Running the pre-launch marketing compliance check
Go-live is a Monday, 09:00. On Thursday afternoon the regulatory reviewer strikes the headline sentence on the launch email, the congress panel and the pre-roll cut, because all three were built from the same claim. Nobody argues the strike is wrong. The open question is who decides what ships instead, and whether that decision lands before the email platform locks its deploy queue on Friday at 17:00.
The pre-launch check is that decision made in advance, with names and dates attached. Everything below assumes the rest of the module: the authorities that hold the pen, what a defensible evidence file contains, the fairness duties owed to prescribers and patients, and how a claim is drafted inside the label. Here we run the sequence.
Work backwards from go-live
A review calendar is built from the launch date backwards, in working days, with one named owner per milestone. Holidays and congress print deadlines are on the same calendar, because they are what actually removes your slack.
| Working days before go-live | Milestone | Owner |
|---|---|---|
| T-60 | Asset inventory frozen: every piece has an ID, a channel, a market and a booked review date | Brand lead |
| T-55 | Reference pack locked (the evidence file the substantiation lesson describes) | Medical information |
| T-45 | Core claims and the safety-information treatment approved; every derivative asset builds from these | Regulatory |
| T-40 to T-15 | Rolling review of derivative assets, batched by channel | Committee chair |
| T-15 | Translations and affiliate adaptations enter as new assets, never as copies of an approved parent | Local regulatory |
| T-10 | Final sign-off, version IDs and expiry dates assigned | All three reviewers |
| T-5 | Media bookings, distribution lists and field kits reconciled against approved IDs | Brand ops |
| T-0 | Form FDA 2253 filed for US materials at first dissemination | Regulatory |
Two dates sit outside your control. If the product carries accelerated approval, 21 CFR 314.550 requires promotional materials to go to FDA ahead of first use during the early post-approval window, which puts a fixed 30-day block into a calendar you were already compressing. And if you want OPDP advisory comments on the core launch pieces, the queue runs in weeks, so the submission has to be made before the pieces are creatively finished, not after.
Who signs, and in what order
Sequence matters more than committee size. Medical reviews first: scientific accuracy and whether the data actually says what the sentence says. Regulatory second: label conformity, safety-information prominence, channel rules, 2253 obligations. Legal last: liability, competitor response, IP, contractual exposure. Running legal first burns your most expensive reviewer on claims that medical will delete anyway.
Four rules keep the sequence from collapsing:
- One named approver per discipline, with one named delegate. "The medical team" is not an owner. On a launch, someone is always on annual leave in week three.
- The brand lead has no vote on a strike. Marketing owns the substitute, not the verdict.
- Deadlock goes to a chair with a 24-hour clock, usually the country medical director. Unresolved disagreement is a decision to ship nothing.
- Every comment resolves in the review system of record, not in email threads. Veeva Vault PromoMats, which sells exactly this kind of review and asset-management software, is the common choice in pharma precisely because the audit trail, version ID and expiry field survive people leaving the company. A shared drive does not.
The triage log
Review output is a log, not a conversation. Each claim gets a disposition, a fix owner and a date. From a mock injectable launch:
| Claim | Disposition | Fix owner | Due |
|---|---|---|---|
| "Lowers A1C by up to 1.5%" | Pass, annotated to primary endpoint | Medical | Closed |
| "Once-weekly dosing" | Pass | Regulatory | Closed |
| "The safest option for diabetics" | Kill, no head-to-head data behind a comparative superiority claim | Brand, substitute required | T-12 |
| "Reverses insulin resistance" | Kill, outside the label | Brand, substitute required | T-12 |
| Patient hiking image | Rework, implies a quality-of-life benefit the file does not support | Agency | T-8 |
The third column is the one that gets skipped. A claim marked "fail" with no owner and no date reappears three assets later, because the copywriter reused the deck.
The fallback ladder at T-48 hours
When a claim fails two days out, you work down a ranked list and stop at the first rung that holds. Decide the order now, in writing, while nobody is panicking.
