Has AI delivered on its promise in pharma supply chains?

This week's Expert View comes from Steve Brownett-Gale (pictured below), marketing lead at British pharmaceutical packaging firm Origin.

The use of AI across all industries and sectors has been heralded as driving efficiency and accelerating innovation, with some experts even citing the technology as the biggest disruptor of our lifetime.?

This is certainly true for big pharma, with generative AI expected to generate $60 billion to $110 billion annually for pharmaceutical sectors.

A year on from AI’s big breakout year, I ask whether AI truly lived up to its promise to improve pharmaceutical supply chains, and where are the areas for improvement.

Improved quality control

In high-stakes markets, such as the pharmaceutical industry, precision and reliability are crucial. Quality control stands as a cornerstone of the pharma supply chain, and with advancing technologies enabling businesses to conduct quality testing faster and more accurately, solely relying on manual inspection is no longer adequate.?

As a result, AI has become an integral part of quality control processes. Machine vision and natural language processing, in particular, have revolutionized the detection of defects and anomalies in pharmaceutical packaging, meaning AI systems are now surpassing the capabilities of traditional inspection methods.

Read on The Pharma Letter

Take, for example, Amgen’s integration of AI for addressing air bubbles in viscous injectables. The switch to AI-driven solutions has led to a 70% increase in particle detection rate and a 60 percent reduction in false detections. Not only has this saved Amgen time in accelerating quality control processes, helping to bring the syringes to market quicker, but also saving money in preventing product recalls and minimising waste.?

It's AI’s ability to address complex challenges that were previously insurmountable that makes it supremely efficient for quality control. But despite this, concerns still linger. There is always the risk of unintended consequences or biases creeping into AI’s decision-making processes since it is autonomous.?

There’s no doubt that pharmaceutical companies should invest in integrating AI technologies into their quality control systems to stay competitive and ensure product safety and compliance. However, it’s equally important they also establish robust governance frameworks and ethical guidelines to mitigate the risks associated with AI-driven decision-making processes.?

Brands must keep in mind that prioritizing the well-being of patients and adhering to regulations should first and foremost be the most important factor ahead of efficiency.?

Improved agility and supply chain transparency

Due to its globalized nature, it’s commonplace for pharmaceutical companies to have numerous stakeholders across different countries. Many outsource practices overseas, meaning varied regulatory requirements must be met. As a result, the pharmaceutical industry naturally has a very complex supply chain and, because of this, achieving full visibility and being able to track every step of processes is often difficult.

However, AI has significantly enhanced supply chain visibility by providing real-time insights and enabling proactive risk management.?

By leveraging advanced algorithms and predictive analytics, AI-powered systems can analyse vast amounts of data to identify potential bottlenecks, forecast demand fluctuations and optimize inventory management. This improved agility allows pharmaceutical companies to respond rapidly to changing market dynamics, minimise supply chain disruptions, and meet customer demands more effectively.

However, achieving end-to-end transparency remains a challenge thanks to data silos and difficulties in integrating systems. While AI facilitates greater visibility into certain aspects of the supply chain, disparate systems and fragmented data sources can hinder efforts to achieve comprehensive transparency.?

Personalised medicine packaging

One of the most exciting possibilities for AI in pharmaceuticals is personalised medicine packaging. From custom dosages to personalised instructions, AI-powered packaging could transform patient care, making medications more effective and easier to use.

However, we’re yet to see this become mainstream for a number of reasons. The biggest roadblock is the concern surrounding privacy.? AI needs access to sensitive patient information, raising questions about how to protect privacy while still delivering personalized solutions.?

And, from a regulatory standpoint, AI-driven packaging solutions would need to meet strict standards to ensure they're safe and effective. Navigating the rules and regulations can be tricky, especially as they vary from one place to another. Staying compliant with GDPR guidelines and other regulations is crucial for keeping patients safe and avoiding legal trouble.

Accelerated drug discovery and development

Not only is AI useful for pharma packaging design, but it has also proven to be a game-changer for drug discovery and development, particularly in the form of machine learning.

By crunching vast amounts of data on molecular structures, biological pathways and clinical trial results, AI algorithms can spot potential drug candidates quicker and with pinpoint accuracy through modelling, simulation and virtual screening.?

An example of this is Pfizer’s utilization of AI to accelerate the development of Paxlovid (nirmatrelvir/ritonavir), an antiviral treatment for COVID-19. In doing so, Pfizer reduced computation time by an impressive 80%-90%, significantly expediting the drug research process and bringing Paxlovid?to market in record time.

Yet, the challenge remains that AI algorithms rely on robust datasets for training. As mentioned earlier, biases or inaccuracies can make AI less reliable. However, regulatory bodies like the UK's MHRA are recognising AI's potential to streamline the approval process for cutting-edge healthcare tech. The recent launch of the AI-Airlock sandbox offers a platform for developers to test AI-driven healthcare innovations, promising faster regulatory approval and quicker access to ground-breaking therapies for patients.

This could serve as a valuable resource for collecting diverse and representative datasets for AI training purposes. By leveraging a wide range of data sources and scenarios, developers can ensure that AI algorithms are robust and resilient to potential biases or inaccuracies.

It’s worth keeping in mind that drug development is complex and requires human expertise and collaboration across disciplines. AI can crunch numbers and spot trends, but human intuition and domain knowledge are irreplaceable for interpreting results and making informed decisions.