It’s astonishing to think that, in the highly-regulated pharmaceuticals and medical devices industry, labelling mistakes – and the recalls that follow – are as commonplace as they are.
Then again, the pace at which the industry is moving means we can also see why it happens. Frequent updates from the regulator, demand for new treatments to be accelerated, drug reformulations, business acquisitions and a reliance on manual processes can all open manufacturers up to labelling errors.
In 2022, 166 manufacturing sites generated 912 drug recalls according to the FDA and so far this year, the FDA has also recalled one product per month relating to mislabelling – which is one of the most common reasons for recall, surpassed only by undeclared ingredients and contamination.
Meanwhile, in the UK, printing, labelling and packaging errors are implicated in a number of safety alerts and recalls. Of the seven safety messages recorded by the Medicines and Healthcare products Regulatory Agency (MHRA) in September, four relate to missing or incorrect information, or labelling errors.
Nurses are usually the first to spot a problem with a drug or device, including any labelling concerns. As well as the unacceptable risk to patient safety, dealing with quality issues can seriously impede efficiency in a busy clinical setting.
The cost of a recall can run into millions for pharma companies, on top of the reputational damage that is particularly difficult to repair in an industry built on safety, efficacy and trust.
Prevention better than cure
Sometimes, it’s only when a manufacturer is faced with a recall that they realise traditional labelling systems are no longer fit-for-purpose in today’s world.
It’s easy to adopt an ‘if it ain’t broke’ mentality to labelling simply because you have not been faced with any major issues until now.
But the scale and complexity of modern manufacturing, along with the need for full traceability, is such that firms must proactively prepare for sudden changes, before a product reaches a clinician.
It’s not just recalls that expose a broken labelling system either. Regulatory changes and reformulations can also require labels to be changed at speed while maintaining compliance.
It could take 10 employees several weeks to manually change 1,000 labels. Multiply that over millions of units, sold across the world, and you soon see how this can push up costs up to unsustainable levels and reduce the availability of critical treatments.
On top of that, labels may need to be changed typically around five times a year because of updates, putting a significant strain on resources.
As well as being highly-inefficient, manual labelling and artwork creation also heightens the risk of human error, so the process of correcting previous mistakes could be counter-productive.
If your company is reliant on manual labelling then you are far from alone. The same applies to some of the most innovative pharmaceutical and medical device manufacturers around the world.
Pharmaceutical and medical device manufacturers are becoming more ambitious in their digital transformation strategies as the benefits become more widely known.
A truly connected supply chain may still be a long way off, but the direction of travel is clear. The appointment of chief digital and technology officers (CDTOs) is evidence of this, as this report from McKinsey suggests.
Its authors go on to say that top players in the industry are already incorporating analytics into research and development (R&D) and transforming how companies work with healthcare providers and patients.
The problem, highlighted in this report and many others, is that digital initiatives are often siloed when what’s needed is large-scale transformation, with technologies like artificial intelligence (AI) and machine learning (ML) being integrated into every business process.
Without that vision, there’s a risk that pockets of manufacturing will remain manual, and eventually they will become an unsustainable barrier to efficiency. Labelling is a good example of this.
Given how long it takes to manually update just 1,000 labels, it’s easy to see how one task could slow down an otherwise fast-moving and data-driven supply chain.
The benefits of AI and ML in life sciences are now widely-recognised even if adoption is still in its infancy. Given the power of these technologies to make decisions that impact human life, it’s only right that companies should tread with caution.
As one technology expert puts it, adoption of new technologies isn’t particularly slow in life sciences – companies need to undertake robust due diligence checks before making critical decisions.
Technology vendors must play their part in building trust in new technologies, like AI. Along with relevant ISO certifications, pilot projects between industry and academia can provide a proof-of-concept that the technology could be used in a compliant way.
Labelling and artwork management (LAM) software might not be the first area that springs to mind when thinking about the processes that could be transformed by AI and ML.
In fact, our collaboration with Aston University – part of a Knowledge Transfer Partnership (KTP) – is only the first instance of the technologies being integrated within labelling software.
Academics at Aston already had experience embedding AI and ML algorithms into software for other sectors, so it wasn’t a great leap to apply it to labelling.
Our LAM software was developed to guarantee uniformity and support the most up-to-date legal, regulatory, marketing and manufacturing standards in all aspects of packaging and labelling.
Going a step further, the experts at Aston University demonstrated how AI and ML could automate and improve the artwork creation stage of the labelling process.
With intelligent, pre-approved templates, artwork can be generated without human intervention yet with full control and visibility of label components and content. This not only frees up capacity but reduces the chance of human error creeping in when updates are made.
We’re already seeing evidence of how this could enable companies in regulated sectors to gain a competitive advantage. In one case, a manufacturer was able to cut its product’s time to market by half, while also reducing the risk of human error.
Avoiding recalls at all might be the goal of any pharmaceuticals or medical device manufacturer – but change is a fact of life. Reformulations of drugs and new regulatory requirements are designed to improve patients’ quality of life and longevity.
Good labelling processes, that leverage proven AI and ML algorithms in a responsible way, can help prevent recalls due to labelling issues, and create a standardised workflow where updates are managed swiftly and accurately.
Operational improvements like these can remove a major barrier to efficiency, taking the average times it takes to produce labels from weeks to seconds.
While there is still work to be done to win the trust of decision-makers, we are already starting to see how the technology could reduce time to market and give manufacturers a competitive advantage.
Bob Tilling is a VP at Kallik. Go to kallik.com