Strengthening the EU’s Medical Supply Chains

3. New technologies and innovations to better monitor medical supply chains and forecast demand

Abstract

This chapter examines how new technologies and data-driven innovations can improve the monitoring of medical supply chains and support earlier detection of shortages of medical countermeasures. It reviews public track-and-trace frameworks in the European Union, the United States, Türkiye and Canada, alongside private sector initiatives using artificial intelligence, predictive analytics and other new technologies to monitor supply chains. These examples show that end-to‑end visibility is technically feasible but challenging to scale. The chapter argues that an effective public monitoring mechanism requires common standards, proportionate data sharing arrangements, trusted public-private governance and escalation mechanisms for public health emergencies.

Key findings

• Existing public track-and-trace systems provide useful traceability for medicines but were not designed for anticipatory shortage monitoring and do not cover all medical countermeasures. Public and legal frameworks in the EU, the United States, Türkiye and Canada primarily aim to prevent falsified medicines, verify product authenticity and support recalls. They can improve security of supply indirectly, but they focus on finished pharmaceutical products and provide limited visibility on upstream risks such as API and KSM dependencies, manufacturing capacity, supplier concentration, transport bottlenecks or demand surges.

• End-to‑end visibility is technically feasible but difficult to implement at scale. Public systems and private initiatives demonstrate that near real-time monitoring of pharmaceutical products, shipments and selected inputs is technically possible. However, scaling such systems across countries, products and actors requires more than technology. It depends on common standards, interoperable infrastructures, high-quality and granular data, proportionate reporting obligations, trusted governance and incentives for participation.

• Current systems remain fragmented and insufficiently interoperable. Despite growing convergence around global standards such as GS1 DataMatrix codes, existing public systems are not designed to exchange data internationally or to connect systematically with information on stocks, production capacity, international flows and demand. Effective monitoring would require stronger alignment between product identification, regulatory master data and supply-chain event reporting, including links between finished products and selected upstream inputs.

• Private‑sector innovations show what is possible, but do not provide a ready-made public model. Firms and logistics providers use AI, predictive analytics, Internet of Things (IoT) sensors, Radio-Frequency Identification (RFID), cloud platforms and other supply chain mapping tools to improve visibility and responsiveness. These experiences demonstrate operational feasibility, but private systems are usually designed for firm-specific purposes and may rely on proprietary platforms, commercially sensitive data and uneven incentives. They can inform public monitoring but cannot simply be transposed to a public preparedness system.

• Monitoring medical countermeasures requires a broader scope than medicines alone. Preparedness for health emergencies requires monitoring of selected medicines, vaccines, diagnostics, medical devices, personal protective equipment and critical inputs. The EU List of Critical Medicines provides a useful starting point, but MCM monitoring should also draw on DG HERA’s prioritisation of health threats and the forthcoming list of priority MCMs. Most existing monitoring mechanisms and shortage instruments are focussed on critical medicines and there is a gap to cover the full range of MCMs.

• Data sharing requires both trust and legal obligations. Voluntary co‑operation can improve data quality where firms receive clear benefits and confidential business information is protected. However, for critical MCMs and crisis situations, public health objectives can justify proportionate mandatory reporting and escalation mechanisms.

3.1. Introduction

Efforts to secure medical supply chains have accelerated globally in recent years, driven by concerns over shortages of critical medicines and medical devices, particularly in the context of health crises requiring medical countermeasures. To assess risks and prevent shortages, countries have implemented a mix of policy measures. An important policy response consists in enhancing the visibility of medical supply chains and in harnessing information across the whole supply chain on participants, flows, stocks and shortages (OECD, 2024[1]). While firms closely monitor their supply chains to ensure the continuity of supply and to match demand, governments have also started their own information collection to assess vulnerabilities and to manage shortages.

For example, several OECD countries have created national registers to monitor shortages or require MAHs to notify supply disruptions in advance (Chapman, Dedet and Lopert, 2022[2]). In the European Union, the European Shortages Monitoring Platform (ESMP) is used by the European Medicines Agency (EMA) to detect and manage shortages. Information on the availability, supply and demand of medicines is collected in situations that require close monitoring, such as crises or preparedness actions (EMA, 2026[3]).

Supply and demand of medicines is also more systematically monitored in the context of track-and-trace systems designed to fight falsification and counterfeiting.1 Türkiye’s İlaç Takip Sistemi (ITS), launched in 2010, was the world’s first national pharmaceutical track-and-trace covering every medicine from production/import to dispensation. Notably, Turkey’s system evolved from a point-of-dispense verification model to full end-to‑end traceability by 2012. In the United States, the Drug Supply Chain Security Act (DSCSA) of 2013 similarly mandated an interoperable electronic track-and-trace system, rolling out item-level serialisation and full supply chain traceability by 2023.

In the European Union, the Falsified Medicines Directive (FMD) established in 2011 the legal framework to introduce mandatory serialisation and an EU-wide verification system to ensure medicine authenticity. These public frameworks demonstrate that end-to‑end visibility is technically feasible (at least for finished pharmaceutical products) at a national or regional scale. However, they were not designed to address shortages.

In parallel, private‑sector initiatives are piloting innovative tools to monitor supply chains and forecast demand. For example, A.P. Møller-Maersk and IBM’s TradeLens platform sought to digitise global trade logistics using blockchain, creating tamper-proof shared records of shipments (e.g. bills of lading, customs clearance) to improve transparency and speed. While TradeLens attracted dozens of partners, it ultimately failed to attain the critical mass of industry adoption required and was discontinued in 2023 – underscoring that even a viable platform cannot succeed without broad ecosystem participation and trust.

Pharmaceutical companies have also leveraged big data and AI for demand forecasting and supply chain resilience. Johnson & Johnson and Pfizer, for instance, use real-time analytics and predictive modelling to anticipate production adjustments and potential shortages before they occur. Global logistics providers such as DHL have introduced RFID smart tags to track pharmaceutical products. Private initiatives are not limited to firms involved in the production or distribution of pharmaceutical products and medical devices. The U.S. Pharmacopeia (USP), a non-profit organisation involved in standards setting, has also developed an AI-driven platform to map medical supply chains and to assess risks of shortages in the United States, extending the mapping to APIs and key starting materials (KSMs).

These private innovations highlight the potential of digital technology to improve supply chain monitoring and resilience (Box 3.1). They also demonstrate, however, that scaling such solutions to an entire industry requires overcoming major challenges in interoperability, data-sharing, and governance. There is also a cost associated with enhanced monitoring, leading to trade‑offs with the expected benefits in terms of resilience and security of supply.

Building on these experiences, this chapter examines the challenges to developing a public monitoring mechanism and early warning system for critical medical products and medical countermeasures in the EU. It assumes an ambition to move beyond existing mechanisms towards a more integrated framework that can track critical medicines and medical devices (and possibly key intermediate inputs and raw materials) across borders in near real-time2 and predict or detect shortages.

To investigate the potential of new technologies in improving medical supply chain monitoring and demand forecasting, existing tools and current legislative frameworks intended to trace finished products are explored, together with a selection of private initiatives employing innovative approaches. The objective is to provide information on solutions that enable end-to‑end visibility of the supply chain and early detection of disruptions.

The chapter combines a qualitative literature review with stakeholder interviews from both the public and private sectors. Based on this research, it presents tables summarising the characteristics of each initiative, allowing readers to compare relevant information at a glance. Detailed descriptions of each initiative are included in the subsequent sub-sections. The chapter then provides a discussion of key challenges in establishing an EU early warning system for shortages of MCMs that could not only monitor pharmaceutical and medical device supply chains within the EU but also incorporate information from third countries or be part of a global monitoring mechanism. The chapter concludes with policy recommendations and concrete steps for the EU and DG HERA to move towards better monitoring mechanisms, focussing on international and public-private co‑operation.

Box 3.1. Digital technologies transforming supply chains

Businesses and policymakers are increasingly looking to artificial intelligence (AI) and other digital technologies to strengthen supply chain efficiency and resilience. This box outlines how digital technologies can contribute to more transparent and resilient supply chains, with a focus on applications that improve visibility, enable early disruption detection, and enhance demand forecasting.

Machine learning and predictive analytics

Machine learning (ML) algorithms can analyse vast datasets to find patterns and make predictions. In supply chains, ML-driven predictive analytics improve decision making in planning and operations. Companies report that AI adds value to production planning, inventory management, and distribution by processing real-time data and improving the accuracy of demand forecasts (McKinsey, 2022[4]). ML models enable predictive demand forecasting, dynamic inventory optimisation, and proactive maintenance of equipment to reduce downtime. By learning from historical trends and external factors, AI predictive tools help firms anticipate market changes and adjust supply chain plans accordingly.

Computer vision and IoT sensors

Computer vision (CV) applies AI to interpret visual data (images and video), offering powerful capabilities in logistics and trade. In warehouses and factories, CV systems can automatically detect product defects or packing errors, monitor worker safety, and guide autonomous robots. At border checkpoints and customs, AI-driven image analysis of X-ray scans can identify anomalies or contraband in shipments far faster than manual inspection. One AI model developed for cargo screening was 98% accurate in spotting prohibited items on X-ray images (Miller, 2025[5]). Coupled with the Internet of Things (IoT) – networks of sensors on containers, vehicles, and infrastructure – AI vision and sensor analytics provide real-time data on the location, condition, and status of goods. These technologies greatly enhance situational awareness, enabling a live view of supply chain operations that was previously impossible.

Digital twins and simulation

A digital twin is a virtual replica of a physical supply chain (or its parts) that allows companies to simulate scenarios and stress-test their operations. By integrating real-time data and predictive AI into these simulations, organisations can create self-updating models of their supply networks. Digital twins enable predictive and prescriptive analytics, effectively becoming “self-healing” supply chain systems that can anticipate problems and recommend solutions. For instance, a retailer using a detailed digital twin of its distribution network can dynamically adjust inventory and production in response to localised demand fluctuations, optimising the whole chain. In one case, such granular digital twin modelling improved on-time delivery performance by up to 20%, while cutting logistics costs and labour needs (McKinsey, 2024[6]).

Natural language processing and automation

AI is also transforming the information flows in supply chains through natural language processing (NLP) and intelligent automation. NLP techniques allow AI systems to read and extract data from unstructured documents (such as invoices, purchase orders, or customs declarations) which often come in various formats and languages. By automating document processing, AI tools accelerate administrative tasks in trade. For example, AI can auto-complete customs forms or classify products under Harmonised System codes, reducing errors and delays in cross-border shipments (Klotz, Barr and Sotelo, 2024[7]). Chatbots and machine translation tools driven by AI further facilitate global trade by enabling multilingual communication between suppliers, logistics providers and customers. These technologies free up human workers from repetitive paperwork and help standardise compliance, contributing to faster and more efficient trade facilitation. Sentiment analysis (i.e. the identification of sentiments, emotions and connotations from texts) can also be used on medicine‑related content disseminated by news media to improve demand forecasting (Nguyen et al., 2022[8]).

Key applications and benefits

Digital applications in supply chain management translate these technologies into concrete benefits for resilience:

• Improved end-to‑end visibility: digital tools have the potential to increase the visibility of products and shipments as they move across the supply chain. By analysing data from IoT sensors and logistics systems in real time, AI platforms can provide a “control tower” view of supply chains. This enhanced visibility lets firms trace goods in transit and monitor inventory levels at each stage. With better transparency, stakeholders can ensure compliance with standards (e.g. verifying a product’s origin) and respond faster to any issue, ultimately facilitating smoother trade flows.

• Early detection of disruptions: One of AI’s most powerful contributions is the early warning of supply chain disruptions. AI-powered platforms today analyse data from shipments and suppliers to identify deviations from normal patterns – detecting issues before they escalate. In practice, this means a company can get advance notice of a likely delay or shortage and activate contingency plans (such as switching to an alternate supplier or rerouting a shipment).

• Better demand forecasting: Anticipating variations in demand is a perennial challenge in trade and logistics. AI-driven predictive analytics greatly improve demand forecasting by processing far more data (historical sales, market trends, even online consumer sentiment) than traditional methods. This leads to more accurate predictions of what products will be needed, where, and when. Moreover, AI can continuously adjust forecasts as new data comes in and then recommend production or ordering changes in response.

• Logistics optimisation and reduced frictions: digital technologies are streamlining physical logistics by optimising routes, schedules, and resource allocation. In transportation, AI algorithms can evaluate factors like traffic, weather, fuel costs, and delivery constraints to find the most efficient routing for shipments. This dynamic route optimisation saves time, helping carriers and freight forwarders avoid delays. AI is also powering smarter warehouse and port operations. All these improvements translate to lower frictions in trade logistics: goods flow faster and more predictably from origin to destination.

• Trade facilitation and customs efficiency: AI’s impact extends to border agencies and customs processes, which are key to international trade facilitation (OECD, 2025[9]). Customs authorities are increasingly adopting AI to manage growing volumes of trade and complex regulations. Over 130 customs organisations worldwide have begun implementing automated or AI-enhanced systems to expedite border clearance (Klotz, Barr and Sotelo, 2024[7]). These systems use AI for risk analysis – for example, analysing import data to target high-risk shipments for inspection while allowing low-risk cargo to pass swiftly. AI can also automate document review and compliance checks. All these applications mean faster processing at borders and fewer delays for legitimate trade.

Source: Based on a review of the literature; Cohen and Tang (2024[10]); Ferencz, González and Oliván García (2022[11]).

3.2. Public frameworks

This section reviews four public frameworks implemented across different jurisdictions to understand the mechanisms and technologies they employ to monitor medical supply chains. These are the EU Falsified Medicines Directive (FMD),3 the US Drug Supply Chain Security Act (DSCSA), Türkiye’s Pharmaceutical Track and Trace System (ITS), and Canada’s DataMatrix initiative. These four frameworks were chosen as they already enable the traceability of finished medical products. They represent different approaches, i.e. centralised versus decentralised systems, top-down versus bottom-up approaches. This allows the analysis to be based on the understanding of what is already possible under current legislation, agreements, and technology deployments. To facilitate comparison, the table below (Table 3.1) summarises the key features of each framework across 11 criteria. These criteria have been developed through a literature review and semi-structured interviews with public stakeholders.

The lessons learned from the review of these public frameworks are the following. First, they were not designed to monitor shortages or inform policymakers and regulators about vulnerabilities in supply chains or potential disruptions. While they can contribute to security of supply, their main objective is the traceability of medicines for patient safety and the identification and recall of counterfeit or falsified medicines. All the track-and-trace systems reviewed focus on finished pharmaceutical products, in the distribution chains. They exclude APIs and KSMs, as well as medical devices (with the exception of Canada where the inclusion of some medical devices is optional). They focus on all medicines or a sub-group of medicines (e.g. prescribed medicines).

Second, there are three types of public track-and-trace systems:

• The EU system is a “point-of-dispense check” where finished products are only scanned at the beginning and at the end of the value chain, i.e. when the medicine is manufactured and when it is dispensed. The system can be described as ensuring “end-to‑end visibility” in the sense that the dispensed medicine can be traced back to its manufacturer, but it does not provide full visibility through the entire distribution network.

• The US and Turkish systems provide “full” track-and-trace allowing “real-time” tracking along each stage of distribution. However, having to scan products all along the distribution chain creates challenges for manufacturers, wholesalers and pharmacies and the data management can be either centralised (Türkiye) or decentralised (the United States).

• A third type of public system, as illustrated with Canada, can be fully decentralised and left to stakeholders with the government only participating in setting the standards (the creation of a unique product identifier) with the management of the track-and-trace left to manufacturers, distributors and pharmacies.

The third main lesson from the analysis is that while global standards seem to emerge (such as GS1 2D matrix codes4), the public frameworks put in place are not interoperable and not designed to exchange information across countries. Consistent with their focus on finished products, they organise a track-and-trace for the medicines dispensed in their jurisdiction and are not designed to monitor global supply chains and flows of APIs and KSMs across countries. As such, these public frameworks cannot be regarded as first steps towards a global monitoring or early warning system. Their analysis is nonetheless useful to understand what such monitoring would involve and how a decentralised system building on existing track-and-trace infrastructures could be designed.

