Our medical writing team excel at making complex information easier to understand and keeping our clients up to date with the latest innovations. In this blog, they take an in-depth look at the potential of artificial intelligence in healthcare.The ability of artificial intelligence (AI) is evolving rapidly. This technology has the potential to address the biggest global healthcare issues while shifting the emphasis from treatment to prevention. With non-communicable diseases being the leading cause of mortality globally, coupled with ageing populations, the cost of delivering adequate healthcare is rising considerably year-on-year. Therefore, both public and private sectors must invest in health AI to enhance clinical decisions and so, help improve future care provision.
What is AI?
AI can be defined as the replication of human intelligence by machines or computer systems. Based on the principles of biology, computer science, logic and maths, these technologies are able to undertake processes such as reasoning, learning, interaction and sensory understanding.
For a more detailed explanation of artificial intelligence
Multiple forms of AI are used in healthcare today; the majority of current applications rely on human-created algorithms to analyse data or perform specific tasks. However, the most successful variant of AI in recent years, “machine learning”, allows computer systems to recognise patterns and apply their own sets of rules for interpreting data, rather than adhering to pre-programmed instructions.
Who is investing in AI?
Although AI has been studied for decades, the 2016 global healthcare market was estimated to be worth $1.4 billion, with it set to reach $2.3 billion by 2023. Significant growth in the AI healthcare sector has been fuelled by its potential to address major health challenges, reduce human error and improve clinical outcomes. A significant rise in venture capital investment has also contributed to its growth.
Major technology companies, like Google, Microsoft and Facebook, are already investing heavily in developing AI that can operate within the data-rich environment of healthcare.
How can AI be used in healthcare?
With the immense amount of data generated in healthcare, the argument for adopting AI is gaining strength – not only could health AI improve clinical outcomes, it could also reduce costs by improving processes and ensuring effective use of resources. Below are six examples of how AI could change the way healthcare is delivered:
1. Analysis of healthcare system performance
AI has the potential to analyse vast amounts of data to identify workflow inefficiencies and mistakes in treatment, and to provide suggestions on avoiding unnecessary re-admissions to hospital. In the Netherlands, 97% of clinical invoices are digital; Zorgprisma Publiek analyses these digital invoices to identify hospital or staff-associated treatment errors and suggest solutions that reduce the incidence of re-admissions.
2. Drug discovery
Developing new pharmaceuticals is both costly and time-consuming. AI can be used to analyse large, complex clinical and scientific data sets to identify compounds for screening during the initial stages of drug discovery. It can also be used to identify compounds or medicines that can be re-engineered for treating other diseases. Its uses also include “personalising” medicines based on predictive analytics and the analysis of individual medical data. AI also has the potential to identify appropriate candidates for participation in clinical trials.
AI could improve diagnostic accuracy and reduce the cost and time required for a diagnosis. In the UK, DeepMind has developed a promising AI that can screen retinal scans for age-related macular degeneration, diabetic retinopathy and glaucoma. In the USA, the FDA has approved an AI diagnostic device called IDx-DR that can detect various eye conditions by analysing retinal scans.
4. Risk identification and prediction
By examining historical clinical data, AI technology could provide real-time information on at-risk patients and those most likely to be readmitted to hospital. For example, the Karolinska University Hospital in Stockholm and Sygehus Lillebælt in Kolding, Denmark are together currently developing AI technology to identify admitted patients at risk of acquiring nosocomial infections. AI also has the potential to predict communicable disease outbreaks and their sources, as well as predicting the incidence of adverse drug reactions.
Robot-assisted surgery has significantly improved the clinical outcomes of many surgical procedures. Recent research has focused on] developing reasoning algorithms to increase the autonomy levels of AI-assisted robots so that they can perform surgical procedures independently. In 2017, at the Maastricht University Medical Centre in the Netherlands, an AI-assisted robot successfully sutured microscopic lymphatic vessels to blood vessels to alleviate swelling in a patient suffering from lymphoedema.
6. Digital consultation
AI-assisted apps such as virtual nursing assistants, which provide personalised health advice, are already available to the public. These apps allow enquiries about medical issues to be answered quickly, reducing the stress placed on primary care providers. They also have the potential to provide medical consultation to populations who would not normally have adequate access to healthcare. The AI-assisted chatbot GP at Hand, developed by Babylon Health and currently being trialled in London, aims to combine patient-reported symptoms with available medical data to propose diagnoses.
What can we conclude?
AI-assisted technologies are already being explored and employed within healthcare settings. The recent advancements in this sector show huge promise in delivering solutions to global health demands and alleviating the strain placed on global healthcare systems. However, to optimise the algorithms that govern AI, vast quantities of high-quality data are needed; unfortunately, most health-related data is not currently standardised or digitised, so this restricts the development of AI within healthcare. There is also debate about AI’s ability to replicate human traits such as compassion and to read social cues. In addition, there are social and ethical issues associated with data use.
Despite the remaining questions about the practical application of AI in healthcare, and the predictions by high-profile scientists such as Professor Stephen Hawking and technology entrepreneurs such as Elon Musk that developing fully autonomous AI could lead to the end of humanity, the widespread integration of AI into healthcare models seems certain. Even with many issues still to be addressed, the potential benefits of AI could outweigh the associated risks.
Questions or comments about this blog?
If you have any queries on this article, reply to Phil below in the comments.
Was this article helpful?