15 years of labeling: what’s changed and what’s next
Serialization mandates. UDI and 2D matrix codes. E-labeling rolling out market by market. In the past 15 years, pharmaceutical and medical device labeling has been rebuilt from the ground up and most teams are still catching up to what “accurate” now requires.
This RAPS webinar brings together Peter Muller (Director, Americas) and Marc Chaillou (Head of Sales Europe & Global Strategic Projects) at Schlafender Hase — a combined 31+ years of labeling and compliance experience — to walk through what’s changed, why “good enough” accuracy isn’t, and how TVT catches what a visual check alone would miss.
The standard no one questions
Your review process works. Until the standard itself is the problem.
Labeling teams in pharma and med device have never had more checks in place. Serialization rules, barcode standards, artwork sign-off, pre-press review — the process has only gotten more rigorous.
But most of that process is still built around a human eyes-on-page check, and regulators have historically accepted human-level accuracy — around 97 to 98% — as good enough. Applied to an average 2,600-word patient information leaflet, that standard still leaves roughly 52 words that could be wrong.
For labeling, which functions as a safety component, that gap isn’t a rounding error. One wrong word can trigger a recall or, in the worst case, put a patient at risk. This webinar covers why the industry is moving past “good enough” accuracy, and what actually closes the gap.
What we covered:
1. Fifteen years of regulatory change
The rules didn’t just get stricter — they got more complex
Peter and Marc walk through the shifts that have reshaped labeling since [webinar year]: serialization and track-and-trace under laws like the U.S. Drug Supply Chain Security Act, the move to package-level UDI and 2D matrix codes, anti-counterfeiting measures like holographic seals and tamper-evident closures, and the arrival of e-labeling (EPI) in HTML and MHTML formats. Each shift has meant a packaging redesign, tighter process controls, and more regulatory affairs involvement earlier in the artwork cycle.
2. Why “good enough” accuracy isn’t good enough
Human-level accuracy was never built for patient safety
Peter breaks down the FDA’s 97–98% accuracy benchmark, what it actually means applied to real label content, and why regulated labeling needs to be 100% accurate and validatable. The discussion also covers where AI genuinely helps today — drafting, translation, consistency checks — and why the core verification step still needs deterministic accuracy with a human in the loop, not an AI judgment call.
3. Live demo: catching deviations in real time
What TVT actually looks like, error by error
Marc runs a live walkthrough comparing label and artwork versions side by side, flagging text deviations, a mislabeled decimal (0.0025% vs. 0.25%), a repeated “do not” warning, and formatting differences that a visual check would likely miss. The demo shows how each flagged difference routes back to the person responsible for the update — with a full audit trail.
4. Audience Q&A: languages, file types, and what TVT does (and doesn’t) do
Straight answers on translation, XML, images, and more
Peter and Marc field live questions from attendees: whether TVT translates content (it doesn’t — it compares, regardless of language), whether it can flag content that should have been updated in a new document version, which file types it supports (including image comparison and XML for EPI), and what TVT Artwork — the graphics-focused version of the platform — actually does.
"Having a 98% accuracy level is not good enough. You have to have 100%, because even one word wrong can result in a recall — or, in the worst case, compromise the patient's health."
Peter Muller
Director, Americas, Schlafender Hase
Key takeaways
- Serialization and 2D barcodes reset the bar for packaging. Unique, package-level codes — not batch-level — mean tighter process controls, and catching errors earlier saves both time and cost.
- E-labeling adds regulatory complexity, not less. HTML, MHTML, and XML are now part of the content mix, and regulatory affairs involvement in artwork updates has only grown.
- 98% accuracy is a passing grade for humans, not for labeling. At a 98% accuracy rate, a standard 2,600-word patient leaflet still hides roughly 52 potential errors. Regulated content needs 100% — and validatable software to prove it.
- AI has a role in labeling — just not in core verification. Drafting, translation, and consistency checks can be accelerated with AI, but the core compare-and-catch step still needs deterministic, 100% accuracy with a human in the loop.
- Every deviation, in any format, is fair game for review. TVT compares Word, PDF, XML, HTML, and image/artwork files side by side — including barcodes and translated content — to catch what a visual check alone would miss.
Meet the Speakers:
Peter Muller
Director, Americas, Schlafender Hase
25+ years working with Fortune 500 companies across pharma, medical device, and consumer goods. Partners with international clients to define organizational goals and drive productivity gains, process improvements, and cost savings.
Learn More
Marc Chaillou
Head of Sales Europe & Global Strategic Projects, Schlafender Hase
Working in pharmaceutical and medical device labeling since 2011, specializing in pre-press — how processes and technology reduce costs and delays while reinforcing patient and consumer safety.
Learn More