Table 3.1. Summary of public frameworks reviewed

Criteria	EU – FMD	United States – DSCSA	Türkiye – ITS	Canada
Purpose	Prevention of falsified medicines entering legitimate supply chains; verification of authenticity; comprehensive legal framework	Counterfeit prevention; identification and tracing of medicines; recall control	Counterfeit and parallel trade prevention; recall and reimbursement control; prevent shortages	Patient safety; traceability; forecasting; recall management
Incentives	Compliance incentive	Compliance incentive	Compliance incentive; reimbursement	Regulatory alignment; recall facilitation
Supply Chain Comprehensiveness	Post-production tracking; excludes APIs and KSMs	Post-production tracking; excludes APIs and KSMs	Post-production tracking; excludes APIs and KSMs	Post-production tracking; excludes APIs and KSMs
Product Coverage	Prescription medicines; optional for high-risk over-the‑counter	Prescription medicines	All medicines (prescribed and over-the‑counter)	All medicines (prescribed and over-the‑counter); optional for medical devices
Product Identification	Unique product identifier; 2D DataMatrix with serial number, expiry and lot number	Unique product identifier; 2D DataMatrix with NDC; serial number; lot; expiry; GLN and EPCIS encouraged	Unique product identifier; GS1 2D DataMatrix with GTIN; serial number; lot; expiry; RFID	Unique product identifier; GS1 2D DataMatrix with GTIN; serial number (optional); lot; expiry
Product Tracing	Partial track-and-trace; “book-end” approach	End-to‑end track-and-trace with verification of all transactions	End-to‑end track-and-trace through RFID; real-time visibility	End-to‑end track-and-trace
Product Verification	Verification against national databases, barcode decommissioning at dispense	Verification against FDA-compliant source, transaction statement	Verification against web portal, real-time scan at dispense	Verification against stakeholder-managed databases
Data architecture	EU-harmonised hub with national databases; EMVO governs EU-hub; nonprofits manage national databases	No centralised database; stakeholder managed	National database, governed by Turkish Medicines and Medical Devices Agency (TİTCK)	No centralised database, stakeholder managed
Interoperability	GS1 based, ISO16022 compliant	Stakeholders responsible for implementation; GS1, GLN and EPCIS recommended	GS1 based; EU FMD-compatible	GS1 based; EU FMD and US DSCSA-compatible, ISO16022 compliant
Implementation Cost	Borne by industry; moderate to high investment	Borne by industry; exemptions for small stakeholders; high compliance costs for serialisation, scanning, IT systems	Government-backed with industry heavily invested in tech infrastructure and software	Borne by industry; costs unknown due to ongoing implementation
Alert System	Automatic alerts on falsified medicines	Decentralised alerts on falsified medicines generated at stakeholders	Automatic alerts on falsified medicines	Forecasting; predictive analytics; traceability pilots ongoing

Notes: APIs = Active Pharmaceutical Ingredients; EMVO = European Medicines Verification Organisation; EPCIS = Electronic Product Code Information Services; GLN = Global Location Number; GTIN = Global Trade Item Number; KSMs = Key Starting Materials; NDC = National Drug Code; RFID = Radio Frequency Identification. See Box 3.4 for a description of GS1 and ISO standards.

Source: Based on interviews and literature review.

3.2.1. EU Falsified Medicines Directive (FMD)

The EU established the Falsified Medicines Directive (FMD) in 2011 to prevent falsified medicines from entering the legitimate supply chain (Vajda et al., 2021[12]). It introduced a medicine authentication and verification system to enhance supply chain security (European Commission, 2024[13]).

The FMD replaced previous national systems in some EU member states. As Belgium, Greece and Italy had previously introduced national serialisation systems that permitted the validation and tracking of medicines, these countries were granted extensions for the implementation of the directive, though only Greece and Italy made use of the extension. Box 3.2 provides a short overview of these legacy systems. Since February 2025, all EU member states apply the unified verification system.

Box 3.2. Legacy national verification systems in Belgium, Greece and Italy

Belgium

In 2004, Belgium introduced the serialisation of all reimbursed medicinal products. The purpose was to avoid double billing of insurers and to better control the supply through pharmacies (Association Pharmaceutique Belge, 2018[14]). Each retail package of reimbursable prescription medication was equipped with a unique CNK code. The CNK code consisted of a linear bar code, encoding a six‑ or seven‑digit number, as well as the printed number (Honeywell, 2009[15]). Manufacturers were obliged to submit the code to the authorities, while pharmacists submitted the code when dispensing the medication. Despite the transition to the 2D DataMatrix under the FMD, the CNK code remains relevant for reimbursement (Association Pharmaceutique Belge, 2018[14]).

Greece

Since 1989, Greece has been using physical authenticity labels on medication to prevent counterfeit medications as well as enable insurance reimbursement. In 2004, a ministerial decision introduced the mandatory usage of a linear barcode in the EAN format and a unique identification number. The unique identification number is based on a 12‑digit serial number (Pagonidis, Sapuric and Lois, 2020[16]). The barcode was placed as a sticker on the package and assigned by the National Organization for Medicines (Greece, 2004[17]). Since the implementation of the EU FMD in February 2025, the authenticity labels have been phased out.

Italy

In 2000, Italy introduced the so-called Bollino system to prevent fraud against the public health system. Bollino refers to the unique identifier that was initially introduced for each reimbursed medicinal product but extended to all medicines in 2002 (Italian Medicines Agency, 2020[18]). A special sticker containing a serial number and a linear bar code was placed on each unit of sale to increase the traceability of prescription medications. It was further mandatory to record and archive each serial number (Pagonidis, Sapuric and Lois, 2020[16]). In 2014, the usage of a 2D DataMatrix was introduced (Italian Medicines Agency, 2020[18]).

Following the adoption of the Commission Delegated Regulation (EU) 2016/61, all prescription medicines5 must contain a 2D DataMatrix that serves as a unique identifier.6 In addition, each package must be fitted with an anti-tampering device (Merks et al., 2022[19]). Manufacturers must print the DataMatrix on the outer packaging and upload the data of the unique identifier into the EU hub from which it is subsequently redistributed into the respective national repository. Wholesalers may scan the DataMatrix to verify the authentication based on risk (European Commission, 2024[13]). Medicines returned from pharmacies and medicines that have a heightened risk of falsification need to be verified systematically by the wholesaler. Pharmacies are required to scan each package upon dispensing it to verify the authenticity and thereby decommission it from the repository (Bouvy and Rotaru, 2021[20]). If the verification fails, an automatic alert is issued (European Commission, 2024[13]).

National repositories allow product verification. They are established by non-profit organisations and are connected to a central EU hub run by the European Medicines Verification Organisation (EMVO) to ensure data exchange. The data uploaded by the manufacturer and the information on whether the product has been decommissioned is accessible to all stakeholders while other dynamic data is restricted in access. The National Competent Authority (NCA) can access the repository without limitations to investigate possible falsifications (Bouvy and Rotaru, 2021[20]).

Implementation cost was borne by industry. Manufacturers adapting packaging lines and investing in IT systems faced additional costs of EUR 0.016 per pack. Wholesalers required sophisticated scanning equipment, which was estimated at EUR 1 200 each, while pharmacists standard scanning equipment was estimated at EUR 250‑300 (EAEPC, EFPIA, GIRP, PGEU, 2012[21]). Pharmacists further reported that verifying products increased their workload (Vajda et al., 2021[12]).

As previously emphasised, the FMD was not designed to monitor shortages. In the EU, the ESMP (under EMA) is the platform for preventing and managing shortages of medicines. Section 3.5.2 below addresses the question of whether information collected under the FMD or its existing infrastructure could be used for monitoring purposes in the context of health emergency preparedness.

3.2.2. The US Drug Supply Chain Security Act (DSCSA)

Prompted by incidents of counterfeit and falsified drugs infiltrating legitimate supply channels, the Drug Supply Chain Security Act (DSCSA) was enacted by the United States in 2013. The Act seeks to prevent counterfeit, adulterated, misbranded, or otherwise illegitimate medicines from entering the pharmaceutical supply chain. Additionally, it aims to improve the detection and efficient removal of compromised products, facilitate rapid response and recall procedures, and ensure greater transparency and accountability among supply actors.

The DSCSA establishes a comprehensive system for tracking pharmaceutical products (prescription drugs) across the supply chain (U.S. Food and Drug Administration, 2025[22]). Central to this framework is the serialisation of drug packages, where each package carries a unique identifier consisting of a National Drug Code (NDC), serial number, lot number and expiration date. The Act mandates an electronic, interoperable traceability system, facilitating the exchange of critical transaction information, transaction history, and transaction statements among trading partners. This system provides a detailed record of each transaction from (final) production through dispensing.

Manufacturers, wholesale distributors, dispensers and repackagers must implement systematic verification measures to confirm the authenticity and legitimacy of medicines. These measures include validation of transaction documentation and procedures to identify and quarantine suspect or illegitimate products promptly.

The DSCSA implementation has been structured in phases over a ten‑year period (2013-2023), providing stakeholders time to adapt to compliance requirements. The initial implementation stages concentrated on lot-level tracing, eventually progressing to full serialisation and electronic interoperability by November 2023. Throughout this phased approach, regulatory oversight from the Food and Drug Administration (FDA) has been critical in guiding stakeholders toward compliance through inspections, issuing guiding documents and engaging with stakeholders to resolve implementation challenges. In particular, the FDA delayed the enforcement of the verification of prescription drugs at the smallest unit level (i.e. a package) until November 2024 (Kannarkat, Denham and Sarpatwari, 2024[23]). In October 2024, while acknowledging progress made in the implementation of the DSCSA requirements, the FDA further extended the period before compliance becomes mandatory.7

Stakeholders across the pharmaceutical supply chain have responded proactively, investing significantly in technological infrastructure upgrades, staff training, and cross-sector collaboration to ensure compliance with DSCSA mandates. An independent public-private platform was established to promote co‑operation (the ‘The “Partnership for DSCSA Governance”). One characteristic of the US track-and-trace system is that the adoption of standards such as GS1, GLN and EPCIS (see Box 3.4 for a description) is recommended but not compulsory. Stakeholders are free to select the technical solutions they implement as long as they fulfil their commitment for interoperability. This approach was useful to decrease implementation costs and to allow all stakeholders to be part of an ambitious tracking system, but it also created challenges. The industry acknowledges the need to enhance standardisation and interoperability (Partnership for DSCSA Governance, 2025[24]).

While the DSCSA can advance the security of the US pharmaceutical supply chain by combating increasingly sophisticated counterfeit activities, there is also the question of how the infrastructure could be leveraged to address drug shortages and to bolster overall supply chain resilience. As part of the “DSCSA Pilot Project Program” (established in February 2019), this question was addressed by different stakeholders, highlighting that the information collected can also be used to optimise supply chains. The pilot projects focussed more on the technical side of the implementation of DSCSA’s interoperability requirements, but they also included proposals for new standards for data exchange across companies (including through blockchain technologies) that provide useful references for the design of monitoring mechanisms going beyond the verification of medicines.8

3.2.3. Türkiye’s pharmaceutical track-and-trace system (ITS)

Operational since 2012, Türkiye’s İlaç Takip Sistemi (ITS) is the world’s first national pharmaceutical track-and-trace system for pharmaceuticals covering every medicine (prescribed and over the counter) from production/import to dispensation (Parmaksiz, Pisani and Kok, 2020[25]). The system delivers real-time, end-to‑end traceability for all packaged pharmaceuticals, tracking over 2 billion product movements every year for all finished medicines to ensure each drug’s authenticity, origin, and legal status (Altunkan et al., 2012[26]). By connecting reimbursement approvals to real-time scanning, Türkiye aims to block fraudulent claims, prevent parallel trade and ensure compliance.

Although the system is managed centrally by the Turkish Medicines and Medical Devices Agency (TİTCK) and covers all the key stakeholders like manufacturers, wholesalers, pharmacies, hospitals, and reimbursement agencies like the Turkish Social Security Institution (SGK), the majority of the implementation costs were borne by the industry (Bulut and Bilgener, 2022[27]; Parmaksiz, Pisani and Kok, 2020[25]). Whereas the government led the system’s design (but mostly outsourced its development), manufacturers were responsible for major investments in serialisation infrastructure, wholesalers for adapting logistics systems, and pharmacies for acquiring 2D barcode scanners (Parmaksiz, Pisani and Kok, 2020[25]).

ITS uses GS1‑compliant 2D DataMatrix codes that contain a product’s Global Trade Item Number (GTIN), serial number (SN), expiration date, and batch number (Yorulmaz et al., 2012[28]). These elements allow the system to uniquely identify and monitor each unit. Web services ensure interaction between the stakeholder systems and the central database. Notably, ITS does not allow the transfer of any pharmaceutical product without a valid code. A complementary Package Transfer System (PTS) ensures the grouping of product units under a single Serial Shipping Container Code (SSCC), which streamlines logistics and warehouse operations (Bulut and Bilgener, 2022[27]).

The system has been designed as an interoperable tool that aligns with international standards such as GS1. Both Türkiye’s ITS and the EU FMD therefore share fundamental technological principles, including the use of GS1 standards, 2D DataMatrix encoding, and serialisation using GTIN, batch, expiry, and serial number data. One key difference, however, lies in the governance of the initiatives. Whereas ITS is a centralised system under the control of a national authority (TİTCK), the EU operates a decentralised model with each member state having its own national repository linked to the central EMVO hub. In addition, the Turkish system goes one step further by providing full end-to‑end tracking from manufacturing to dispensation, rather than point-of-sale verification only. The centralised database also allows for instant recalls and blocks the movement of recalled or expired products in real time.

3.2.4. Canada’s community-led DataMatrix initiative

Canada does not have a system in place to track and trace medicines along the supply chain (Safemedicines, 2021[29]). However, a community-led initiative aims to align Canada’s pharmaceutical supply chains with global standards (GS1 Canada, 2025[30]). Contrary to the top-down approach of the EU, the United States and Türkiye, Canada is following a bottom-up approach. Industry stakeholders came together to move towards a comprehensive medicines track-and-trace system. The community-led initiative is therefore included in the analysis as this diverging approach helps to understand different implementation modalities.

The Canadian health sector seeks alignment. A healthcare community project, supported by numerous Canadian and international manufacturers, pharmacists and hospital associations, and purchasing organisations, has jointly decided to introduce 2D DataMatrix codes for medicines. The Public Health Agency of Canada has endorsed the project. As Canada exports more than 50% of its pharmaceutical production and imports 78% from the United States and the EU, aligning its products and packaging systems is advantageous. Utilising the GS1 standard would align Canada with the EU, the United States and India (GS1 Canada, 2025[31]).

The 2D DataMatrix printed on the primary and secondary packaging allows for easy identification of products. Each product is assigned a unique GTIN, comprising the company prefix, the item reference number, and a check digit. In addition, the lot number and expiry date of the product are contained in the matrix. Inclusion of the serial number is optional (Institute for Safe Medication Practices Canada, 2012[32]).

The DataMatrix improves patient care and supply chain planning. One core reason for introducing a new tracking system is to improve patient safety. Through the DataMatrix, medications can be traced from the manufacturer to the patient, which shall reduce medical errors. Further, scanning the DataMatrix along the supply chain will provide visibility of the available inventory and therefore will also permit some forecasting analysis that may be utilised to mitigate shortages (GS1 Canada, 2025[31]).

Implementation was to be completed by the end of 2025. Manufacturers and distributors had to adapt their systems to the DataMatrix by the end of 2023. Pharmacies (retail and hospital) had until end of 2025 to invest in scanning equipment and prepare their database management (GS1 Canada, 2025[31]). The system should become fully operational by the end of 2027, when all retailers should be ready to scan the 2D barcodes at each point of sale.

3.3. Private initiatives

Following a similar approach and methodology as applied to the analysis of public frameworks, this section reviews five private initiatives to improve the monitoring of medical supply chains: Maersk’s blockchain-enabled shipping solution developed in partnership with IBM (TradeLens), Pfizer’s use of big data to optimise supply chain operations and forecast demand, Johnson & Johnson’s automated supply chains, DHL’s application of Radio Frequency Identification (RFID) tags and smart labels, and the US Pharmacopeia’s Medicine Supply Map, which is a drug shortage intelligence platform. To facilitate comparison, the table below (Table 3.2) summarises the key features of each initiative across 11 criteria.

While the initiatives target the companies’ individual needs9 and have different objectives, common factors can be observed in terms of monitoring supply chains. First, across all cases, these initiatives leverage the digital technologies outlined in Box 3.1. These include IoT sensors and RFID technologies for real-time tracking of location and environmental conditions, cloud-based data architectures enabling multi‑actor data access and secure information exchange, machine learning and predictive analytics for demand forecasting, anomaly detection and scenario modelling, digital twins and control towers to simulate disruptions and optimise supply-chain responses, or blockchain pilots to ensure tamper-proof traceability and enhance trust across actors. Importantly, these technologies are not used in isolation. They are often combined to ensure continuous monitoring, automated alerts and predictive decision support.

A second common feature is the emphasis on interoperability. Private actors operate across jurisdictions and regulatory environments. Their systems must therefore connect heterogeneous IT infrastructures and data standards. Interoperability is achieved through reliance on international standards (notably GS1 identifiers and EPCIS formats), use of cloud connectors and data solutions for supply chain information (see Box 3.4), modular architectures that allow integration with regulatory platforms (e.g. DSCSA or FMD-compliant serialisation systems) and data-sharing models that protect commercially sensitive information. The private experience demonstrates that multi‑actor, cross-border data integration is operationally achievable, provided common standards and governance arrangements are in place. This insight is reinforced by the fact that in the case of TradeLens, the platform had to be discontinued due to the reluctance of some stakeholders to share sensitive business information and the perception that the benefits were not equally distributed among participants. Without strong incentives or legal obligations and adequate governance, it is challenging to establish long-term co‑operation.

Lastly, one important lesson is that (near) real-time monitoring across the full supply chain (production, distribution and in some cases upstream inputs) is already implemented at scale in the private sector. Pharmaceutical companies have established “control towers” combining production data, shipment tracking and demand forecasts across global networks. Logistics providers operate real-time monitoring of multimodal shipments, and some platforms extend visibility upstream to APIs and KSMs. Private initiatives do not provide a blueprint for public monitoring; their objectives are firm-specific and they are part of the management and optimisation of supply chains by the companies that operate them. However, they highlight that supply chain visibility is technologically feasible and that data exchange across stakeholders, while challenging, is possible. Data sharing with governments or public institutions has its own set of challenges further discussed in Sections 3.4.5 and 3.4.6 in the report.

