Track & Trace Roundtable

 Track & Trace Roundtable
Darryl Peterson
Business Development Manager
Antares Vision Group


Eric Henefield
Director - Traceability Business
OMRON Automation Americas


Udit Singh
CEO of ACG



What can companies do to improve their Track and Trace capabilities? 

Peterson: Right now, it’s really about upgrading to aggregation, which will be mandated starting November 2023. Doing so early presents a sizable advantage for suppliers and vendors, especially when dealing with several major pharma companies. 

This is because, at this point, many of the most prominent pharma companies are strongly favoring aggregated orders; in fact, they’re even penalizing suppliers and vendors for not aggregating. For example, they may charge a separate fee for the additional supply chain hassle on their end. This goes to show that, with aggregation a done deal, companies are best served to incorporate it sooner rather than later. 

Henefield: In general, for any IIoT Digital Transformation journey, it is easiest to start at the plant floor level. Many manufacturing facilities have some level of automation (smart devices) within their factories. Take advantage of the data you have available and build from there. Specific to Track and Trace capabilities, the best option for improving transparency within the manufacturing process is to mark all assembly components with a unique serial number and read those component marks throughout the manufacturing process. This allows the user to incorporate a timestamp for each activity, which can then lead to connecting the production data with other key elements of their organization.

At Omron we refer to this methodology as MVRC – Mark/ Verify/Read/Communicate: 

  • Mark every component 
  • Verify the marks to be correct/robust (correct information and high quality to last the life of the finished assembly) 
  • Read every component as it moves through the manufacturing process (every assembly, every quality check, every system test...) • Communicate this information throughout your organizational ecosystem (Production, Quality, Procurement, Supply Chain...) 

Singh: To improve their track and trace capabilities, companies can integrate their business processes, such as manufacturing, packaging, supply chain and distribution into a single track and trace application. This gives visibility and analytics to each party or department in the supply chain. 

 

What is meant by “aggregation” and “serialization” in the Track and Trace space, and why are these concepts important? 

Peterson: Serialization, which has been mandated on the unit level for some time, entails assigning an individual tracking number to a bottle, cartoned blister pack, or other pharma product. 

Aggregation can be defined as the “next level up” from serialization, as it entails assigning a separate-yet-related marker to, for example, a shipping box that includes several units. This “parent-child” relationship allows companies to quickly identify what’s inside a shipped box without having to open and scan the individual items into inventory. 

As the next DSCSA deadline looms in November 2023, aggregation is becoming the biggest drumbeat in traceability. From what AVG is seeing, some of the companies who went with a unit-dose serialization solution to meet the initial DSCSA mandates are having mixed results with those same vendors when returning to them for aggregation solutions. 

It’s important to realize that the initial decision and actions to serialize was the toughest hurdle. Aggregation isn’t simple, of course, but it’s comparably less complicated than the initial push toward unit-level serialization, because manufacturers are basically building upon established infrastructure rather than starting from scratch. Still, obviously manufacturers who decided to incorporate aggregation along with the serialization infrastructure naturally have an advantage, as they’ve been aggregation-ready for a few years by now. 

Unsurprisingly, as a major traceability player, AVG has been involved in several projects outfitting aggregation lines for companies skeptical of reengaging their initial serialization provider. 

Henefield: Aggregation is just a fancy term for combining data from several different (often disparate) sources. Aggregating data from several different sources should provide a better overall view as opposed to viewing that same data in individual silos. Serialization is the idea of creating a unique identification (serial number) for every single component and finished good. Serialization allows for differentiation between like items in order to provide more detailed transparency. An example of serialization in action would be in the event of a recall within the automotive industry. If an issue is uncovered, the automakers would need visibility down to the component level in order to properly identify the impacted vehicles. 

Singh: Serialisation is where each product or its individual packaging is assigned a unique serial number to provide it with a unique identification in a series. 

When multiple serialised products are packed into a case, such as a box, the case is given a code to identify the individual packaged-level products inside. Then multiple cases are packed on pallets and the pallet is given a code that identifies the cases. This is aggregation, which builds a parent-child relationship between the individual packaged product, the cases and the pallets. This makes it possible to understand what is inside a case or pallet by scanning the outside—which is necessary in distribution-chain transactions such as shipping, dock-side checks, third-party logistics through to goods in, when the order is received by the customer. Serialisation and aggregation are important in providing visibility of products at every stage of the distribution chain. This traceability provides the manufacturer to respond quickly to problems, analyse root causes, facilitate continuous improvement, improve quality and reduce waste and the cost associated with spoilage. 


To what extent has the pandemic made any changes to inspection/track and trace practices if any? 

Peterson: There were some exceptions that were made during COVID that are unlikely to continue. For example, one company AVG dealt with was manufacturing a vaccine syringe pre-filled in a novel way, for single-dose use all over the world. They wanted to gain FDA approval, and proved that their process of filling and producing it were all in “one step”—in other words, it wasn’t segmented. This is important, as it limits foreign particle potential.

