Setting the Stage/Framework/Scope of Advertising and Labeling
Stephen L. Lacey, Division Counsel, Abbott Laboratories
I came in near the end, but Lacey emphasized that a marketer needs to know the difference between labeling (FDA authority) and other material (FTC)—if you detail to pediatricians about foods with nutritional benefits, for example, FDA largely prohibits nutrient claims for kids under 2. However, he pointed out that the difference in formal standards—misleading v. materially misleading—was not the key; a marketer debating whether its communications are only misleading or materially misleading is already in a bad place.
State AGs are also active, often working jointly with the federal government; widespread in false claims matters for drug and device companies. Oregon AG settled with Kellogg, requiring destruction of packages, among other things.
Private lawsuits: increasing use of violation of federal regulation as basis for a complaint under state consumer protection laws, despite absence of private cause of action under FDCA or FTCA. Example: litigation over color additives used in salmon feed. FDA said any product made had to be labeled “color added.” Very few companies complied. The FDA didn’t bother to do anything about it. Plaintiffs sued major chains in California.
Christopher Hagenbush, FDLI Board: GMA’s new program for front of package information has received mixed reviews: mix of general nutrition information and promotional material. His perspective: FDA can’t stop you from putting info on the front of your package, but if it suggests your food is more desirable than others then FDA may respond badly to that. Media/consumer groups are suspicious of front of package labeling. The front of package statement may suggest that there are gradations of healthy/healthier foods.
Q: can FDA act against advertisers who encourage searchers to search for a particular term?
Lacey: FDA has taken issue with sponsored links for, say, rheumatoid arthritis. Not aware of the same thing for a food product (e.g., whole grain oats and cholesterol). Is the search term itself labeling? Most likely not, but the advertiser’s website might be depending on what statements it makes.
The Changing Landscape for Marketing Health and Nutrition Benefits, Part 1
Steven B. Steinborn, Partner, Hogan Lovells LLP
Understand the messages you’re conveying to your audience; substantiate your claims; but sometimes, labeling rules become a proxy for consumer protection. USDA has a lot of definitions. Some are archaic, most of them make sense. USDA doesn’t regulate ads, but you have to worry about FTC/private plaintiffs who will rely on USDA standards when challenging ads.
Short definition of advertising law: Liability depends on the claim you’re making and the support you have for that claim. Almost every FTC case involves a dispute over what claim was conveyed; not just what the advertiser intended but how the reasonable consumer reacted—with the FTC playing the role of the reasonable consumer. So, you need to know how reasonable consumers behave, and you also need to know how to evaluate science—as scientists do and as the FTC does. Most scientists are in love with their science; they are believers. Keep in mind that a regulator may look at the matter differently.
FTC consent orders: policy or unique to the facts? Historically companies have looked to consent orders to figure out what the FTC expects. So when do you need to do double-blinded trials given recent FTC consent orders—are they fact-specific or broader?
Keep in mind marketing, technical, regulatory, and legal folks all have to work together to cover the bases.
Richard Cleland, Assistant Director, Bureau of Consumer Protection, Division of Advertising Practices, Federal Trade Commission
Managing supervisor on Iovate, Nestle, and Dannon cases. Will offer discussion of thought processes/results in Nestle and Dannon, though his comments are his own opinions and not those of the FTC or of any individual commissioner.
Is the FTC targeting probiotics? No, not at all. There are lots of emerging claims about probiotics, but the principal focus that brought attention to Nestle and Dannon were the cold and flu claims. Compare to Airborne a few years back.
Nestle: Kids Essentials, probiotic straw and drink for children. Nothing wrong with the product; the issue was the claims. Claims: prevents upper respiratory tract infections, protects against cold and flue viruses, reducses absence from daycare or school due to illness, reduces general incidence of illness in children, reduces the duration of acute diarrhea in kids up to 13, etc.
Issues that came up for FTC on substantiation: while there were studies, the decent ones involved immune system related endpoints, not cold prevention benefits. The diarrhea claim for kids up to 13 required extrapolation that went beyond the population; the studies covered children under 2 (who are very different!). Confounding variables: studies tended to study a bunch of things at once; when you looked at claims about days absent from school, you couldn’t separate out other issues like fever from diarrhea. Core claim about reducing duration of diarrhea in children under 2 was really the best you could get. Also, studies on probiotics are strain-specific; submitting studies about strains not in your product is not going to be persuasive.
Dannon: DanActive and Activia, dairy drink and yogurt with probiotics. Challenged claims: DanActive: reduces the risk of colds and flu. Activia: relieves temporary irregularity and helps with slow intestinal transit time.
Issues that came up: again, endpoints. If your evidence relates to biomarkers like T cell counts, we don’t view those as substantiating a claim of reduced likelihood of cold and flu. There is also a problem with self-reported events. Some studies rely on parents’ reporting; not persuasive v. studies where a doctor confirms that the event occurred. Placebo controls are also very important, along with randomization and blinding. That can raise issues with probiotics. Dose-response is also critical. If you have a study where the doses do not match with what’s in the product, that will be an issue.
