Consequences, and the Counter-Case

Part 5 of The State of DHCW. Numbered citations resolve at the sources page.

Consequences for clinicians and patients

Senedd Research states the systemic risk plainly: “Across Wales, hundreds of separate digital systems mean that patient information which should flow seamlessly between services frequently does not, with potentially serious consequences for patient safety.”^[26]

• Cross-border records. A Welsh Government written answer of 7 June 2024 confirms: “Unfortunately, GP to GP electronic transfers of the medical record cannot take place between England and Wales due to differences in IT system architecture.” Paper records are sorted and forwarded by courier instead.^[11] BMA Cymru Wales told the Welsh Affairs Committee that restrictions “can therefore lead to a reliance on paper-based communication [that] often delay the flow of vital clinical information. Digital solutions must be prioritised.”^[57]
• Fragmented clinical systems. The Royal College of Podiatry told the 2023 joint scrutiny: “To date, the lack of strategic planning and coordination by DHCW has seen a proliferation of different systems which do not communicate with each other (for example, WISDM, WCCIS, WCP, PARIS, WNCR, Therapy Manager).”^[56]
• Independent corroboration. The Bevan Commission warned in August 2023 that “for too many years the only health and care data that the NHS trusted was the data it generated itself … Counterparts internationally, including in England have acknowledged this, with Wales now at risk of falling behind,” identifying silos, incompatible platforms and unclear accountability.^[55]
• Front-line frustration. The Royal College of General Practitioners described “frustration” among members over information sharing and referrals; consultant respiratory physician Dr Simon Barry cited a “lack of ownership and management of workstreams”; and the National Endoscopy Programme clinical lead told the Health and Social Care Committee that changes to the bowel screening information management system had been requested “10 years ago, and that still hasn’t happened,” identifying “bottlenecks” within DHCW.^[22]
• Usability at the bedside. An independent mixed-methods evaluation of the Welsh Nursing Care Record found negative experiences “mostly related to IT and hardware issues, such as slow log in processes and lack of available devices,” compounded by “issues related to running paper-based and digital systems in parallel.”^[58]
• Coronial concern. Responding to a Prevention of Future Deaths matter in February 2024, the Welsh Government acknowledged “difficulties for clinicians in providing appropriate diagnosis and treatment for patients, as a result of information not being readily accessible or available within patients’ clinical records in the absence of electronic records.”^[59] In the Senedd, Rhys ab Owen MS cited 36 coroner alerts issued across England and Wales in a year about lack of information sharing between computer systems.^[30]
• Digital exclusion. An estimated 4% of Welsh adults do not use the internet at all, with exclusion concentrated among older, rural and low-income people; Audit Wales estimated in March 2023 that around 32% of over-75s were digitally excluded, against 7% of over-16s.^[26][22] A digital front door that is not universal shifts, rather than removes, access barriers.

The counter-case

A fair assessment has to record what DHCW has delivered and where blame is misdirected. On the evidence:

• Corporate governance process is well run — but the audit verdict is narrower than it looks, and the public record is thinner. Audit Wales has repeatedly found effective board conduct, livestreamed meetings, papers published a week ahead, live Welsh translation, BSL on request, robust risk management, unqualified accounts and effective financial control.^[12][13] That is real and should not be dismissed. Three qualifications matter, however — set out below.
• Real delivery successes exist — the ones that are actually DHCW’s. Wales was the first UK nation to share clinical portal records across all health board boundaries.^[26] EPS rollout is running a year ahead of its original timetable.^[4] The Welsh Nursing Care Record went live across all eligible inpatient wards in 58 hospitals. The Dental Access Portal enrolled over 83,000 patients in its first year.^[4] The Welsh Immunisation System was reconfigured fast enough to let Wales start RSV vaccination “at the earliest feasible opportunity,” with test automation covering over 80% of the codebase supporting 14 releases in one quarter.^[67] More than three-quarters of DHCW’s resource goes to running over a hundred operational services at high availability^[4] — though DHCW itself notes that some of the 61 major incidents in 2024–25 “were due to issues with third party suppliers or infrastructure problems at health board premises,” so the headline figure is not purely a measure of DHCW-controlled performance.^[67] The EPS claim also needs its scale stated: “a year ahead of the original programme timetable” describes a rollout that was live in 135 GP practices — 37% — at January 2026.^[4]
• Some failures are structurally not DHCW’s to fix. DHCW hosts programmes but cannot compel adoption. On WCCIS, some health boards publicly stated they had “no intention” of using the system^[22]; the Auditor General’s diagnosis was that a “patchwork” of local implementation choices — not central incompetence alone — destroyed the integration benefits.^[16] On RISP, the master services agreement lets each organisation unilaterally move its own go-live date.^[4] Both LIMS and RISP were established and run by the NHS Wales Collaborative from 2019 and only transferred to DHCW in January 2023, already requiring remediation.^[4] The INPS collapse was a supplier insolvency. Audit Wales itself notes that DHCW’s escalation plan contains “milestones that require input from outside partners.”^[12]

Six claims the evidence does not support

Not supported by the evidence

1. A qualified or adverse audit opinion. None found. DHCW’s accounts have received unqualified opinions, including for 2024–25 on 27 June 2025.^[12] Criticism concerns delivery, benefits and programme governance — not financial regularity.

2. A 2023 DHCW patient-data breach. No evidence was found that the MOVEit/Zellis supply-chain incident affected DHCW, NHS Wales or Welsh patient data. The frequently cited Welsh patient-data publication incident — 18,105 people’s Covid-19 data exposed for about 20 hours — was Public Health Wales in August 2020, a different organisation and a different year.^[55a] No ICO enforcement action against DHCW or NWIS was located.

