potential impact on population health, the
hypothesized biological mechanism, and the time lag between consumption and the production of
the demonstrated beneficial effects should not be optional parts of an application. This
information should be required on every application for a health claim authorization. See, section
5.4, part III.
h. Methodological Advice: In general, we recommend that Health Canada officials seek
technical assistance from Statistics Canada, or other independent sources, in establishing and
applying criteria for assessing the statistical and methodological merits of research, especially
research involving observational data.
7. Restricting use of health claims to otherwise healthful foods
Health Canada must prohibit the use of health claims on foods that are generally low in
a. Inadequacy of the Canada Food Guide as a regulatory tool: Qualifying nutritional criteria
for health claims must be precise and enforceable; it is inadequate to merely prohibit these claims
for foods that fall into the Canada Food Guides "Other Foods" category because that documents
categories lack legal certainty, are not clearly defined, and may not be enforceable. Also, Health
Canadas proposed approach does not prevent the use of health claims promoting even more over-consumption of foods that are already consumed at levels in excess of the nutrition
recommendations set out in the Canada Food Guide. For instance, most Canadians already
consume more than 2-3 daily servings from the "meat and alternatives" category. Health claims
on eligible meats could encourage additional and unhealthful over-consumption of those foods,
while inadequate consumption of, for instance, fruits and vegetables remains a problem.
b. Disqualifying maximum levels of unhealthy nutrients: A health claim must provide
consumers with the assurance that the food does not have properties that are potentially harmful
to health. Therefore, a product should not be allowed to highlight a characteristic that may help
reduce the risk of a disease, but remain silent about another characteristic that may increase the
risk of the same or another disease or health condition. Foods that contain high levels of fat,
saturated fat, cholesterol, sodium, trans fatty acids, or added sugars are all linked directly to
certain forms of cancer or heart disease or, indirectly to heart disease and diabetes by
contributing to obesity (a major risk factor). Therefore, foods that exceed prescribed levels of
unhealthful nutrients should be prohibited from making health or health-related claims.
c. Minimum qualifying levels of healthful nutrients: Similarly, foods bearing health claims
should have at least 10% of the recommended daily intake, before fortification, of protein, fibre,
vitamin A, vitamin C, calcium, or iron so that less-than-nutritious foods, like coffee, are
prohibited from making health claims.
8. Health-related claims
a. Structure/function claims, biological role claims and logos/endorsements of third party
Health Canada must require that biological role claims, so-called "structure/function
claims," and endorsements or logos of third party health charities are subjected to the same
standards for claim substantiation, safety, and labelling as those set out for disease risk-reduction
If a significant market develops for the use of health claims on foods in Canada, we may
see expanded use of "biological role" claims (pursuant to sections B.01.311, D.01.006, and
D.02.004 of the Food and Drug Regulations), structure/function claims, and third party claims
(such as the Heart and Stroke Foundations "Health Check" logos) for foods which cannot meet
the more rigorous pre-approval requirements for disease risk-reduction claims. Because
consumers have considerable difficulty appreciating the nuanced distinctions between the
language used in health claims and the language used in biological role claims or other types of
health-related claims, they may mistakenly choose less healthful foods on the impression that they
deliver the same benefits as foods that have successfully met pre-approval standards of evidence
for risk-reduction claims. Even consumers able to differentiate the four types of claims would
probably not be sufficiently familiar with the various sets of applicable qualifying criteria to
intelligently distinguish the meaning of each type of claim.
These kinds of mistaken impressions and faulty inferences could undermine confidence in
the whole health claims program. Accordingly, we recommend that, contrary to section 1.1 of the
Consultation Document, all third party endorsements/logos, structure/function claims, biological
role claims, and disease risk-reduction claims be subjected to the same standards for safety, claim
validity, and labelling requirements. This measure will prevent less rigorously regulated claims
from being "passed-off" as health claims.
b. Medical foods
Health Canada must ensure that the channels of distribution for foods encompassed by the
proposed regulatory category "medical foods" are limited by prescription and/or by venue (i.e.,
only available from a pharmacist) consistent with the proposed label warning that such foods
should only be consumed under "medical supervision." See table 3 of the Consultation
9. Transparency of applications
Health Canada must ensure that all documents considered in assessing the validity of
health claims are placed on the public record, in their entirety, and are available for comment by
interested persons before a decision is made to permit a health claim. Food industry proponents
of health claim applications should be required to sign a notarized affidavit attesting to the fact
that they have provided copies of all relevant research evidence including published and
unpublished research or data in their possession or power (whether that evidence or data supports
or does not support their application), and that they provide identifying information related to
other evidence not in their power or possession (i.e, research conducted by other food companies).
