Assessing the Feasibility of AI-Based Forecasting to Prevent Medicine Shortages — Outcomes of the REMEDY Project

Research Center Pharmaceutical Engineering GmbH 

Pharmaceutical supply shortages represent a significant global challenge. According to the current statistics, around 460 drug products are either not available or only available in limited quantities in Austria. The FFG-funded REMEDY (REsilient MEDication supply through multimodal Data sYnthesis) project (01.09.2024 to 30.11.2025) addressed these challenges and investigated whether the legal and technical requirements in Austria allow the development of an AI-supported forecasting system to mitigate drug shortages.  

The project was led by Fraunhofer Austria in collaboration with the Research Center Pharmaceutical Engineering (RCPE), the Faculty of Computer Science, Economics and Law at the University of Oldenburg, and the Institute for Civil and Civil Procedure Law at Vienna University of Economics and Business (WU Wien), combining technical and legal expertise to map the feasibility space. In particular, REMEDY focused on three interlinked questions: i) which data are required to predict demand and supply disruptions; ii) whether those data can be collected under Austrian and EU legal frameworks; and iii) whether heterogeneous data sources can be prepared and harmonized for AI and large-scale predictive methods.  

The key outcomes of the project include: 

• Overall feasibility: AI-based shortage forecasting is feasible if technical hurdles, legal safeguards and stakeholder coordination are resolved. 

• Technical requirements: A viable concept for an AI forecasting system was established, but requires data harmonization, interoperable interfaces, secure stages/access controls and model validation. 

• Legal requirements: regulatory and data-protection constraints were clarified; lawful use demands anonymization/pseudonymization, clear purpose limitation, governance and contractual arrangements for cross-actor data sharing. 

Overall, the project highlights two central challenges: the heterogeneity and accessibility of essential datasets, and the need to reconcile innovative data-driven methods with stringent legal and privacy safeguards. REMEDY did not yet implement an operational system; rather, it provided the evidence base to decide whether and how an ethically and legally compliant, technically robust AI forecasting solution can be pursued to strengthen public health preparedness. In future projects, once these prerequisites have been met and confirmed in targeted pilots, an operational system will be deployed to reduce critical drug shortages and strengthen Austria’s supply resilience.