Due to the pandemic’s urgency, they were able to convince the FDA albeit temporarily, that the inspection process did not need to be as stringent (in fact, particle inspection was completely bypassed in this instance) given the circumstances. However, this was truly an extenuating circumstance, and as the pandemic wanes the FDA is unlikely to grant such exceptions.

Henefield: In general, the pandemic has forced accelerated changes in many manufacturing environments. Manufacturers are facing challenges from the workforce shortage, aging infrastructure, and growing global demand from the middle class. With fewer humans available to perform certain tasks, automation has definitely ramped up to help manufacturers maintain productivity. Within the pharmaceutical industry, inspection of product has evolved beyond the abilities of human vision. Small, sometimes microscopic anomalies, such as scratches and other imperfections are often undetectable with the human eye. On the topic of traceability, the pandemic has also accelerated the need for data utilization, data transparency and data sharing.

Traceability data can help connect the entire manufacturing ecosystem to create more efficient processes. As manufacturing facilities try to squeeze every ounce of efficiency out of their existing factories, the idea of traceability has become more of a necessity than ever. In general, if a manufacturer is not currently tracking their internal manufacturing processes (including every component, every quality check, every assembly process, every test) it would be difficult to achieve the highest levels of OEE (overall equipment effectiveness). 

Singh: If a manufacturer is relying on the manual inspection of their products, their inspection methods are limited. This increases the chance of rejecting an entire batch because of an anomaly in a single tablet or capsule from that batch. 

Likewise, there is also a higher chance of one defective table/capsule being released into the market and reaching the end consumer/patient. This could result in reputational damage, damage to the brand and ultimately revenue loss for the organisation. 

The pandemic has opened the opportunity for inspection technology to be adopted. Although it has been challenging, it has made it more evident that human inspection of the quality of products must be reduced to the bare minimum by removing processes such as hand-picking products to test products. It makes automated inspection a “must-have” to circumvent challenges like COVID-19. If there has been any hesitation around this automation, the experience of COVID reinforces its importance. 

Inspection technology performs at high speed, and in real-time, meaning samples do not have to be taken off the line to be tested. The real-time intelligence regarding rejections and verification guarantees the quality of the products in the batch and drives efficiencies for the manufacturer. 


Is AI playing a role (or could in the future) in inspection and/or track and trace? 

Peterson: Most definitely, and brand owners are looking at AI/deep learning as a true differentiator. It’s shown tremendous potential toward advancing visual inspection especially. As the technology gets more sophisticated, this effect will grow even more. 

One promise of AI has been reducing false reject rates. It’s already showing the potential to significantly cut down on false rejects, which gets more good product more efficiently into the supply chain. This is because as an algorithm accrues learning capabilities, it’s able to identify tolerances increasingly accurately. 

Notably, adjustments to AI/deep learning in pharma need to be targeted, since they often count as a production change and therefore need to be validated. So the role of AI/deep learning needs to be smartly applied for the right balance of efficiency and inspection improvement.

Henefield: AI is certainly playing a role in traceability. AI is used to analyze data throughout the entire manufacturing process in order to find improvement opportunities (eliminate bottlenecks, scrap reduction, increased productivity, predictive/prescriptive maintenance). AI can also help to identify trends that could lead to reduced quality or increased assembly time. 

Singh: Artificial intelligence algorithms are designed to make decisions, often using real-time data. Using sensors, digital data or remote inputs they combine information from various sources, analyse in real-time and act on the insights provided by the data. 

Utilising AI, manufacturers can currently access and provide highly sophisticated analysis and therefore adopt fast and accurate decision-making from the point of manufacture through the entire supply chain to the end customer. As AI relies on multi-sources of data, it can only be utilised in conjunction with digital inspection tools. 

It is anticipated that as AI becomes more widely understood, it will be adopted in many applications. 


What are the challenges to incorporating AI into inspection and Track and Trace? 

Peterson: AI in inspection can pay dividends by significantly reducing false rejects. This is because, as production continues and AI components get more familiar with a project’s parameters and products, they can develop a highly attuned sense of what is acceptable and what isn’t. Here, it’s the products in the “gray” area – the ones that show certain abnormalities but aren’t truly out of specification – that AI can successfully differentiate between. This element of hypersensitivity would typically elude manual inspection (especially at high speeds) and can greatly reduce false rejects. 

The challenge, however, is that it typically takes quite a while for AI to really take root, because these gray area issues need to play out over an extended period for AI to truly become optimally attuned to a particular inspection process. Also, validation can become a hurdle, since AI technically means changing a program/system and therefore mandates re-validation. That can be both expensive and time-consuming, so the best approach is to tightly specify AI’s role – putting guardrails around that process so that re-validations can be kept to a minimum and at smaller scales. 

In track & trace, AI is probably more aspirational at this point. But there’s potential there – for example, AI could potentially monitor for serialization trends and alert companies to problematic issues. 