Companies that refuse to provide internal research findings should be prohibited outright from
obtaining the approvals sought or using approved claims.
Health Canadas explicit commitment to protecting manufacturers rights to proprietary
information (see page 18, footnote b, para. 3) does not, at this time, bode well for a transparent
process. Similarly, in section 7 (page 36), Health Canada shows a fundamental failure to
acknowledge the interests of consumer-citizens, health professionals and other independent
observers when it describes the relationship between regulator and food company applicant as,
essentially, a private bilateral affair:
"Regular consultation between applicant and review agency prior to and after submission
will also help facilitate the evaluation process."
10. Potential for bias in industry research and industry expert opinion
Health Canada must recognize the potential for bias in commercially-motivated research.
In particular, when weighing the evidence for a proposed claim, Health Canada must consider (a)
the source of funding for the research, and (b) whether the research was published in an
independent, peer-review scientific journal.
The same considerations should apply to weighing advice given by industry experts
whether it be from a single scientist or a team of scientists. Health Canada should always inquire
about the funding sources for expert opinions and research materials supplied by food companies.
Of course, any available studies showing potential safety (as opposed to claim validity)
problems arising from the use of certain health claims or functional ingredients should be
carefully considered regardless of the these factors.
Plainly, the political stimulus for designing a regulatory scheme for health claims on food
has primarily come from the U.S. government (that currently permits such marketing claims) and
Canadian and multinational food companies eager to use them to boost product sales. We are not
aware of any Canadian health or consumer organizations (that are not substantially funded by the
food industry) advocating for health claims on foods. The first survey evidence demonstrating
any explicit interest among consumers in seeing health claims on food labels was not published
until four years after Health Canadas consultation commenced. Even so, its data were collected
and published (in summary form only) by a food industry funded research body, the National
Institute of Nutrition.
11. Trivial health claims
Health Canada must ensure that health claims concerning insignificant public health
problems are prohibited. Trivial health claims may divert consumers away from eating healthful
foods (including foods that do not carry health claims on their labels) that may help reduce the
incidence of more pressing public health problems. In section 8.1 and the last bullet of section 2,
the Consultation Document proposes that "at least initially, only claims (regardless of type) with
the potential for major public health benefit...be given priority for evaluation." We submit that
there is never any need to approve applications for health claims promising insignificant public
Grocery stores filled with food labels touting minor health benefits (such as aiding
digestion) will create confusion, draw consumers away from consuming foods with benefits
related to major health problems, and trivialize more important health claims. We urge Health
Canada to limit health claims to diet-disease relationships that involve serious public health
12. Food- and product-specific claims
Health Canada must prohibit the use of food-specific and product-specific health claims
(so-called, "magic bullet" claims) because they undermine nutritional recommendations to eat a
balanced diet. Moreover, allowing health claims for individual foods or products (as opposed to
generic claims) would allow other less healthful products that contain those foods or products as
ingredients to ride on the coattails of the foods that qualify to make health claims.
The inverse relationship between research quality and claim specificity/marketing
potential is another reason why health claims for particular foods and products should not be
considered for approval. Also, highly specific health claims (product-specific claims being the
most specific) may have the greatest commercial potential for influencing consumers purchasing
decisions, but they are also least likely to be supportable by good quality, independent scientific
research. For instance, scientific studies examining the health effects of a single food or product
are likely to be few in number and exclusively funded by, or conducted by, food manufacturers.
When a research issue is so narrow as to only be of interest to commercially-motivated scientists,
problems associated with publication bias (e.g., under-representation of published studies
showing no beneficial effects and adverse effects) become much more severe.
13. Health claims in advertising
In Health Canadas Consultation Document on Generic Health Claims for Food, it
proposes that when a health claim is made in advertising, the entire claim must be presented, but
that information required to be given in relation to the claim may be provided in the
advertisement or on the label. The Consultation Document on the Standards of Evidence for
Evaluating Foods with Health Claims notes that, in some cases, caution statements,
recommended usage levels, or statements indicating the need for medical supervision may be
required to accompany authorized health claims. This type of information would be required
when health risks are associated with the use of the product.
This important information should always be stated in advertisements and on labels where
the health claim is used. Failing to do so is both misleading and potentially dangerous.
