(5) Ban on exaggeration and misleading expression | JPMA Guide Commentary | Advertising Regulation in Japan (Promotional Material Review) and Artificial Intelligence

Pharmaceutical information is subject to two distinct requirements that are alike in appearance but different in reach. The first is not to make false statements. The second is not to mislead. The second sets a higher bar than the first. A text can be free of every factual error yet still leave a reader with a wrong impression, through the selection of evidence, the placement of emphasis, or the cropping of context. What Chapter 1's prohibition on exaggerated and misleading expression targets is precisely that family of techniques — factual statements that nonetheless mislead.

The provisions list many specific prohibitions, but they resolve into four types: emphasis and guarantees of safety, impression management through authority cues, the leap from non-clinical evidence to clinical conclusions, and generalizing from exceptional data. None of these require a false statement. Each works by exploiting the reader's cognitive tendencies to produce a degree of confidence that the evidence does not support. The prohibition on exaggeration goes further than the prohibition on falsehood because it brings that cognitive manipulation within its scope.

01"Do not lie" and "do not mislead" — the asymmetry of the two requirements

The prohibition on falsehood is a minimum requirement: do not invent numbers, do not claim indications that were not approved. It is the floor that no one may breach. The prohibition on exaggeration asks what lies beyond that floor. An 83% response rate can be perfectly true; if the material does not say which population, which endpoint, and which trial it comes from, a reader will come away thinking "this drug works broadly." No lie is needed to plant that impression.

Understanding this asymmetry changes how the whole of Chapter 1 must be read. Beyond asking "is what is written correct," one must also ask "what will remain in the reader's mind." The same facts, presented differently, leave different impressions. The Guide regulates the manner of presentation, not only its content.

02Emphasizing and guaranteeing safety — the most frequently repeated prohibition

No expression that emphasizes or guarantees the safety of a drug is permitted, and in particular no statement that contradicts the warnings and contraindications in the prescribing information. This prohibition recurs throughout Chapter 1 in various forms. It is the single most frequently stated constraint in the chapter.

"Few adverse reactions" may be true — but few compared to what, which adverse reactions, and over what observation period? The type and frequency of adverse reactions are disclosed in the package insert. Any expression that contradicts that content, or steers the reader away from it, constitutes an exaggeration of safety. Phrases such as "excellent tolerability" or "patients can use it with confidence" carry the same risk. The first question is whether, held against the warnings and contraindications in the package insert, the expression creates a contradiction.

The prohibition on emphasizing safety does not bar any mention of safety. Recording a drug's safety profile as fact is required. What is prohibited is emphasis strong enough to push the risk information in the package insert into the background, and guarantee language that instills "this drug is safe" as a conclusion. Safety-related information must be disclosed even when it is unfavorable to one's own company. That asymmetric duty has a corollary: favorable safety information may not be made to look better than it is.

03Authority as impression management — what the photograph prohibition reveals

No false, exaggerated, or potentially misleading expression is permitted with regard to efficacy, safety, or quality. One of the detailed rules under this provision states that materials whose page layout centers on portrait photographs of healthcare professionals must not be made. At first glance this looks like minor guidance on image use. It encodes a more important principle.

When a white-coated physician or prominent specialist occupies the main visual space of a material, a sense of trust in the drug is created before the reader reaches the content. This is a textbook application of authority bias. Nothing exaggerated has been said in words; the look of the page does the work, creating the impression that "authoritative physicians endorse this drug." The detailed rule closes off this technique as a form of misleading expression.

The same logic governs catch phrases, photographs, and illustrations more broadly. Those that could mislead about efficacy or safety, or that damage the dignity of the drug, are not permitted. The regulation of expression is not confined to language. Visual elements fall under the same rules.

04The leap from non-clinical to clinical — crossing the levels of evidence

No expression that directly connects the results of animal or in vitro studies to clinical efficacy or safety is permitted. This prohibition translates the hierarchy of scientific evidence into a rule of expression.

A large proportion of compounds that show an effect in animal models do not produce the expected result in human clinical trials. Mechanism, metabolism, receptor distribution, and immune response all differ across species. In vitro experiments reflect conditions more remote still. Such data can support a hypothesis about mechanism of action; they cannot be directly converted into a claim of clinical efficacy or safety.