- Swap in a sentence from the approved core claims set. Fastest, because the claim itself needs no new review. Only the layout change gets a regulatory eye, and only to confirm safety-information prominence did not shift.
- Delete and reflow. Cheap on paper, but if deleting the headline pushes the safety information onto a second screen or below the fold, you have created a new problem while fixing the old one. Re-check, do not assume.
- Pull the asset, keep the launch. Works when the piece is one of forty. It fails when the piece is the congress panel that has already been printed and freighted, or the media buy that is non-refundable.
- Move the go-live date. The expensive rung, and the correct one when the failed claim is the campaign's central idea rather than a line inside it.
The reason to rank these in advance is that at T-48 the cost gradient is invisible from the brand side. Only ops knows that the email deploy is reversible until Friday 17:00 and the print run is not.
Knowledge check
1. In the opening example, a sales aid claimed a therapy could 'reverse tumor progression' when the approved label did not include the word 'reverse.' What core MLR failure does this illustrate?
2. A reviewer asks: 'Is every clinical claim in this brochure accurate and backed by the trial data?' Which lens of MLR review is this reviewer applying?
3. A drug is approved for 'moderate to severe plaque psoriasis in adults.' A marketing team wants to promote its use for a mild pediatric skin condition based on anecdotal physician feedback. Why is this prohibited?
4. Select ALL correct answers about how MLR review authorities and scope work.
Select all the correct answers.
5. Select ALL correct answers describing what each MLR lens is responsible for evaluating.
Select all the correct answers.
Worked example: does your review capacity fit your calendar?
Planning arithmetic, using ordinary launch numbers rather than any one company's:
- 74 assets in the launch inventory
- 2.3 review rounds per asset on average, counting the round that produces changes and the round that verifies them
- Committee sits twice a week and clears roughly 10 assets per two-hour session
That is 74 × 2.3 ≈ 170 review passes against 20 passes per week of capacity, so about eight and a half weeks of pure committee time before rework, before annual leave, before the affiliate translations arrive. A six-week runway does not fail gracefully. It fails by rubber-stamping in week five, which is how a claim that three reviewers technically approved ends up in an untitled letter.
Three levers exist and only three: raise throughput (pre-review offline so the session handles exceptions only), start earlier, or cut the inventory. Cutting is the honest lever and the least used. Forty considered assets beat seventy-four approved at speed.
Failure modes that survive a good process
Translation drift. A company running launches across dozens of affiliate markets, the way Novartis does, sees the same sentence re-approved locally in fifteen languages. Translated copy routinely comes back stronger than the English it came from, because "reduces" becomes the local equivalent of "eliminates". Adaptations enter review as new assets or they do not enter it at all.
The asset already in the field. If a claim fails after dissemination, you are no longer editing, you are withdrawing: kill the version in the review system, recall printed leave-behinds from reps, and expect FDA to expect corrective communication to the same audience that saw the original. The recall of a sales aid from 400 territories takes longer than the review that would have caught it.
Expiry cascade. Every approved asset carries an expiry date, and a mid-year label update invalidates every piece built on the previous safety information at once. If your expiry field is empty, the update becomes a manual hunt through channels nobody has inventoried. Set the date at sign-off, not later.
Legacy congress material. The slide deck from last year's symposium, still on a rep's laptop, is the single most common source of shipped non-compliant content. Version control that stops at the review system and never reaches the field is not version control.
Key takeaways
- The check is a dated calendar with one named owner per milestone, running backwards from go-live, not a checklist someone applies at the end.
- Review in order: medical, then regulatory, then legal. Marketing owns the replacement sentence, never the verdict.
- Rank your fallbacks before you need them, because the cost of pulling an asset is known to ops and invisible to brand at T-48 hours.
- Do the throughput arithmetic. Assets multiplied by review rounds against weekly committee capacity tells you whether your runway is real, and rubber-stamping is what happens when it is not.
- Assign a version ID and an expiry date at sign-off, so a label update becomes a query rather than a manual hunt through the field.
This lesson is educational and not legal advice. Always route real assets through qualified regulatory counsel and your internal review process.