Label change impact assessment, explained.

What a new drug label sets off across every ad, email and brochure, in plain words

One new label, from FDA's approval letter to the last brochure accounted for

1. FDA approves the new labelVeltrimab Rev. 10/2026, on October 2. The approval letter says what FDA wants done about promotion, and by when.
2. Someone compares old and newSix changes, each rated for what it means to promotion. A new liver warning is the serious one.
3. Every live piece is read against the changes312 pieces in nine channels, by exact text, numbers, references and meaning. This is where a piece gets missed.
4. A person decides, piece by pieceWithdraw now, revise, update the safety text, or no change, with the two passages side by side and a signature.
5. Each decision is followed until it is doneEvery system where the piece lives is checked until it shows the piece gone or fixed. Revisions go back through review and to FDA on Form 2253.

All five steps are the subject of this guide. The people, companies and figures in this example are fictional.

Why "promptly" has no number, and three more surprises in label change work

Four words this guide uses: piece, claim, safety module, read-back

Fourteen terms are defined in the summary and glossary.

What a label change impact assessment covers, from the redline to the four decisions

In shortIt starts with a redline, a comparison of two label versions that marks each change, and ends with a recorded decision on every live piece. It covers everything a company uses to promote the medicine. Each piece gets one of four answers: withdraw now, revise, safety text update or no change.

FDA's rules use two words for promotion. Advertisements are ads in journals, magazines and newspapers, and ads broadcast on radio and television. Promotional labeling covers nearly everything else: brochures, mailing pieces, detailing pieces, letters, films, exhibits and "similar pieces of printed, audio, or visual matter descriptive of a drug".4 The duty after a label change names both, so an assessment that stops at ads misses most of the inventory.1

The assessment sits downstream of the label work. Regulatory affairs negotiates the label text with FDA. Here, the text is settled and the question becomes what it does to everything already in use. New safety information may be why the label changed. This work starts from the sentence that puts it in the label.

The four answers a promotional piece can get after a label change

The four answers are common practice as this guide reads it. FDA's rules require consistency with the label and name no such categories; your own procedure sets them.

A company that also sells outside the US may run a similar cycle in each country, on that country's own label. This guide keeps to the US case and notes the European and UK wording in section 3. Getting a company-wide label change into each national label is separate work.

The FDA promotion rules behind a label change: 314.70(a)(4), Form 2253 and what changed since 2023

In shortTwo near-identical sentences carry the duty: one for drugs, one for biologics. They say "promptly" and give no number. Form FDA 2253 carries each new piece to FDA at first use. Since 2023 the rules on broadcast ads have tightened and FDA's enforcement has jumped, while a rule to end the "adequate provision" route, which lets a TV or radio ad point people elsewhere for the full labeling, is still only planned.

Three words in that sentence do the work. All means every channel. Consistent sets a higher bar than checking that a piece has no false sentence. Implemented reaches changes the company made through the changes-being-effected route, which it may use for some stronger safety wording before FDA approves it.1

Each revised piece is then a new piece for FDA's purposes. A drug company submits "specimens of mailing pieces and any other labeling or advertising devised for promotion" at the time of first use, on Form FDA 2253, with a copy of the current professional labeling.6 Biologics follow 601.12(f)(4).2 Submissions in the scope of section 745A(a) of the Federal Food, Drug, and Cosmetic Act have had to go electronically since June 24, 2021.7

The rules that govern a label change in promotion, with their status in November 2026

How Europe and the UK word the same duty

Europe reaches the same place with fewer words. "All parts of the advertising of a medicinal product must comply with the particulars listed in the summary of product characteristics," says Article 87(2) of Directive 2001/83/EC.14 That summary is the officially approved description of the medicine and its use, Europe's counterpart to the US label.15 The UK carries that wording into regulation 280 of its Human Medicines Regulations, as the Blue Guide from MHRA, the UK regulator, explains.16

The industry's UK code, the ABPI Code, requires a signatory to certify that material is "not inconsistent with" the marketing authorization, which is the approval to sell the medicine, and the summary of product characteristics, and to recertify anything still in use at intervals of no more than two years.17 The EU's new pharmaceutical package is not law: Parliament's file read "Awaiting Parliament 2nd reading" on November 3, 2026.18 We found no European equivalent of FDA's seven-day letter sentence.

