Artificial intelligence and machine learning are already transforming the delivery of healthcare, including within occupational health. The good news, argues Dr Julian Eyears, is this revolution is unlikely to put practitioners out of a job. Do, however, expect OH’s role, remit and (more positively) reach to change as a result.
Machine learning is already replacing white collar jobs. So will occupational health doctors and nurses be out of a job anytime soon?
The good news: this is unlikely. When we look at history, technology has changed doctors’ jobs but never put doctors out of work.
For example, in the 1980s, newspapers predicted that the new invention of the CT scanner would make doctors redundant because no fancy tools (such as stethoscopes) would be needed to see inside the body. The eventual reality was CT allowed doctors to do more. That meant the public wanted more and more CT scanners and radiologists to interpret their images.
‘Artificial intelligence’ (AI) is a very general term given to the emulation of human intelligence using computers. ‘Machine learning’ is a form of AI that typically uses a computer model of a primate brain and teaches that brain by providing example problems and solutions.
The computer taught in this manner can quickly exceed the expertise and judgement of an individual human because its learning is drawn from the experience and expertise of many humans.
OH and the future
Thus, the computer can, for example, copy and store the collective expertise of many senior medical consultants by being shown hundreds of chest radiographs, spirographs or audiograms. Computers are able to read handwriting and can, for example too, produce a medical summary of handwritten notes.
Technological change ‘inevitable’
It is inevitable that AI and machine learning will change what we do within occupational health. In some cases the technologies will threaten our livelihood because they will enable other people to take what traditionally we have regarded as ‘our’ OH business.
As ever with new innovations, however, there are opportunities as well as threats. If there is an aspect of your job that you find repetitive and mundane, then machine learning may well be poised to relieve you of that burden. So be careful what you wish for!
When one looks at other industries that have been transformed by new technologies, such as the travel industry, one can see that the threat did not come from human-like computer robots replacing travel agents sitting at desks. Instead it came from these jobs and desks disappearing altogether when the internet allowed customers to book flights and hotels directly.
So far, health and education stand out as industries that have not undergone such dramatic transformation. However, many observers think that may be about to change.
This is not because of the advent of machine learning alone. There is an explosion of cheap electronic sensors that sniff the air, your breath, your sweat electrolytes and the strain in your muscles.
Sensors that measure things in your sweat have obvious medical application, but sensors that measure things in factories and offices will revolutionise OH too. It will be possible to identify patterns of air pollution or joint strain, for example, that give rise to industrial disease. HSE department are already using such techniques to predict and prevent accidents.
Threats and opportunities
The unique ‘threat’ to occupational medicine probably comes from the fact that we interact with other agencies such as managers, HR, lawyers, GPs and hospitals on a regular basis.
This introduces the possibility that these parties we currently deal with might appropriate some sectors of our business.
For example, OH clinicians currently interpret hygiene data ranging from noise measurements through to benzene and recommend work modifications based on that data.
Algorithms will do this better and more consistently. So an industrial hygiene or health and safety provider could appropriate this role simply by owning the algorithm.
Similarly, HR can already purchase ‘stress-watching’ algorithms that monitor employees’ work stations and flag ‘atypical’ employee behaviours. Hospitals and GPs could, similarly, own algorithms that advise on return to work based on clinical data.
For those of us who think that our UK OH consultation model coupled with opining on legal nuances of disability make us indispensable: we should be wary.
Machine learning is already invading the white collar job market. Junior lawyers are being put out of work by machine-learning algorithms that read legal tomes and search for legal precedent.
It is entirely likely that machines will learn return-to-work strategies for stroke or myocardial infarction for all industries from historical OH data and clinical data from hospitals and be able to advise on liability under disability legislation. It is important, however, that we view these new tools as opportunities and not threats.
Replaced by a computer?
In my view it is unlikely that the model in the UK of occupational health physician consultations could be entirely replaced by computer.
Technology will allow SMEs to implement occupational health for the first time and cheaply and this in turn will increase demand for our opinions.” – Dr Julian Eyears
What is entirely possible, however, is that the need for these consultations could be diminished over time because employers themselves will be able to implement clever computer algorithms that prevent occupational disease. On the other hand, other algorithms and industrial sensors could in time flag up diseases that were not known to exist before, therefore probably resulting in more OH referrals.
It is likely that, in time, our role will become more about supervising (and designing) algorithms and less about individual day-to-day contact with employees.
In my view, this evolution is to be welcomed because it will extend the reach of occupational health to workplaces where we don’t currently have a presence.
We are repeatedly informed that our specialty is too small and will eventually cease to be relevant.
In my view, technology is our unique opportunity to vastly extend our reach both by the use of productivity-enhancing tools such as sensors and machine learning but also remote video consulting.
Technology, too, will allow SMEs to implement occupational health for the first time and cheaply and this in turn will increase demand for our opinions.
Finally, we also need to think globally. Although the UK OH discipline can be at times a little introspective, it is paradoxically regarded as a model around the world.
The newly-arrived ability to carry out consultations via video link greatly extends our reach and I, for one, can envisage a future where UK occupational health providers are providing services to many jurisdictions. This is not blue skies thinking; this is already happening in many multi-national organisations.