The numbers are right. The conclusion is wrong. — How data repurposing and manipulation create false impressions

②-1 Anticancer drug / Slides (in-hospital briefing)
A trial enrolled patients who had shown resistance to first-line treatment. A graph comparing response rates between the overall population and the Japanese subgroup was presented, with the explanation that "the drug is more effective in Japanese patients." When the company representative was questioned, the response was that "this may be because the first-line treatments differed between the overall population and Japanese patients." The regulatory review report confirmed this: in the overall population, Agent A was used in roughly 50% and Agent B in roughly 30% of cases, whereas in the Japanese subgroup, Agent A was used in roughly 20% and Agent B in roughly 70%.
Core violation: Data with different preconditions (first-line treatment content) was used to create a false impression that the drug is more effective in Japanese patients. Original text: "explained that the drug was more effective in Japanese patients compared to the overall population, without disclosing the difference in preconditions between the two groups"

②-2 Antipsychotic / Oral explanation (in-hospital study session)
When asked whether QT prolongation had been observed, the representative answered that "in overseas trials, QT prolongation was not observed at a dose of 12 mg/day." However, the regulatory review report showed that QT prolongation had in fact occurred in domestic trials at doses lower than those used overseas.
Core violation: Results from domestic trials unfavorable to the product were not disclosed; overseas trial results alone were used to assert safety.

②-3 Dyslipidemia drug / Slides (in-hospital briefing)
The original trial was a three-arm comparison (low-dose, high-dose, placebo). The slides showed only the low-dose arm and placebo arm for graphs including TG change rate, concealing the finding that increasing the dose produced little additional effect.
Core violation: From a three-arm trial, results from only one or two arms were extracted to construct the graph.

②-4 Anticancer drug / Oral explanation (new-drug evaluation)
The appropriate use guidelines stated clearly that "superiority over the control arm has not been established." Despite this, the representative emphasized only the portion of the Kaplan-Meier curve where the treatment arm was ahead of the control arm, and claimed superiority. Unfavorable information — that OS was shorter in the treatment arm compared to the control arm among patients who had received hematopoietic stem cell transplantation — was also withheld.
Core violation: Only the portion of the KM curve where the product appeared superior was highlighted, and superiority was claimed despite no superiority having been established.

②-5 Psoriasis treatment / New-drug evaluation materials
A safety comparison table between the product and a comparator was translated from the original paper for inclusion in the materials, but the row for "malignant tumors" was missing entirely. Malignant tumors were a particularly serious adverse event — and one that occurred with the product but not with the comparator.
Core violation: The one adverse event category that was both most serious and most unfavorable to the product was excluded from the citation.

②-6 Diabetes drug / Product brochure
The brochure featured the primary endpoint graph alongside the claim that "a superior reduction in HbA1c was observed," but the vertical axis had been changed from the "HbA1c change from baseline" used in the original paper and regulatory review report to "absolute HbA1c value." The primary endpoint was the change from baseline; converting to absolute values does not constitute citation of that endpoint.
Core violation: The graph was constructed by changing the primary endpoint from change from baseline to absolute value.

②-7 Diabetes drug / Video on a healthcare professional information website
A graph showing blood glucose trends for the treatment arm and placebo arm had the legend displayed in reverse. The aspect ratio had also been altered to make the graph wider, creating the impression that blood glucose fluctuations were small and glycemic control was good.
Core violation: An incorrect legend and an altered aspect ratio together produced a misleading visual impression.

②-8 Analgesic / Product brochure
The brochure cited a graph from the original paper showing no change in blood pressure in the treatment arm. However, the original paper also included a graph for the comparator, which showed no statistically significant increase in blood pressure either — even though the comparator was being used at several times the domestic approved dose. By omitting the comparator graph and showing only the product's graph, the brochure implied that only the product carried no risk of blood pressure elevation.
Core violation: The comparator graph was omitted, leaving only the product's graph and creating the impression that only the product carried no blood pressure elevation risk.

②-9 Antiviral / Product brochure
A brochure focused on "virus reduction effect" gave the primary endpoint (illness duration; no significant difference from comparator) only a few lines of text with no comparator comparison. Secondary endpoints (change in viral titer and time to cessation of viral shedding) were presented across two full pages with comparative graphs and commentary.
Core violation: Only the secondary endpoints that showed a favorable result were described in detail; the primary endpoint, which showed no significant difference, was given minimal treatment.

