I Am Not a Salesperson — Redefining the Role of Scientific Evidence Steward

People who build materials carry a second job title, I think. Not product manager, not marketer, not MR — evidence steward. The person entrusted with the responsibility of delivering the data on this medicine accurately to the world. When that sense of responsibility wavers, the largest violation category in seven years of monitoring — "explanations unsupported by evidence," 60 cumulative cases — begins, quietly. This article of conscience starts with a redefinition of role.

So What / So Why — Why This Article of Conscience Matters

The Ministry of Health, Labour and Welfare collected questionable promotional activity cases over seven years, from fiscal 2019 through fiscal 2025 — roughly 186 cases in total. The most frequent category is "explanations unsupported by evidence": 60 cumulative cases, approximately 32% of the whole. "Presented personal experience as data." "Declared the product safer than competing drugs based on mechanism of action alone." "Used a non-inferiority trial result to claim superiority." Reading through the case details, there is almost no sign that the representatives involved intended to deceive. If anything, a prior conviction — "this is a genuinely good drug" — was already in place, and the search for something to support that conviction led them to reach for whatever thin evidence was at hand.

The danger lies in how physicians use our information: as the basis for prescribing. A physician who hears "significantly safer than competing drugs in terms of bleeding risk" will build a prescription on top of a foundation that does not exist — because no head-to-head trial exists. Saying "there was no supporting evidence" after the fact cannot undo the prescription. The quality of information and the quality of patient care are continuous with each other.

Facing that reality, how you define your own role becomes the fork in the road. Begin the job as "someone who communicates why our product is good," and data becomes raw material to support a predetermined conclusion. Begin it as "someone who accurately stewards evidence and delivers it," and data becomes something you hold in trust on behalf of others. Which starting point you choose changes the judgment you make when you face the same piece of data.

The Article of Conscience — Who I Want to Be

To put this article of conscience in a single sentence: I want to be an evidence steward, not a persuader.

When I work as an evidence steward, my question is not "how can I use this data to move the physician" but "what does this data say, and what does it not say?" If the trial demonstrating efficacy is a non-inferiority trial, I do not claim superiority. If there is no comparative trial, I start with "there is no head-to-head data." If safety information is listed in the RMP, I communicate it even when no one asks. These are not ideals to aspire to — they are, I believe, the default behavior of someone who stewards evidence.

I will be honest: moments of doubt will arrive. There are instants when I think, "the competitor is taking share. Is there anything else I can push?" When that happens, I treat it as a signal that I am beginning to drift from "evidence steward" toward "persuader" — deciding that in advance is the practical meaning of this article of conscience. I have come to think that upholding the article is not about stopping the drift; it is about noticing it.

One more thing: the role of evidence steward is rooted somewhere other than loyalty to the product or the company. It is a sense of responsibility toward the entire mechanism by which physicians prescribe on the basis of accurate information. If the information I hand over is distorted, one part of that mechanism breaks. The awareness that I am doing something that consequential is what keeps the axis steady under pressure.

Daily Practice — Concrete Actions and Checks

• Write "what this data does not say" first. When building materials or slides, start by mapping out the distinction between primary and secondary endpoints, the patient populations excluded from the trial, and whether any comparative trial exists. Confirming what cannot be claimed before confirming what can is the habit that prevents later violations.
• Say the comparative basis out loud. Whenever using language like "compared with other drugs," verbalize whether a head-to-head trial exists. The moment you drop the prefix "there is no head-to-head data, but" is often the moment a factually unsupported claim enters the room.
• Ask yourself: "Who collected this data, where, and for what purpose?" A conference poster without peer review, a case report with fewer than ten patients, a single-institution survey — when data you want to use is at hand, check where it was generated. "There is usable data" and "there is evidence" are different claims.
• Distinguish between "communicated" and "included in the materials." Important precautions listed in the RMP may not be discharged simply by placing them in small print at the end of a document. Communicating them without being asked is the intent of the guidelines that require balanced presentation of both efficacy and safety.
• Test in advance: "Could I show the basis immediately if asked?" When building materials, confirm which page of a journal article, package insert, or review report answers the question "what is the evidence?" — so that if a physician asks, you can answer on the spot. If you cannot, that claim does not go into the materials.

When It Is Tested — Under Pressure

This article of conscience is tested hardest when deadlines, quotas, pressure from management, and competitive developments converge. "The competitor is getting formulary adoptions with this data — if I don't present it the same way, we lose." "My manager approved this slide structure." "There is no room left to miss this quarter's adoption target." In circumstances like these, the judgment of an evidence steward is easily eroded.

The monitoring reports for fiscal 2025 and 2026 recorded eight cases of competitor disparagement in total — a signal that intensifying market competition is reshaping the form violations take. On disparagement, the guidelines are explicit: even scientifically accurate information constitutes a violation if the purpose is disparagement. When the impulse arises to bring up a competitor's data, it is worth asking whether the purpose is to present evidence or to damage an impression.

"It's just one slide, there's no problem." "I qualified that part as a general observation." "The physician asked, so I answered." When phrases like these surface in your own thinking, they are a signal of local rationalization — a fading sense of responsibility for the overall impression the physician receives. The detailed mechanics of how judgment shifts under pressure are examined in "Inside the Maker" Vol. 1: The Map of Pressure — How Far Good Intentions Travel Before Bending. Cross-reference your own situation against it.

Anchors — Where to Return When You Are About to Break

When you are close to breaking, there are two places to return to.

The first is the language of the guidelines. The Guidelines on Promotional Activities for Prescription Drugs (enforced 2018) specify as principles of information provision: that information must be "based on scientific and objective evidence, with that evidence shown"; that information on safety, not only efficacy, must be provided; and that "arbitrarily inducing prescriptions without scientific basis is inappropriate." These are obligations — but read differently, they are also the guidelines' own endorsement of the role: your job is to steward evidence. Reading them as a role definition rather than as regulation gives the work a more settled center of gravity.

The second is what actually happened. The 60 cumulative cases in the "explanations unsupported by evidence" category include: a case where a representative qualified data as "for reference only since there is no evidence" and then used thin grounds to explain superiority; a case where safety was declared on the basis of mechanism of action alone — "gradual Hb rise means low thromboembolic risk"; a case where the same "more effective than competing drugs" claim was repeated across multiple sites (see Past Incidents Vol. 3: "Explanations Unsupported by Evidence"). The representatives in those cases do not appear to have been malicious people. And yet the violations occurred. The recognition that I could enter that same chain is the most realistic brake available.

Standing as an evidence steward is not a lofty declaration. It is the unglamorous habit of confirming, each time, what this data can and cannot claim. That habit, accumulated over time, makes it harder for the axis of judgment to shift under pressure. That is, I think, the most practical effect this article of conscience has.