Form of presentation	Problem	Permitted wording
"Confirmed tumor reduction in rat studies. Antitumor effect in humans is anticipated."	Animal study results are directly linked to clinical efficacy.	"Reduction effect shown in rat tumor model (rat). Clinical efficacy requires separate confirmation."
"Potent antibacterial activity in vitro. High efficacy against infectious disease expected."	In vitro results are used to guarantee human efficacy.	State only "showed antibacterial activity (in vitro)" and keep the clinical data strictly separate.
"No serious adverse reactions in toxicity studies. Safety confirmed."	Animal toxicity data are used to guarantee human safety.	"No serious findings in [species] toxicity studies," labeling the species, and separating any statement about clinical safety.

When animal or in vitro data are included in a material, the species or condition must always be shown in parentheses — "(rat)," "(in vitro)." That one notation keeps the level of evidence visible to the reader. Narrating non-clinical and clinical data in the same flow, blurring the boundary between them, is the archetypal violation of this prohibition.

05Appeals to emotion — anxiety, fear, and material that damages dignity

Catch phrases, photographs, and illustrations that could mislead about efficacy or safety, or that damage the dignity of the drug, must not be used. Expressions that cause anxiety, fear, or discomfort, and expressions that damage the credibility of pharmaceuticals, are also prohibited.

Anxiety and fear have the power to make a modest benefit look like an urgent solution. Expressions such as "before it's too late" or "without treatment, this cannot be reversed" drive acceptance independent of scientific evidence. This too is a form of exaggeration. Pharmaceutical information is supposed to support considered decision-making. Techniques that manipulate emotion and short-circuit judgment deprive the reader of the chance to evaluate the drug on the evidence.

The prohibition on language that damages dignity is also directed at disparaging comparisons with competing products. Exaggerating a competitor's adverse reactions, or presenting clinical outcomes in a way that unfairly positions a rival drug as inferior, is the mirror image of overstating one's own drug. References to competitors are confined to the facts of data from appropriately designed trials.

06Generalizing from exceptional data — the prohibition on cherry-picking

Taking exceptional or limited data and presenting it so as to give the impression that it represents a general fact is prohibited. This is the prohibition on cherry-picking. From the many data points a trial produces, the best result from a subgroup, a time point, or a particular analysis method is selected and placed in the foreground, as though it reflects the overall picture. The numbers are real; the presentation, stripped of context, misleads.

The recurring patterns are recognizable. A subgroup analysis not specified in the original protocol is run on a trial where the overall population showed no significant difference, and only the subgroup result appears in the material. Data that looked favorable at one specific time point are shown in a way that suggests durability. When the primary endpoint showed no difference but a secondary endpoint did, the secondary result is made the lead finding.

Chapter 2's requirements — show only the result of the primary endpoint as the confirmatory analysis item; distinguish exploratory analyses from confirmatory ones — are the concrete implementation of this prohibition. A pre-specified answer to one pre-specified question is confirmatory evidence; a favorable number found after the fact is a reference finding at best. Making that distinction legible to the reader of a material is what the prohibition on generalization requires in practice.

In closing

The prohibition on exaggerated and misleading expression has the widest reach of any constraint Chapter 1 imposes. Where the prohibition on falsehood asks whether the content stated is true, the prohibition on exaggeration asks whether the impression conveyed is accurate. Prohibiting the emphasis and guarantee of safety, prohibiting authority cues through portrait photographs, prohibiting the direct connection of non-clinical results to clinical conclusions, prohibiting the generalization of exceptional data — each is a fence that protects the healthcare professional's ability to evaluate a drug accurately.

The three pillars of the foreword — the package insert is the original document; inform accurately without misleading; disclose information unfavorable to one's own company — take concrete form here as specific prohibitions. What the author is not permitted to do is not only to lie. Selection, emphasis, omission, and visual framing that steer the reader's judgment are equally forbidden.

Case studies — the "it isn't written, so…" mindset, and how each lands

"Don't lie" is the floor; "don't mislead" is a ceiling one level higher. Selecting facts, staging authority, transposing animal data into clinical claims, and passing exceptions off as general findings — all involve accurate numbers and a distorted impression. The four levels trace that progression.

GRAY 1 — Straying from the spirit (untouched by the wording, but against the foreword's spirit)

The scheme: "We didn't write 'excellent tolerability.' We wrote 'discontinuations due to adverse events were limited to 2.1% of the study population.' The figure is accurate. We made no mention of the package insert's serious adverse events section, which lists events that can follow a fatal course."

Verdict: Conveying a favourable safety impression in a context that conflicts with the warnings and contraindications in the package insert is the pattern the provision flags most repeatedly. Citing a low discontinuation rate while omitting serious adverse events makes the safety profile appear better than it is. Any safety statement must be checked against the package insert's risk information for discrepancies before it is written.