Henefield: Artificial Intelligence is the ability for machines to make intelligent-like decisions. AI at its core is the ability to sort through data in order to make informed decisions. The more high-quality data you have, the better the chance to make a decision that creates a successful outcome. In most cases, the data will lead to direct correlations or early detection of trends, which are often not easily recognized by human analytics. 

I believe the biggest challenge to incorporating AI in general is the actual gathering of robust digital data. Once this digital data has been gathered and aggregated with other data, the decisions tend to come easier and the outcomes often present themselves more clearly. An AI system is only as good as the data it has to analyze. Gathering the correct data, at the correct time, and presenting this data to the correct consumers, is the secret to success. The best way to implement a data gathering solution on the factory floor is through MVRC track and trace techniques. 

Inspection has been using various forms of AI Machine Learning for many years. Most vision systems have the ability to detect anomalies that are beyond human capabilities. These vision systems have the ability to “learn” from previous image scans in order to become more effective in detecting anomalies over time. It is easy to imagine that combining additional data from other resources (smart devices, ambient plant conditions, supplied materials/ components, quality measurements) would lead to a better understanding of anomaly detection overall. Ideally, the AI system could potentially “learn” how to anticipate (and prevent) anomalies before they occur.

Singh: Although the pandemic has sped up the adoption of AI in the pharmaceutical market, the initial challenges include having a culture that recognises the need for AI, lack of data, lack of skilled people and difficulties identifying appropriate business cases. 

AI requires an intensity of integration to connect elements of software that will then communicate with each other - in cases where a customer relies on multiple vendor solutions to fulfil their business processes, the level of complexity to achieve integration increases. 

ACG’s solution to this has been to develop standardised plug-and-play modules that simplify the integration of multiple solution providers on a single platform, making it easier to implement AI in inspections and track and trace. 


What changes and improvements do you see coming to Track and Trace? 

Peterson: Getting much deeper into individual tracking of tablets and capsules – up to and including RFID tags and security elements in the ink and even the formulation – is driving the next stages of track & trace. The technology exists to get to the point where undetected counterfeiting is all but impossible. There’s even a “fingerprinting” technology that offers a pixel-by-pixel “before and after” of pills as they leave the line and upon return. This can help with both patient safety and legal issues. Edgyn, a recent Antares Vision Group acquisition, can perform this next-gen technique. 

Brand awareness also is becoming a subset of serialization. On top of individual tracking, QR codes are allowing companies to engage more intimately with consumers, as well as learn from them and, of course, about them. This is made possible, of course, since everyone has a “scanner” in their pockets in the form of smart phones. 

Henefield: The most significant changes we see coming for traceability is the utilization of data to make real-time production decisions. Track and trace is a great solution to help mitigate recalls, improve quality, and protect brand reputation. But this same traceability data could be utilized in many other ways. After the traceability (MVRC) implementation has been completed, manufacturers must leverage this data for more than simply storing in a database for reactive purposes. This same data can be instrumental in providing complete transparency to the overall manufacturing process. High quality data is vital to achieve predictive maintenance, reduced scrap, increased productivity, and improved OEE. One could argue that without detailed traceability manufacturing data, it would be difficult or impossible to achieve true manufacturing excellence. 

As mentioned previously, another great way to utilize traceability data is to share/combine with other data sources. Aggregating manufacturing data together with other data sources (such as MES, ERP, Procurement, Quality) can lead to new insights. However, data is often not easily shared between these disparate sources. This is referred to as the gap between IT and OT. Omron recognizes this gap as a challenge for manufacturing organizations and has developed some unique features to help bridge this gap between IT and OT. Omron has incorporated the ability to communicate a variety of data protocols to help better match the data with the intended source. Omron IIoT Gateways, and some Omron Smart Devices, have the ability to communicate information via OPC UA, MQTT, SQL, and other data protocols. Omron is able to achieve this data sharing without the use of any additional hardware or middleware required. This allows users to more easily connect data from various sources without having to reformat the source data. 

Singh: Regulations are emerging and changing, therefore traceability will become a necessity for businesses going forward. Manufacturers need to maintain their compliance with regulations according to their activity and markets they are servicing. 

Many manufacturers are looking to improve efficiency, supply chain visibility, customer engagement and their carbon footprint management. This has created an opportunity to include additional elements to integrate with track and trace solutions. These include processes such as overall equipment efficiency (OEE), integration with 3 PL partners, web-based customer engagement, goods tracking in the supply chain for complete supply through manufacturing to distribution visibility. 

Track and Trace requirements are also emerging from other markets such as FMCG, agriculture, production of alcohol and beverages and nutraceuticals. All Track and Trace applications produce data, however, data mining is required to generate usable analytics. The next era would be to use data to generate analytics, which can ultimately improve human life and healthcare. For example, with track and trace analysis, a food import company will be able to report on the effect of its goods on public health. 



References 

1. https://www.netapp.com/blog/ai-for-pharma/


Subscribe to our e-Newsletters
Stay up to date with the latest news, articles, and events. Plus, get special offers from Tablets & Capsules – all delivered right to your inbox! Sign up now!