Table 3.2. Summary of private initiatives reviewed

Criteria	Maersk/IBM	Pfizer	J&J	DHL	USP
Purpose	Efficiency; unified view of shipments; single contact point	Efficiency; transparency; resilience; compliance	Efficiency; transparency; resilience; compliance; data-driven decision making	Enhanced efficiency; regulatory compliance	Resilience, risk analysis, data-driven decision making
Incentives	Cost reduction; visibility of shipments; customs compliance	Efficiency gains; cost reduction	Productivity increase; cost reduction; reduced product loss and spoilage	Visibility of shipments; compliance; enable interventions	Visibility, predictive analytics
Supply Chain Comprehensiveness	Only transport stage	Post-production tracking; partial tracking of intermediates	Full lifecycle tracking of finished goods and partial tracking of intermediates	Distribution stage and dispensation	Finished products, APIs and KSMs
Product Coverage	Shipments; consignment; transport equipment	Medicines	Medicines	Finished products from customers	Medicines authorised in the US
Product Identification	Blockchain; GTIN	Unique product identifier; 2D DataMatrix; GTIN; serial number; expiry	Unique product identifier; 2D DataMatrix; GTIN; serial number; expiry	RFID; smart labels	National Drug Code (FDA’s identifier)
Product Tracing	Near real-time monitoring (GPS)	Real-time monitoring (IoT; GPS)	Real-time monitoring (IoT; GPS; RFID)	Real-time monitoring (IoT; GPS; RFID)	Updated every month
Product Verification	Blockchain-solution	Internal web and mobile solution, AI-driven analytics	AI-driven analytics; exploration of blockchain	Integration with serialisation	AI-driven analytics
Data architecture	Platforms and cloud computing, interoperable private network	Platforms and cloud computing, interoperable private network	Platforms and cloud computing, interoperable private network	Platforms and cloud computing, interoperable private network	Platforms and cloud computing, interoperable private network
Digital technologies	Blockchain	IoT; AI/ML	IoT; digital twins, AI/ML, Robotic Process Automation	RFID, IoT; smart labels	AI and predictive analytics
Interoperability	Interoperable with enterprise systems; applies UN/CEFACT standards; contends with varying customs regulations	Interoperable with logistics partners; Infor Nexus and Fivetran; compliant with US, EU, and WHO standards	Interoperable with logistics partners, enterprise systems; compliant with US, EU, and Good Manufacturing Practices	Interoperable with customers’ logistics and supply chain platforms; compliant with Good Distribution Practices	Interoperable with customers’ systems; aligned with FDA standards
Implementation Cost	Unable to achieve commercial viability	Undisclosed; significant savings for the company from the investment in monitoring and forecasting	Undisclosed; significant savings for the company from the investment in monitoring and forecasting	Feasible cost-effective monitoring for customers	Investment of several million USD; builds on existing USP network.
Technological Neutrality and Competition	Proprietary technology; competitive concerns from stakeholders	Controls core systems; collaborates via neutral platforms for partner integration	Controls systems end-to‑end; designed to interface but not governed by third parties	Proprietary technology	Based on proprietary data, public datasets and licensed commercial datasets

Notes: AI/ML = Artificial Intelligence and Machine Learning; GPS = Global Positioning System; GTIN = Global Trade Item Number; IoT = Internet of Things; RFID = Radio Frequency Identification, UN/CEFACT = United Nations Centre for Trade Facilitation and Electronic Business. See Box 3.4 for a description of Infor Nexus and Fivetran.

Source: Based on interviews and literature review.

3.3.1. Maersk/IBM

The TradeLens platform, a joint initiative by A.P. Møller Maersk (an integrated transport and logistics company) and IBM was launched in December 2018 (Ahmed and Rios, 2022[33]). TradeLens’ objective was to create a consolidated view of shipment movements based on real-time, immutable and decentralised tracking information (George, 2025[34]). TradeLens could greatly reduce tracking costs and make regulatory compliance easier and faster through co‑operation with customs authorities (Jovanovic et al., 2022[35]).

Blockchain technology ensured data security and encouraged data sharing. The blockchain’s distributed ledger10 made TradeLens’ data immutable and provided encryption (Louw-Reimer et al., 2021[36]). Trade documents were secure, unalterable and traceable, which enabled TradeLens to establish a reliable audit trail (Ahmed and Rios, 2022[33]). Ocean cargo liners were the primary data providers who uploaded their shipment documents onto the platform. Shippers and cargo owners consumed the uploaded data. For data providers, a seamless integration of their internal IT systems could be challenging, while data consumers simply subscribed to the platform and did not require sophisticated system integration (Louw-Reimer et al., 2021[36]). TradeLens possessed permissioned access which permitted each carrier to decide who could access their data (Rukanova et al., 2021[37]). For example, a truck driver moving a single container was provided with only the information necessary for the task, while a cargo owner had access to more information (Louw-Reimer et al., 2021[36]).

Using international standards supported interoperability. To ensure interoperability with the various stakeholders, the data format was aligned with the UN/CEFACT standards11 and the bill of lading standards of the Digital Container Shipping Association (Louw-Reimer et al., 2021[36]). In addition to the platform, the incorporated marketplace enabled the launch of specific and customised applications that supported operations (Ahmed and Rios, 2022[33]).

Maersk’s role in TradeLens raised competitive concerns. TradeLens was developed as a neutral industry platform (Rukanova et al., 2021[37]). However, as the initiator of the platform and the largest ocean carrier, Maersk’s dominant position on the platform was viewed with scepticism (Jovanovic et al., 2022[35]).12 Nonetheless, by July 2019, five of the six largest cargo liner companies had committed to join the platform (Louw-Reimer et al., 2021[36]). As the number of participants increased, the benefits that members could realise increased as well (Jovanovic et al., 2022[35]).

A lack of incentives led to the discontinuation of TradeLens. According to Maersk (2022[38]), insufficient commercial viability and scale made the closure of the platform in March 2023 inevitable. George (2025[34]) identifies several further causes. Lacking financial incentives for key stakeholders undermined widespread industry support. The perception that benefits were unequally distributed, and Maersk was positioned to gain an unfair competitive advantage, exacerbated competitive concerns. Regulatory barriers, including different customs requirements, hindered TradeLens’ adoption as the blockchain-based documentation could not be harmonised across jurisdictions. Further, interoperability issues paired with high costs of adapting legacy systems deterred participation.

3.3.2. Pfizer

In recent years, Pfizer went through a significant supply chain transformation through its Horizontally Orchestrated Supply Universe Network (HOSuN) and Digital Operations Center (DOC); during the COVID‑19 pandemic, this underpinned its work to deliver over 2 billion doses of vaccines worldwide by 2022 (Pfizer, 2022[39]). One of the key innovations of the firm was the use of IoT sensors embedded in shipping containers to collect real-time data regarding environmental conditions, e.g. temperature, humidity, exposure to light, and GPS location (Pfizer, 2021[40]) for the transport of temperature‑sensitive COVID‑19 vaccines, which proved invaluable in maintaining a vaccine spoilage rate below 0.1% globally.

The internal digital supply chain infrastructure of Pfizer combines a variety of technologies, including cloud computing, IoT, AI-driven forecasting, and data-sharing networks with end-to‑end security (European Pharmaceutical Manufacturer, 2025[41]). Through these platforms, the firm is able to monitor the flow of its pharmaceuticals across more than 175 countries.

Furthermore, Pfizer’s monitoring system relies heavily on modern cloud platforms like Snowflake and SAP (Shikha Yug Sachdeva, 2024[42]). The Snowgrid architecture of Snowflake enables synchronisation of data in real-time among various regions, firms, and external collaborators, whereas SAP is being used for stock and production planning to synchronise data between forecasts and supply chain logistics (European Pharmaceutical Manufacturer, 2025[41]). Data dashboards offer immediate information on product movements, warehouse status, and exceptions such as delay in shipments or breach of conditions.

Beyond that, the system also includes forecasting in its monitoring framework by integrating demand forecasts with operational planning. According to Pfizer, advanced machine learning and deep learning models are applied to internal sales history as well as internal and external market information, including public procurement trends. These forecasts are then translated into executable demand signals to drive production and logistics planning (Shikha Yug Sachdeva, 2024[42]).

Pfizer has controls in place to ensure compliance with FMD and DSCSA, including through inter alia the adoption of GS1 standards and tamper-evident packaging. The internal monitoring system goes further by allowing for full end-to‑end visibility, real-time monitoring of movement and automated exception handling using serialisation data available for many countries across the world.

Box 3.3. Technologies supporting interoperability

Interoperability is the ability of IT systems to exchange and make use of information. Companies also face interoperability challenges and need specific tools to connect platforms and exchange supply chain information. During the review of private monitoring initiatives, two data solutions were mentioned by companies: Infor Nexus and Fivetran.

Infor Nexus

Infor Nexus (formerly GT Nexus) is a cloud-based multi‑enterprise supply chain platform that connects companies with their suppliers, manufacturers, logistics providers and other trading partners on a shared network. Over 90 000 companies are using the platform that provides a single source of information for suppliers, manufacturers, retailers, and logistics providers (Infor Nexus, 2024[43]). Infor Nexus monitors multi-tier global supply chains and provides real-time visibility from raw materials to the finished product. Goods can be tracked both on an item and lot basis. The usage of AI and machine learning enables the platform to estimate arrival times, to anticipate delays, and thereby to identify bottlenecks (Brennan, 2025[44]). Companies can utilise this data to optimise their capacity planning and to respond quickly to supply disruption, which increases resilience and operational efficiency (Software Connect, 2025[45]). Infor Nexus’ customers come from the manufacturing, distribution, healthcare, hospitality, public, and fashion sectors (Levinas-Ménard, 2025[46]).

Fivetran

Fivetran is a tool for integrating and transforming data. Its core functionality is to extract, load, and transform (ELT) data within its cloud-based platform, which aims to simplify data integration (Tobin, 2026[47]). Fivetran acts as the bridge between companies that need to consolidate and transform data. Automated data pipelines can be set up that extract data from different sources. The extracted data can be cleaned for entry into the company’s selected data warehouse and is updated automatically. Fivetran offers pre‑built connecting features that allow companies to set up extracting functions with very little coding or maintenance required (Smith and Watts, 2024[48]).

3.3.3. Johnson & Johnson

Johnson & Johnson (J&J) began digitising its supply chain during the mid‑2010s by piloting Industry 4.0 technologies like IoT, AI, and robotics for smart manufacturing, digital control towers, and automation. Over time, with the rollout of enterprise systems for pharmaceuticals, consumer healthcare, and MedTech businesses, the company has established a fully digitised, end-to‑end traceability framework including production facilities, warehouses and distribution channels.

Like all manufacturers, the firm has implemented total serialisation for all drugs under US DSCSA and EU FMD regulations, as well as many other markets. These identifiers are linked to J&J’s enterprise resource planning and warehouse systems to trace finished goods throughout their entire product life cycle in real-time. The tracking infrastructure also includes IoT sensors, RFID, and GPS technology to track location, condition, and temperature while in transit.

Verification is done through automated quality control machinery and anomaly detection software based on AI (Muppalla, Devi Maddi and Maddi, 2025[49]). J&J has implemented real-time release testing across pilot sites, authorising batch quality through predictive models and without full manual inspection. Moreover, blockchain pilots are in development in some high-risk markets to ensure authenticity. These technologies combined reduce inspection time, improve traceability, and enable compliance verification.

Similar to Pfizer, J&J’s supply chain is based on enterprise‑class systems like SAP, Oracle, Snowflake, and Microsoft Azure. These systems are integrated internally across control towers and externally with third-party logistics providers like DHL and UPS. Their architecture supports integration with public regulatory networks like DSCSA and EU FMD. Although J&J remains the owner of all its assets, its systems are designed to be interoperable with third-party platforms when necessary.

The company manifests compliance with the EU FMD in its serialisation policy and verification tools. Beyond that, it offers functionality above FMD requirements, including certain upstream input traceability (APIs and excipients), factory smart monitoring, and real-time dashboards. It has also added digital labelling capabilities that through a simple scan of the serialised product labels can connect the physical product to digital content (e.g. electronic product information leaflets, product quality vigilance integration, etc.). J&J is one of the leaders in the industry for this work that offers patient-focussed improvements in access to digital content. These extensions add value by making environmental tracking, early warning of quality deviations, and advanced recall protocols possible. For instance, IoT-based shipping data exchange or supply-demand forecast alarms reinforce shortage anticipation.

According to J&J, significant performance improvements have been achieved since its digital transformation: 40% productivity gains, 24% cost savings in smart factories, and significant reductions in spoilage and production errors. These advancements validate the power of smart automation, real-time monitoring, and harmonised data management. A dedicated “Intelligent Automation Council” oversees system integrity, access policies, and regulatory alignment, allowing for compliance with GDPR, HIPAA, and GMP guidelines.

Overall, J&J’s digital supply chain exemplifies private‑sector innovation in end-to‑end traceability, AI-powered quality control, and real-time monitoring of logistics. Its scalability and flexibility allow for collaboration with regulatory initiatives like the EU FMD. In that regard, public-private alignment can potentially enhance medical security (Schoelmann, 2024[50]).

3.3.4. DHL

As one of the world’s largest logistics providers, DHL has been a leader in adopting radio frequency identification (RFID) and smart label technologies13 to improve traceability, efficiency, and compliance in medical and pharmaceutical supply chains. DHL has important activities in healthcare, as it operates healthcare distribution centres, manages cold chain logistics and delivers temperature‑sensitive products across borders.

DHL began integrating RFID into its healthcare logistics operations in the mid‑2000s. Together with IBM, it developed a temperature tracking system for its pharmaceutical customers combining temperature sensors with RFID technology (IBM, 2007[51]). Under the pressure from the US FDA to guarantee the temperature of shipments in transit, the solution was successfully deployed to track container temperature in real time along supply chain “checkpoints” where the shipment changes its mode of transportation (e.g. from road to air).

However, the benefits of RFID technology come at a relatively high cost. A DHL report notes that the infrastructure needed is complex and challenging to scale (DHL, 2025[52]). RFID only works in a controlled environment with a dedicated infrastructure, as part of a “closed-loop system”. High costs also come from the use of special readers and routers built on proprietary systems. During an interview with a pharmaceutical wholesaler, it was also mentioned that medicine packages have aluminium blisters (a leakproof material) that interfere with RFID signals.

While cheaper RFID tags have been developed over time, there is no specific advantage over barcodes when RFID systems are “passive” (i.e. they do not transmit information). Relevant use cases are for high-value products where there is a need for the tag to be active and communicate some information. This is the case for pharmaceutical temperature‑sensitive products. Through its Thermonet network, DHL provides global cold chain services designed specifically for life sciences shipments (including pharmaceutical products). Both air and ocean shipments integrate RFID sensors that automatically log temperature readings at key checkpoints and transmit them to DHL’s central monitoring platform.

Through DHL’s LifeTrack system, a proprietary web-based track-and-trace platform for healthcare shipments, healthcare clients (including pharmaceutical manufacturers and hospitals) can view detailed shipment status, environmental history, and chain-of-custody records. The platform integrates data from the RFID readers and sensor labels to offer real-time visibility and event logging. Its aim is also to reduce paperwork and to improve regulatory compliance documentation.

During COVID‑19, stringent cold temperature requirements were applied for the distribution of RNA vaccines (‑70 degrees Celsius). DHL also worked with RFID sensors attached to vaccine packages to continuously monitor temperature conditions throughout multimodal transport (air freight, last-mile delivery). If any temperature deviation occurred, the system triggered automated exception handling, such as diverting shipments for inspection or re‑routing them to cold storage. This real-time monitoring prevented loss of efficacy in the shipped vaccines (DHL, 2020[53]).

In addition, DHL has deployed RFID-enabled smart shelves and racks within its pharmaceutical distribution centres to track the movement of tagged medicines, medical devices, and hospital supplies in real time (DHL, 2020[54]). These RFID systems support automated inventory control, reducing manual scanning and lowering the risk of picking and shipment errors.

DHL’s experience demonstrates that large‑scale RFID deployment is both technically and economically feasible in medical supply chains, although there may be specific use cases. RFID and other smart label technologies can improve transparency and visibility, automate compliance with serialisation and Good Distribution Practices (GDP), and support public-private co‑operation (e.g. with health authorities during vaccine rollouts) through shared data on shipment conditions and chain of custody.

3.3.5. USP Medicine Supply Map

The USP Medicine Supply Map (MSM) is an intelligence platform built by the U.S. Pharmacopeia (USP) to increase visibility into upstream pharmaceutical supply chains to identify, characterise and quantify vulnerabilities that can lead to drug shortages. It has near-complete coverage for medicines authorised in the US market and covers more than 96% of generic medicines. USP positions the platform as an early warning capability: it offers predictive analytics that can alert stakeholders when a medicine is at elevated risk so they can take mitigating actions (e.g. diversify suppliers, build buffer inventory, adjust contracting and sourcing). The platform is used by distributors, hospitals, manufacturers and governments for advance planning and emergency response.

USP’s MSM maps upstream supply chains for medicines, linking finished drug products to APIs and, in its most recent expansion, deeper upstream inputs such as KSMs where data are available. The focus is on the manufacturing-heavy upstream segment, where many shortage drivers originate (Van Beusekom, 2022[55]). Through modelling and additional data collected (such as shortage data or EU and FDA inspections), the platform creates a medicine risk profile. It analyses about 200 shortage risk factors that were identified with experts. The platform thus provides actionable insights on vulnerabilities at the product level that can be used by stakeholders in operational decisions such as product planning and risk mitigation.

What is noteworthy about USP’s platform is the combination of the supply chain structure with a series of indicators that are useful for vulnerability assessment, such as manufacturing location intelligence (the platform covers 85% of API volumes mapped to facilities and 81% of APIs mapped to KSM country of origin), concentration exposure (through an analysis of market shares and geographical concentration), manufacturing complexity and handling constraints (e.g. how hard a product is to make or store), economic and market signals (e.g. price levels, with low prices a structural driver of shortages for off-patent medicines) and quality risk signals (using information on EU and FDA inspection outcomes). While “predictive analytics” and “AI tools” are generally part of the hype around supply chain mapping platforms, the risk analysis is here based on a robust series of indicators that were found in the literature to be linked with medicine shortages.

