Award-winning AI deployment engine
Understanding the problem
Artificial intelligence has the power to transform the healthcare industry. However, the time taken to move from clinical research to live clinical implementations of AI is currently very slow.
While progress is being made in areas such as imaging, projects often lack scalability, with the focus being on implementing a single solution at a time. Additionally, the NHS has strict rules on data sharing, meaning that it is difficult to access the large data sets needed to train AI algorithms efficiently.
As the NHS faces a 7.6% vacancy rate which includes a 29% shortfall of radiologists, and a backlog of over 7.5 million appointments, appetite for the benefits of AI is high. However, with legacy tech and creaking infrastructure also hindering progress, there are many areas requiring a share of limited budgets.
AI in healthcare is at an inflection point. PACS and RIS suppliers are hastily developing their own costly proprietary AI add-ons which tie trusts into the strategy of their PACS provider. This also limits the focus on other modalities outside of radiology.
At the other end of the spectrum is an emerging ecosystem of innovative AI model creators each trying to capture their own corner of the market with bespoke point-to-point integrations and proprietary infrastructure. While this provides flexibility for trusts, it is not scalable.
Unfortunately, both of these approaches lock NHS trusts and networks into single vendors, misusing the limited investment available for the NHS to adopt AI at scale and prepare it for future innovations in the AI sector.
A collaborative, NHS driven approach
Working with global tech leaders NVidia, world-renowned academic institutions, and innovative NHS healthcare providers as part of the Medical Open Network for AI (MONAI), we led the project to build a first-of-its-kind AI orchestration engine.
The technology, called MONAI Deploy, aims to become the de-facto standard internationally for developing, packaging, testing, deploying, and running medical AI applications in clinical production.
MONAI Deploy creates a set of intermediate steps where researchers and model developers can build confidence in the techniques and approaches used to deploy AI — allowing for an iterative workflow until the overall AI inference infrastructure is ready to move to clinical environments.
Award-winning approach
The approach is groundbreaking, shaping the AI deployment process for the NHS and providing a global exemplar for the application of AI at scale.
This award-winning approach was rolled out at seven NHS trusts, including Kings College Hospital NHS Foundation Trust (KCH), University College London Hospitals NHS Foundation Trust (UCLH) and Guys and St Thomas’ NHS Foundation Trust (GSTT).
An award-winning approach to accelerating NHS AI deployment
Commissioned by the AI Centre for Value Based Healthcare, Answer Digital developed the award-winning AI Deployment Engine (AIDE), based on MONAI Deploy technology. AIDE provides a safe, scalable, affordable way of deploying multiple, world-leading AI models, which successfully navigates the strict data sharing standards in the NHS and delivers value quickly to clinicians and patients.
Benefits include:
01
Democratised access to AI in the NHS
02
Transformed the way NHS organisations can deploy and manage AI models
03
Ability to trial models in Shadow Mode prior to live deployment to build confidence in the clinical teams whilst ensuring workflows aren’t disrupted until the teams are ready to deploy clinically.
04
Supported the deployment of a brand new stroke model into live minimum viable product in three months at KCH
05
Once live, it provides the capability to analyse the outputs of models in real-time, ensuring they are performing as they should.
06
Optimisation of cloud-based infrastructure.
07
Enabling the NHS to move with the ever-evolving AI landscape.
The results
The work of Answer Digital and the expert multidisciplinary consortium of partners not only led to the creation of AIDE, it also created a blueprint for information governance, clinical safety assessment, and scalable deployment processes.
Following this approach, Answer Digital worked with a team of clinical scientists at KCH to take an AI model for analysing stroke CT scans from ideation to live minimum viable product in just three months.
Over the next two years of the project, six other London and south-east trusts deployed AIDE, including East Kent Hospitals University NHS Foundation Trust that integrated the deployment engine in just four weeks.
Best-in-class CE marked AI models.
Following the completion of the AIDE project, funded by Innovate UK, we have gone on to form relationships with many best-in-class CE marked AI models.
The models which can now be rapidly deployed using Answer Digital’s AI Platform and integration blueprint cover chest CT, chest X-Ray, musculoskeletal, 2D/3D mammogram and prostate services.
By having access to multiple models that can be deployed in as little as one week, trusts choosing to follow an open platform architecture model are at the forefront of innovation and able to realise clinical benefits much more rapidly, rather than being locked out of it by being tied into a single provider.
Our AI platform, built on open standards, enables rapid deployment and scalability, even for the least digitally advanced organisations. With the same NHS resources and integration investment as proprietary solutions, our platform unlocks access to the full AI radiology marketplace. Trusts can trial and select the best models, with flexibility to expand into genomics, digital pathology, and clinical coding as their AI maturity grows.
Find out how to deploy AI quickly and at scale
Contact us to start your journey.
Leave your details and our team will be in touch within 24 hours.