A familiar pattern repeats across the NHS. Something is not working well enough. A solution is chosen, often the same kind of solution that was tried last time, and it is rolled out. Some time later, the problem is still there, or a new one has appeared in its place.

Professor Mary Dixon-Woods, Director of THIS Institute, names this pattern precisely in her 24 June 2026 article, “How can improvement and innovation save the NHS?” Her word for it is a “zombie solution”: an intervention repeated without properly understanding why it did not work before, without designing it for real-world implementation, and without a plan to evaluate whether it worked this time either (Dixon-Woods, 2026).

This article is not a summary of her piece for its own sake. It sets out, in three clearly separated parts, what Dixon-Woods argues, what VeinTrain and HALO already do that sits close to that argument, and what it means in practice for a service trying to improve IV Cannulation, venepuncture or phlebotomy performance. The three should not be read as one voice. They are kept deliberately distinct below.

Part one: what Dixon-Woods argues

Dixon-Woods’ central claim is that the NHS is not short of talented people or good ideas. Its recurring failure is procedural:

it reaches for solutions before it has properly diagnosed the problem.

 

Understand the problem before selecting the intervention. Some problems look like capacity problems (not enough staff, not enough appointments, not enough beds) when they are actually problems of process design. She points to time-and-motion evidence from general practice showing staff working hard while large amounts of their time are lost to “operational failures and compensatory labour”: extra, often invisible work created by poorly designed everyday processes (Dixon-Woods, 2026).

Adding staff, funding or technology on top of a badly designed process does not reliably fix it.

Avoid zombie solutions. Repeating a familiar intervention without asking why it failed before, how it will actually be implemented, and how success will be evaluated, wastes effort and distracts from the real problem (Dixon-Woods, 2026).

Co-design with patients and staff, and engineer the work system. Successful change is “rarely just a technical fix”; it needs deliberate attention to the social, cultural and emotional work involved, and to equity, fairness and trust. Technology should be treated as part of a sociotechnical system, people, tools, culture, practice, policy and infrastructure interacting together, not bolted on as a standalone fix (Dixon-Woods, 2026).

Design once at scale, then adapt locally. Unwarranted variation, differences in care that cannot be explained by patient need, is, in her words, “the bane of the NHS.” Rather than many local teams each reinventing a solution, she argues the NHS should use its position as a national system to design well once, centrally, and then support local adaptation (Dixon-Woods, 2026). 

Sarah Phillips would add to this that adaptation is a revisit to the double diamond adn expanding again the thinking togethor and opening possibiities - someone holding the dissoance and then enabling the convergence. Once all things have been considered then a test and learn, then scale locally, led by the leaders in the location who were part of the adaption. 

Look for positive deviance. Sometimes a workable solution already exists inside the system. The task is to identify high-performing units, understand what they are actually doing differently, and turn that into something transferable, with support, rather than assuming every service must start from nothing (Dixon-Woods, 2026).

Address culture, voice and institutional conditions. Some of the hardest problems, bullying, racism, harassment, fear, secrecy, credibility gaps, are relational and cultural, not technical, and are unlikely to be solved by a new system or piece of technology (Dixon-Woods, 2026).   

Build evaluation in from the start. She identifies evaluation as a particular NHS weakness: too little learning is formally captured, and what is captured is not reliably used to shape what happens next. She calls for evaluation to be designed into implementation from the beginning, with methodological and infrastructural capacity to deliver it closer to real time (Dixon-Woods, 2026).

Use digital collaboration to widen participation, without pretending everything can happen remotely. Online tools can widen who takes part in co-design and shared learning, but Dixon-Woods is explicit that understanding problems, identifying solutions and evaluating them “will always need to happen in person” as well (Dixon-Woods, 2026).

Part two: what VeinTrain and HALO already use

VeinTrain has not built its own competing improvement methodology, and this article does not claim to. Where our existing practice sits close to Dixon-Woods’ argument, it does so because it draws on established improvement and co-design thinking already in use across VeinTrain and HALO: the Double Diamond (discover and define the problem, then develop and deliver the response), PDSA cycles for small, safe tests of change, and systems thinking that looks at people, processes, culture, policy and infrastructure together, not in isolation.