MSM is maintained by a non‑profit, independent scientific organisation and relies on a combination of public datasets (including those made available by the US FDA) and licensed commercial data. It does not rely on proprietary company‑submitted supply-chain data, helping to address competition and neutrality concerns often associated with supply-chain monitoring initiatives. USP also publishes an Annual Drug Shortage Report drawing on insights from MSM, summarising recent shortage trends and key drivers.

3.4. Key challenges in creating effective monitoring mechanisms for medical supply chains

The review of public track-and-trace systems and private initiatives to better monitor supply chains and forecast demand has already highlighted a series of challenges in creating effective monitoring mechanisms and securing medical supply chains. This section further discusses these challenges. The discussion is organised around six key challenge areas – from standards and technology infrastructure to data governance and incentives and obligations. The focus is on practical obstacles and enabling conditions for an EU-led system that could ultimately interface globally to bolster collective supply chain resilience.

3.4.1. Standardisation and international interoperability

Harmonising product identification and serialisation standards is a foundational challenge. An international monitoring system would require that each medicine item, batch, manufacturer, and other entities carry unique identifiers recognisable across all participating jurisdictions. The good news is that most countries already converge around global standards, particularly the 2D “DataMatrix” bar codes such as GS1 (Box 3.4).

2D barcodes provide benefits with regard to the data they can store. A regular linear barcode is limited to encoding static information. Dynamic data, such as batch numbers and expiry dates, cannot be included. However, modern systems rely on these dynamic data for the automated detection of abnormalities, as well as inventory management (Klein and Stolk, 2018[56]). A 2D DataMatrix can encode such elements, making it superior to other formats (WHO, 2024[57]). Moreover, these codes are smaller in size than linear barcodes, allowing for more space on the packaging (Klein and Stolk, 2018[56]).

A WHO (2024[57]) survey of a limited number of member countries indicates the importance of international standards in tracking pharmaceutical products across the supply chain. The majority of responding members indicated that they exclusively utilise international standards, while a few use them in addition to their domestic standards. The standards are mostly utilised to identify the individual product units, and the survey responses show a clear trend towards the usage of a 2D DataMatrix as most responding member states use it.

Legislation regarding the introduction of h2D barcodes has been passed in Argentina, Brazil, Canada, China, the EU, India, New Zealand, Norway, Pakistan, Korea, Switzerland, Türkiye, the United Kingdom, and the United States (Rasheed, Höllein and Holzgrabe, 2018[58]). This list includes China and India, both major suppliers of pharmaceutical inputs and medicines. However, several countries had to postpone the implementation due to cost and lacking expertise on the barcode system (Rasheed, Höllein and Holzgrabe, 2018[58]).

Box 3.4. Global standards for product identification and tracing

GS1 standards

GS1 is a global non-profit standards development organisation. GS1 provides standards for unique identifiers, data carriers, and information sharing that can be applied to a wide range of products and sectors. More than 2 million companies in 150 countries are using GS1’s standards as they facilitate interoperability (GS1, 2023[59]). Over 70 regulatory authorities mandate or recommend GS1 barcode for medicine products. In the United States and EU over 16.5 billion medicines include GS1 barcodes annually, which facilitates tracing and tracking at different supply chain stages (GS1 UK, 2025[60]). The Global Trade Identification Number (GTIN) is a unique identifier that helps to identify individual products (Jayaraman et al., 2011[61]). The GTIN can consist of eight, 12, 13, or 14 digits and is encoded into barcodes and data matrices. It serves as the key to retrieve predefined information (GS1 US, 2025[62]). It is part of GS1’s identification standards and frequently used in combination with the Global Location Number (GLN) and the Global Data Synchronization Network (GDSN). The GTIN facilitates communication of product information between the manufacturer, distributors and healthcare providers (Jayaraman et al., 2011[61]). The Electronic Product Code Information Systems (EPCIS) is a GS1 standard that captures and shares information about the movement and status of goods. It defines the interfaces and applications through which supply chain data can be shared among different supply chain stages and stakeholders. EPCIS is used increasingly in the healthcare sector to facilitate tracking and tracing of products and to adhere to regulatory requirements concerning the product’s chain of custody and availability (GS1 US, 2025[62]).

ISO standards

The ISO’s Identification of Medicinal Products (IDMP) is a set of five standards specifying the definitions for the identification and description of medicinal products. The standards provide data elements and structures with regard to the substance, pharmaceutical dose forms, units of measurement, regulated pharmaceutical product information, and regulated medicinal product information (EMA, 2026[63]). This provides a consistent international terminology and facilitates the exchange of information between the regulating agencies, manufacturers, and suppliers. Using IDMP therefore improves data quality and information sharing by increasing transparency and fostering interoperability. As the standards allow the identification of equivalent products across regions, it could also support mitigating shortages (FDA, 2022[64]). IDMP and GS1 standards complement each other, as IDMP defines what the product is, while GS1 enables the tracing of products along the supply chain (GS1, 2021[65]).

Nonetheless, despite broad use of GS1 identifiers, differences in data formats and registries impede interoperability. Each jurisdiction maintains its own product databases and coding systems for manufacturers and facilities. Behind each GS1 identifier, there are national registries and different systems to organise the information. Within the EU, a step towards harmonisation is the implementation of the ISO IDMP (Identification of Medicinal Products) standards via the EMA’s SPOR database (Substance, Product, Organisation, Referential data). SPOR provides a common EU master data hub for all authorised medicines, intended to streamline data exchange and facilitate the reliable exchange of medicinal product information across member states. It will enable cross-border recognition of product identifiers and support functions like pharmacovigilance, e‑prescriptions and shortage management.

However, globally there is no equivalent master data system. To develop a co‑operative monitoring mechanism involving EU and non-EU countries, agreements on mutual recognition of identifiers would be needed and the interoperability of systems would have to be organised. While encouraging partners to adopt global standards can help (such as GS1 and ISO IDMP), a broader harmonisation would involve a combination of government-to-government agreements and public – private collaboration. For instance, regulatory co‑operation forums (like ICH14 or ICMRA15) could endorse common serialisation and data standards, while industry coalitions ensure implementation on the ground. Without such harmonisation, an EU system might need to build costly mapping solutions to translate between different code systems used in, say, the United States, India or China – an approach that is less robust and prone to errors.

A robust cross‑border monitoring mechanism would need not only convergence around GS1 DataMatrix codes, but also explicit alignment with IDMP/SPOR for medicinal‑product master data and EPCIS for upstream supply-chain event reporting. A practical next step would be the development of a GTIN to IDMP mapping guideline, ensuring that each serialised pack (GTIN, batch, serial) is reliably linked to its IDMP Product and Manufactured Item entries. This mapping is a prerequisite for interoperable exchange and avoids costly one‑off translation solutions. EPCIS‑based event capture would further allow monitoring of API and intermediate‑level flows, which extends the traceability beyond finished products.

The implementation of the DSCSA in the United States illustrates the challenges of converging towards effective serialisation (Partnership for DSCSA Governance, 2025[24]). While the legislation had strong requirements to create interoperability and the FDA worked closely with stakeholders to provide guidance and address challenges, up to 30% of labels could not be read consistently across platforms (Schweihs, 2024[66]) leading to postponing the deadline for full implementation (from November 2023 to November 2024).

Moreover, when it comes to covering the whole supply chain, the scope of standardisation remains limited. As highlighted in Section 3.2, existing regimes primarily serialise finished pharmaceutical products at the saleable unit level. But an effective early-warning system may need to track deeper into the supply chain (e.g. API batches, key excipients, or even precursor chemicals). These intermediate inputs often lack a standardised global identification scheme. The FMD introduced tougher rules on importation of APIs (requiring documentation of GMP compliance) but did not extend unique serial numbers to API packages. Creating a consistent coding for critical intermediate inputs and raw materials and linking those to finished product codes remains uncharted territory.

3.4.2. Technological compatibility and infrastructure

Integrating diverse technologies and systems across countries is a second major hurdle and challenge. As illustrated in Section 3.2, different countries have different monitoring systems in place, whether it is end-to‑end verification networks or specific shortage monitoring platforms (such as the ESMP in the European Union). Interoperability within jurisdictions (such as within the United States in the context of the US DSCSA or within the EU in the context of the EMVS) is already challenging. Technological compatibility across jurisdictions opens another layer of complexity, even if firms and other stakeholders already operate in a global environment where they have to comply with or contribute information to the reporting and monitoring mechanisms of various jurisdictions.

For policymakers, the question is how prescriptive they should be on technology choices. One strategy is setting common data standards and interoperability requirements but allowing companies and solution providers to implement any compliant system (“technology-neutral” approach). This can spur innovation – for example, multiple vendors might offer interoperable software that connects to the network – but it risks fragmentation if not carefully co‑ordinated. The DSCSA in the United States is essentially taking this route: rather than a single centralised database, it envisions an interoperable “network of networks” connecting manufacturers, distributors, pharmacies, etc. via standardised electronic data exchange. This distributed model requires robust interface standards and governance to ensure all parties (including many private IT systems) can communicate seamlessly.

Alternatively, a common public platform could be established – a centralised “data lake” where all critical supply data must be uploaded. Such a platform would simplify data integration and analytics, at the cost of greater initial infrastructure investment and potential industry pushback on centralised data control. Notably, the EU’s EMVS is based on a public-private partnership model: the system is managed by stakeholders (pharmaceutical industry and pharmacists) under supervision of authorities. A similar approach could be conceived for a shortage‑oriented monitoring system, balancing public oversight with industry operational know-how. But operationalising it across several jurisdictions would require strong international co‑ordination.

Emerging technologies like blockchain have been touted as solutions for multi-party supply chain transparency, but real-world results have been mixed. Blockchain’s appeal lies in creating a tamper-proof, shared ledger of transactions visible to authorised participants, which could enhance trust in the data’s integrity. The TradeLens initiative showed how a blockchain platform can digitise trade documents and events, but it also demonstrated that technology alone cannot overcome reluctance to share data. TradeLens struggled because bringing the entire ecosystem together around a common approach proved elusive.

In the pharmaceutical context, some pilot projects (e.g. the EU’s PharmaLedger consortium) have experimented with blockchain for tracking medications and verifying authenticity across companies and countries.16 While technically viable, the governance model is critical – participants need clear rules on data ownership, access permissions, and liability. Blockchain might best be applied in specific areas (such as validating provenance of high-value or highly regulated medicines or enabling smart contracts for automatic supply re‑routing when certain conditions trigger). However, it is not a panacea; other cloud database technologies may achieve interoperability with less complexity if trust among parties can be ensured through legal agreements.

A further option is to rely on federated learning or, more broadly, federated analytics (Bersin, Swartz and Karlsson, 2025[67]). In traditional data systems, information is pooled in a central repository before it is analysed. This can be problematic where the relevant data include commercially sensitive information on production volumes, inventories, orders, supplier relationships, manufacturing locations or shipment flows. Federated approaches offer a different model: participating firms retain their data locally, while a common algorithm is sent to local databases to generate agreed outputs, model parameters, indicators or alerts (Li et al., 2020[68]; Zhang et al., 2021[69]).17 In principle, this makes it possible to identify abnormal patterns in supply, demand or inventories without requiring all underlying data to be transferred to a central database.

Federated learning can therefore be useful where the objective is to detect common risk signals across distributed datasets. For example, a model could be trained to identify unusual changes in demand, production capacity or stock movements across several companies or countries, while the underlying firm-level data remain within each participant’s own environment. This could be particularly relevant for international co‑operation, where governments and firms may be willing to contribute to a common early-warning capability but reluctant to transfer detailed operational data outside their jurisdiction.18

However, federated learning should not be presented as a complete solution. It has several limitations. First, it requires a high degree of standardisation before it can be useful. Without common data models, federated computation may simply produce non-comparable outputs. Second, federated learning is technically complex and costly to implement, particularly when it comes to eliminating all confidentiality risks and ensuring secure aggregation and query controls. Third, federated learning is better suited to statistical learning and pattern detection, which can be useful for early-warning systems but not for other dimensions of preparedness where authorities need to verify specific stock levels, investigate a supply disruption or co‑ordinate emergency procurement. For this reason, federated analytics should be considered as one component of a wider data-governance architecture, with alternative decentralised models that can also ensure data privacy and protection, such as trusted data intermediaries.

In summary, the technological challenge is twofold: achieving technical interoperability (systems talking to each other) and choosing a system architecture (centralised vs. distributed vs. hybrid) that best balances effectiveness with trust. Governments can encourage harmonisation by funding integration efforts, convening stakeholders to agree on protocols, and even mandating certain connectors or hubs. They can also support experimentation (e.g. sandbox environments) with technologies like blockchain or federated learning to see how they might augment the monitoring mechanism. Crucially, the technological infrastructure must be reliable, cybersecurity-hardened, and scalable – any perception of insecurity or unreliability would discourage use.

3.4.3. Product and input coverage scope

Defining the scope of products and inputs to be covered by a monitoring system is a third challenge and critical policy decision. From the perspective of securing supply in the context of health emergency preparedness, medical countermeasures (MCMs) include vaccines, therapeutics, diagnostics and other medical devices, as well as personal protective equipment (PPE).

For medicines, an approach observed in different initiatives is to focus on a list of critical finished products (e.g. essential medicines that are lifesaving or have no alternatives). The WHO Model List of Essential Medicines offers an internationally agreed foundation for identifying such products. The EU has also begun such prioritisation efforts: in December 2023, the EMA published the first Union List of Critical Medicines, with the most recent version released in January 2026 including 299 medicines (EMA, 2026[70]).

However, limiting the system to finished products may miss upstream fragilities. Shortages often originate from upstream supply chain disruptions – for instance, an API plant closure, scarcity of raw materials, or packaging component issues can choke off production of the final medicine. Thus, many experts argue the monitoring scope should extend to key intermediate inputs: APIs, KSMs and even critical medical devices or consumables (like syringes or vials) that are essential for delivering care. The COVID‑19 crisis illustrated the importance of items like reagents for diagnostics and ancillary supplies for vaccination, which were not medicines per se but were bottlenecks when in short supply (Behnam et al., 2020[71]; Fritz et al., 2022[72]).

Including intermediate products in an international monitoring system poses challenges in identification and prioritisation. Not every input can feasibly be tracked – regulators would need to identify which inputs are truly critical or single‑sourced.19 For example, if 80% of the EU’s supply of a generic antibiotic depends on one precursor chemical from a single country, that precursor is a good candidate for tracking. Some countries have begun mapping these vulnerabilities; the EU’s pharmaceutical strategy work and Critical Medicines Alliance have noted the geographical concentration of certain production chains as a major risk factor for shortages. Those mappings can inform which inputs to cover. Once identified, tracking them could leverage trade data and customs codes, but more granular serialisation might be needed.

Bulk raw materials do not have unit-of-use packages to sticker like finished drugs do. Instead, batches of API could be assigned lot IDs that link to the finished product GTINs they go into. This is conceptually possible – indeed, the FMD’s safety features include batch numbers that could serve as a link – but it would require new data collection at the manufacturing level and possibly agreements with API and KSM manufacturers globally to participate in data sharing.

Moreover, to be adequately prepared in terms of medical countermeasures, monitoring will have to go beyond the List of Critical Medicines and include medical devices and equipment, as well as medicines that may not be part of the List. The European Commission, in co‑operation with EU member states, is currently developing an EU List of Medical Countermeasures for Priority Threats, to be published in 2026. As a first deliverable under the EU Medical Countermeasures Strategy (COM(2025)529 final), the Comprehensive Health Threat Prioritisation Assessment for Medical Countermeasures identified four priority threat categories: 1) respiratory or contact-based viruses with pandemic potential, 2) vector-borne or animal-reservoir viruses with epidemic potential, 3) armed conflict-related threats and chemical, biological, radiological and nuclear (CBRN) threats, and 4) antimicrobial resistance.

As such, it goes beyond medicines and also includes vaccines, therapeutics, diagnostics, personal protective equipment, and CBRN-specific countermeasures (European Commission, 2026[73]). For example, if a shortage of ventilators or dialysis filters could imperil patients similarly to a drug shortage, a comprehensive system might monitor those as well. The EU’s existing frameworks (like EUDAMED for devices) are not as advanced in traceability as the medicines system, though UDI (Unique Device Identification) standards are being implemented for devices. An integrated critical supply monitoring effort might at least link to device supply databases or include a subset of devices deemed critical for emergencies.

The Critical Medicines Act (CMA), on which the EU Council and Parliament reached a provisional agreement in May 2026, is intended to complement the EMA-led shortage framework and avoid duplicative data collection. The agreed text relies on supply chain vulnerability evaluations carried out under the revised pharmaceutical legislation, using aggregated data across all medicinal products that have the same active substance, route of administration and formulation. These evaluations identify dependencies on a single or limited number of third countries or sites for active substances, key inputs or finished dosage forms (European Commission, 2025[74]). The draft CMA also gives national competent authorities and, for certain purposes, EMA, targeted powers to request information from actors across the supply and distribution chains of critical medicinal products, their active substances and key inputs in specific circumstances (Council of the European Union, 2026[75]).

Ultimately, there is a trade‑off between breadth and manageability. Given the critical role of both medicinal products and medical devices, monitoring arrangements should not be confined solely to medicines. Recent health crises have underscored the importance of addressing medical countermeasures in a holistic manner as part of crisis preparedness. In practice, an EU‑level monitoring system that incorporates upstream inputs could begin with a pilot phase focussed on a limited set of essential active pharmaceutical ingredients (APIs) and key starting materials (KSMs) linked to a small number of priority medical countermeasures. Such an approach would allow proof of concept and operational refinement, with scope for gradual expansion over time.