Manufacturers should never be permitted to make promotional use of health claims in
advertisements without also being required to communicate qualifying health warnings in the
advertisement and on the product labels. Of course, any required usage or warning information
should always appear on product labels that bear the health claim, regardless of whether that
claim also appears in advertisements.
14. Baby food
Health Canada has not demonstrated a clear position on the use of health claims on infant
foods. For instance, in section 4.1 (para. 3, Terms of Reference) of the Consultation Document
on Standards of Evidence for Evaluating Foods with Health Claims, Health Canada states:
"This means that for food products that are subject to control under [regulations
concerning infant formulas, novel foods and food additives], premarket safety assessment
will precede any evaluation of their potential health benefits, if health claims are made for
these products." [emphasis added]
By contrast, in section III (para. 2, General Requirements for Claims) of the Consultation
Document on Generic Health Claims for Foods, Health Canada states:
"Health claims are not appropriate on the labels of foods intended for infants and children under the age of two."
Health Canada should prohibit the use of health claims on food labels, advertisements and
promotional materials directed at children under two years of age to prevent abusive marketing
practices targeted at vulnerable populations. Caring parents want to feed their children the most
healthful foods possible. As such, parents and, by extension, their young children, are particularly
vulnerable to marketing information that promises health advantages not promised by other foods.
The infant food industry does not have an enviable record in the area of product
promotion. In many cases, these marketing practices may discourage the consumption of the
almost universally more healthful infant food, breast milk. INFACT Canada has reported to
Health Canada and the Canadian Food Inspection Agency that numerous aggressive and
potentially illegal health claims are currently being used on infant product labels in Canada.
INFACT has noted, with dismay, the failure of regulatory officials to take corrective measures to
stop these practices. We are concerned that the failure to prohibit health and health-related claims
on infant food labels, advertisements and promotional materials and enforce such rules may
exacerbate this serious problem.
15. Genetically modified foods
Health Canada must ensure that all foods containing genetically modified ingredients are
labelled as such so that it is possible to effectively conduct post-market surveillance. The
Consultation Document recognized the importance of conducting post-market surveillance to
"ensure the long-term safety" of novel foods. See section 4.4, para. 3 of the Consultation
Health Canada has yet to demonstrate that it has sufficient scientific and administrative
capacity for product safety assessments, especially concerning novel foods and novel ingredients.
Health Canada should ensure that sufficient resources are available for properly conducting safety
assessments of all novel foods, regardless of whether those foods carry health claims.
16. Health claims on imported foods
Health Canada must take steps to ensure that functional foods imported into Canada,
regardless of the type of regulatory system in place in the country of origin, are held to high
standards of safety and labelling appropriate to Canada. Only foreign health claims that advance
Canadian national nutrition policies should be considered for approval in Canada. For instance, a
Japanese food bearing a label stating "improves gastrointestinal condition" should not be eligible
for approval as a health claim in Canada. In addition, the significant regulatory problems that
exist in the U.S. and Japan should not be "imported" into Canada. Canadian expectations for high
food safety and labelling standards should not be weakened by trade liberalization measures.
The U.S. General Accounting Office recently criticized the Food and Drug Administration
(FDA) for failing to adequately regulate health claims and the FDA is currently considering a
series of petitions (filed by CSPIs Washington office) to prohibit the sale of 75 food products
presenting serious safety and misleading labelling problems. A recent U.S. Court of Appeals
decision (allowing health claims based on inconclusive scientific evidence to be placed on dietary
supplement labels) suggests that legal prohibitions on misleading claims for food may not be able
to withstand legal challenges based on free speech provisions of U.S. constitutional law.
Currently, the U.S. FDA is also considering the possibility of permitting 21 claims for over-the-counter drugs to be also used for dietary supplements without adequate controls for safety,
efficacy or good manufacturing practices.
In addition to considerable enforcement problems and regulatory loopholes, Japans
product-specific "Foods for Special Health Use" (FOSHU) claims approval system allows labels
to bear claims concerning mainly insignificant public health problems. Japanese law does not
prohibit FOSHU claims for candy or other foods of low nutritional value. Claims are often
approved on the basis of low standards of scientific evidence (an endemic problem with food-specific and product-specific claims) through a process that is heavily dominated by food industry
Any Canadian program established for validating U.S. or Japanese health claims for use in
Canada should engage those countries in discussions about raising their standards to meet higher
Canadian standards or, for instance, should invite Japanese and U.S. officials to help flag "export
quality" claims that would meet Canadian public health standards.