VP115 Practical Issues Of Using Real-World Data In Effectiveness Research

Abstract

INTRODUCTION:

The Innovative Medicines Initiative, IMI-GetReal project aimed to explore incorporation of robust methods for real-world data (RWD) collection and synthesis earlier in the medicines development process, both by pharmaceutical companies and healthcare decision makers. The focus was on the potential use of RWD, alone or in combination with randomized controlled trials (RCTs), to demonstrate effectiveness of new interventions. Four case studies were conducted in multiple disease areas to examine methods for predicting drug effectiveness and the perspectives of different stakeholders on these methods. This study aimed to identify practical obstacles in accessing and using RWD and RCT data for effectiveness research conducted as part of these case studies.

METHODS:

Qualitative content analysis was conducted to identify and characterize key issues relating to accessing and analysing study data from external sources, both RWD and RCTs.

RESULTS:

Accessing RWD from registries proved difficult due to multiple reasons, including: complex and non-transparent application procedures, resistance from registry owners to discuss applications and datasets not being research-ready within project timeframes. There were also issues with the RWD eventually accessed, including a lack of individual participant data (IPD) and incomplete data. Where access to IPD from RCTs was obtainable, there were restrictions imposed on how it could be used. For example, it could not be used to target analysis on an individual product, but rather explore methodologies for data synthesis in a product-anonymised setting. This condition encouraged additional data sharing by other stakeholders.

CONCLUSIONS:

Despite the collaborative, multi-stakeholder nature of IMI-GetReal and proper disclosures with data owners, access to data proved challenging. Such barriers to data accessibility can delay effectiveness research, restrict opportunities for the development of methods incorporating RWD and diminish the potential use of RWD in decision making. Where data is intended to be used for this purpose, sufficient attention should be paid to these potential barriers.