AI in Healthcare — help or hindrance?
By Luke Kenworthy on April 26, 2019
There can be no question that Artificial Intelligence (AI) is set to alter the face of various industries, changing the landscape of both labour and skill based employment tasks. The paradigm shift set to impact the Healthcare industry is suggestively powered not only by the increase in investment but also the rapid progression of analytics techniques.
How is AI set to change the face of healthcare?
The ability for AI to be applied to both structured and unstructured healthcare data sets is said to be imperative in regard to wide-scale industry application, supporting an increase in early detection rates of major disease and aiding in patient diagnostics for industry professionals. Initial industry literature recognises AI as a viable and realistic treatment for a number of ailments, including, but not limited to eye disease, sepsis, skin cancer, heart disease, lung cancer, Parkinson’s, strokes, despite current application examples being small-scale pilots or research projects.
Medical practitioners and those working in the healthcare industry would no doubt have been inquisitive if not cautious of AI development; are the suggestions of job loss and insecurity through automation legitimate or simply fearmongering? Whilst it is certain that AI will become apparent in healthcare over the next 10–15 years, even considerably routine in NHS practice, it will merely tweak the daily tasks of healthcare professionals rather than entirely replacing them, as the mundane and data-driven activities lessen.
Consider it a collaborative approach. Imagine the endless capabilities across industry if the creativity and problem-solving skills of humans were to be combined with the infinite computing power and cognitive resource of technology that comes with the implementation of AI. Recent studies have suggested that the aforementioned collaborative effort could be increasingly important for industry progression, with trials suggesting that in image classification alone, localisation scores and diagnosis increased significantly with not only increased efficiency but a decrease of 85% in human error rate with the aid of intelligent agents. A similar study from Stanford University also recognised that the accuracy of current algorithms were able to match the accuracy of trained professionals in cancer detection, also suggesting that not only could this again decrease human error, but lower the cost per patient, bringing accurate and affordable healthcare to the masses.
A valid point referenced by many is the inability to replicate the human touch and empathy which could potentially be seen as an issue if your local GP were in the future to be replaced with an intelligent agent. There is no doubt that in times gone by, a local family doctor could be seen as the sole avenue for medical advice, however, in recent years there has been an increasingly open-minded approach to all things tech in regard to patient care, with 75% of U.S. consumers surveyed citing that AI technological advances (including mobile apps, wearable monitoring devices, and smart scales) were important to them to help them manage their health and an even larger amount being happy to rely on AI-based systems for symptom diagnosis.
It should be recognised that there are both pros and cons to the advancement of AI and it’s implementation in the healthcare industry. Future collaborations could see an increase in accurate diagnosis, detection of disease, lower readmission rates and lower human error which in turn would lower the financial burden and also increase efficiency, creating widescale, accurate and affordable accurate patient care. Of course, this is far from ‘application-ready’ and there are larger issues to think of including ethics, privacy, security and regulation.
The development of AI in healthcare is ongoing and will no doubt play a huge part in the future of industry progression, however, at this time it seems that the potential positives far outweigh the negative possibilities of industry-wide application.
Interested to hear more about the latest industry research and application methods in healthcare? The Deep Learning in Healthcare Summit is taking place next month in Boston.
Confirmed speakers at the summit include:
- Farhan Siddiqui, Advanced Analytics Architect, Pfizer
- Sadid Hasan, Senior Scientist, Philips Research
- James Cai, Head of Data Science, Roche
- Wei-Lun Hsu Alterovitz, Senior Bioinformatics Data Scientist
- See more on the agenda and list of speakers here.