The Rare Bleeding Disorders Network

Prospects for incorporating machine learning in ERNs

A new article has been published in the Orphanet Journal of Rare Diseases which discusses how emerging machine learning (ML) technologies might be applied within European Reference Networks (ERNs) to shorten diagnostic delay. The study was conducted in the framework of the Screen4Care project.

As ML-based tools become more sophisticated, they appear to have great potential to analyse large amounts of genotypic and phenotypic data accurately, precisely, and efficiently. This has significant implications for how they could be applied to rare disease diagnosis, which often requires the time-consuming analysis of large amounts of such data. Implementing ML-based diagnostic support technologies could greatly reduce the amount of time RD patients are forced to wait before obtaining a diagnosis, therefore leading to more timely access to appropriate care.

As representatives of key opinion leaders in the rare disease field at the European level, the ERNs are central to implementing ML technologies in RD diagnosis. In order to determine their readiness for this step, the authors explored ERN members’ expectations towards and acceptance of ML and its potential application through an online survey and focus group.

The survey asked participants about the relative importance of different potential benefits of ML technologies, as well as their perspectives on the feasibility of implementing them in ERNs. A lack of training in such tools was identified as the most substantial barrier to their adoption and implementation. Both the survey and focus group results indicated that ERN members believe ML to be a promising tool to help improve timely diagnosis of RD, but found that further research is needed on the opinions of other stakeholders, particularly patient organisations.

Screen4Care is a public-private-patient consortium funded by the Innovative Medicines Initiative 2 Joint Undertaking, which receives support from the EU’s Horizon 2020 program and the European Federation of Pharmaceutical Industries and Associations (EFPIA). The five-year project aims to accelerate rare disease diagnosis through ML technologies and genetic newborn screening.