3. Radiology system outages as a DHCW failure. DHCW’s board reporting explicitly excludes Welsh Radiology Information System outages from its availability figures because “WRIS is not a service hosted by DHCW (they are hosted in local organisations).”^[43] Attribution needs care.

4. That NHS Wales’s AI diagnostic “firsts” belong to DHCW. They do not. Brainomix 360 Stroke — the first national commissioning of stroke AI in the UK — was procured by NHS Wales Shared Services Partnership “acting on behalf of Cardiff and Vale University Local Health Board,”^[73] and rolled out from October 2023 under the National Stroke Programme, led by the National Clinical Lead for Stroke with health boards building the centre network; the announcement came from NHS Wales Performance and Improvement.^[74] Ibex Galen was deployed by Betsi Cadwaladr UHB — “Betsi Cadwaladr University Health Board has rolled out an ‘AI pathologist’” — with five of its own consultant pathologists, and funded via the Welsh Government Innovation Fund and the SBRI Centre of Excellence, not DHCW.^[75][76] DHCW appears nowhere in either attribution chain. Senedd Research describes both as NHS Wales firsts; folding them into DHCW’s credit column is a misattribution the sourcing does not permit. See the corollary below.

5. That the SAIL Databank is a DHCW asset. It is not. SAIL “is based within the Medical School, Swansea University” and is “Funded by Health and Care Research Wales.”^[83] DHCW’s own evidence pack treats it as a third-party service its planned secure data environment will complement — describing the SDE as “an anonymised service supporting research and innovation activity, complementing existing services like Swansea University’s SAIL Databank.”^[4] Wales’s strongest asset in trusted population-data research is a university facility, not a DHCW product — the same misattribution pattern as the AI diagnostics at point 4, and a further instance of the argument below.

6. Leadership misconduct or a substantiated whistleblowing scandal. No employment tribunal judgment, named union dispute or substantiated bullying finding involving DHCW or NWIS was located, and no published document states what any of the whistleblowing concerns alleged. That absence must not be read as exoneration — for the reasons set out in Part 3, the public record is structurally incapable of showing live proceedings. What is documented is a cluster of internal-control and culture findings: three raising-concerns investigations “related to values and behaviours and processes being adhered to,”^[60] a Limited Assurance recruitment audit commissioned in response to a whistleblowing concern,^[40] a 15-recommendation culture and wellbeing review,^[62] and a sharp fall in staff confidence that concerns would be addressed.^[64] That is a governance and culture problem evidenced in DHCW’s own records — not a misconduct allegation, and it should not be reported as one.

The awkward corollary: what succeeded, succeeded locally

The attribution corrected at point 4 above does more than tidy a credit column — it exposes a pattern that runs against DHCW’s founding premise. Set the two columns side by side.

The contrast is not incidental. In its 2024–25 annual report DHCW records that the Digital Maternity “delivery model for the Programme has changed, and Welsh Government is now funding individual health boards to implement locally. Therefore, the national programme, managed by DHCW, will close in 2025-26,” retaining only a data-standards project.^[67] The lessons-learned review confirms the procurement itself was abandoned: evaluation completed December 2024, “the outcome was a decision to terminate the procurement process,” with supplier feedback that “the service management requirements were disproportionate compared with what the market could realistically provide.”^[71]

Connecting Care went the same way: “The programme faced challenges in establishing effective collaborative governance … It was agreed to devolve responsibility for social care delivery back to the local authorities,” and for mental health “it was agreed that Betsi Cadwaladr University Health Board would progress with procurement as a pathfinder for other health boards.”^[67] On CareDirector itself, DHCW concedes: “the system may not have fully met the needs of all service users.”^[67]

Reading the governance verdict correctly

“Good governance” is the most-cited defence of DHCW, and it needs handling carefully, because the audit finding is narrower than the phrase implies and the published record is thinner than the transparency claim suggests.

First, Audit Wales explicitly excluded delivery from scope. Both structured assessments say so in terms: “We have not reviewed DHCW’s operational arrangements as part of this work” (2024) and “We did not look at the body’s operational arrangements” (2025).^[13][12] The four questions asked were about board conduct, assurance systems, planning and financial management. So the positive verdict certifies that DHCW runs its board properly — not that the board is being told the right things, and not that delivery is sound. That is why Audit Wales can rate corporate governance effective in September 2025 while the Welsh Government escalates the whole organisation for governance, leadership and culture seven months later.^[1] The two are not contradictory; they are measuring different things, and only one of them was measuring the thing that failed.

Second, the public minutes are AI-generated summaries. Every page of DHCW’s public board minutes carries the header “Copilot assisted.”^[70][72] The resulting record is a bulleted précis of conclusions rather than an account of discussion: positions taken, challenges made and disagreements aired are largely absent. A reader cannot tell from the minutes who questioned what, or what the executive said in response. Livestreaming mitigates this for anyone who watches in real time, but the durable, searchable, citable record of how DHCW’s board reasons is a machine summary.

Third, the substance is routinely in private session. DHCW publishes “PRIVATE Abridged” minutes, and the pattern of what goes into them is consistent: the Limited Assurance recruitment audit and its whistleblowing origin;^[40] the three raising-concerns investigations and the culture review;^[60] the recruitment spot-checks follow-up;^[63] a confidential counter-fraud investigation;^[61] and four private audit actions.^[61] Two further gaps are notable. The Limited Assurance recruitment audit report has not been published, although DHCW publishes other internal audit reports in full on its website — the Declarations of Interest, Follow-up of Recommendations and NDR reports are all openly available.^[41][39] And the 2024–25 Duty of Quality “Always On” reports were drafted but, in Audit Wales’s words, “remain unapproved and unpublished.”^[12]