Inquiry on using algorithms for decision-making

30 October 2017

Algorithms are becoming ubiquitous in healthcare as their role in answering complex questions and solving sophisticated problems in research, diagnosis, prognosis, and monitoring increases. In our written evidence to the House of Commons Science and Technology inquiry into their use in decision-making we note:

  • Algorithm performance depends on data inputs: quality algorithms require quality data. It is therefore imperative that data flows are integrated and streamlined
  • Algorithms represent unique challenges in regards to transparency. Emphasis on technical and non-technical solutions early in algorithm development is necessary to ensure algorithms remain transparent and their developers sufficiently accountable
  • Public perception and trust are an often ignored yet important consideration. Public engagement and consideration is imperative to ensure access to health data necessary for the development of healthcare algorithms

While recognising that the Information Commissioner’s Office (ICO) has made important steps in providing data controllers with advice as to how to comply with the General Data Protection Regulation (GDPR) transparency requirements, we argue more specific advice is highly desirable to foster good  practice, which will require  that the ICO should engage with the algorithm development sector. We also point to the In Vitro Diagnostic Devices Regulation (IVDR) as another source of potential regulation, which may be a means of ensuring greater transparency and that the Medicine and Healthcare products Regulation Agency have responsibility for regulation in this area.

Read the full response here