Policymakers should also consider modularity – the system could have tiers or modules: one for finished medicines (with high granularity of tracking), one for bulk inputs (perhaps at batch or shipment level monitoring via customs/import data), and one for devices or other relevant medicinal products and PPE (maybe monitored via a simpler stock-level reporting). Each module might use different data sources but ultimately feed an overall early warning dashboard. A key challenge lies in ensuring effective linkages across the system so that, for example, a disruption in an active pharmaceutical ingredient (API) supply generates an early alert for the associated medicine before evolving into a patient‑level shortage. Monitoring systems that do not cover upstream inputs risk remaining largely reactive, detecting problems only once pharmacy stocks begin to decline. A comprehensive approach seeks to be proactive by identifying vulnerabilities earlier in the supply chain; this, however, requires extending monitoring beyond finished medicinal products to include critical upstream components.

3.4.4. Information requirements and data collection mechanisms

A fourth main challenge is to establish a robust monitoring and early warning system, which hinges on gathering the right data in a timely manner. The key information domains include production and supply data, inventory/stock levels (with expiration dates), demand or consumption data (or forecasts), and distribution flow data. Collecting such information systematically at EU-wide (eventually international) level is unprecedented in scope and would require novel data-sharing arrangements. Provisions in the EU Emergency Framework Regulation (COM(2021)577 final) currently foresee this kind of data collection for MCMs only during health emergencies.

For production and supply, authorities would ideally know how much of a critical medicine (or input), medical device or PPE is being produced, where, and any impending changes (e.g. planned maintenance shutdowns or capacity expansions). Currently, this data resides with manufacturers and is often considered commercially sensitive. Companies do report some information to regulators – for example, when MAHs foresee a shortage, EU law obliges them to notify national authorities,20 but this is a reactive, manual process and not uniform across countries.

A monitoring system might instead regularly pull capacity and output data from firms (perhaps under confidentiality) or use proxy indicators. One proxy could be regulatory batch release data (for medicines that require official batch release by authorities) or data from CMO (contract manufacturer) output. Another proxy is customs/export data for APIs, medicines, medical devices and PPE, which can signal volumes moving in supply chains. Integrating customs trade data could help flag drops in import of a key ingredient, for instance.

For inventory and stock, one needs visibility into both public stockpiles (e.g. strategic reserves, hospital stocks) and private inventories (e.g. manufacturer and wholesaler warehouses). During COVID‑19, many EU countries found they lacked basic visibility on inventory levels of MCMs nationally. Following the pandemic, the EU created the Health Emergency Preparedness and Response Authority (HERA). In the event of a public health emergency at Union level and the activation of the Emergency Framework Regulation, HERA can deploy temporary crisis measures. These include the collection of inventory data from MAHs, distributors, and manufacturers, as well as member state reporting. Hospitals may be asked via their national authorities to supply situational information, though they are not subject to mandatory reporting as economic operators are. A monitoring system could require periodic reporting of stock levels from suppliers and major distributors. This could be facilitated by digital inventory management systems – for instance, wholesalers might provide automated feeds of days-of-stock remaining for certain drugs. However, such reporting must be carefully designed to avoid overly burdensome requirements, especially on small firms.

The volume of data is significant, raising questions about how it is collected. One could envision a centralised data platform where firms upload required data fields (production volumes, inventory levels, etc.) at set intervals. This central platform would need to reconcile and analyse the data to produce alerts. The EMA’s extended mandate since 2022 is moving in this direction with the creation of the Medicines Shortages Steering Group (MSSG), the Voluntary Solidarity Mechanism to co‑ordinate stock sharing among Member States and the launch of the ESMP (EMA, 2025[76]) (see Box 3.5 for further information on the ESMP). These efforts rely initially on manufacturers’ notifications and member states’ input, rather than full automated tracking. The new system could greatly enhance this by linking in supply chain data feeds. Further relevant data may come from the obligatory shortage prevention plans that MAH will have to provide according to the new EU pharmaceutical legislation (“Pharma Package”). The shortage management plan shall include, among others, potential alternative products, quantities delivered per month per member state, manufacturing capacity globally per manufacturing site and forecasts of supply and demand (European Commission, 2023[77]).

Box 3.5. The European Shortages Monitoring Platform

Reporting under the European Shortages Monitoring Platform was launched in February 2025. The specific requirements vary depending on the applicable scenario. Under the scenario of “normal circumstances”, MAHs must report potential or actual shortages of centrally authorised products. Under the scenario of “MSSG-led preparedness”, closer monitoring is enacted. EMA’s Medicine Shortages Steering Group (MSSG) creates a specific list of medicines which defines the products and reporting frequency. Following the notification, all MAHs must report on centrally and nationally authorised products. Under the “crisis” scenario, extensive reporting is required. After the European Commission has recognised the crisis, EMA publishes a list of related critical medicines. Based on the products in scope and the reporting frequency defined by MSSG, MAHs and national competent authorities (NCAs) must report the supply, demand and availability of their respective medicinal products (EMA, 2026[3]).

The platform enables efficient information exchange between the EMA, NCAs and MAHs. To ensure interoperability with NCAs and industry stakeholders, the platform utilises a machine‑to-machine communication interface which enables automatic data submission. Manual tabular data submission is also possible. By offering three user interfaces for different stakeholders, secure access can be established, yet selected shortage information may also be communicated publicly by the EMA (EMA, 2025[78]).

The agreed CMA text adds a separate transparency mechanism for contingency stocks. It requires EMA to establish and maintain a digital platform providing an overview of contingency-stock requirements imposed by national law, including the critical medicinal products covered and the size of required stocks. When the Voluntary Solidarity Mechanism is activated, member states will have to provide stock data for relevant critical medicinal products that they identify as available for reallocation.

Another aspect is demand forecasting data. Forecasting demand for medicines can be complex, influenced by epidemiological trends, seasonality, and physician prescribing behaviour. A monitoring mechanism might incorporate external data like disease incidence (e.g. flu surveillance data to anticipate antiviral demand) or even novel sources (Google search trends, wastewater data for disease prevalence – as Pfizer did for COVID‑1921). At minimum, the system should capture historical consumption patterns as a baseline, so that it can detect when supply is deviating from expected demand. This implies collecting healthcare utilisation data (e.g. pharmacy sales or hospital dispensation data). Some of this can be gleaned from existing databases like the EU’s ESAC-Net dashboard (for antibiotic use) or national health insurance reimbursement data. Big Data analytics could then be applied to model future demand under various scenarios.

While the breadth of data to be collected seems daunting and may encounter many regulatory and technical challenges, one should keep in mind that effective demand forecasting does not require all existing data points. The experience of Pfizer is that efficient prediction models can be developed based on incomplete datasets or a selection of key variables.

The same applies to the monitoring of shortages. One has to decide how much data is enough to effectively predict or detect shortages.22 There is a danger of information overload if every data point on supply and all transactions along the value chain are collected “just in case.” One strategy could be to initially focus on exception reporting – i.e. set thresholds and only collect/flag data when something is out of normal range. For example, only alert if a manufacturer’s output falls X% below its normal level, or if a country’s stock falls below Y days of cover. This could reduce noise and respect competitive sensitivities by not exposing normal operations.

At the other extreme, a minimalist approach could be to rely mainly on declarations of shortage by companies (the status quo in many places) and simply improve the sharing of those notices internationally. However, that is a lagging indicator; by the time a formal shortage is declared, patients may already be affected. Indeed, current public datasets of shortages (such as the EMA’s shortages catalogue or ASHP’s list in the United States) are valuable, but they are essentially compilations of issues already manifest. Interoperability of these datasets (sharing shortage alerts among jurisdictions) is useful for managing crises – for instance, if the US FDA announces a shortage of a drug, EU officials would benefit from knowing in case the same product is used in Europe. But true prevention requires moving upstream to predictive data.

Therefore, designing data collection for an EU monitoring system likely means a multi-tiered approach: continuous collection of a few critical metrics (production, inventories, consumption rates) for critical MCMs, plus event-triggered reporting (e.g. an issue at a factory) and integration of external indicators for forecasting. Automation will be crucial – wherever possible, tapping into existing digital records (inventory databases, manufacturing systems, customs trade data, pharmacy dispensing data) will be more sustainable than manual reporting. The challenges are ensuring data quality (common definitions, avoiding double counting), timeliness (reporting lag must be short to enable early action), and securing industry co‑operation (firms must trust that proprietary data won’t be misused, as discussed next).

3.4.5. Incentives, obligations and public health benefits

For any monitoring system to succeed, both companies and governments must be willing participants. Creating the right balance between co‑operation, incentives and legal obligations is therefore a central and fifth key challenge. From the perspective of national governments and regulators, the incentive to join an international monitoring mechanism is the promise of fuller situational awareness beyond their own borders. No country wants to be caught off-guard by a MCM shortage that could have been anticipated by looking at supply-chain signals elsewhere. EU member states may benefit from knowing that, for example, production of a drug in Country X outside the EU has halted – allowing them to procure alternatives or urge ramp-ups before patients are affected.

A shared system can also foster solidarity: if Country A sees a surplus while Country B has a shortage, information sharing can enable voluntary reallocations. The agreed CMA text reinforces the EU’s existing Voluntary Solidarity Mechanism by providing for the exchange of relevant contingency-stock data and voluntary reallocation when that mechanism is activated. In essence, a collective monitoring provides an insurance mechanism for governments, helping prevent or mitigate crises that no single country can handle alone. Additionally, governments have an incentive to avoid purely national solutions that might conflict. If every country hoards stock or imposes export bans due to lack of information, the overall outcome is worse, as shortages are exacerbated elsewhere. A co‑operative monitoring approach backed by agreements (for instance, that countries will share data rather than suddenly cut off exports) can reduce mistrust among nations during emergencies. However, governments will only invest in such co‑operation if they trust the system and see reciprocal commitment.

For firms and supply chain actors, participation incentives and legal obligations in the name of public health benefits need careful calibration. The preferred model for firms would generally be one based on clear scope, limited reporting burdens, protection of confidential business information, and tangible operational benefits. Companies are understandably cautious about sharing data on production volumes, inventories, orders, sourcing strategies, manufacturing relationships or demand forecasts. Such information may reveal competitive strategies, bargaining positions or vulnerabilities. Firms may also be concerned that data provided for shortage prevention could later be used for other regulatory or enforcement purposes. These concerns are not only private‑sector preferences; they are also relevant to the credibility and effectiveness of the system. If firms believe that information will be misused, reported inconsistently or exposed to competitors, the quality and timeliness of reporting will suffer.

A co‑operative design can therefore improve data quality. Firms are more likely to provide useful information when the system is purpose‑limited, technologically neutral, proportionate and embedded in clear governance arrangements. The mechanism should define what information is required, from whom, at what frequency, at which level of aggregation, and under what confidentiality safeguards. It should also minimise duplication by reusing existing reporting channels where possible, including regulatory shortage notifications, manufacturing and quality data, stock and distribution information and relevant digital traceability systems. This is particularly important for small and medium-sized enterprises, wholesalers, pharmacies and other actors that may not have the same compliance capacity as large multinational manufacturers.

However, voluntary co‑operation alone is unlikely to be sufficient for an EU monitoring system whose purpose is to prevent or manage shortages with public health consequences. The public health benefits of early warning create a legitimate basis for requiring certain information from firms, especially for products included in a MCM list or products under preparedness or crisis monitoring. Legal obligations, as they already exist in current EU shortage prevention and preparedness mechanisms, may be necessary to ensure completeness, comparability and fairness across actors. Without a legal basis, firms that co‑operate may bear costs and disclose information while competitors do not, with authorities receiving partial data that understate risks.

Incentives remain important even where legal obligations exist. Reporting should not be experienced only as a compliance burden. Firms that provide early and accurate information should see that this information can trigger constructive public action. Possible responses include regulatory flexibility where safety and quality standards are maintained, support for alternative sourcing, facilitation of imports or public procurement arrangements that reward reliability of supply. In this sense, incentives and obligations are complementary.

During COVID‑19, for example, some regulators gave regulatory exemptions (like batch import waivers or adjustments to oxygen allocation quotas) when industry alerted them early to shortfalls (Bolislis et al., 2021[79]; Klein et al., 2022[80]). A formalised monitoring system could incorporate such triggered support: e.g. if a company’s data indicates a shortage risk, it might automatically qualify for certain regulatory flexibilities or assistance in sourcing alternatives (subject to oversight to prevent abuse).

Additionally, participating firms could receive aggregated market intelligence from the system. While company-specific data would remain confidential, the system could share back insights like overall demand trends or supply gaps that all firms can use for better planning (and that unlike supply data are not likely to infringe any competition rule). For example, if the data show EU-wide consumption of a drug class rising significantly above forecast, all manufacturers in that class can ramp up appropriately. Without a shared system, each firm might only see its own order data and miss the bigger picture.

Finally, reputational and legal incentives play a role. If the EU (and partner countries) strongly signal that ensuring medicine continuity is a priority, companies may face reputational damage if they are seen as non-co‑operative or secretive in the face of patient needs. Conversely, being a “reliable supplier” that works with authorities to avert shortages could be a market advantage.

In summary, a credible monitoring mechanism needs to balance private‑sector concerns with the public interest in preventing and mitigating shortages of critical medicines and medical countermeasures. It should combine trusted governance, proportionate mandatory reporting, meaningful incentives, safeguards for confidential information and escalation powers when patient access or health security is at risk.

3.4.6. Data privacy and security concerns

Last but not least, any public monitoring system for medical supply chains must uphold high standards of data privacy and security, especially in the EU context with stringent regulations such as the General Data Protection Regulation (GDPR). While much of the supply chain data (e.g. production volumes, stock levels) is not personal data, privacy in this context refers more broadly to the protection of confidential business information and the assurance of data sovereignty. Companies will be reluctant to share detailed supply data if there is a risk it could be exposed or misused (for example, competitors learning each other’s detailed supply or inventory positions). Likewise, governments will insist on cybersecurity measures given the sensitivity – a breach of a medicine supply database could potentially be exploited (e.g. by malicious actors knowing where shortages are emerging). The rise of cross-border data sharing intensifies these risks, driving many governments to implement or consider localisation measures to ensure access, oversight, and protection of sensitive data (OECD/WTO, 2025[81]).

While the monitoring system would not involve personal data, it should be noted that certain categories of non-personal industrial data may still be subject to protection under Union law, in particular where such data qualify as trade secrets within the meaning of Directive (EU) 2016/943. This is especially relevant for information on production volumes, stock levels, supplier relationships, orders, manufacturing locations, lead times or shipment flows. Protecting such information is not only a legal requirement; it is also necessary to ensure trust and the quality of data shared by companies and other supply chain actors.

One approach is to ensure that data are shared as far as possible in aggregated or anonymised form. For instance, the system’s wider outputs could indicate risk levels, percentage shortfalls or expected supply constraints without disclosing exact figures from individual companies. Detailed raw data could be accessible only to a limited number of authorised analysts under strict confidentiality or federated learning approaches could be used to avoid sharing centrally the sensitive information (see Section 3.4.4). Technical measures like data encryption, access control, and possibly homomorphic encryption (allowing computations on encrypted data) or secure multiparty computation can reinforce that even system operators cannot easily extract individual secrets. The principle of data minimisation from GDPR is useful: collect only what is necessary for the purpose. For example, if real-time granular data at batch level is not needed to predict shortages, then perhaps collecting weekly aggregate figures is sufficient and less invasive.

Whether a secure central platform, a trusted data intermediary or federated approaches are used, clear legal frameworks for data sharing are needed. The EU Data Act (Regulation EU 2023/2854, which has applied since September 2025) includes provisions for Business-to-Government (B2G) data sharing in situations of public emergency. A severe medicines shortage could arguably constitute such an emergency, activating mandatory data access. However, reliance on emergency provisions is reactive. It might be preferable to have an ongoing arrangement that defines what data companies should share routinely and how it will be safeguarded. Explicit clauses about protecting trade secrets and limiting use for public health purposes will be important to get buy-in.

Additionally, agreements with international partners for data exchange would need to address how data are protected on each side, respecting the EU’s high standards. Any global monitoring effort might consider using trusted intermediaries – an international organisation or independent non-profit structure that could receive certain data and provide anonymised global indicators, which might be easier for companies to accept than directly handing data to foreign governments. Part of building trust will also be to demonstrate that the platform is secure. The involvement of bodies like the European Union Agency for Cybersecurity (ENISA) could be considered.

In conclusion, addressing privacy and security is both an ethical obligation and a practical requirement to make the monitoring system viable. The system must only share information that is necessary and beneficial (e.g. warning signals) and do so in a way that preserves confidentiality of business-sensitive data. Modern data science approaches offer ways to achieve this balance. Successful track-and-trace implementations (like Türkiye’s) were facilitated by aligning with privacy laws and giving stakeholders confidence that data are secure.

3.5. Recommendations for improving medical countermeasures supply chain monitoring and enhancing international co‑operation

The EU has made significant progress in strengthening its capacity to monitor shortages of medical products. Regulatory obligations on MAHs, enhanced co‑operation through EMA and the development of EU-level platforms for information gathering, analysis and sharing -such as ATHINA (Box 3.6)- have the potential to improve transparency and crisis co‑ordination. As highlighted in Section 3.2, the EU also has a track-and-trace system in place that, while aimed at verifying the authenticity of medicines, can also offer a basis for further monitoring stocks and supply. However, it was not designed to provide full end-to‑end visibility (from KSMs to finished products) or anticipatory monitoring of MCMs.

