Medical articles

Algorithm errors: an acceptable risk in healthcare?

Published on · Updated on · By Dr Sam Ward

In this opinion piece on Medi-Sfeer, Dr Sam Ward and lawyer Karolien Haese respond to a computer error in the UK breast cancer screening programme. The piece grew out of a debate within Building HealthCare for Tomorrow (BHCT).

The authors ask a different question than usual. Not: should we fear artificial intelligence? But: how many women would not have received a screening invitation without the algorithm, and what margin of error do we find acceptable? Zero risk does not exist, and using the risk of error as a reason to block innovation is, in their view, counterproductive.

Legally, responsibility lies with those who programme or use an algorithm: the errors remain human errors. The authors also describe the tension between managers, who must weigh the system as a whole, and doctors, who face the individual patient. Big data and AI can bring both perspectives together, but only with complete and reliable databases. They see sharing anonymised data as a first step.

Published on Medi-Sfeer on 18 May 2018. Read the full article on Medi-Sfeer (in Dutch).

A question about your situation?

A few questions are enough to see whether a consultation with Dr Ward is right for you.

Read also