Tariq Ahmad, James V. Freeman and Folkert W. Asselbergs

European Journal of Heart Failure (2018) 21(1): 86-89

https://doi.org/10.1002/ejhf.1370

We have come to rely on a high degree of speed, precision, and accuracy from the modern world. Algorithms can instantaneously map out all routes to our destination and adjust the path according to real-time traffic information; restaurant recommendations are transmitted to our devices as we walk through a neighbourhood relying on incredibly granular Global Positioning System (GPS) and personal preference data; millions of songs, books, and movies can be accessed on handheld devices with just a few clicks. How-ever, in medicine, no less a data science than other field, the norm is uncertainty, imprecision, and inaccuracy. A rather dramatic illustration of this is the case of implantable cardioverter defibrillators (ICD) with the capacity for cardiac resynchronization therapy (CRT) — despite the expense and invasive nature of the therapy, our ability to predict who will benefit from it remains archaic and inexact.

This article discusses ‘Machine learning-based phenogrouping in heart failure to identify responders to cardiac resynchronization therapy’ by M. Cikeset al., published in the same journal issues (pages 74–85).