Big Data, fewer doctors

Algorithmic driven software is changing how doctors work as well as the patient and doctor relationship.

How important is that relationship if a machine can do the same job cheaper, quicker and more accurately? How much has medical technology colonized a clinician's free will?

This weekend the Big data analytics: Leveraging capability in healthcare conference in Sydney will discuss the practical applications of algorithms and predictive analytics in individual and population health. I have reservations about this brave new world of machine thinking.

An algorithm is a list of instructions that leads a user to a conclusion, e.g., 'If this, then that.' All information processing machines from a Fitbit to an ultrasound scanner use algorithms. They have been around since the Persians.

Computer automation has swept through shop floors and factories, transforming businesses and it is now revolutionising the professions. Knowledge-based jobs were supposed to be safe career choices, as the years of study it took to become a doctor or lawyer, in theory, guaranteed a lifetime of lucrative employment.

Now expert radiologists are routinely outperformed by pattern-recognition software, diagnosticians by simple computer questionnaires. US studies show that when an electronic medication management system is implemented, the rate of adverse drug reactions was reduced by 40.9 per cent and prescription errors by 99.4 per cent saving millions of dollars in legal actions, not to mention lives.

Sun Microsystems cofounder Vinod Khosla predicts that diagnostic machines driven by algorithms will replace 80 per cent of doctors within a generation.

There are real time predictive analytics programs which identify sepsis, a notoriously difficult-to-diagnose infection. It affects 1 per cent to 2 per cent or all hospital patients and kills roughly half of those who contract it. The application monitors medical data and alerts doctors when a patient is at risk.

Surely doctors and computer driven clinical analytics can work together? Not if history is any guide. The great artisans and tradesmen of the 18^th century were over thrown by the steam driven machines of the industrial revolution. Besides, giving a doctor veto power over algorithmic systems introduces human bias. The promise of big data decision-making, after all, is that decisions based on data and analysis - more science, less gut feel and rule of thumb - will yield better results.

The rise of nuclear and genomic medicine, advances in processing speed, metadata storage systems and breakthroughs in artificial intelligence (especially voice recognition software) have changed the equation in favour of the machine. Millions of people will live longer, healthier and happier lives because of it. But there are problems that algorithms can't anticipate.

Patients sometimes present to doctors with no idea of what ails them. They are frightened or have a mental illness. They ramble and sometimes lie. It is up to the clinician to make sense of this and then make a diagnosis.It is an organic and narrative process - something entirely human - that algorithms cannot replicate as yet.

If machines in the future can provide cheaper healthcare – isn't that what we're all after? But something niggles - something that can't be measured or quantified. If I scan my conscience, surely someone who has dedicated years of their life to a profession should also be the masters of the technology, not its hand maidens. I want a human to talk to, to relate with and who will tell me they have my best interests at heart. I want a confessor and a clinician.