Chief Data and Analytical Officer Network position on the Federated Data Platform (FDP)

Marc Farr
Chair – Chief Data and Analytical Officers Network
Association of Professional Healthcare Analysts
20th February 2025

Ming Tang
Chief Data and Analytics Officer
NHS England

Dear Ming

As the Chief Data and Analytical Officer’s Network (CDAON) we thought it would be helpful to provide a position on the Federated Data Platform (FDP) which is a source of much discussion across the NHS. There appears to be two main dimensions to the discussion: the need, purpose, objectives, and implementation of the FDP; and the supplier(s) chosen to provide the technology upon which FDP relies. It is important to make a distinction between these two dimensions. In respect of the first main dimension of discussion – the need for FDP – the CDAON recognises and fully supports the need to establish a capability to federate data from many sources at the national level. For too long health data have been stored in different places with no means of connection and linkage and therefore we have been unable to realise the true potential of the NHS’s unique data assets. A national capability that enables us to connect our many national health datasets, and make these readily available to local and national stakeholders outside a proprietary system is absolutely essential.

The CDAON is entirely supportive and actively engaged in achieving the fundamental objectives of the FDP. However, we have also expressed some concerns with the implementation and approach to adoption of it. These concerns centre on three aspects. The first relates to programme drift. The original business need for FDP is about creating a data connection capability – it is about federating data and interoperable standards. It should not be about imposing or advancing adoption of specific software solutions and yet the current rollout of FDP appears to be entirely associated with the adoption of specific tools or modules on a single software platform rather than the integration of data, we know that you have acknowledged this yourself.

This presents an issue for many CDAOs because we already have similar tools in use that presently exceed the capability and application of what the FDP is currently trying to develop or roll out at a system level. The second concern is that the FDP is an NHS solution for health, predominantly NHS, data. At a system level we recognise that the health needs of the population are met through a combination of services both healthcare and non-healthcare related. A nationally commissioned platform (FDP) will not be able to meet all the bespoke requirements of a local system to support the health of its citizens through better data and intelligence. This is best done through locally designed and managed infrastructure and CDAOs and other partners across the system are in discussion with a range of government departments about how best to integrate wider public service data at a system level.

Lastly, the national investment made in the FDP has been almost exclusively set against technology and trialled in environments primed to benefit from this technology adoption. Digital industry research highlights that adoption costs often outweigh the costs of technology itself, a point that I believe you would sympathise with as well. The adoption and ongoing maintenance costs as part of a total cost of ownership that will sit with those who choose to use FDP needs to be made very clear before they can commit – especially with many systems under significant financial stress – and at present this clarity and an independent evaluation of the cost-benefits of adoption is not available.

The second dimension to the general discourse on FDP – the choice of supplier(s) to provision the FDP – is not a matter over which CDAOs have a say [although we believe as a general point that it is a missed opportunity to not have been more involved in the procurement as potential future users]. It is for NHSE to address any public concerns about procurement or its choice of supplier(s). However, public trust around the use of sensitive data is hard won and easily lost. This relationship of trust with our citizens requires constant and careful attention and we receive regular expressions of interest and concern about health data use including specifically the extent of our engagement with FDP and its supplier(s). We must take into consideration in the development of our local data assets and demonstrate responsiveness to these concerns if we are to safeguard and preserve this trust.

In summary and to be clear, the CDAON is fully supportive of the need to federate national NHS data and make this available for the benefit of patients across the country. We are using our experience of having done this locally to support the national developments and are providing constructive feedback on how to ensure that local and national infrastructure can co-exist in supporting meaningful innovation that retains the public’s trust.

We hope that you and your team will find the points below useful in your planning and would like to offer any assistance that we can in developing your roadmap further.

Key Questions for Comment

1. Existing Investments and Capabilities

ICBs have made significant investments in local data infrastructure as previously directed by NHSE, including:
· Carefully designed relational databases and data linkages
· Information governance arrangements for lawful data linkages and secondary use of GP data
· Trained analysts proficient in SQL, R, Power BI DAX, and Python, supported by reusable scripts for reporting
· Established contracts, software, dashboards, and outputs tailored to local needs
Transitioning to FDP raises concerns about potential disruptions and the risk of losing existing functionality and value. Clarity is needed on whether a partial adoption model is feasible and how local capabilities can be integrated into the platform.

2. Skills and Workforce Impact

There is uncertainty about the usability of key skills (e.g., SQL, R, Python) within FDP. How can we mitigate the risk of deskilling our analytical workforce while ensuring they remain equipped to deliver value in the new environment?

3. Data Access and Integration
· Will SUS data remain available through the DSCRO, or will FDP become the mandated source?
· What level of access will ICBs have to SUS data at both patient and aggregate levels?
· If CDS can be on FDP, will ICBs be able to federate access to each other’s CDS, as data owners to enable benchmarking?
· Will there be any new national datasets available to ICBs through FDP in the next three years? It would be helpful to have early sight of the plan for these so that we do not duplicate these additional collections and flows locally.
· How will local datasets be ingested and linked at a patient level, and what pseudonymisation processes will be in place?

4. Compatibility with Existing Tools
Many of us rely on tools like Power BI, Tableau, and QlikView. Will FDP allow these tools to connect to its back end, or will we need to rebuild dashboards within FDP? If the latter, adequate transition time and support will be critical.

5. Product Quality and Engagement
The Health Inequalities Improvement Dashboard, an early FDP product, has raised concerns regarding usability and quality. Feedback from recent sessions did not result in clear actions or reassurances, leaving us uncertain about how local needs and input will be accommodated.

6. Transparency and Governance

We are concerned about potential reliance on proprietary products and AI tools developed by Palantir, particularly if these are inaccessible to local analysts or lack transparency, contrary to the principles of the Goldacre Review.

7. Communications and Engagement
Current communication channels and engagement forums are insufficient to address technical questions or foster meaningful collaboration. Greater transparency around steering groups, membership, and decision-making processes is needed to ensure inclusive and representative discussions.

8. Relationship to CSU
Many of the ICBs are reliant on CSU for their data warehousing and data engineering, we do not have the resources to run two systems in parallel. Given the move towards one DSCRO, is that FDP? We need clarity as to the future state of NHS data warehousing so we invest in the right solution.

Moving Forward

We believe FDP has significant potential and want to work collaboratively with you and NHS England colleagues to realize it. To this end, we respectfully request:
1. A written response addressing the technical details and commenting on the points raised.
2. Further opportunities for technical discussions with your team to ensure alignment with local needs.
While many of us have signed up for onboarding, this does not equate to full endorsement. Instead, it reflects our recognition of FDP’s potential and our desire to contribute constructively to its development.

Thank you for considering our feedback. We look forward to your response and the opportunity to engage further on this critical matter.

Kind Regards
Dr Marc Farr LFEDIP
Chief Data and Analytical Officer

East Kent Hospitals NHSFT and NHS Kent and Medway

National Chair, Chief Data and Analytical Officer’s Network, APHA
https://www.aphanalysts.org/nhs-chief-data-and-analytics-officer-network/
E: marc.farr@nhs.net