If you put a kid in a food ad about colds and flu, the FTC will probably interpret the ad as making a claim about reduced likelihood of colds and flu.
FTC orders: standard FTC order prohibits particular kinds of health claims unless the defendant possesses and relies on competent substantiation. What constitutes competent and reliable scientific evidence will vary based on product and claim. That doesn’t provide much guidance to defendants; it’s hard to enforce and led to protracted enforcement cases. Courts have misinterpreted the language.
FTC is making transparent in orders the analysis it already does to determine whether a claim is substantiated. One size does not fit all. Dietary supplement orders may look different from food orders. For some claims, FTC is using FDA determinations as a proxy for substantiation.
Three tiers (not every order has all three): Specified disease claims; other specified health-related claims; residual fencing-in provision for any other health-related claims. First tier: where defendant made claims that product could treat, cure, prevent or reduce the risk of disease, order will require such claims to be approved by FDA, whether through OTC drug monograph or approved NDA, or by regs under NLEA for foods. If there is a legal impediment to complying with FDA provisions, those are issues to discuss in negotiations with FTC. Bottom line: if the claim wouldn’t be permitted in labeling, the order won’t permit it in advertising.
His position: this is not so new. FTC has long said that the absence of an FDA determination that a health claim is scientifically valid will be a significant factor in the FTC’s assessment of substantiation’s adequacy. FTC regards “significant scientific agreement” in NLEA and FDA regulation to be the principal guide to what experts in the field of disease-diet relationships would consider sufficient substantiation. Case law says FTC has wide discretion to determining type of order necessary to remedy challenged practices; may go beyond illegal practices challenged so long as orders bear a reasonable relationship to the unlawful practices found.
Second tier: other specified health-related claims that have been investigated: based on the record of the investigation, and based on what experts would require, “competent and reliable scientific evidence” will be tailored, depending on the product/claim. One or more clinical trials may be required for particular claims/products.
Treat those orders the same way you’d treat any other guidance from the FTC; if a company not under FTC order comes in with a large, multicenter study and makes the case that it has a reasonable basis, that’s what the law requires. The inquiry will always be on the adequacy of substantiation, not prejudged. Various FTC orders have required studies conducted by different researchers, independently of each others—to correct against conscious or unconscious bias, or even fraud, all of which we’ve seen. Studies must conform to acceptable designs and protocols; matched to the right endpoints. Studies’ results must be considered in light of the entire body of relevant and reliable scientific evidence—if you have 2 studies your way and 8 the other, that may not be enough.
Two definitions of adequate and well-controlled human clinical studies: first, where it’s possible to conduct a randomized, double-blinded, and placebo-controlled study; second, where it may not be possible to double-blind, you have to do the other things right.
Third tier: broad categories of products and claims. Claims about health benefits, performance, or efficacy need sufficient substantiation—the quality and quantity of the evidence must meet accepted scientific norms, and the studies must be evaluated in the context of the entire body of relevant evidence.
James E. Hoadley, PhD, Senior Consultant, EAS Consulting Group, LLC
Types of scientific evidence: lab studies—in vitro, experimental animals, studies of absorption, distribution, metabolism, and excretion (ADME). Vital foundational information. For evaluating health claims, we need direct evidence of the relationship between the substance and the disease etc. in humans. In vitro and animal studies provide useful background only.
Human trials: descriptive, observational, interventional. Descriptive: anecdotal/case reports; population studies where you compare population statistics, e.g., “the Mediterrranean diet.” Cross-sectional studies: compare some measurable factor among groups with and without a disease, such as blood cholesterol. No scientific conclusions can be made based on descriptive reports.
Observational studies: retrospective and prospective. Retrospective: enroll subjects diagnosed with a disease and look at their histories. E.g., case-control studies where each diagnosed person is matched with a disease-free control subject (matched on age, gender, socioeconomic status, etc.). They’re fast, but they provide little information for health claims because it’s hard to get information on diet patterns. People’s recollections on what they ate are influenced by what they’ve been eating recently, and people who are sick change how they eat.
Prospective studies: enroll healthy subjects and follow them for extended periods of time. May have them keep diaries to track what they eat. Cohort studies: a population picked from a relatively homogenous group that will commit to following the study and have accessible medical records for confirmation. Male physicians, female nurses in Iowa, California 7th Day Adventists. A nested case-control lets you analyze disease cases in the cohort versus matched non-disease cases.
Observational human evidence shows associations only, not cause and effect. Good observational studies produce questions: why is green tea consumption associated with cancer risk?
Intervention: exposure to the test substance is controlled by the investigator. Clinical trials can provide evidence of cause and effect. Gold standard: placebo-controlled, double-blind, randomized clinical trial. Other methods can suggest cause and effect but are less reliable.
