Intended for healthcare professionals


Clinical applications of machine learning algorithms: beyond the black box

BMJ 2019; 364 doi: (Published 12 March 2019) Cite this as: BMJ 2019;364:l886

Re: Clinical applications of machine learning algorithms: beyond the black box

Perhaps one would be kind enough to direct us/me to a website which has utilised data or big data to create algorithms -statistical or computer driven, in common biological diseases. Clinical approach and bedside manner in a profession which empathises with patient having a disease either in isolation or in concurrence with other diseases, is unlikely to be solved by machine driven apps or artificial intelligence. Computer based data can give statistical data of prevalence, specificity, sensitivity and other indicators but a very unlikely to solve a biological problem or its complications. Robotic surgery for precision medicine is using tools for precision delivery of technique, drug delivery or diagnosis but the underlying procedure remains the same,- designed and developed by man. So does that mean one has do away with all apps and algorithms and robots? What is going to be the prevalent undergraduate clinical medicine study structure- is there going to be a robot doing all the teaching- pointing out all the pathological and histological specimens and the fine nuances ? How is translational medicine being encorporated in conventional teaching ---- is medical student going to learn that, all that he has learnt in his undergraduate years hogwash and to be replaced by newer drugs which has hit the market. Has the medical faculty disposed of all the older generation drugs and replaced their newer analogues with knowledge of their adverse effects and side effects and therapeutic indices. Educational tools have enhanced learning but one still has to fall back on older methods of bedside teaching, laboratory medicine and anatomy dissection.

Competing interests: No competing interests

14 March 2019
SR Dasari