Paris, 9 April 2019

In its response to the public consultation on big data organised by the HMA-EMA Joint Big Data Taskforce, Prescrire considers that regulation on the authorisation and supervision of health products cannot be based on hope and promises but needs to rely on facts, trustworthy evidence and transparency.

> Click here to download Prescrire's comments (pdf, 158 Ko)

Summary

It is undisputable that some big data are important elements to be considered in the surveillance of the harm-benefit balance of a medicine throughout its entire lifecycle. However, to consider their utility and to avoid being lured in the hype around them, big data weaknesses and limitations have to be addressed: huge data heterogeneity and fragmentation, poor quality control, costs, sustainability…

As a prerequisite, Prescrire urges European and national regulators not to weaken marketing authorisation requirements by shifting the provision of evidence to real world data after marketing authorisation. Instead, it needs a critical review.

Research on the utility of big data needs to be focused and steered by clearly stating which problems are intended to be solved. Priorities should be fixed considering which big data, if any, might be helpful to improve the regulatory and decision-making process. It is of outmost importance to consider very carefully the source and quality of big data, and their internal validity and utility for new approaches in the regulatory context.

Research is needed into the clinical effectiveness and related cost and cost-efficiencies of wider use of big data in health products regulation. Major methodological challenges of bias in real life data have to be addressed. Like for clinical studies, research and studies using big data (including protocols) need to be registered in a publicly accessible and centralised European Registry that complies with reporting requirements. Reproducibility, validated statistical analyses and transparency throughout the whole process (from data collection to data use) also need to be addressed carefully. Transparency on conflicts of interests of interested parties has to be ensured. Recommendations should be based on robust evidence generated through independent rigorous evaluation free from vested financial conflicts of interests.

> Click here to download Prescrire's comments (pdf, 158 Ko)

©Prescrire April 2019

Share