FDA has issued guidance: Evidence-Based Review System for the Scientific Evaluation of Health Claims (2009). Describes FDA’s systematic procedure for evaluation of the strength of the credible evidence to support a health claim, and establishes written criteria for what evidence is credible.
Step one: eliminate studies from which no substance/disease conclusions can be made, looking only at how a study was done, not at the results. Questions: was the substance that was the subject of the claim measured? One study of selenium, for example, used an estimate of selenium content in the body based on regional variations; this was excluded. Did the subjects have the disease that is the subject of the claim? Was the disease that is the subject of the claim specified and measured? Was the disease measured as a primary endpoint? Did the study include a concurrent control group? Was the study designed to identify the independent role of the substance on disease risk (e.g., the substance is zinc, but the intervention was a multivitamin with zinc)? Were baseline data different between control and intervention groups? How were the results statistically analyzed? What type of biomarker of disease risk was measured? Was the study long enough to ensure that any change in endpoint was in response to the intervention/any change in endpoint had time to show up? When the intervention is dietary advice, was there follow-up to show that the advice changed intake? Where was the study conducted? (If in an area where there are chronic shortages not present in American diets, for example, that will be inappropriate.)
Step 2: rate the methodological quality—high, moderate, or low—of remaining studies. Low quality studies are excluded. Was the study randomized, blinded, and placebo-controlled? Were the inclusion/exclusion criteria and other key study populations identified? Was subject attrition assessed, explained, and reasonable? How was compliance with study protocol verified?
Step 3: evaluate the totality of credible scientific evidence. Number of studies, consistency, quality, relevance to general US population.
Step 4: assess significant scientific agreement. Where there is a sufficient boty of relevant, publicly available scientific evidence showing consistency across studies and researchers, FDA looks at opinions of experts outside FDA for evidence of SSA, including recommendations from federal bodies like NIH and opinions of independent expert bodies like the American Cancer Society.
Step 5: determine language for qualified health claims.
Case study: selenium/cancer claim: 221 scientific documents available for review; three qualified health claims allowed, each beginning “one (or two) weak study suggests …” Attrition was significant!
Q: what about health claims for specific populations, like women or racial groups?
Hoadley: acceptable with substantiation for that specific population.
Q: a lot of times researchers are more interested in studying effects in a diseased population—probiotics in people who suffer from indigestion. Can you look at a disease population study as substantiation for the general public?
Hoadley: most heart disease claims are about lowering cholesterol; FDA has included people who have high cholesterol in relevant population, accepting that lowering cholesterol in heart disease-free people with high cholesterol works in the same way as helping people with heart disease does. But if you make a cancer claim, that’s unlikely to be accepted in the same way with people with cancer.
Q: are probiotics good for me?
Cleland: the right ones are wonderful!
Q: Dannon claims immunity, FTC thought the evidence wasn’t there?
Cleland: we never challenged the immunity claim. We challenged it in the context of the ad, when it made an implied prevention/reduction of risk of cold and flu claim. From our perspective, the studies that showed positive benefits of the strains of probiotics used in the products on the immune system did not substantiate the claim that it would reduce the risk of cold and flu.
Q: but you guys are the ones that made that connection.
Cleland: look at the ads. Little kids running around getting sneezed on; images of umbrellas coming out to protect the kids. Why else as a consumer do I care about the immune system of my kids? Because it will prevent illnesses, especially cold and flu, the usual problems for kids. We didn’t think the ad claimed that the product prevented cancer.
Q: so if the product label just said “immunity” without the ad campaign, would the FTC have reacted?
Cleland: he thinks not.
Q: using the consent order as guidance, but someone could come in with a good multicenter study—is this a rebuttable presumption?
Cleland: for advertisers in the first instance, it’s still an issue of reasonable basis. In most of our cases, we never reach the issue (before an order) of what counts as generally accepted substantiation among relevant experts because the evidence on the table is just not good enough. Iovate: Generally accepted standard in the area of weight loss is two well-controlled studies. In a de novo enforcement action, we’d have to address the specific substantiation; FTC would have to show the substantiation was not viable. If you’re looking for guidance, you can look at what the FTC says in these orders and have comfort in complying with the standards for those specific claims.
Q: isn’t there too bright a line between observation and intervention, at least when you have a prospective study? If the epidemiological study is conducted properly, it can control for confounding variables like socioeconomic status, other dietary habits, etc. Doesn’t this exalt the intervention study over the observational, when both have their defects (clinical trials have human variation).
Hoadley: he would agree, but we have to look at the substance that’s the subject of the claim. If it’s whole grains, observational studies may be helpful. But if the subject is some specific strain of probiotics, observational studies are not going to work. Some FDA approved claims are based wholly on observational studies—fresh fruit etc.