Box 3.6. ATHINA: The EU operational platform for medical countermeasures preparedness and response

The Advanced Technology for Health Intelligence and Action IT System (ATHINA) is DG HERA’s information system supporting the preparedness, procurement and deployment of medical countermeasures at the EU level. Its role is operational rather than regulatory. ATHINA is designed to:

• Consolidate information relevant for medical countermeasures preparedness, including supply risks, procurement data and stockpile status;

• Support anticipatory decision making through early-warning indicators and scenario analysis;

• Facilitate co‑ordination between EU institutions, Member States and external partners during health emergencies.

ATHINA is still under development at the beginning of 2026 and does not replace existing regulatory systems managed by the EMA. Instead, it complements them by translating validated supply-chain signals into operational preparedness and response actions. By acting as a central integration and analysis layer, ATHINA can be seen as a key tool for DG HERA to move from reactive crisis management to proactive risk mitigation.

The success of ATHINA will depend on its integration with national IT infrastructures. The EU Interoperability with HERA’s IT Platform (EU-HIP) consortium addresses this issue, involving public health institutes, health authorities and ministries of health from 15 EU Member States.

Source: European Commission and EU-HIP consortium.

3.5.1. Limits of existing EU monitoring tools for medical countermeasures supply chains

Current EU monitoring mechanisms are primarily regulatory in nature. They focus on authorised medicinal products and rely on legally defined events, such as confirmed or anticipated shortages, manufacturing disruptions, or quality incidents. Information is typically collected once a risk has crystallised, through notifications submitted by firms or EU member states.

This approach is appropriate for regulatory oversight and co‑ordination, but it is not well suited for early detection of systemic risks affecting MCMs. Upstream vulnerabilities, such as dependence on single API suppliers, constrained manufacturing capacity, or logistical bottlenecks, may remain invisible until they materialise as shortages.

The current EMVS is not intended to monitor shortages or to provide a real-time tracking of medicines. It was established to protect patient safety and prevent falsification; it provides confirmation of product authenticity and legal supply, but does not offer continuous insight into inventories, flows or capacity constraints. While some EU member states and stakeholders have explored secondary uses of verification data for availability monitoring (see below), these initiatives remain partial and heterogeneous. At EU level, there is no systematic mechanism to translate traceability data into actionable preparedness signals.

Moreover, monitoring of MCMs currently spans multiple products (medicines, vaccines, devices, PPE), actors (manufacturers, distributors, health authorities) and data systems. Information is dispersed across regulatory databases, procurement systems, national reporting tools and ad hoc crisis instruments. As previously highlighted, the revised pharmaceutical legislation and the new Critical Medicines Act have the potential to strengthen the information base for critical medicines through vulnerability evaluations, targeted information requests, co‑ordination on available information about existing or planned manufacturing capacity, and greater transparency concerning national contingency-stock requirements.23 However, not all MCMs are critical medicines (in the sense of the EU Union list) and MCMs include medical devices and PPE for which shortage notifications, serialisation and other monitoring tools are not as developed as for medicines. These products are covered by the Emergency Framework Regulation and there are more shortage‑monitoring obligations for medical devices as part of EU Regulation 2024/1860 amending the Medical Devices Regulation, In Vitro Diagnostic Medical Devices Regulation and EUDAMED (the European Database on Medical Devices).

Still, the landscape remains fragmented, and this fragmentation complicates timely aggregation and interpretation of signals at EU level. It also limits the ability to assess cross-product or cross-country dependencies that are particularly relevant for MCMs during large‑scale health emergencies.

Lastly, most existing monitoring tools focus on supply-side events. Demand dynamics -such as sudden epidemiological shocks, changes in clinical practice or patients’ behaviour- are only indirectly captured. As a result, mismatches between supply and expected needs may only be detected late.

From a preparedness perspective, effective monitoring requires joint consideration of supply, demand and operational constraints, including logistics, deployment timelines and stockpile mobilisation. This is how J&J or Pfizer address resilience in their own supply chains (see Section 3.3).

These limitations in existing EU monitoring tools do not reflect weaknesses of individual instruments, but rather the fact that current tools were mostly developed for regulatory co‑ordination, not for anticipatory crisis preparedness. Moreover, mechanisms are not as developed or comprehensive for all types of MCMs. For MCMs, the EU requires a complementary capability focussed on: i) early identification of emerging supply chain risks; ii) integration of heterogeneous data sources; and iii) translation of signals into operational decision making. The rest of the section introduces recommendations to further develop these capabilities, building on existing EU tools and initiatives, as well as reinforcing international co‑operation.

3.5.2. Leverage and extend existing frameworks for the early detection of emerging supply chain risks

As highlighted in the previous section, the EU is not starting from scratch and despite the fragmentation of tools, there are proven systems and technologies already in place for critical medicines, if not for all categories of MCMs.

The EMVS set up under the FMD, for example, already connects thousands of manufacturers, wholesalers, and pharmacies across Europe via a shared database and 2D barcode scans. It was not designed to monitor shortages, but with some adaptation, EMVS data could feed shortage monitoring: for example, spikes or drops in the decommissioning (dispensing) rates of a product could indicate abnormal demand or supply issues. Patterns in “alerts” (failed verification scans) might even reveal distribution bottlenecks. Importantly, the FMD repository holds data on each pack’s status (active, dispensed, expired, etc.). By aggregating this, one could estimate stocks in the system (Box 3.7). Integrating this with other data (like manufacturer-reported supply) would greatly enrich the picture.

Supply chain stakeholders have mixed views on extending the FMD to a full track-and-trace system. Manufacturers generally support the extension of the verification requirement as they would benefit from better visibility of the distribution channels and they have already invested in systems that provide the required information (Medicines for Europe, 2023[82]; EFPIA, 2024[83]). Wholesalers oppose this as scanning all products would be impractical and costly. Pharmacists are divided on the topic. While it may support public health, additional reporting obligations and a potential loss of control over their data are feared (European Commission, 2023[84]). Ten out of the 19 surveyed National Competent Authorities (NCAs) where explicitly in favour of extending the monitoring system while only one was explicitly against it (European Commission, 2023[84]).

It should also be noted that while it is theoretically possible to calculate available stock levels, gaps in the verification chain would remain and the exact location of the products would be unknown. Further, wholesalers may hold back products in certain regions which may cause a shortage in another region, even though the stock at national level is sufficient to satisfy demand. A reliable monitoring system for the purpose of preventing shortages therefore requires more information and analysis than are currently available under the FMD. To monitor shortages of MCMs effectively, the scope would also need to be extended to further products.

Box 3.7. Calculation of net stocks of medicines using FMD repositories

National repositories set up in the context of the EU FMD include enough data at the product level to calculate net stocks of prescription medicines in each Member State. The formula below is based on variables that are all available in the national repositories and accessible to National Competent Authorities.

$National\mathrm{ }stock\mathrm{ }level=\mathrm{ }{U}_{\left(m\right)}+U_{\left(pt\right)}-D_{\left(n\right)}-E_{\left(n\right)}-{IMF}_{\left(o\right)}$

Where:

$U_{\left(m\right)}=$ number of unique identifiers uploaded into the national market by the original manufacturer (i.e. number of physical packs released on the market).

$U_{\left(pt\right)}=$ number of unique identifiers uploaded into the national market by parallel traders (i.e. number of physical packs imported in the national market).

$D_{\left(n\right)}=$ number of unique identifiers decommissioned at national level (i.e. number of physical packs which have been dispensed to patients within the territory).

$E_{\left(n\right)}=$ number of unique identifiers decommissioned for export at national level by parallel traders (i.e. number of physical packs exported from the national market).

${IMF}_{\left(o\right)}=$ number of unique identifiers destined for the national market but decommissioned in other markets via the Internal Market Functionality of the system (an Internal Market Functionality is triggered when a medical product destined for market X has been decommissioned for dispense in market Y, without having been parallel traded).

As the National Competent Authorities have access to the data contained in the system via an API, this calculation can be done in real time. There are, however, different factors that can complicate the above calculation, such as the issue of multi-market packs or multi-source products.

Source: Bouvy and Rotaru (2021[20]), “Medicine Shortages: From Assumption to Evidence to Action – A Proposal for Using the FMD Data Repositories for Shortages Monitoring”, https://doi.org/10.3389/fmed.2021.579822.

The other existing tool is ATHINA (Box 3.6), which is already a platform focussed on MCMs, and would be the most relevant to receive information for an extended monitoring. It is not a track-and-trace tool and it should not duplicate existing regulatory or verification systems. But it should interface with them, extracting risk-relevant signals and translating them into preparedness-oriented outputs (See Section 3.5.4). What seems to be missing in the current EU framework is a risk-based and proportionate monitoring that would support ATHINA by providing:

• An early-warning system for MCMs based on threshold-based indicators and escalation triggers (rather than continuous high-granularity reporting);

• A dynamic assessment of upstream dependencies with high systemic relevance (to cover the main inputs of MCMs);

• Information on stocks and international flows relevant for emergency response. This information would come from external sources that would be developed either within the EU or through international co‑operation and private‑public co‑operation.

Such monitoring is technologically feasible but there are important institutional and operational challenges, starting with the architecture of the system that should integrate heterogeneous data sources from various countries and companies, while providing incentives for all stakeholders to co‑operate and share information.

3.5.3. Evaluate alternative system architectures to integrate cross-border data sources

Different systems can be considered for the extended monitoring and early-warning system envisaged in the previous section for MCMs. For a monitoring system that would not be limited to the EU and would be able to integrate information from different jurisdictions, we can outline three potential architectures, each with pros and cons.

A “world control tower”: Centralised global monitoring

In this model, a single international platform (possibly managed by an entity like an international organisation or a coalition of regional agencies) collects data from all participants and provides a global dashboard of supply and demand for MCMs. Alerts would be generated centrally and disseminated globally based on a collection of agreed data, such as production figures, shipment tracking, inventory levels, and health data. The strength of this approach is complete visibility – because all data is aggregated, the system can use powerful analytics to detect issues and co‑ordinate responses in one place. It also simplifies standard-setting and interoperability by having a unique agreed IT system.

However, the weaknesses are significant. It concentrates a lot of sensitive information in one system, raising trust and security issues. It is feasible within the EU but at a more global level nations may be reluctant to cede control to a single “tower” that might be outside their jurisdiction. Also, ensuring that all relevant information is regularly collected from every stakeholder requires massive investment and compliance. Feasibility-wise, this might be something to aspire to long-term, perhaps for a subset of products needed in a global health emergency. But in the short term, it is likely too ambitious and not feasible. As exemplified with TradeLens, it is very hard to bring the entire ecosystem together. Unless there is unprecedented international consensus, this model could falter due to lack of buy-in.

Privacy-preserving distributed architecture: Shared analytics with controlled data access

This model builds on the privacy-preserving approaches previously discussed, including but not limited to federated learning. Instead of pooling all data by default, participants (companies, EU member states or third-country partners) keep most data within their own systems and share agreed outputs, indicators or selected datasets with a central hub. The architecture could combine several tools: federated analytics for specific risk models, aggregated reporting for routine monitoring, secure data spaces or trusted intermediaries to manage permissions and control access to more granular data when justified. For instance, companies or national systems could regularly transmit standardised indicators on stocks, demand, production, interruptions or lead times, while more detailed information would be available only under predefined escalation conditions.

The strengths of this model are in data protection and scalability. It avoids the continuous centralisation of all commercially sensitive data, while still allowing authorities to build a collective picture of emerging risks. This can encourage participation since information does not need to be systematically transferred to a central database. It can also be adapted to different levels of co‑operation: within the EU where ATHINA and related systems may require more granular information for preparedness and crisis response; and internationally, where partners may initially be willing to exchange only indicators, alerts or anonymised data trends.

On the downside, the weaknesses include technical complexity, uneven data quality and the risk that insufficiently granular information may lead to reactive rather than preventive action. Each participant must apply compatible definitions, reporting formats and security standards. Where only alerts or high-level risk scores are exchanged, important early signals may be missed, and authorities may lack the information needed to plan mitigation measures before a shortage materialises. Co‑ordination can also be difficult and even aggregate indicators or alerts can be commercially or politically sensitive.

This model is more realistic than a fully centralised global control tower, but it would need to clearly distinguish between routine monitoring, preparedness analysis and crisis management. It could be piloted first for a limited set of MCMs, combining regular aggregate reporting with secure access to granular data when public health needs require it.

Co‑operative framework: Decentralised, interoperable national systems with information exchange

In this architecture, each country or region maintains its own monitoring system (track-and-trace, shortage databases, etc.), but they agree to share information through a co‑operative framework. Here, the EU would have its system, the United States would have its own, etc., and a protocol (or alliance) would allow them to notify each other of certain triggers. For instance, regulators might exchange shortage notifications in real time or allow querying of each other’s databases for specific critical items.

The strength of this model is that it respects sovereignty and existing investments – nobody has to give up their system, they just have to make them talk to each other. It could be built on mutual recognition agreements or data-sharing memorandums of understanding (MoUs). It also localises data governance issues; each system follows its national laws and only agreed summary data crosses borders. This model might be more politically palatable and easier to start with – e.g. the EU and a few non-EU regulatory agencies could pilot a secure info exchange on shortages and supply status for a few MCMs of mutual interest, building trust gradually.

However, the weakness is that it may be slower and less comprehensive. If every regulator only shares what they choose, some early signals could fall through the cracks. It may end up being more reactive – basically a network of alert systems rather than one proactive system. Interoperability issues are also non-trivial: different data formats and terminologies need translation. And in a crisis, the co‑ordination might still be clunky if there isn’t a central command. The feasibility of this approach is relatively high in incremental steps (we already see ad hoc exchanges during crises, and forums like ICMRA that facilitate information sharing). It might not prevent all shortages, but it can improve collective response. Over time, this could evolve towards more integration if trust deepens.

In practice, the optimal solution might combine elements of these models. For example, the EU could implement a regional control tower (central within the EU) that communicates with other regional towers (North America, Asia, etc.), thus mixing model 1 and 3. Or a privacy-preserving architecture (model 2) could exist within the EU to gather industry data on specific MCMs, and then the EU shares aggregated outcomes with other jurisdictions (model 3 externally). It is not either/or. A clear-eyed analysis of strengths, weaknesses, and practical feasibility (including cost and governance complexity) should inform which model to pursue. The centralised approach is powerful but demands high trust and considerable political will; the decentralised and co‑operative approaches are more immediately achievable but less comprehensive. A staged strategy could start with enhancing exchange of shortage alerts (co‑operative framework), then progress to sharing of certain data or models (distributed networks), aiming ultimately for a more unified near real-time oversight via a central hub as confidence and infrastructure grow.

To make any of these architectures work in practice, policy tools to enhance co‑operation are essential. One key lever is incorporating supply chain transparency provisions and information exchange mechanisms into trade agreements, mutual recognition agreements or new types of bilateral or regional agreements covering MCMs, such as strategic partnerships (see Chapter 2). The provisions to be explored are the following:

• Commitments to adopt global medicine traceability standards (such as GS1 or ISO) and to co‑operate on interoperability;

• Mutual recognition of track-and-trace data;

• Commitments to share critical supply data during emergencies;

• Timely notification of supply disruptions among parties;

• Joint stress testing and scenario planning.24

Such commitments are likely to be non-binding and are useful only if there is some willingness to co‑operate. This is why international co‑operation also requires some type of forum to bring together governments and the industry in a co‑operative network working on the improved monitoring of MCM supply chains. Whether it is under the umbrella of WHO, OECD, the G20 or any other governmental organisation, the inclusion of industry is vital. An international version of the EU Critical Medicines Alliance would need to support discussions to yield the best solutions.

Specific trade‑oriented co‑operation might also be considered as part of a plurilateral agreement at WTO-level on essential goods supply chain transparency – akin to an “agreement on medical supply chain cooperation” where signatories agree to certain principles (e.g. no export bans without notification, sharing of shortage data, co‑operation in emergencies). While WTO agreements are challenging in the current international context, a coalition of willing countries could start it outside WTO and later multilateralise it.

Another existing framework is the WHO’s global monitoring initiatives. WHO manages databases for substandard/falsified medical product alerts and has been working on a more systematic approach to shortages. It has issued guidance encouraging countries to monitor and report shortages.25 While a global monitoring platform or dashboard is not discussed in the context of WHO, it is also a possible avenue for efforts to connect existing early warning systems and shortage reports globally.

Lastly, interoperability pilots should be funded. The EU could co-finance projects to connect the EU’s system with tracking systems or shortage monitoring platforms in third countries, whether under development or already implemented. For partners in developing countries, this could be part of development aid where the EU supports financially the establishment of medicine verification systems while creating a wider network of surveillance for emerging issues. If a shortage starts in one region, it often propagates globally. Early awareness is key and information from smaller partners in the developing world could also provide relevant alerts and data. Prospective EU member countries are also candidates for enhanced co‑operation in this area (DG HERA, 2025[85]).

3.5.4. Translating signals into operational decision making: Enhancing predictive analytics and demand forecasting

Experience from private supply-chain platforms and recent empirical research shows that early-warning signals can be generated from partial and imperfect data, provided that analytical tools are appropriately designed. For public authorities, the objective is not the granular tracking of individual products, but the identification of abnormal patterns and emerging stress.

As it is done by USP in its Medical Supply Map, DG HERA could further invest in models that can effectively detect anomalies and generate shortage risk indices for MCMs, updated regularly. Through ATHINA, it will have access to multiple sources of information and types of data that can be used to apply modern AI techniques and predictive analytics. One lesson from the work of USP is that the risk assessment should be based on health and supply chain expertise, rather than just data science. The AI and Big Data experts need to closely work with regulators, health authorities and stakeholders with experience in the manufacturing and distribution of MCMs.