What the statement of intent to comply says, and how often FDA's letters ask for one

In shortThe statement is a short commitment to FDA, asked for in a sentence near the end of many supplement approval letters, due within seven days of receipt. It is a request in a letter, and the regulation itself sets no such time. In our own random sample of 2025 and 2026 letters on labeling supplements, 53 of the 118 we could read asked for it, about 45 in 100 give or take 9 points. Read each letter.

Many supplement approval letters carry a heading, "Promotional materials", near the end. Here is what it said in a letter signed September 30, 2026.

A 2014 letter uses almost the same words and cites 601.12(a)(4), and it gives a mailing address and fax number for the statement.20 Two things stand out in both. The clock starts on receipt, so a team needs to record when the letter arrived. And "prominent disclosure" sets a higher bar than adding the new warning somewhere in the safety text.

We drew that sample because nobody publishes the figure. From openFDA's copy of Drugs@FDA, FDA's database of approved medicines, we took every approval letter attached to an approved labeling supplement dated 2025 or 2026. With duplicates removed, that left about 1,300. We drew 120 at random and read their text. Two links returned an error page instead of a letter. Of the other 118, 53 carried the phrase "statement of intent to comply", 45 in 100.3

This is one sample of our own, with a margin of error of about 9 points. It covers only letters openFDA attaches to supplements in its labeling class, and openFDA warns that its data are unvalidated. We did not sort the letters by the kind of change, so we cannot say which changes draw the sentence.

A team that assumes the seven days will sometimes file a statement nobody asked for, and a team that assumes there is no clock will sometimes miss it.

What goes into a statement of intent to comply, and who sends it

We found no FDA template, and the letters ask only for a statement of intent. Most of the content below is therefore a company's own choice.

A useful statement says which label it refers to and quotes the regulation, commits to prominent disclosure of the named warning, and gives the counts and dates the company is prepared to stand behind. Regulatory affairs usually files it, since that team holds the correspondence on the application. A short statement can be drafted on day one. The counts behind it are the hard part, because they require the assessment to be well along by day seven.

Two more clocks a label change can start: 14 days and 10 days

The same letter sets another clock for the label itself: the content of labeling goes to FDA in structured form "as soon as possible, but no later than 14 days from the date of this letter".19,20 And where FDA itself required the change under section 505(o)(4), its guidance expects the new labeling on the holder's website within 10 calendar days of approval, or of FDA's receipt of a changes-being-effected supplement.10 Teams that link every safety block to a web version of the label depend on that page being current.

Which label changes send a piece to withdrawal, and which only to revision

In shortThe rating asks what the new sentence does to what your pieces say. A new warning can contradict a claim outright. A changed number can make a piece inaccurate, or worse if the piece uses it to reassure. A renumbered table only makes a citation wrong. Rate each change before you search, because the rating decides the clock.

Comparing the labels is quick if both exist as structured text and slow if they do not. FDA's Recent Major Changes list at the top of the label is a help: it names changed major sections and their dates and stays for at least a year.5 It is far from enough, since it says nothing about side effect tables, clinical study numbers or patient information.

A sound comparison covers every section and also the Medication Guide, the FDA-approved leaflet patients receive with certain medicines,21 because the supplement rules treat that as labeling too.1

Then each difference gets a type and a rating. The rating is about promotion, and it is a separate judgment from how medically important the change is. A new indication is good news medically and low risk here: pieces that list two indications are out of date, and nothing they say is false. A one-point drop in an efficacy number sounds trivial and still forces a revision of every piece that quotes it.

Six changes in one fictional label, rated for promotion

Fictional. 39 + 9 + 6 + 11 + 14 + 8 = 87 impacted pieces, each counted under its main change.

Look at the §6.1 row. Two pieces quote the same old 3% figure, and one gets a revision while the other gets withdrawn. A chart that says "Injection-site reactions: 3% vs 1% placebo" is out of date. A post that says "only 3 in 100 had a reaction" uses the figure to play down a risk, and FDA's draft guidance on presenting risk treats minimizing as a problem in its own right.9

So a reviewer has reason to look at how a number is used as well as whether it is right. Section 11 counts how often recent FDA letters raise it.

Nine channels where an out-of-date promotional piece stays live after a label change

In shortPromotion lives in many systems, and few talk to each other. The content system knows what was approved. It often does not know what is on a device today, which template a rep can still send, or how many brochures sit in a warehouse. Each channel has its own meaning of "gone", and its own way to check.

