Despite artificial intelligence’s (AI) considerable potential to improve the way we deliver healthcare in the UK, its implementation may prove premature if infrastructure adjustments are not prioritised first. In a recent report compiled by the Department of Health, NHS England, and AHSN Network, NHS IT infrastructure was found “not fit for AI” due to a lack of “standards to facilitate data sharing and the development of appropriate commercial models to leverage the value of public/NHS data”.
At the same time, much has been said about the potential of AI to be truly transformative for healthcare delivery, and isolated examples of innovation are beginning to appear.
Alder Hey Children’s Hospital, for example, is one of the most cited instances where AI is being used within an NHS context, with an AI patient app launched in 2016 in partnership with IBM Watson to improve understanding and communication between patients and doctors within the hospital. The plan is that later iterations of the application would even be able to offer diagnostic information; however, progress in this area is limited by challenges around patient data governance.
More recently, in May 2018, Great Ormond Street Hospital (GOSH) announced a partnership with Microsoft and University College London (UCL), which will see the hospital experiment with new proof of concept AI-driven tools that Microsoft and UCL create. Already in operation is the use of recording equipment and AI transcription to analyse environmental data during surgeries to improve insights into best practices that can be used in training.
However, it seems that such examples remain the exception rather than the norm, as public sector IT decision makers see the potential, but not the practicality, of implementing such systems on a mass scale. This was echoed a recent study conducted by SolarWinds on A World Powered by Tech Pros, where IT professionals in UK public sector organisations selected AI as third in a ranked weighting of the top new technology advancements that they would use to solve challenges within their IT environment if they had more time.
So, if investment in AI technology could be putting the cart before the horse, what should the priority be? Scoring first in the survey above for the top technology that UK public sector IT professionals would implement if they had more time is the cloud.
Compared to AI, it’s easy to see why cloud might be considered a less glamorous, less transformational technology in the eyes of the business leaders who control the budget, but in reality, it could be exactly the type of bedrock that would make NHS IT infrastructure ready to support the demands of AI. Cloud is critical in providing not only the data sharing required for AI, but the public cloud is also one of few viable options to support the huge amount of compute needed for real-time, at-scale data processing.
As mentioned earlier, implementation of the cloud has been delayed, in part by challenges around patient data governance, like those experienced by Alder Hey Children’s Hospital. Research earlier this year by Healthwatch England suggested that, while 73% of U.K. adults would be happy for the NHS to use their information to improve the healthcare treatment of others, there are rightfully stringent controls on data sharing and privacy within the organisation.
The good news
The good news is that embracing the public cloud in a way that is efficiently managed, accounting for the right data sharing and privacy requirements, also provides a solid basis for the NHS to then be able to deploy AI on top of. Both systems boil down to the same principles: ensuring the right people and systems have access to the right information in the right format at the right time.
Achieving a secure cloud future can depend on a smooth transition between current legacy environments to a more agile cloud system, both of which require monitoring tools capable of working across both existing and new platforms. With data in the NHS hopefully being used to save lives, it would be good to check that the cloud environment is actively monitored to ensure that systems are reliable and secure. System owners would benefit from including tools that look at resiliency and redundancy to provide a benchmark of how well the system is performing.
In addition, the NHS could benefit from implementing endpoint security tools and patch management, all of which can be monitored from one central dashboard. This can help make sure that the devices accessing the cloud data platform do not become vulnerabilities (by checking each connecting device has the right antivirus, firewall, and operating system settings to reduce risk).
Finally, IT team visibility into which systems or individuals are accessing patient data records can help ensure that sensitive information is only being used for approved purposes, and that individuals or systems are not misusing the access to this data, either maliciously or by mistake.
With these controls in place to help ensure a reliable, secure cloud solution, the NHS will then likely be well on the way to having an infrastructure able to support AI. As a result, while Alder Hey Children’s Hospital and GOSH will remain leaders in terms of AI deployment, we will be able to see broader use of AI across different departments and NHS trusts. This means we will be one step closer to seeing the considerable benefits that AI promises, but only if we put the horse first, and build from a secure cloud foundation.