01What the "Features (Characteristics)" section is for

Placed third in the comprehensive product information summary, the "Features (Characteristics)" section is where a drug's identity is impressed on the reader in a few words. It comes right after the development history (②) has set the background, and just before the clinical results (⑤), the heavy chapter of evidence. In other words, it sits at the point where the outline of the drug is handed over before any detailed figures have been shown. That is exactly why writers are tempted to compress: "in short, this is what the drug is."

That temptation to compress is the source of the danger. The reason the Creation Guide singles this section out and warns that it is the section most prone to exaggeration is structural: it tends to become the place where the conclusion is pre-empted and stated flatly. The shorter the assertion, the easier it becomes a claim severed from its evidence.

02Balancing efficacy and safety — why one side alone is a violation

In the features section, if you speak of efficacy you must present safety in the same place, in balance. This is not a matter of mere appearance. The duty the foreword raises — to convey accurately without misleading — has two layers: the duty not to lie, and the duty not to mislead. Even if every individual statement is factual, lining up only efficacy and pushing safety off to the margins forms in the reader's mind the image of "a drug that works with little danger" — an image different from the sum of the facts. This is the textbook case of "factual yet misleading," the gap the Creation Guide consistently seeks to close.

So the features section avoids the emphasis or guarantee of safety with the same severity it avoids false or exaggerated efficacy. Flat assertions such as "few side effects" or "highly safe," even when they capture one face of the data, are forbidden because they turn safety into a selling point. Safety is the object of an asymmetric duty — to be disclosed even when unfavorable — not a favorable sales line.

03Not a single step outside the approved scope

The efficacy and usage that may be written as features are confined to the approved indications and dosage, and the so-called qualifying (restrictive) wording (qualifying conditions such as "only when ○○ is ineffective or inappropriate") must be reflected accurately without omission. In a short feature sentence the qualifying condition is exactly what one wants to drop, but removing the limit turns it into the description of a different drug.

At the same time, the features section must not conflict with the warnings and contraindications. To state on the cover the patients who "must not use" the drug, yet give in the features section a broad impression as though such patients were included, sets the most important safety information against the body text. Because the reader carries forward the impression of the features seen first, this contradiction readily leads to real harm.

04If you assert it, put the basis inside the same material

If you touch on efficacy or safety as a feature, place the supporting results inside the same material and note the page. This is how the guide's pillar of structurally guaranteeing verifiability appears in the features section. Since a feature pre-empts the conclusion, you must always provide a route by which the reader can check that conclusion. If you write "it works," build a bridge by page number to which trial's which figures justify saying so.

Confirmatory analysis, or a nominal p-value?

When touching on efficacy in the features section, make clear whether the claim rests on a confirmatory analysis result or on a nominal p-value. A confirmatory analysis that confirms a single pre-specified primary hypothesis and a nominal p-value (any p-value derived from analyses other than pre-specified confirmatory analyses) bear entirely different weight as conclusions. Promoting a nominal p picked from a post-hoc analysis or subgroup into a feature, as though it were a verified effect, is a misstep down the evidence hierarchy.

05Comparison with other companies' products — the permitted range and the ban on evaluation

Touching on a comparison with another company's product in the features section is not itself forbidden, but the permitted range is narrow. If you mention a comparison, state the trial title, and you may show only the result of the primary endpoint (the confirmatory analysis item). You cannot speak of superiority by picking up secondary items or post-hoc figures.

Moreover, you must not add evaluation or commentary on the other product's results. Presenting figures as fact is one thing; laying meaning or a judgment of superiority over those figures is another. The latter touches the ban on disparagement. Name the other company's product by its generic name; do not identify it by brand name.

06How to write safety — consistent with the package insert, and no control-drug side effects

When touching on safety in the features section, keep it consistent with the (electronic) package insert, state the serious side effects and the main side effects, and note that the (electronic) package insert and the clinical-results safety data are to be referred to. Do not build a material-specific picture of safety; act as a window onto the approved safety information.

Here a brake specific to the features section applies. Do not record the side effects of placebo or the control drug.

07Do not link animal or in vitro data directly to the clinic

When touching on non-clinical findings in the features section, mark animal results with "(animal species)" and in-test-tube results with "(in vitro)", and do not link them directly to clinical efficacy or safety. Given differences in species, dose, and system, animal or in vitro findings cannot pre-empt a conclusion in humans. In the evidence hierarchy, peer-reviewed clinical research sits above; non-clinical work is placed below. Blur the source in the assertion-laden features section, and the reader mistakes a preclinical finding for a clinical fact.

08Brakes on charts and tables (detailed rules)

When a chart accompanies the features section, the manner of presentation itself is constrained. Each rule blocks "letting the same data form a different impression through presentation."

Do not turn a small number of cases into a percentage or graph

When the number of cases is under 10, do not graph the response rate or use a % notation; show it as an actual count (●cases / ■cases). Writing 2 of 3 cases as "66.7%" erases the smallness of the denominator and makes the figure look robust. Because a proportion from few cases swings widely by chance, an actual count is the only way to convey the uncertainty to the reader honestly.

Hazard ratio and risk reduction rate when there is no significant difference

When there is no significant difference between groups, the hazard ratio itself may be stated, but the risk reduction rate must not be stated from it. No significant difference means the direction or size of the effect does not exceed the range of chance. Translating that into "an ○○% risk reduction" presents an unconfirmed effect as a definite benefit. A hazard ratio is an index to be read together with its confidence interval and the absolute difference — not a tool for asserting a benefit on its own.

09"Significance ≠ clinical meaning" — to be kept in mind precisely here

Because the features section hands the reader a value judgment, a flat claim that mistakes the meaning of statistics does the most harm here. In a large trial, even a clinically tiny difference can become statistically significant; conversely, the absence of a significant difference is no proof of "equivalence." A feature sentence that swaps "it was significant" for "clinically meaningful," or "no significant difference" for "equivalent," states facts throughout yet breeds misunderstanding. Do not assert merely "it works" while leaving outside the section the context that matters — the range of effect shown by the 95% confidence interval, the size of the absolute difference.