In practical terms, ATHINA could progressively further integrate:

• Trend monitoring on supply, stock and utilisation indicators, with the challenge of going beyond information collected within the EU to better cover supply chains upstream;

• Anomaly detection to flag deviations from expected patterns;

• Scenario analysis to assess the system-wide implications of specific disruptions.

Big data can help answer “what-if” scenarios: e.g. if one supplier goes offline, can global supply from others fill the gap? Scenario modelling with AI could identify where vulnerabilities would bite hardest, allowing pre‑emptive actions (like temporarily boosting stock of substitutes). Partnering with external technology providers or developing its own capacity in this area will be crucial for EMA and HERA to foster preparedness.

While much of the chapter has focussed on tracking supply and identifying disruptions upstream, an equally important component is forecasting demand to know whether supply will meet needs. Improved demand forecasting can turn data into actionable prevention: it’s not enough to see that X units of drug are available – one must judge if X is sufficient given likely patient needs. Therefore, a recommendation is to invest in big data and AI-driven forecasting as a parallel initiative to the supply monitoring, learning from the private sector experience.

Modern AI techniques can dramatically improve demand predictions by analysing diverse data sources. Healthcare utilisation data (like prescriptions, hospital admissions, epidemiological surveillance) is fundamental. Machine learning models have been successfully trained on pharmacy dispensation data to predict drug shortages in Canada, accurately flagging risks for the majority of drugs in test cases (Pall et al., 2023[86]). Such models consider historical usage patterns and can learn the precursors to a shortage (e.g. a slow decline in supplies or unusual spikes in use). The EU could aggregate de‑identified prescribing data through its health data space initiative and feed that into forecasting models. Additionally, real-time indicators can be incorporated: for example, Google search trends for flu symptoms can predict flu outbreaks ahead of clinical reports, which in turn predicts surge demand for antivirals and fever reducers. Social media sentiment or news feeds might foreshadow panic buying or public health threats that drive demand.26 Another big data source is logistics data – how quickly products are moving through distribution. AI can pick up subtle changes like longer replenishment cycles or inventory turnover rates which might indicate incipient shortages.

To harness AI effectively, data sharing is again crucial – particularly sharing of demand data across borders. While concrete steps for EMA and DG HERA could involve investing in people and skills (e.g. hiring data scientists, building computational models, creating networks of experts, launching pilot projects), the successful deployment of AI technologies would also involve data agreements and data sharing mechanisms as discussed in the previous sections.

Finally, these forecasting tools should loop back into strategic stockpile management. If an AI model predicts a likely shortage, that could trigger pre‑emptive releases from stockpiles or reallocation of manufacturing resources (e.g. asking a company to prioritise a predicted shortfall drug). Over time, as models improve, the hope is to transition from reacting to shortages to preventing them entirely by adjusting supply chain inputs ahead of time. This requires again strong public-private co‑operation and platforms to exchange data. ATHINA already has collaboration features to jointly work with external stakeholders that can be used in this context.

To conclude, the concrete recommendation would be for DG HERA to establish a dedicated Supply and Demand Forecasting Centre that uses big data and AI to support the EU monitoring system. This centre can collaborate with academia and industry data teams to continuously refine predictive models. It should also engage internationally to share methodology and even jointly develop global predictive dashboards. Given that demand surges can be global (as in a pandemic) or regional, a network of forecasting centres sharing insights could dramatically enhance preparedness.

3.6. Conclusion on new technologies and innovations to better monitor medical supply chains

The growing frequency and severity of medicine shortages in Europe and globally underscore the urgent need for a more co‑ordinated, predictive and transparent approach to monitoring MCM supply chains. The European Union, with its regulatory capabilities, digital infrastructure, and long-standing commitment to cross-border health co‑operation, is well positioned to not only create efficient intra-EU monitoring mechanisms but also to take a leading role in promoting international co‑operation and supply chain monitoring at a more global level.

This chapter has outlined the challenges in designing a public monitoring system for MCMs covering the whole supply chain, both geographically (including beyond the EU) and in terms of products (covering not only final pharmaceutical products but also APIs, critical components and raw materials, as well as medical devices and PPE).

The main challenges are the lack of globally harmonised serialisation standards, the fragmented and often non-interoperable technology landscape, the difficulty of including upstream inputs and raw materials in monitoring systems, and the scarcity of real-time data on production, inventories and demand. Overcoming these challenges requires not only innovative technological tools but also regulatory co‑ordination, political commitments, the right balance between incentives and legal obligations, and stakeholder trust.

It is clear that there is no one‑size‑fits-all solution and that different system architectures (centralised platforms, distributed networks, and interoperable national systems) will have to be combined, as each offers distinct trade‑offs in terms of data privacy, feasibility and predictive capacity. Whichever model is pursued, progress will depend on embedding strong incentives for participation and aligning data governance frameworks across jurisdictions. Both with firms and other governments, the challenge lies in finding the right mix of partners to turn good ideas into operational solutions.

Moreover, the integration of big data and artificial intelligence to improve demand forecasting is no longer a luxury left to advanced private firms but a necessity for public authorities to pre‑emptively address mismatches between supply and need. Coupling a public supply monitoring mechanism with cutting-edge demand forecasting would give policymakers a powerful toolkit. It is the combination of visibility (knowing where things are and how much) and predictive insight (knowing what will be needed and where) that truly enables proactive management of medical supply. The recommendations in this chapter aim to move the EU and its partners towards that comprehensive capability, through smart use of technology, co‑operative governance, and strategic policy measures. By addressing the challenges outlined – standardisation, interoperability, data governance, the balance between obligations and incentives, escalation mechanisms when there is a crisis –, the EU can shift from shortage response to shortage prevention for MCMs, strengthening both health security and policy credibility.

Main recommendations of the chapter

• Complement ATHINA with early-warning indicators covering the whole supply chain, first for a limited set of high-priority MCMs and on a pilot basis. This new monitoring mechanism would build on existing EU tools and additional data from public and private sources, involving public-private and international co‑operation.

• Build dedicated analytical capacity within DG HERA, focussed on forecasting, scenario analysis and risk assessment. Partnering with external technology providers can also support this objective.

• Develop confidentiality-preserving data sharing arrangements with industry and distributed network approaches, focussing on risk signals rather than raw commercial data for routine monitoring and incorporating escalation mechanisms to access more granular data when there is a crisis.

• Explore how ATHINA could also be a co‑ordination node for international co‑operation, supporting information exchange with trusted partners during health emergencies and contributing to a global early-warning platform developed with these partners.

References

[33] Ahmed, W. and A. Rios (2022), “Digitalization of the international shipping and maritime logistics industry”, in The Digital Supply Chain, Elsevier, https://doi.org/10.1016/b978-0-323-91614-1.00018-6.

[26] Altunkan, S. et al. (2012), Turkish Pharmaceuticals Track & Trace System.

[14] Association Pharmaceutique Belge (2018), “Implementation of the Falsified Medicines Directive in Belgium: from a linear (unique) barcode to a 2D-datamatrix”, https://www.cibh.be/uploads/files/presentations/20181029-datamatrix1542269437.pdf.

[71] Behnam, M. et al. (2020), COVID-19: Overcoming supply shortages for diagnostic testing, McKinsey & Company.

[67] Bersin, A., P. Swartz and L. Karlsson (2025), “Managing Supply and Value Chains in a New Era of Global Trade: The Promise of Federated Learning”, World Customs Journal, Vol. 19/1, https://doi.org/10.55596/001c.133998.

[79] Bolislis, W. et al. (2021), “Regulatory Agilities in the Time of COVID-19: Overview, Trends, and Opportunities”, Clinical Therapeutics, Vol. 43/1, pp. 124-139, https://doi.org/10.1016/j.clinthera.2020.11.015.

[20] Bouvy, F. and M. Rotaru (2021), “Medicine Shortages: From Assumption to Evidence to Action - A Proposal for Using the FMD Data Repositories for Shortages Monitoring”, Frontiers in Medicine, Vol. 8, https://doi.org/10.3389/fmed.2021.579822.

[44] Brennan, C. (2025), “Control Tower Technology Value Matrix 2025”, Nucleus Research, https://dam.infor.com/api/public/content/da3be50a26574c09b82476bd9a567a92?v=406efd85.

[27] Bulut, E. and E. Bilgener (2022), “Evaluation of the Turkish Track and Trace System (ITS): perspective of community pharmacists”, Brazilian Journal of Pharmaceutical Sciences, Vol. 58, pp. 1-8, https://doi.org/10.1590/s2175-97902022e20407.

[2] Chapman, S., G. Dedet and R. Lopert (2022), “Shortages of medicines in OECD countries”, OECD Health Working Papers, No. 137, OECD Publishing, Paris, https://doi.org/10.1787/b5d9e15d-en.

[10] Cohen, M. and C. Tang (2024), The role of AI in developing resilient supply chains, Georgetown Journal of International Affairs, February 5, 2024, https://gjia.georgetown.edu/2024/02/05/the-role-of-ai-in-developing-resilient-supply-chains/.

[75] Council of the European Union (2026), Proposal for a Regulation of the European Parliament and of the Council establishing a framework to strengthen the security of supply and teh availability of critical medicinal products, ST 11059/26, 30 June, https://data.consilium.europa.eu/doc/document/ST-11059-2026-INIT/en/pdf.

[85] DG HERA (2025), Pharmaceutical industry study in prospective EU member countries, European Commission.

[52] DHL (2025), Smart printables: What the future of smart labels means for logistics, https://www.dhl.com/global-en/delivered/innovation/from-smart-labels-to-smart-printables.html.

[53] DHL (2020), Delivering pandemic resilience. How to secure stable supply chains for vaccines and medical goods during the COVID-19 crisis and future health emergencies, DHL White Paper.

[54] DHL (2020), Next-generation wireless in logistics, DHLTrend Research, https://www.dhl.com/content/dam/dhl/global/core/documents/pdf/next-generation-wireless-in-logistics.pdf.

[21] EAEPC, EFPIA, GIRP, PGEU (2012), Coding & serialisation - Delegated Act on the detailed rules for a unique identifier for medicinal products for human use, and its verification, https://health.ec.europa.eu/document/download/c4a93148-8f85-4994-ad87-233412fc5126_en?filename=efpia-eaepc-girp-pgeu_en.pdf.

[83] EFPIA (2024), Medicine shortages. EFPIA proposal for action.

[63] EMA (2026), Data on medicines (ISO IDMP standards): Overview, European Medicines Agency, https://www.ema.europa.eu/en/human-regulatory-overview/research-development/data-medicines-iso-idmp-standards-overview.

[3] EMA (2026), “European Shortages Monitoring Platform (ESMP)”, https://www.ema.europa.eu/en/human-regulatory-overview/post-authorisation/medicine-shortages-availability-issues/european-shortages-monitoring-platform-esmp.

[70] EMA (2026), “Union list of critical medicines”, https://www.ema.europa.eu/en/human-regulatory-overview/post-authorisation/medicine-shortages-availability-issues/availability-medicines-during-crises/union-list-critical-medicines.

[76] EMA (2025), European Shortages Monitoring Platform (ESMP), https://www.ema.europa.eu/en/human-regulatory-overview/post-authorisation/medicine-shortages-availability-issues/european-shortages-monitoring-platform-esmp.

[78] EMA (2025), “European Shortages Monitoring Platform (ESMP) Informational brief: why, who, what, when, where, how”, https://www.ema.europa.eu/en/documents/other/european-shortages-monitoring-platform-esmp-informational-brief_en.pdf.

[73] European Commission (2026), “Comprehensive 2025 Health Threat Prioritisation Assessment for medical countermeasures”, Commission Staff Working Document, https://health.ec.europa.eu/document/download/147630e7-835a-40cc-96d4-e0dc44707396_en?filename=hera_wd-health-treat-prioritisation-assessment_en.pdf.

[74] European Commission (2025), “Proposal for a regulation of the European Parliament and of the Council laying a framework for strengthening the availability and security of supply of critical medicinal products”, https://health.ec.europa.eu/document/download/2abe4fc8-059e-47d9-a20a-d9e3bfc5dc2c_en?filename=mp_com2025_102_act_en.pdf.

[13] European Commission (2024), Study supporting the report on trends in the falsification of medicinal products and measures provided according to Directive 2011/62/EU, Publications Office of the European Union, Luxembourg, https://health.ec.europa.eu/publications/study-supporting-report-trends-falsification-medicinal-products-and-measures-provided-according_en.

[77] European Commission (2023), “Proposal for a Regulation of the European Parliament and of the Council laying down Union procedures for the authorisation and supervision of medicinal products for human use and establishing rules governing the European Medicines Agency”, https://eur-lex.europa.eu/legal-content/EN/TXT/HTML/?uri=CELEX:52023PC0193.

[84] European Commission (2023), Study supporting the report to the European Parliament and to the Council on trends in the falsification of medicinal products and measures provided according to Directive 2011/62/EU, Publications Office of the European Union, https://health.ec.europa.eu/document/download/59efb019-57a1-4161-99a2-9b5cc63d3d26_en?filename=trends_medicinal-products-falsification_study_en.pdf.

[89] European Commission (2022), Future-proofing pharmaceutical legislation - Study on medicine shortages: Country fiches, European Commission - Directorate-General for Health and Food safety, Brussels, https://op.europa.eu/en/publication-detail/-/publication/f54a3d7c-7743-11ec-9136-01aa75ed71a1/language-en.

[41] European Pharmaceutical Manufacturer (2025), Pfizer Global Supply embarks on digital transformation, European Pharmaceutical Manufacturer, https://pharmaceuticalmanufacturer.media/pharma-manufacturing-news/latest-pharmaceutical-manufacturing-news/pfizer-global-supply-embarks-on-digital-transformation (accessed on 20 July 2025).

[64] FDA (2022), Identification of Medicinal Products (IDMP), U.S. Food & Drug Administration, https://www.fda.gov/industry/fda-data-standards-advisory-board/identification-medicinal-products-idmp.

[11] Ferencz, J., J. López González and I. Oliván García (2022), “Artificial Intelligence and international trade: Some preliminary implications”, OECD Trade Policy Papers, No. 260, OECD Publishing, Paris, https://doi.org/10.1787/13212d3e-en.

[72] Fritz, J. et al. (2022), “Syringes must be prioritized globally to ensure equitable access to COVID-19 and other essential vaccines and to sustain safe injection practices”, Human Vaccines & Immunotherapeutics, Vol. 18/7, https://doi.org/10.1080/21645515.2022.2077580.

[34] George, B. (2025), “Blockchain in Global Trade: Insights from the TradeLens Experiment”, Center for Open Science, Vol. No. apsz9_v1, https://osf.io/download/67bd5afb3f268cc20855646d/.

[17] Greece (2004), “Joint Ministerial Decision ΔΥΓ3/ΓΠ 7139: [Authenticity label of medicinal products]”, Government Gazette Β’ 1202/04.08.2004, https://www.e-nomothesia.gr/kat-ygeia/farmakeia/ya-dug-3gp-71349-2004.html.

[59] GS1 (2023), GS1 Healthcare Strategy 2023–2027, GS1 Global Office, https://www.gs1.org/docs/healthcare/Strategy/GS1‑Healthcare‑Strategy‑Final.pdf.

[65] GS1 (2021), Position paper: GS1 and ISO IDMP, https://www.gs1.org/docs/healthcare/position-papers/IDMP-Position-Paper-June-2021.pdf.

[88] GS1 (2018), “GS1 DataMatrix Guideline: Overview and technical introduction to the use of GS1 DataMatrix”, https://www.gs1.org/standards/gs1-datamatrix-guideline/25#1-Introduction-to-GS1-DataMatrix+1-2-Technical-characteristics.

[31] GS1 Canada (2025), 2D DataMatrix Implementation Roadmap, GS1 Canada, https://gs1ca.org/gs1ca-components/documents/2D_DataMatrix_Implementation_Roadmap.pdf.

[30] GS1 Canada (2025), “GS1 DataMatrix Barcodes on Pharmaceuticals in Canada”, Implementation Roadmap, https://gs1ca.org/pharmacy/roadmap/ (accessed on 1 July 2025).

[60] GS1 UK (2025), “GS1 barcodes have made patients safer for 20 years”, https://www.gs1uk.org/insights/news/GS1-barcodes-have-made-patients-safer-for-20-years.

[62] GS1 US (2025), Benefits of EPCIS, GS1 US, https://documents.gs1us.org/adobe/assets/deliver/urn:aaid:aem:8b0c1f79-eed3-4371-aad0-c72470a78f42/Benefits-of-EPCIS.pdf.

[15] Honeywell (2009), “Belgian Pharmaceutical products / CNK code / Pharmanet”, https://sps-support.honeywell.com/s/article/Belgian-Pharmaceutical-products-CNK-code-Pharmanet.

[51] IBM (2007), DHL breaks new ground with RFID-based real-time tracking of sensitive shipments, IBM Corporation, https://public.dhe.ibm.com/software/solutions/pdfs/ODC00298-USEN-00.pdf.

[43] Infor Nexus (2024), “Supply chain resiliency: Modern technology helps direct smart supply chain orchestration” White Paper, https://dam.infor.com/api/public/content/56211fcaf6994f49b2552cdd6d2edd93?v=1bde5b22.

[32] Institute for Safe Medication Practices Canada (2012),  Canadian Pharmaceutical Bar Coding Project: Joint Technical Statement (Version II) on Pharmaceutical Automated Identification and Product Database Requirements, https://www.ismp-canada.org/barcoding/download/JTSv2/JTSv2.pdf.