Content systems were built to manage approval and expiry, so they hold the record of what was approved. They may not hold the truth about what is in use. A piece approved in the content system gets copied: to the customer relationship management (CRM) system that holds rep email templates, to the closed-loop marketing (CLM) app on rep tablets, to a website's content system, to a media agency's trafficking sheet, to a print warehouse. Expiring it in the first place starts the work. Everything downstream still has to follow.

What "gone" means in each channel, and how you check it

Our reading of the common setups, grouped so the two websites share a row. Your systems will differ.

The channel mix decides how long the last mile takes. A website can change in an hour. An iPad changes when its owner next connects it. Paper changes when a rep drives back to the office. So one signed decision produces very different tasks, and the last copy of a piece is usually the one nobody can see from head office.

Why a keyword search of the claims library misses the pieces that matter most

In shortSearch finds the words you already know. Agencies adapt claims to the channel, numbers sit inside images, and references point to tables that have moved. A thorough search reads every piece four ways: exact claim text, changed numbers, old references and meaning. No public study gives a miss rate for keyword search, so measure yours on a past label change.

The usual first step is to search the content system for the affected claim and its known variants. It finds pieces that quote the claim. It has trouble with three other kinds. A patient-friendly rewrite says "No lab tests needed between visits" where the library says "No routine laboratory monitoring is required". A banner carries "3%" inside an image file. A footnote says "Source: USPI Table 5", citing the US prescribing information. It matches no claim at all, yet that table number just changed.

Four ways to read one piece, and what each catches

In that fictional case, 13 of the 87 pieces, about 15%, were found by meaning alone. That figure belongs to the example. We found no published miss rate for keyword search on real inventories. You can make one in a week: take your last label change, list every piece you finally changed, then run only the keyword searches your team ran on day one. The ones they miss are your number.

Text inside images deserves its own check. A web page's words can be searched. A slide exported as an image, a banner or a chart cannot, until something reads it. Any process, by hand or by software, should state which formats it reads, and a test set should include those formats.

Why the pieces found last are often the riskiest

The reworded pieces tend to be written for quick reading, which often means written to reassure: "simple", "nothing to check", "only". Those words are the ones a new warning contradicts. So the sentences a keyword search misses are often the ones that would need to come out first. All four of the withdrawals still waiting in the fictional case were found by meaning, while each of the five already handled had been found by exact text or a number.

How to clear a promotional piece without opening it, and what a spot-check can prove

In shortA shared safety module lets one update reach many pieces, so a piece that carries it and quotes no changed text can be cleared by rule. That rule needs checking, and a sample of 12 proves less than it seems: 12 clean results still allow up to about 22% wrong at 95% confidence. Size your sample for the error rate you are willing to live with.

Modular content, in plain words, means building pieces from approved blocks. The block most useful after a label change is the safety block: one approved text with versions, used in many layouts as a web footer, a detail aid slide or an email block. When the label changes, the block is updated once. A piece that carries it, or links to a safety page that carries it, gets the new text without being reworked.

That allows a defensible shortcut. A piece can be cleared without being opened by a person if two tests hold: it uses the current module, and a text and image read finds none of the changed text in the rest of the piece. Both tests matter. The module fixes the safety text. It does nothing about a claim elsewhere on the page.

How many cleared pieces to re-read

Any rule that clears pieces can be wrong, so a good process re-reads a sample. The sum you need is an old one. If you check n items and find nothing wrong, the true error rate could still be as high as 1 − 0.05^1/n at 95% confidence, which is about 3 divided by n for larger samples.22 Twelve out of twelve correct sounds complete. It means only that the error rate is probably below about 22%. On 225 cleared pieces, that upper limit is roughly 50 pieces.

So choose the sample from the consequence. A clearing rule on a warning that changes how doctors monitor patients justifies a large sample, stratified across channels and formats, at least until a few label changes have shown the rule holds. A small sample still has uses, since it catches a rule that is badly broken. Record what it does and does not show.

When the shared module is out of reach

If your pieces have safety text typed into each one, no clearing rule applies and each piece needs its own update, so the rework grows with every piece. The fictional case has 32 pieces like this, spread across eight of its nine channels. Moving pieces to a shared module is worth doing with or without any software, and it shortens every later label change.

Working the clocks on one October approval: 7 days, 5 business days, 30 days

In shortA label change runs several promotion clocks at once, and only one comes from FDA's letter. The rest are your own procedure, because the rule says only "promptly". Write them down before you need them, count from the right day, and compare your plan with the count still live each morning.

