Written evidence from Dr Emilia Vann Yaroson (MED0041)

Written evidence from Dr Emilia Vann Yaroson (MED0041)

I am a Lecturer and Researcher in Operations and Supply Chain Management at the Sheffield University Management School, University of Sheffield. My research focuses on the resilience and integrity of pharmaceutical supply chains, with particular emphasis on how advanced technologies such as artificial intelligence (AI), blockchain, and big data analytics can be used to address risks such as counterfeit medicines and drug shortages.

This submission draws upon findings from a research project partly sponsored by the University of Sheffield Management School on the use of AI in mitigating medicine shortages. These findings have not been widely disseminated and therefore offer original insights. My reason for submitting evidence is to provide research-informed perspectives on the systemic causes of UK medicine supply chain issues, current and emerging threats, and how technology—particularly AI—can strengthen monitoring and predictive capabilities. This evidence is offered to support policymakers in developing effective, resilient strategies for safeguarding medicine supply in the UK.

Medicine supply issues significantly influence patient care and undermine the efficacy of healthcare supply chains. Some of the underlying issues include understanding the causes, exploring future trends and identifying the best technologies that can mitigate the impact. These have been discussed in briefly in the sections below.

1. Causes of medicine supply chain issues

1. What are the causes of medicine supply chain issues in the UK?

Increased incidences of shortages in the UK are driven by multiple factors, typically framed as supply or demand issues and as anticipated or unanticipated events. However, such explanations represent only primary-level reporting and offer limited insight into deeper, systemic and indirect causes. Understanding supply chain issues requires viewing the pharmaceutical supply chain as a high-risk system, where the identified causes of shortages (limited manufacturers, strict regulations, complex production and delivery, increased demand, and tightly coupled subsystems) interact and propagate. It implies that a failure in one of these components can trigger shortages, with systemic risks extending beyond single events to contribute to globalised disruptions. It implies that a single shortage abroad could trigger a disruption in the UK. For instance, our analysed data showed that when shortages occurred in the US and the cause was an FDA failure, it took an average of 110 days to reach the UK. In the same vein, if a shortage occurred in Belgium, the likelihood of the same shortage occurring in the UK was about 72 days.

2. What are the current and likely future threats facing the UK medicine supply chain?

1. The UK PSC is characterised by its global interconnectedness, inherent vulnerabilities, and economic constraints, all of which influence the formulation of effective resilience strategies. Furthermore, the PSC operates under substantial pressure and must address various current and emerging threats, which can be classified into structural, geopolitical, technological, and environmental categories.

2. Structural challenges in pharmaceutical supply chains (PSC) refer to fundamental vulnerabilities caused by the design of the supply system. Short-term measures such as switching suppliers or emergency imports do not resolve these issues. Examples include high concentration and dependency, such as significant reliance on India and China for Active Pharmaceutical Ingredients (APIs); a limited number of manufacturers for certain critical medicines, with only one or two global producers for some products like PERT; and inflexible production processes due to the specific nature of pharmaceutical products and strict regulations. Just-in-Time (JIT) logistics can reduce resilience during demand surges. At the same time, the low profitability of essential generics results in limited investment in redundancy, innovation, and resilience. Addressing structural challenges in PSC involves strategic redesign, including enhancing domestic production and diversifying suppliers.

3. Geopolitical instability

Geopolitical instability, driven by international tensions, conflicts, and policy changes, disrupts supply chains by affecting the movement of goods. The UK medicine supply chain is particularly vulnerable due to its reliance on imported raw materials and products. Trade barriers like sanctions and export bans can restrict essential Active Pharmaceutical Ingredients (APIs), while political upheaval and conflict impact manufacturing, logistics, and energy supplies. Sourcing APIs from countries such as India and China increases risk from shifting trade relations. Brexit has further complicated procurement with new customs procedures and regulatory differences, leading to delays. Events like the COVID-19 pandemic saw export restrictions that limited UK access to medical supplies. Additionally, increased energy costs from crises like the Ukraine conflict have raised production expenses and introduced more volatility. These factors compound challenges in ensuring stable medicine provision in the UK.

