Skip to main content
Blog

5 key themes for the future of the NHS Spine

Earlier this year, we talked to our strategic partners across the public and private sectors about how we can evolve our Spine platform to meet the needs of a future health and care system. Ash Raines, Head of Delivery for the Platforms Directorate, shares the key themes and insights we’ve learned.

Our Spine platform remains the backbone for NHS service delivery, providing secure access to many core products including 65 million Summary Care Records (SCRs) and over 90 million Patient Demographic Service (PDS) records.

The author, Ash Raines, at his workstation at home.

This year we marked 6 years since NHS Digital brought Spine in house and we have seen a significant rise in demand. We now regularly see more than 1 billion transactions per month and we’re continuing to add features for health and care professionals.

Spine has performed brilliantly amid the increases in demand we have seen during the pandemic, thanks largely to the dedicated teams behind the scenes. But we now have a new challenge: how do we evolve Spine to continue supporting the needs of a 21st century health and social care service?


Our vision

We want to build and run a more open platform. It will consist of discrete services that can evolve independently to ensure they continue to make use of modern technologies and open standards. It will be supported by robust but efficient governance processes to ensure safety, while making it as easy as possible for new entrants to join the market. We’ll get a richer mix of system providers creating more new and innovative solutions.


Our challenge

We know Spine is pivotal to delivering interoperability of health and social care nationally.

We know there are barriers to the creation of new services for health and care professionals and for patients.  We want to have greater flexibility to scale up quickly, reducing build times. However, as the current service expands, we face hardware and software support challenges as products start to reach end of life. These are compounded by emerging constraints on our data centre estate, floor space, rack space and inter-rack cabling.

So, earlier this year we started discussions with our strategic partners to explore how we could evolve Spine and support cross-government long-term sustainability efforts. We set up a series of workshops to help us understand and challenge different approaches to building future national platform capabilities for the NHS. These are ongoing discussions, but I wanted to share some key themes and insights that are emerging.


1. Hosting

As you would expect, responses from strategic partners focused on the use of public cloud, with some suggesting hybrid or cloud agnostic approaches. This is an approach we support, as the cloud will have a pivotal role in any future Spine service. 

Today, Spine is already using hybrid cloud services to deliver internet access to the services it supports and we’re considering some of our low-risk, high-value services as potential early cloud migration candidates. For example, services such as MESH and the archival store of Summary Care Records, are ideal candidates balancing the scale of processing and storage, with the demand profile of recent coronavirus-related activities.

The use of public cloud also allows more business value to be delivered in the time available, by providing more flexible and on-demand infrastructure and reducing the operational responsibilities.  However, cloud in itself is not a panacea. We must consider the long-term approach to using cloud for critical national services, weighing the benefits of cloud against potential dips in availability and ensuring the system architecture accounts for this.

We did discuss the benefits of using a single region (with multi-availability zone) versus using a multi-region approach. However, the proximity of datacenters is largely irrelevant for Spine hosting. That said, multi-region can have a role to play in achieving higher availability and, in theory, this raises the 99.89% potential maximum single region availability to 99.99% when using multi-region. 


2. Target architectures

Several target reference architectures were discussed, and most partners highlighted the benefits of moving from a monolithic codebase, team and deployments to a microservices-focused approach. This would allow for independent deployments, independent scaling, smaller impacts of change and a wider technology choice.

However, there is some trade off to consider when thinking about adopting a microservices approach, such as a natural tendency toward inconsistencies in approach. Additionally, some partners discussed the potential of moving directly to serverless architectures. It’s certainly the case that smaller teams could select more specific architectures with microservices.

Key design patterns were a large part of the workshops. Spine today follows a mainly broker-model routing, with each domain having events and command query responsibility segregation (CQRS) for create, read, update and delete (CRUD) events. There were 3 main alternative patterns that came out of the workshops:

  1. API-first - a fully synchronous API-first approach, using a service mesh to coordinate across the various microservices.
  2. Mesh App and Service Architecture (MASA) - taking the API-first approach further with key features such as being entirely modular. Services can be developed individually with no cross-dependencies.
  3. Enterprise Service Bus (ESB) - an event-driven approach could be considered, although these do create single points of failure and complexity. This would also allow fully asynchronous events (similar to most updates in Spine today). However, asynchronous events have been noted as one of the features that makes Spine ‘hard’ to integrate with.

3. Transition approach

Multiple transition approaches were discussed. All approaches were mindful of the need to keep the risk profile low and the impact on the existing service minimal. Here’s 4 that we covered:

  1. Strangler Pattern - a pattern designed by Martin Fowler as a way of migrating services. The premise being that individual services are migrated until there is nothing left to move.
  2. Lift, Shift and Evolve - an initial lift and shift of the existing service to the cloud and then an in-place migration away from legacy concerns such as Python 2.7 with an API Façade in place to ensure consistency to end-users.
  3. Red Squad - an approach adopted by Spotify. A new environment is built out on the cloud and then they used a special team to move teams and services across one-by-one. This does require a service that is already built into more discrete services than Spine.
  4. Façade - there exists an opportunity to use the relatively new API Gateway (see Tony Heap’s Tech Talk) as a façade for services as they are migrated.  This could be paired with the Strangler Pattern to minimise disruption.

4. Technology modernisation

Most strategic partners promoted the idea of presenting Spine services over the internet, an approach that aligns with both NHS Digital’s and GDS’s Internet First strategies. Some examples were provided of where we could replace current Spine technologies with cloud native equivalents. For example, we could replace RabbitMQ with AmazonMQ or replace NGINX with the Amazon API Gateway (which now supports TLS-Mutual Auth).  Some partners went into more detail about how technology modernisation could be achieved in stages, again with cloud-native technologies. 

The consensus view from all partners was to prioritise areas of technical, architectural and operational debt without increasing the risk to the live service.


5. Future transformation

There were exciting discussions that took us beyond immediate operational technology concerns toward a bold vision for the Spine data platform. By taking the Spine service on a journey through full cloud migration, artificial intelligence readiness and then into person-aware services, there is the prospect of a truly transformational approach for the NHS.


Just the start

We have lots to do and this is just the start of a long journey but it’s a really exciting one! We will be engaging more partners and organisations over the coming months and all of these contributions will inform the development of our thinking and strategy, before any decisions are made.

We are also really mindful that, as we explore this very exciting future, we must continue doing what we do best, providing a secure and reliable platform that works 24 hours a day, 365 days a year for the health and care system.

We have a fantastic team that sits behind the Spine services, all contributing to serving our healthcare professionals and patients. Some of those colleagues don’t know what we do, but we know they’re relying on us.

Interested in working at NHS Digital? Search our latest job opportunities.


Related subjects

Spine supports the IT infrastructure for health and social care in England, joining together over 23,000 healthcare IT systems in 20,500 organisations.

Author


Last edited: 22 December 2021 3:53 pm