Take the fictional case. FDA approves on Friday, October 2, 2026, and the letter is received the same day. The company's procedure, chosen by the company, gives 5 business days for pieces contradicted by safety information, 10 business days for built-in safety text and 30 days for revisions.

The fictional company's procedure. Yours may differ, and the regulation gives none of the last three. October 12, 2026 is a federal holiday; a procedure that excludes holidays would move October 16 to October 19.

Those 30 days are tighter than they look. One vendor's data from over 350 companies put the average time to review and approve promotional content at 21 days in 2021, a figure the post does not limit to the US.23 Another vendor's 2025 benchmark gives an average job of 14.8 days and an average of 1.5 review rounds; its page does not say whether those are calendar or business days.24 On the older figure, a 30-day revision clock leaves 9 days for the brief and the agency's work.

Reading a burn-down: 62 pieces live on day 5 against a plan of 54

A burn-down is a daily count of the work left, and it tells a team most of what it needs: how many impacted pieces are still live on the old label, against the number your plan says should be left. In the fictional case, the known count starts at 83 and rises to 87 as the search finishes, since the list keeps growing while the work starts. On day 5, 62 pieces are still live against a plan of 54.

Seven ways a label change goes wrong in promotion, and what the public record shows

In shortThe public record is thin, because a company that catches its own miss rarely publishes it. A few cases show the shape: a letter built on an old label, a handout still given out after its withdrawal was confirmed, a page never filed on Form 2253. The seven errors below come from those cases and from how the work is organized.

An untitled letter is FDA's written notice to a company for violations below the threshold of a warning letter.25 FDA's promotion office posts its letters, naming the company and the product.26

Three public cases: an old label, an unfinished withdrawal, an unfiled page

Seven label change errors, why each happens and what catches it

Our analysis, from the cases above and the published rules. We found no public count of how often each one happens.

What an out-of-date promotional piece risks: patients, FDA letters and civil penalties

In shortThe first risk is to a patient whose doctor was told no monitoring was needed. Behind that come FDA letters, which rose from 5 in 2024 to 63 in 2025, civil penalties for false or misleading consumer ads, and complaints from competitors.

Start with the patient. If a detail aid says no lab tests are needed and the label now asks for liver tests at baseline and Weeks 4, 8 and 12, then a doctor may leave those tests out. Promotion exists to shape what doctors do, so a stale claim can change what happens to a patient.

On our count from one law firm's summaries, six of the ten untitled letters posted in the second quarter of 2026 alleged omitted or minimized risk, and at least five challenged efficacy claims.29

FDA's announcement of September 9, 2025 spoke of "thousands of letters" to companies and about 100 cease-and-desist letters, and said FDA "is already implementing AI and other tech-enabled tools to proactively surveil and review drug ads".13 Read that sentence as a buyer of promotion. A web page you forgot about may well be read by a machine at FDA. The FDA table counts only one office's untitled letters, and enforcement can move again, so read 2025 as a change in climate. One year makes no trend line.

Who can act on an out-of-date piece: the penalty statute, FDA and competitors

Consumer ads carry the statutory penalty named above.32 Promotion to doctors draws FDA's letters as well.26 And competitors can act: the UK case in section 10 began with a competitor's complaint, and the PMCPA noted that while undertakings between companies sit outside the Code, "it was important that companies complied with such undertakings".28

How much promotional material reaches FDA each year, and how often labels change

In shortFDA received 172,476 promotional materials on 80,043 Form 2253 submissions in fiscal 2026, up about 18% in materials since fiscal 2022. About 1,000 labeling supplements a year are approved for brand and biologic applications. More pieces, on labels that keep changing, means a growing inventory to check each time.

FDA now publishes the size of the pile. Its promotion office counts each Form 2253 submission and each material inside it, and one submission can hold several. In fiscal 2026, ending September 30, it counted 80,043 submissions and 172,476 materials.33 Materials rose from 146,053 in fiscal 2022, or about 18%.

A separate count shows how often the trigger fires. openFDA's copy of Drugs@FDA lists about 1,000 approved labeling supplements a year for new drug and biologic applications since 2019: 981 in 2025 and 663 in 2026 through late October.3 We left generic applications out, on our judgment that their sponsors do little promotion. Not every supplement touches promotion, and a brand can have several a year.

