What's Happening with CSUs (Data and Analytics)?

One year on from the Government announcing the closure of Commissioning Support Units, we take a look at what's happening with their data and analytics services as they progress with service transition and closure.

DSCRO, Data Management and Tooling

CSUs employ staff in the Data Services for Commissioners Regional Offices (DSCRO) who have a specific role in de-identifying data, plus wider data management functions, reporting platforms ('Aristotle', 'Athena', 'Raidr', 'Insights') and BI tooling. [1]

The future of these elements of CSU provision has been the subject of much concern and angst for ICBs (and others) who rely on CSUs for these crucial data and analytics services.

With the first in what promises to be a monthly bulletin on the future of CSU data and analytics services published at the end of May 2026, some clarity has now emerged. Two key actions confirmed are:

• NHS England will assume responsibilities for DSCRO and data management functions, delivering a nationally scaled service that aligns to the national strategy, with CSU data services transferring to NHS England by December 2026
• Existing CSU DSCRO, data management and tooling services will continue to operate as they do now to ensure continuity for 2027/28

This basically means that the CSU data management 'people' will transfer to NHSE by December 2026 and the existing CSU data 'infrastructure' will remain in place through to March 2028, whilst work continues on a single, national DSCRO/data management service.

As we know, NHSE is on the same closure timeline as CSUs (March 2027) so staff will need to find another home before then.

The Federated Data Platform (FDP) is still very much part of the picture with the latest bulletin confirming that the future national data service will require 'alignment with FDP architecture, standards, and security controls' and 'expectations for ICB adoption of FDP-aligned analytics solutions or alternate provision'. However, as the above confirms FDP is not able to fulfil the role CSUs were doing now, or before March 2028 at the earliest.

The main thing yet to be confirmed is how it's all going to be funded. The two obvious options here are direct charges to ICBs or a top slice of their funding. The tricky bit will be moving from a situation where almost every ICB bought a different service, for a different price from four suppliers to a 'one size fits all' service spec from one supplier. The implications here are both financial – an increase (or decrease?) in what an ICB pays – and operational – a potential loss of/gap in capacity and capability.

The bit that CSUs were really good at doing (but really bad at communicating) was delivering 'analyst ready' datasets. Doing the messy bits, the bits in between, bridging the gaps between raw data and something analysts can work with. De-duplicating large complex datasets like Mental Health (MHSDS) and Maternity (MSDS). Developing SUS and SLAM Reconciliation Tools (SSRT) to facilitate the work of commissioning and finance teams. [2] Support on IG. Setting up local data flows. There is no obvious home for this. Some ICBs will have this capacity in-house; many will not, as they bought this service from a CSU. Whether there is enough money left in the ICB coffers after they have paid for the national data service will be critical to their strategic commissioning work. Even in their larger sub-regional format following clustering and merger (27 ICBs instead of 42) the technical skills of data engineers, data architects, data optimisation and enablement experts, will be scarce, at least in the immediate term.

Analytics

CSUs provide broadly two types of analytical service – those enacted 'at-scale' for multiple customers (ICBs) by a central team and bespoke teams of analysts 100% aligned to one specific client (ICB) only – variously called 'embedded analytical teams', 'spoke team' or 'aligned analytics'.

On the whole embedded analytical teams will have transferred from CSUs, by TUPE to their respective ICB, minus anyone who has taken Voluntary Redundancy (VR) by then end of Q1 2026/27. In many cases those staff are moving from one Management of Change (MoC) process in the CSU to another MoC process in the ICB as ICBs try and meet their (up to 50%) reduction targets. Where this has necessitated ICBs clustering or merging, the job matching and slotting-in process is further complicated by bringing two teams together.

The ICB cost reduction targets were meant to be facilitated by the Blueprint for ICBs and the Blueprint for Regions. It remains somewhat opaque as to what ICBs can actually stop doing? 'Performance monitoring', for example, should be a function undertaken only by regions, according to the Blueprints. However, it is proving a hard habit to kick for ICBs charged with strategic commissioning and transformation and tasked with holding a variety of providers to account.

Central analytical services have an uncertain future. Whilst there's general consensus of economies of scale for data management, there is much less support for at-scale analytics; ICBs, understandably, want to directly manage their own analysts focusing on local priorities and aligning with local governance.

However, there remains a strong case for some central analytical functions, given (1) common areas of interest for all ICBs – e.g. reducing waiting lists, delivering and evaluating neighbourhood models of care, and (2) the requirement for highly specialist but infrequently required analytical expertise, such as geospatial analytics and demand and capacity modelling. There is a high risk of significant loss of expertise, not just from CSUs and NHSE, but from the public sector as a whole. Pushing such expertise out may prove expensive.

Consultancy

Across the CSUs there are several specialist consultancy services which are seeking a new home before the March 2027 closure date. [3] In my CSU – Mids and Lancs – we house the likes of Digital Innovation Unit, Transformation Unit, The Strategy Unit, NHS Horizons, the Health Economics Unit (HEU). South West & Central CSU is home to the Geo-Spatial Team, AGEM CSU have the Advanced Analytics Unit. All will hopefully find a new host and ensure they are able to continue providing services, although I'm only aware of the HEU that have actually confirmed the move as yet (now part of the Imperial College Healthcare NHS Trust).

Overall, much of the CSUs data and analytics services (including specialist consultancy) is hoped to shift to a new home, rather than being deleted altogether. That is not to downplay the huge loss of knowledge, experience and capability leaving en masse through redundancy. Conversations across CSUs, NHSE and ICBs indicate a substantial reduction in the data and analytics workforce – perhaps half this workforce leaving the NHS in 2026/27.

Prior to the announcement of NHSE and CSU closure in June 2025, the four CSUs were already on a pathway to merger and the formation of a single national data and analytics service, with active engagement from national, regional and local stakeholders.

There is a sense of a missed opportunity here for the evolution of a single national data and analytics services, rather than the revolution and turmoil of closure and redundancy. [4]

Revisiting the recommendations of the final report of the CSU merger Analytics Task & Finish Group from October 2024, shows that the recommendations still provide an apposite roadmap for data and analytics in healthcare:

• Establish a Strategic Customer Board to create a strong customer voice
• Optimise investment inside the NHS
• Realise economies of scale & skill by pro-actively sharing resources …with proactive code and product share within the NHS Family [5]
• Appoint national CSU subject matter experts to lead Communities of Practice to support thought leadership, knowledge exchange & capability development
• Pursue a strategy of convergence where there is clearly demonstrable benefit towards a common suite of tools, platforms & applications, leveraging national data assets such as FDP/SDE/CDM
• Establish a sustainable approach to innovation with increased use of open source & data science approaches to analytics (code, algorithms & Reproducible Analytical Pipelines (RAP)
• Provide dedicated Information Governance Subject Matter Expertise capability & capacity to leverage best value from national flows
• Ensure a holistic approach to insight is adopted
• Invest in the continued development of the analytical workforce, through adoption of the National Competency Framework and proactive engagement with professional bodies.

The Data Observatory

Amongst all this churn the Data Observatory has been forged to provide support and leadership for data and analytics in healthcare. It is 'by the people, for the people' and seeks to play an active role in mitigating risk and exploiting opportunities to make health analytics work.

Whilst the CSUs will close by March 2027, the Data Observatory, working alongside professional bodies such as AphA and FEDIP and self-organised groups such as the Chief Data and Analytical Officers Network, can play a key role in driving forward the better, broader and safer data and analytics services that healthcare demands.

References