[18] Italian Medicines Agency (2020), “The Italian drug traceability system: The bollino and pharmaceutical verification”, https://www.aifa.gov.it/documents/20142/1267542/Traceability2020_WEB.pdf.

[61] Jayaraman, R. et al. (2011), “Data Standards in Healthcare Supply Chain Operations”, Proceedings of the 2011 Industrial Engineering Research Conference, https://jenpazour.wordpress.com/wp-content/uploads/2019/05/data-standards-in-healthcare-supply-chain-operations-ierc-2011.pdf.

[35] Jovanovic, M. et al. (2022), “Managing a blockchain-based platform ecosystem for industry-wide adoption: The case of TradeLens”, Technological Forecasting and Social Change, Vol. 184, p. 121981, https://doi.org/10.1016/j.techfore.2022.121981.

[23] Kannarkat, J., M. Denham and A. Sarpatwari (2024), “Improving Drug Supply Chain Security”, JAMA Health Forum, Vol. 5/1, p. e234819, https://doi.org/10.1001/jamahealthforum.2023.4819.

[56] Klein, K. and P. Stolk (2018), “Challenges and Opportunities for the Traceability of (Biological) Medicinal Products”, Drug Saf 41, pp. 911-918, https://doi.org/10.1007/s40264-018-0678-7.

[80] Klein, K. et al. (2022), “Regulatory Flexibilities and Guidances for Addressing the Challenges of COVID-19 in the EU: What Can We Learn from Company Experiences?”, Therapeutic Innovation & Regulatory Science, Vol. 56/2, pp. 366-377, https://doi.org/10.1007/s43441-022-00383-3.

[7] Klotz, S., S. Barr and J. Sotelo (2024), Digitalization is disrupting global trade - here’s how AI can help customs and businesses to respond, World Economic Forum, https://www.weforum.org/stories/2024/06/ai-is-already-helping-businesses-to-buy-and-sell-internationally-heres-how-technology-can-disrupt-global-trade-even-more/.

[46] Levinas-Ménard, L. (2025), “Review of Infor, Supply Chain Management Software Vendor”, Lokad, https://www.lokad.com/review-of-infor-com/.

[68] Li, L. et al. (2020), “A review of applications in federated learning”, Computers & Industrial Engineering, Vol. 149, p. 106854, https://doi.org/10.1016/j.cie.2020.106854.

[36] Louw-Reimer, J. et al. (2021), “Boosting the Effectiveness of Containerised Supply Chains: A Case Study of TradeLens”, in Progress in IS, Maritime Informatics, Springer International Publishing, Cham, https://doi.org/10.1007/978-3-030-72785-7_6.

[38] Maersk (2022), A.P. Moller - Maersk and IBM to discontinue TradeLens, a blockchain-enabled global trade platform, Press release, https://www.maersk.com/news/articles/2022/11/29/maersk-and-ibm-to-discontinue-tradelens.

[6] McKinsey (2024), Digital twins: The key to unlocking end-to-end supply chain growth, QuantumBlack AI by McKinsey, https://www.mckinsey.com/capabilities/quantumblack/our-insights/digital-twins-the-key-to-unlocking-end-to-end-supply-chain-growth.

[4] McKinsey (2022), The state of AI in 2022 -and a half decade in review, QuantumBlack AI by McKinsey, https://www.mckinsey.com/capabilities/quantumblack/our-insights/the-state-of-ai-in-2022-and-a-half-decade-in-review.

[82] Medicines for Europe (2023), Revision of the pharmaceutical legislation, Position paper, July 2023, https://www.medicinesforeurope.com/wp-content/uploads/2024/01/Medicines-for-Europe-Position-paper-_-Pharmaceutical-Legislation-FINAL-1.pdf.

[19] Merks, P. et al. (2022), “Falsified Medicines Directive in a Secondary Care Environment—Impact on Supply Chain”, International Journal of Environmental Research and Public Health, Vol. 19/6, p. 3276, https://doi.org/10.3390/ijerph19063276.

[5] Miller, J. (2025), New AI-enabled technology is 98% accurate at spotting illegal contraband, TuftsNow, https://now.tufts.edu/2025/01/09/new-ai-enabled-technology-98-accurate-spotting-illegal-contraband.

[49] Muppalla, A., B. Devi Maddi and N. Maddi (2025), “Artificial Intelligence in Regulatory Compliance: Transforming Pharmaceutical and Healthcare Documentation”, International Journal of Drug Regulatory Affairs, Vol. 13/2, pp. 73-80, https://ijdra.com/index.php/journal.

[8] Nguyen, A. et al. (2022), “Managing demand volatility of pharmaceutical products in times of disruption through news sentiment analysis”, International Journal of Production Research, Vol. 61/9, pp. 2829-2840, https://doi.org/10.1080/00207543.2022.2070044.

[9] OECD (2025), “The digitalisation of trade documents and processes: Going paperless today, going paperless tomorrow”, OECD Trade Policy Papers, No. 297, OECD Publishing, Paris, https://doi.org/10.1787/64872f25-en.

[1] OECD (2024), Securing Medical Supply Chains in a Post-Pandemic World, OECD Health Policy Studies, OECD Publishing, Paris, https://doi.org/10.1787/119c59d9-en.

[81] OECD/WTO (2025), Economic Implications of Data Regulation: Balancing Openness and Trust, OECD Publishing, Paris, https://doi.org/10.1787/aa285504-en.

[16] Pagonidis, T., S. Sapuric and P. Lois (2020), “Assessing the implementation of serialisation in pharmaceutical industry in Greece: a qualitative approach”, J. for Global Business Advancement, Vol. 13/1, p. 4, https://doi.org/10.1504/jgba.2020.109144.

[86] Pall, R. et al. (2023), “Predicting drug shortages using pharmacy data and machine learning”, Health Care Management Science, Vol. 26/3, pp. 395-411, https://doi.org/10.1007/s10729-022-09627-y.

[25] Parmaksiz, K., E. Pisani and M. Kok (2020), “What Makes a National Pharmaceutical Track and Trace System Succeed? Lessons From Turkey”, Global Health: Science and Practice, Vol. 8/3, pp. 431-441.

[24] Partnership for DSCSA Governance (2025), Foundational Blueprint for 2023 Interoperability, PDG.

[39] Pfizer (2022), Pfizer Accelerates Digital Transformation with the Launch of a Digital Operations Centre, LogiPharma, https://logipharmaeu.wbresearch.com/blog/pfizer-accelerates-digital-transformation-with-the-launch-of-a-digital-operations-centre (accessed on 20 July 2025).

[40] Pfizer (2021), Applying the Latest Digital Technology to Optimize COVID-19 Vaccine Efforts, Pfizer, https://www.pfizer.com/sites/default/files/investors/financial_reports/annual_reports/2021/story/latest-digital-technology-to-covid-vaccine-efforts (accessed on 20 July 2025).

[58] Rasheed, H., L. Höllein and U. Holzgrabe (2018), “Future Information Technology Tools for Fighting Substandard and Falsified Medicines in Low- and Middle-Income Countries”, Frontiers in Pharmacology, Vol. 9, https://doi.org/10.3389/fphar.2018.00995.

[37] Rukanova, B. et al. (2021), “Realizing value from voluntary business-government information sharing through blockchain-enabled infrastructures: The case of importing tires to the Netherlands using TradeLens”, DG.O2021: The 22nd Annual International Conference on Digital Government Research, pp. 505-514, https://doi.org/10.1145/3463677.3463704.

[29] Safemedicines (2021), Canadian drug importation – What’s the state of the issue in June 2021?, https://www.safemedicines.org/2021/06/canadian-drug-importation-whats-the-state-of-the-issue-in-june-2021.html.

[50] Schoelmann, E. (2024), Inside Johnson & Johnson’s Supply Chain Operations, RFgen, https://www.rfgen.com/blog/inside-johnson-johnsons-supply-chain-operations-success (accessed on 15 July 2025).

[66] Schweihs, D. (2024), Unraveling the DSCSA serialization puzzle: implications for drug shortages and patient care, Pharmacy Times, May 22.

[42] Shikha Yug Sachdeva (2024), What does it take to rewrite the playbook on global supply chains? For Pfizer, it was a race against time, a bold leap into cutting-edge technology, and a commitment to collaboration on a scale never seen before, Medium, https://shikhayugsachdeva.medium.com/game-changer-series-episode-3-pfizers-bold-leap-transforming-supply-chains-saving-lives-45795e87381b (accessed on 7 July 2025).

[48] Smith, A. and R. Watts (2024), “Fivetran Review: Features, Pros and Cons”, Forbes, https://www.forbes.com/advisor/business/software/fivetran-review/.

[45] Software Connect (2025), “Infor Nexus”, https://softwareconnect.com/reviews/infor-nexus/.

[87] Thakur-Weigold, B. and S. Miroudot (2024), “Promoting resilience and preparedness in supply chains”, OECD Trade Policy Papers, No. 286, OECD Publishing, Paris, https://doi.org/10.1787/be692d01-en.

[47] Tobin, D. (2026), “What Is Fivetran? A Deep Dive + How Integrate.io Compares”, Integrate.io, https://www.integrate.io/blog/what-is-fivetran-a-deep-dive-how-integrateio-compares/.

[22] U.S. Food and Drug Administration (2025), Drug Supply Chain Security Act (DSCSA), FDA, https://www.fda.gov/drugs/drug-supply-chain-integrity/drug-supply-chain-security-act-dscsa.

[12] Vajda, P. et al. (2021), “Survey of workflow and cost implications of decommissioning regarding the Falsified Medicines Directive in Hungarian hospital pharmacies”, BMJ Open, Vol. 11/11, p. e047193, https://doi.org/10.1136/bmjopen-2020-047193.

[55] Van Beusekom, M. (2022), Supply map created to predict drug shortages, find solutions, CIDRAP News.

[57] WHO (2024), Country Experiences on Implementation of Traceability for Medical Products, World Health Organisation, https://iris.who.int/bitstream/handle/10665/380837/9789240103535-eng.pdf?sequence=1.

[28] Yorulmaz, M. et al. (2012), “Türkiye’de İlaç Takip Sistemi (İTS) Uygulaması ve Çalışma Biçimi”, Akademik Bilişim’12 - XIV. Akademik Bilişim Konferansı Bildirileri, https://ab.org.tr/ab12/kitap/yorulmaz_altunkan_AB12.pdf.

[69] Zhang, C. et al. (2021), “A survey on federated learning”, Knowledge-Based Systems, Vol. 216, p. 106775, https://doi.org/10.1016/j.knosys.2021.106775.

Notes

← 1.  Falsified medicines are fake medicines that are designed to mimic real medicines, while counterfeit medicines are medicines that do not comply with intellectual property rights or that infringe trademark law.

← 2.  Near real-time means that data are made available rapidly to be able to take action. In practice, it can be every day, every week or every month depending on the type of data and how critical it is to get an update in a given timeframe. Real-time monitoring in the sense of detecting where products are “every second” is generally not achievable and not desirable.

← 3.  The purpose of the EU’s FMD is to prevent falsified medicines from entering the legitimate supply chain. While it is not intended as a tool to monitor supply and predict shortages, its analysis provides a starting point since it illustrates how monitoring of individual medicines has been successfully implemented in the EU.

← 4.  The GS1 2D DataMatrix is a two‑dimensional barcode that can be printed on packages as a square or rectangular symbol using an ordered grid of dark and light dots to encode the information. Common data elements contained in the DataMatrix are the Global Trade Item Number (GTIN), batch or lot number, production date, best before date, expiration date, and serial number (GS1, 2018[88]).

← 5.  Some over-the‑counter medications are also subject to the regulation. EU member countries may include further medicinal products, if they are deemed to be at risk of being counterfeited (Merks et al., 2022[19]).

← 6.  The unique identifier contains the product code, serial number, batch number, expiry date, and where applicable national reimbursement number (Bouvy and Rotaru, 2021[20]).

← 7.  May 2025 for manufacturers and repackagers, August 2025 for wholesale distributors and November 2025 for dispensers. The exemption for implementation by dispensers with 25 or fewer employees was already set to November 2026. See https://www.fda.gov/‌media/182584/download?attachment.

← 8.  https://www.fda.gov/media/168307/download.

← 9.  US Pharmacopeia (USP) is a non-profit organisation. The Medicine Supply Map is a solution also developed for individual companies (or other stakeholders) to assess risks and optimise their supply chain management.

← 10.  The distributed ledger model is a “peer to peer” model, where multiple entities together ensure that the data cannot be tampered with as no single actor can make unilateral changes. A transaction is only accepted if all decentralised nodes are identical. This consensus protocol is therefore critical in ensuring trust (Louw-Reimer et al., 2021[36]).

← 11.  The United Nations Centre for Trade Facilitation and Electronic Business (UN/CEFACT) is an intergovernmental body of the United Nations Economic Commission for Europe established in 1996. It develops standards for trade facilitation and electronic business.

← 12.  Maersk responded to the criticism by moving TradeLens to a separate business entity, GTD solutions, and establishing a joint customer advisory board (Jovanovic et al., 2022[35]).

← 13.  Barcodes are printed labels encoding some information. They need to be directly scanned by a reader. RFID tags are tiny radio transponders that transmit digital data using the energy from the radio signal of the reader. They can be read further away without a direct line of sight to the reader. The RFID tag cannot run on its own and has no computing power. Smart labels are labels with sensors. Unlike passive RFID tags, they have their own source of energy and computing power. Smart labels can rely on RFID technology to transmit their information or other wireless technologies (such as Wi-Fi, Bluetooth, etc.) (DHL, 2020[54]). “Smart printables” are the next generation of smart labels that will be both smart and printable (i.e. microchips will be directly “printed” on the label together with printable batteries to power the device). This technology is not yet fully developed and deployed (DHL, 2025[52]).

← 14.  The International Council for Harmonisation of Technical Requirements for Pharmaceuticals for Human Use (ICH) brings together regulatory authorities and the pharmaceutical industry to discuss scientific and technical aspects of pharmaceuticals. Since its inception in 1990, it has developed guidelines that are applied by a growing number of regulatory authorities. ICH now includes 23 members and 41 observers.

← 15.  The International Coalition of Medicines Regulatory Authorities (ICMRA) is a voluntary, executive‑level entity of worldwide medicines regulatory authorities set up to provide strategic co‑ordination, advocacy and leadership. It started through meetings of heads of medicines regulatory agencies in 2012. ICMRA has 24 members and 18 associate members.

← 16.  https://pharmaledger.org/. See also the US DSCSA Pilot Project Program previously mentioned (https://www.fda.gov/media/168307/download). Interestingly, in one project (the “Optimal Solution”), the initial idea of focussing on a blockchain-based solution was abandoned to develop an alternative distributed data mechanism.

← 17.  Federated learning is a machine learning concept that was first developed by Google Labs in 2016 to describe a cross-device scenario where millions of mobile devices are co‑ordinated by a central server while local data are not transferred.

← 18.  This approach is also discussed in the context of the MEDICON initiative launched at OECD by Switzerland to promote the creation of a global monitoring system for key critical medicines.

← 19.  The FDA list of essential medicines published in 2020 included critical inputs (https://www.fda.gov/media/143406/download?attachment).

← 20.  The new EU “Pharma Package” reinforces shortage notification obligations for MAHs. See the final compromise text of the Directive (https://data.consilium.europa.eu/doc/document/ST-6367-2026-INIT/en/pdf) and Regulation (https://data.consilium.europa.eu/doc/document/ST-6366-2026-INIT/en/pdf).

← 21.  https://www.pfizer.com/sites/default/files/investors/financial_reports/annual_reports/2022/story/data-and-ai-are-helping-to-get-medicines-to-patients-faster/ On 10 December 2025, the European Centre for Disease Prevention and Control (ECDC) announced a plan to integrate wastewater-based surveillance into infectious disease surveillance at the EU level (https://www.ecdc.europa.eu/en/news-events/ecdc-lead-integration-wastewater-based-surveillance-infectious-disease-surveillance).

← 22.  EU member countries apply different definitions for medicine shortage. Denmark, Estonia, Finland, Italy, Latvia, Spain, Slovenia, and Sweden utilise the official EMA’s definition, in which “a shortage […] occurs, when supply does not meet demand at national level”. Belgium, Germany, and Romania have national definitions which include references to the time frame of the unavailability of the medicinal product. France and the Netherlands refer to insufficient stock levels. Similarly, definitions in Hungary and Ireland place importance on the inadequacy of supply. In Italy and Portugal, a shortage occurs when the medicinal product is not available at the national level. Czechia does not have a formal definition for shortages (European Commission, 2022[89]).

← 23.  The agreed CMA text also establishes a Critical Medicines Co‑ordination Group (CMCG), composed of the member states and the Commission, with EMA as an observer. Its tasks include exchanging available information on existing and planned manufacturing capacity, facilitating exchanges on national programmes and contingency-stock requirements, advising on priorities for vulnerability evaluations, and conducting strategic-foresight discussions on long-term supply-chain trends and vulnerabilities.

← 24.  For example, the United Kingdom and Korea conducted a joint stress test of the resilience of the global electric vehicle battery supply chain. This table‑top exercise involved the analysis of worst-case disruption scenarios and was supported by industry experts and a range of government departments (https://www.gov.uk/government/publications/uk-critical-imports-and-supply-chains-strategy/critical-imports-and-supply-chains-strategy-html-version). See also Thakur-Weigold and Miroudot (2024[87]) for a discussion of a related concept, the “preparedness conferences”.

← 25.  WHA69.25, “Addressing the global shortage of medicines and vaccines”.

← 26.  The ATHINA call for tender suggests that such data are part of ATHINA with the monitoring of tweets on social media and news on media outlets.