FDA's Safety-related Labeling Changes database covers a narrower set: approved safety-related changes from January 2016 on, for new drug applications and for the biologic applications FDA's drug center regulates (NDAs and CDER-regulated BLAs).34 In 2015 FDA estimated that about 407 application holders would post new labeling once a year in response to a safety labeling change notification.35

The spending behind the pieces, and the trend in how they are made

A peer-reviewed count of US medical marketing, now a decade old, gives $29.9 billion in 2016, up from $17.7 billion in 1997 in 2016 dollars, with consumer drug ads rising from 79,000 to 4.6 million.36 A vendor's benchmark says content production rose 7% globally and 29% in the US in 2023.37 One services firm reports over 18,000 promotional materials submitted a year for one client, a global specialty drug company.38

Two trends pull in opposite directions. Pieces multiply, because each channel wants its own version. Modular content pulls the other way, because a safety block updated once needs no work in each piece. The net effect depends on how disciplined a company has been. We found no public count of live pieces per brand, so 312 in this guide's example is an illustration.

What AI can do in a label change assessment, and what regulators and reviewers expect

In shortComparing labels and matching numbers is largely rules. Finding paraphrases is where a language model helps, and where it can be confidently wrong. No regulator has written rules for AI in promotional review by name, and we found no independent accuracy benchmark on promotional claims. Keep a person's name on every decision, and test on your own past label changes.

What regulators have said, and where it stops

FDA's Computer Software Assurance guidance covers device production and quality system software only, so it does not reach promotional review. ISPE's GAMP guide on AI, published July 2025, is the usual industry reference for validating such a system, which means proving it works as intended.47

The evidence on accuracy, and the limits

The nearest published work is a February 2026 preprint by authors at a drug maker and consultancies, testing compliance checks against a benchmark about EU AI systems, a different task. Its best design reported 97.5% recall, the share of real problems it found, and 72.1% precision, the share of its flags that were right.48

Read that as a pattern: tune a model to miss little and it flags many pieces that turn out fine. That trade-off suits this job, because a false flag costs a reviewer a little time while a miss leaves a stale claim in the field. Your own past label changes give a far better test.

Three limits hold whatever the score. A model can read text confidently and still miss an implication in a layout. It can state a reason that sounds right and is not, so each flag needs both passages shown. And a contradiction with new safety information should never be something a model can clear on its own.

How regulatory advertising teams run a label change today: roles, hand-offs and where the days go

In shortA regulatory advertising lead coordinates, often from a spreadsheet. Content operations, brand, agencies, field and legal each hold a piece of the work. Searching and confirming take the most days. Teams run this by hand, with a services firm, or on a promotional content platform, so check what each one does about the hunt.

Who carries a label change, from the approval letter to the last brochure

Where about 16 working days go on one label change by hand

DIA, the industry's professional association, has given the problem conference time, with a 2020 session on "strategies to address labeling changes: the impact on promotional materials" and a 2023 session on best practices for implementing one.49,50 Beyond conference agendas, we found little published on how the work is done, which is one reason teams find it hard to compare themselves.

Three ways teams run a label change in promotion

Veltrimab Rev. 10/2026: from approval day to every live piece accounted for

In shortThis is the example from the top of the page, inside the product. It is October 7, day 5. Six agents have compared the labels, read every piece and proposed an action for each, with both passages cited. Dana Okafor, the regulatory advertising lead, reviews, decides and signs, and only then do tasks open in every system.

1

The whole label change on one map

Dana opens on the impact map. On the left are the six label changes, in the middle what to do, on the right the nine channels where pieces are live. Thick bands mean many pieces. The line above reads 312 live pieces read, 225 cleared by the shared safety module, 87 impacted: 9 withdraw now, 46 revise, 32 safety text.

The liver warning, §5.4, runs dark red into withdraw now. Four withdrawals wait for her, due October 9, the same day as the statement of intent to comply.

2

Nine pieces to withdraw now

One click on Withdraw now lights its bands and lists the pieces: 2 rep detail aids, 2 rep emails, 2 paid search and display, 1 HCP web page, 1 social post and 1 print piece. Seven come from the liver warning. Two use the old 3% injection-site rate to play down risk, and the label now says 5%.

Each row in the list quotes the sentence the new label contradicts, and each is one click from its review.