GRAY 2 — Exploiting the gap (deliberately working a loophole to invite misreading)

The scheme: "A well-known specialist's photograph appears prominently. To comply with the prohibition on layouts 'centred on' physician portraits, the photo occupies 30% of the spread and product data fill 70%. The centrepiece is data, not the physician."

Verdict: The provision restricts physician photographs to prevent authority-bias impression-making. Even at 30% of the page, a recognisable specialist's image creates trust before a reader engages with the content. The prohibition is assessed by functional outcome — whether it creates a misleading impression — not by a fixed area threshold.

GRAY 3 — Violating higher norms (arguably "not in the text," yet almost certainly out under the PMD Act, Fair Advertising Standards, and the JPMA Code)

The scheme: "Animal model data support the phrase 'potently suppresses inflammation by inhibiting [target].' This is a mechanism-of-action description, not a clinical efficacy claim. Using animal data is not explicitly prohibited."

Verdict: The provision states: "Do not draw a direct link from animal or in vitro results to clinical efficacy or safety." Attaching "potently" to a non-clinical finding leads readers to extrapolate a clinical effect size — that is the prohibited leap. The PMD Act and JPMA Code treat the resulting exaggerated impression as grounds for violation. Mechanism-of-action descriptions must be clearly attributed to animal or in vitro conditions.

BLACK — Breaching the explicit text (a plain violation of the provision itself)

The scheme: "An 83% response rate was observed in the high-baseline-severity subgroup. The overall population reached 63%. The headline reads 'response rate 83%.' That number is real — the subgroup finding is factual."

Verdict: The provision explicitly prohibits "conducting various analyses and presenting only results favourable to one's interpretation." Suppressing the overall population result and presenting a subgroup's higher figure as the product's response rate is textbook cherry-picking — a direct breach. The overall population result must be shown first; the subgroup finding belongs in a separately framed reference-information section.

Improper intent by implication — three scopes of "insinuating" without stating it

The prohibition on misleading expression sets a higher bar than "do not lie." Selection of facts, addition of adjectives, and context-clipping can manufacture a false impression from accurate numbers. The three manipulations below illustrate the point.

Implication 1 — Local: carried in a single word or sentence

How it is built: "The score improvement rate at week 12 was 42%, demonstrating marked improvement." The 42% figure is the correctly reported pre-specified primary endpoint result. "Marked improvement," however, is the author's language — the protocol defines no threshold for that phrase.

Intent that can be read: Appending the clinical-judgment adjective "marked" to an accurate measurement makes statistical significance and clinical importance appear synonymous. The basis for "marked" is nowhere defined in the material.

Verdict: Numerical accuracy does not validate an adjectival clinical appraisal. If "marked improvement" is to appear, the standard that defines it — minimum clinically important difference, pre-specified threshold, or similar — must be stated. An unanchored clinical-appraisal word is a form of exaggeration.

Implication 2 — Adjacent: read across neighboring clauses

How it is built: Sentence one: "A statistically significant difference from placebo was confirmed on the primary endpoint (p=0.032)." Sentence two: "The average symptom score improved by 37% at week 12." The two sentences draw from different trials and different endpoints, yet no attribution note separates them.

Intent that can be read: The reader takes both sentences as results from the same trial and the same endpoint. The statistical significance context of sentence one transfers to the 37% figure in sentence two, generating the reading "a statistically significant 37% improvement."

Verdict: Presenting results from different sources without connectors causes readers to misattribute data. Each result sentence must be labelled — before or after — with its trial name, endpoint, and source, and results from different datasets must be visually separated rather than placed in consecutive unmarked sentences.

Implication 3 — Whole-context: arising from the document as a whole

How it is built: Throughout the material, every adverse-event statement is paired with "mild to moderate, transient." Individual statements are consistent with the package insert; a serious adverse event (incidence 0.8%) is accurately described as "transient." The modifier appears on every adverse-event entry without exception.

Intent that can be read: Even though each statement is accurate, the unbroken pattern of "mild, transient" across the entire material produces the cumulative impression that "all adverse events are minor and temporary." The 0.8% serious event is technically present but practically erased by the surrounding frame.

Verdict: A repeated modifier pattern creates an overall impression that exceeds the sum of its individually accurate parts. Applying a uniform "mild, transient" frame to adverse events of differing severity distorts the perceived severity distribution. Adverse events of different severity grades must each be presented with their actual grade and management guidance, not subsumed under a single reassuring phrase.