4. Technology

Technology-related issues in the UK medicine supply chain arise from rapid digitalisation efforts and continued use of older systems. Numerous manufacturers, wholesalers, and NHS procurement operations continue to rely on legacy IT platforms that are not fully interoperable. This makes achieving real-time stock monitoring, accurate demand forecasting, and effective cross-border traceability challenging. Implementing advanced technologies such as artificial intelligence, blockchain, and Internet of Things (IoT) sensors varies across organisations, with smaller suppliers sometimes unable to invest due to higher costs or limited digital expertise. Cybersecurity presents an additional concern; Medicine supply chains are subject to risks such as cyberattacks, ransomware, and data breaches, which may impact operational continuity and data integrity. The overall goal of establishing a fully digitised and resilient UK medicines supply chain faces constraints, including inconsistent digital standards, limited infrastructure investment, and potential barriers for smaller entities during technological transitions. These factors related to technological adoption and cybersecurity contribute to vulnerabilities in maintaining consistent access to medicines.

5. How can monitoring the medicine supply chain and predicting potential supply chain issues be improved?

1. Artificial intelligence (AI) is increasingly being leveraged to predict medicine shortages. For instance, Pall et al. (2023) developed predictive models using historical data on frequently dispensed interchangeable medicines in Canadian pharmacies. These models achieved 69% accuracy in forecasting shortages one month in advance and could identify 59% of the most significant shortages. Similarly, companies such as Tracelink and LogicStream Health  (TraceLink, 2025) use AI analytics to notify hospitals and pharmacies of supply chain vulnerabilities, enabling proactive interventions based on inventory levels and fill rate disruptions.

2. In the UK, iethico aggregates data across the national health system, providing up-to-date information about medicines experiencing supply constraints via a platform where clinicians can report shortages. Iethico further assists in identifying potential risks before they escalate into critical shortages (iethico, 2025). However, the fragmented nature of healthcare data in the UK presents significant challenges for accurate forecasting. Consequently, it is recognised that AI alone cannot reliably predict medicine shortages, given the complex adaptive nature of the supply chain, multiple interdependencies, global sourcing of active pharmaceutical ingredients (APIs), and unpredictable demand patterns.

3. In our research lab, we have also tried to forecast drug shortages propagation across countries using machine learning algorithms and reported shortages from represented countries’ databases. We were able to fully classify where the next shortages would occur including the time difference at 88% accuracy. It implies that AI algorithms could be used to better manage supply chain issues.

4. However, global dependencies, unique NHS dynamics, and specific reimbursement strategies complicate the UK's pharmaceutical supply chain. Data fragmentation poses a notable limitation for AI-based prediction, as effective models require high-quality and timely information. In the UK, essential data is maintained separately by manufacturers, wholesalers, NHS trusts, and regulators. Moreover, commercial sensitivities and competition constrain real-time data sharing, thereby diminishing the predictive capabilities of AI.

5. We were also able to predict transition rates of medicines. Our analysed data provide valuable insights. For instance, we found that medicines such as Lormetazepam and Haloperidol medications for insomnia and anxiety respectively had high propagation rates across national boundaries. This trend may suggest that global high demand for a drug may lead to shortage possibly due to their therapeutic benefits or established treatment protocols. Drugs like Pegvisomant and Ribavirina also exhibited a significant frequency of transitions from countries where shortages first occurred to various countries, reinforcing the interdependence of global pharmaceutical supply chains.

6. Furthermore, AI models built on NHS prescribing data may not fully address global supply chain risks, such as interconnectedness of the supply chain export restrictions or shortages of raw materials, a key concern for the UK, which is highly reliant on imported APIs and generic medicines.

7.  Supporting technologies, including the Internet of Things (IoT), can provide real-time monitoring to help AI models track actual stock levels and offer a comprehensive view of nationwide supply and demand. Blockchain technology, serving as a secure ledger, enhances transparency by allowing supply chain participants to share data without compromising confidentiality, thus supporting the traceability of imported APIs, given the UK's dependence on global networks. This improved transparency increases AI accuracy and enables regulators to forecast shortages proactively. Digital twins, virtual replicas of the UK's pharmaceutical supply chain, facilitate simulation of its unique characteristics, supporting AI models in predicting how disruptions at individual points propagate throughout the network.

23 September 2025

References

Iethico, (2025) Med Intelligence. Connecting Supply and Demand for Ethical Medicine Access - Iethico. Accessed 21st September, 2025.

Pall, R., Gauthier, Y., Auer, S. et al. Predicting drug shortages using pharmacy data and machine learning. Health Care Manag Sci 26, 395–411 (2023). https://doi.org/10.1007/s10729-022-09627-y

TraceLink (2025) Product Availability Intelligence Product Availability Intelligence | Drug shortage intelligence | TraceLink. Accessed 21st September, 2025.