②-10 Diabetes drug / Roundtable article on a healthcare professional information website
An adverse event summary table was presented as an overview of Phase III trial adverse events, but items that had occurred frequently in the clinical trial were not included.
Core violation: Adverse events that occurred frequently in the trial were excluded from the summary table.

②-11 Anticancer drug / Product brochure
A brochure presenting Phase I clinical trial results showed an adverse event rate substantially lower than a competing product. When the regulatory review report was checked, it contained information about pre-medications that had been administered — information absent from the brochure. The difference in adverse event rates may have been attributable to those pre-medications. Additionally, the trial objective "safety comparison" as described in the regulatory review report had been removed from the brochure; only "verification of pharmacokinetic equivalence" remained as the stated objective.
Core violation: Information on pre-medications affecting adverse event rates was omitted, and the safety evaluation purpose of the trial was deleted from the description of trial objectives.

②-12 Antirheumatic drug / Product brochure
Multiple instances of inappropriate citation from the original paper were identified in the product brochure. ① The original paper listed serious adverse events individually with detailed incidence information; the brochure consolidated multiple items and omitted entries for events requiring particular vigilance in using the product. ② Numerical values cited from the original paper did not match the source. ③ For patient satisfaction survey results, the brochure omitted the statement that patients with unknown responses had been excluded from the calculation.
Core violation: Three layers of inaccuracy were present simultaneously: consolidation and omission of serious adverse event details, numerical discrepancies, and undisclosed exclusion criteria.

②-13 Bronchial asthma drug / Slides and brochure (in-hospital briefing)
For the primary endpoint of a Phase III international joint clinical trial (annual asthma exacerbation rate), the representative presented only the Japanese subgroup analysis (approximately 15 patients per arm) without disclosing the results of the overall analysis (approximately 250 patients per arm). All other secondary endpoints were presented using overall data. When asked why only Japanese data were used for the primary endpoint, the response was: "Because physicians ask for Japanese data." The Japanese subgroup results were more favorable than the overall analysis despite the small sample size, raising concerns about selective presentation. A simplified brochure distributed to physicians also contained only the subgroup analysis results.
Core violation: For the primary endpoint, only the small-sample Japanese subgroup analysis was presented, and efficacy was claimed on that basis alone.

②-14 Hemophilia drug / Presentation slides
A product presentation displayed side by side: (1) the proportion of patients in the once-weekly dosing group who experienced no bleeding, and (2) the proportion of patients who were able to continue once-weekly dosing and experienced no bleeding. The second figure does not appear in the distributed integrated product information summary or the regulatory review report. It excludes patients who could not continue once-weekly dosing for any reason, making it a figure affected by attrition bias whose favorable appearance is not a reflection of the drug's true performance. The presentation slides were not distributed to attendees.
Core violation: A selective analysis excluding patients who discontinued was placed alongside official data, creating a risk of falsely favorable impressions.

②-1 Antibacterial / Product briefing slides
A slide used by the MR combined results from multiple clinical trials. Primary and secondary endpoint results were presented on equal footing; trials with different eligibility criteria (age) and dose-finding studies with different objectives were mixed in; individual patient counts were hidden; and the control arm was not shown despite the trials being non-inferiority designs. Taken together, these choices created an overall impression that the product was superior to other agents.
Core violation: "Results from multiple clinical trials were aggregated, and data excerpting, manipulation, and presentation methods that created the impression the product was superior to other drugs were employed."

②-2 Diuretic / Product brochure
A table of adverse event rates by treatment period used a threshold of "≥1.0%." The original paper used a threshold of "≥0.5%," and this difference in cutoff caused three adverse event items that appeared in the original paper to disappear from the brochure table.
Core violation: "By showing only adverse events above a certain frequency, the presentation made it appear that no other adverse events existed — an excerption of data from the original paper."

②-1 Antirheumatic / In-person visit (interview form and integrated product information summary)
The clinical trial described in the regulatory review report involved switching patients between the product and another drug (a crossover design). However, the trial design diagrams used in the interview form and integrated product information summary omitted any reference to the other drug, making the study look like a simple switch to the product as monotherapy.
Core violation: "The diagrams contained no mention of the other drug, and were liable to create the misimpression that the trial results applied to a simple switch to the product as monotherapy."