3

225 pieces cleared, and why

Show cleared pieces adds them to the map in green: 139 carry the shared safety module and 86 link to the shared safety page. The module was updated once, to v10/2026 on October 3, and none of these pieces quotes changed text.

Dana runs a spot-check. The inventory searcher re-reads 12 cleared pieces from 7 systems, text and images, and all 12 are confirmed. The result goes on the audit trail. Section 8 shows what a sample of 12 can and cannot prove, so set the sample your own procedure calls for.

4

Slide 4: "No lab tests needed between visits"

The detail aid opens on the flagged slide. Beside it are both passages: the line in the piece, and the new §5.4 text asking for liver tests at baseline and at Weeks 4, 8 and 12. It was found by meaning, with a confidence of 97, because the claims library says the same thing in other words.

The left side shows where the piece lives: 412 rep iPads, a CRM template sent 1,284 times in 30 days, and 4,180 printed copies in stock, about 2,020 of them with reps. The agents propose Withdraw now. Dana can choose Revise, Safety text only or Not impacted, which needs a reason and Legal's second approval.

5

Tasks open only after she signs, and the proof comes from each system

At 9:42 am Dana signs: her name, the time and "I approve withdrawing this piece from every channel". Now tasks open. The content system expires the document so it cannot be used or ordered. iPads remove it at the next sync, and this screen shows 398 of 412 done. The CRM template is switched off. The field recalls printed copies and confirms destruction, and the brand decides whether a new version is needed.

The completion tracker reads back from each system until every channel shows the piece gone, and re-opens the task if it reappears.

In the solution's design targets, 225 of 312 live pieces are cleared without being opened by a person, 13 pieces are found by meaning alone, and a withdraw-now piece is out of every channel within 2 days of the decision. All three are targets on sample data, to be measured on your own label change.

The redline, the impacted-piece list, the fix tracker and the intent-to-comply draft on screen

In shortBehind the map sit the label comparison and sweep, the list of impacted pieces, the review of each one with its signature, the channel-by-channel tracker, the draft statement to FDA, a dashboard and the settings your procedure lives in.

Six agents, one signer and a second approval, and what connects to your content systems

In shortSix agents compare, map, read, propose, open tasks and read back. Each has things it never does. The regulatory advertising lead signs every decision and the statement of intent, Legal gives a second approval for "not impacted" on a safety change, and everything is on the record.

What each of the six label change agents does, and will not do

The people on a Veltrimab label change, and who signs what

Promotion controls that support your compliance

What the promotion team hands over, and what comes back

It connects by category to what you already run: your promotional content system and claims library, the closed-loop marketing and CRM systems behind rep iPads and emails, your website content system and email platform, media agency sheets and social scheduler, your print warehouse and speaker bureau. Clocks, approvers, rules, sources and matching are settings, versioned on the audit trail.

What a missed promotional piece costs you, and how to find your own number in a week

In shortSection 11 found no public figure for a late withdrawal. What is published marks the outer limits: FDA's clock in the letter, its civil penalties, and the climate of its letters. The rest you can measure in a week from your last three label changes.

Finding the cost of your last three label changes

Label change costs that no calculator counts

Your return on automating a label change: reading hours, revision rounds and days live

In shortThe return you can count comes from four levers: fewer pieces opened by hand, fewer revision rounds, fewer days a withdrawn piece stays live, and less printed stock wasted. The calculator counts the first two in dollars and the third in days, because a day of exposure has no honest price. Replace every illustrative input with your own.

Four places a label change assessment pays back

Your hours back across every label change a year

Is your promotion team ready? Eight scored questions on inventory, safety modules and read-backs

In shortEight questions, scored 0 to 2, and one for reflection. A low score points to data worth fixing first, with or without software. Your answers stay on this page.

Questions to ask any promotional content or label tracking vendor about impact assessment

In shortAsk these of anyone selling this, including us. Good answers come with a demonstration on your own past label change. A vendor who answers with a feature list, or who claims a certified product, has told you something too.

An eight-week label change pilot on past approvals, with results you can score

In shortPilot on two or three past label changes whose answers you know, so you can score every result. Add planted paraphrases and images to test the hard cases. Follow one live label change in parallel if the first results hold. Keep your people signing, and decide on numbers you set before you started.

Measures that decide a label change pilot

Label change impact assessment, statement of intent and Form 2253 questions, answered

In shortTwelve common questions about label changes and promotion, the FDA rules and software, each with a short answer.