②-1 Dermatitis drug / Online visit materials (integrated product information summary and company website displayed the same content)
The trial design compared the product arm vs. placebo and Drug A arm vs. placebo separately — it was not designed to directly compare the product to Drug A. Nevertheless, the materials, the integrated product information summary, and the company website all included a graph to which "an estimated between-group difference not present in the original paper had been added." The inserted estimate, placed between the product arm and Drug A arm, created the impression that the product was superior in a direct head-to-head comparison.
Core violation: In a trial not designed for direct comparison, a between-group estimate absent from the original was added, making the graph look like a direct comparison favoring the product.

②-2 Joint function improvement drug / Online product briefing slides and oral explanation
A single slide displayed graphs for both the primary and a secondary endpoint, but the secondary endpoint graph was presented at a larger size than the primary endpoint graph. During the explanation, the representative "described efficacy using the secondary endpoint without mentioning that it was a secondary endpoint."
Core violation: The visual weight of primary and secondary endpoints was reversed, and efficacy was asserted on the basis of the secondary endpoint.

①-1 Diabetes / chronic heart failure treatment / Web product briefing
A drug price comparison slide with a competing product was presented. For heart failure, only the 10 mg dose is approved for both products. However, the slide used the dosing classifications for the diabetes indication — where the company's product has two approved doses (5 mg and 10 mg) and the competitor has two approved doses (10 mg and 25 mg) — to label the company's product as the "standard dose" and the competitor's as the "high dose," and used this framing to claim a price advantage. The precondition that both products are indicated only at 10 mg for heart failure was not disclosed, leaving the audience susceptible to the impression that 10 mg of the company's product was the standard dose while 10 mg of the competitor's was a high dose.
Core violation: "A graph liable to cause factual misrepresentation was shown, describing the company's own product as the 'standard dose' and a competitor's product as the 'high dose.'" The dosing classification from the diabetes indication was transferred to the heart failure indication — a substitution of preconditions.

①-1 / ②-1 (same incident) Hyperlipidemia drug / Online product briefing
The product is administered by subcutaneous injection at a medical institution every six months; the conventional drug allows at-home self-injection every two to four weeks. At the briefing, patient and physician survey data were presented showing "patients preferred administration every eight weeks over every two or four weeks, and preferred injection by a healthcare professional over self-injection." The survey, however, covered patients with severe asthma (not the product's indicated population), conducted overseas in a country with substantially different health insurance arrangements. Furthermore, the original source contained a five-item comparative graph, but the briefing materials showed only the two items favorable to the product.
Core violation: "Favorable items were excerpted from an overseas publication with different healthcare systems and patient demographics, and the resulting materials were used as the basis for information provision." The product's convenience was argued using data from a different disease, a different country, and a different healthcare system.

②-2 Antiviral / Mass email distribution (direct mail)
An MR sent a mass email to healthcare professionals introducing the product. A graph comparing drug interactions between the product and another drug was included, but the original paper documented a group of cases classified as "non-interactors" — patients for whom either drug would produce the same outcome. In the distributed graph, this group had been removed.
Core violation: "Excerpting, modifying, or combining data when citing figures and tables from original papers constitutes a guideline violation." Cases where no difference between drugs existed were deleted, exaggerating the apparent difference.

④-1 Peripheral nervous system drug / In-person explanation (new-drug briefing)
The representative described the incidence of sleep disorder as an adverse event, stating: "Competitor Product B: over 10%; our Product A: 2.4%." A healthcare professional later checked the interview form and found that the figure for Drug B was the incidence of all adverse events, while the figure for Drug A was the incidence of somnolence alone. A proper comparison would require comparing somnolence rates between the two drugs. Using figures with different scopes made the product look safer. In fact, the somnolence rate for Drug B was below 1% — the opposite of what was implied.
Core violation: Figures measured on different definitions were placed side by side, "constituting an explanation liable to cause factual misrepresentation."

④-2 Peripheral nervous system drug / In-person explanation (same drug)
In a non-inferiority trial, a secondary endpoint (duration of treatment success) for which no significance test had been conducted was used to claim: "compared with the *** group, the drug shows a superior difference at 168 hours and therefore produces a prolonged effect." When asked by a healthcare professional for the significance test result, the representative replied repeatedly: "No significance test was conducted, but a superior difference was observed." Describing a numerical difference in a secondary endpoint that was never tested for statistical significance as a "superior difference" misrepresents the interpretation of the trial design.
Core violation: A secondary endpoint from a non-inferiority trial for which no significance test had been performed was used to assert superiority without statistical basis.