In practice, Sarah’s working method for a piece of improvement or co-design work has followed a consistent sequence: identify exemplars (services or practitioners already getting something right), align strategically and diagnose the real local problem, run a structured co-design meeting with those affected, review and test, including simulation-based testing of the pathway rather than only the procedure, evaluate, and then adapt what works locally rather than mandating a single fixed template. This sequence is internal working practice, described here as VeinTrain’s existing method, not as a finding from Dixon-Woods’ article and not as a validated protocol.

Dr Amar Shah’s public NHS improvement work is relevant context here. He is the first National Clinical Director for Improvement for NHS England (a part-time secondment from January 2024) and has led quality improvement as Chief Quality Officer at East London NHS Foundation Trust for over a decade (ELFT, 2024; ELFT QI, n.d.). His work prompted an earlier internal discussion thread at VeinTrain about returning to core co-design and the Double Diamond rather than moving straight to a solution. That LinkedIn post itself has not been independently verified and is not quoted here; it is credited only as the prompt for the internal thinking that follows.

Part three: our interpretation, applied to vascular access

None of the above is a reason to add another improvement framework. It is a reason to be disciplined about the order operations happen in.

Applied to vascular access, training and implementation, the working sequence looks like this: find what works (identify exemplars and any positive deviance already present in a service) → understand the problem properly → co-design with those affected → engineer the system around the skill → test and simulate → adapt locally → evaluate → learn → scale. Evaluation sits inside this sequence from the start, not bolted on at the end.

The clearest illustration is the difference between two starting questions.

The first question is: “We need more IV Cannulation training.” The second is: “Why is IV Cannulation performance, or patient experience of IV Cannulation, actually poor here?”

The first question reaches straight for a solution. The second stays with the problem long enough to find out what is actually driving it. In a given service, the answer may genuinely include a gap in competence development. It may equally include, illustratively and not as an assumption about any named organisation: equipment or device selection, unclear escalation pathways for difficult venous access, local policy that has not kept pace with current standards, workflow and staffing pressure, thin supervision, low practitioner confidence, team culture around asking for help, or a lack of usable data on where difficulty and complications are actually occurring. Training may be part of the answer. It is rarely the whole answer, and reaching for it first, before the diagnosis is done, risks becoming exactly the kind of zombie solution Dixon-Woods describes: repeated because it was tried before, not because it has been shown to fix the actual problem this time.

This is where HALO’s emerging role in champion-led practice change fits. HALO is not positioned here as a delivery mechanism for more courses. It is positioned as support for local clinical champions: people inside an organisation who are equipping themselves, with evidence, tools, implementation support and a structured route, to change practice from within, rather than waiting for an external mandate. That is consistent with Dixon-Woods’ point that national systems should design well once and support local adaptation, and with VeinTrain’s existing method of identifying exemplars and co-designing with the people who will actually use the change.

What is not being claimed. This article does not claim that HALO has demonstrated measurable outcomes from this approach, that every organisation using HALO already runs a full diagnose-before-you-train process, or that Dixon-Woods, THIS Institute or Dr Amar Shah endorse VeinTrain or HALO. It is a statement of method and direction, offered for Knowledge Hub use.

Competency and governance statement

A VeinTrain certificate documents learning or training completion. It does not replace local employer competency sign-off, local policy, supervision or scope of practice. Final competency confirmation remains with the employer, supervisor or clinical organisation. Where this article discusses diagnosis, implementation and evaluation of practice change, that work sits alongside, and does not substitute for, existing employer governance.

AI has supported the finding, organising and translation of evidence for this article. It has not made, and does not make, clinical or governance decisions; those remain with qualified humans.

Conclusion

Better NHS improvement, on Dixon-Woods’ account, starts by

staying with the problem before reaching for the solution.

VeinTrain’s existing improvement and co-design practice, and HALO’s emerging support for champion-led practice change, are built in that spirit: find what already works, understand the problem, design with the people affected, test before scaling, and evaluate throughout rather than at the end.

Applied to vascular access, that means treating “we need more training” as a hypothesis to test, not a starting assumption.

To talk about diagnosing a vascular access practice problem before choosing a solution, including how HALO supports local champions through implementation:  

Read more on the route from training to confirmed competence in the VeinTrain Knowledge Hub:

Dixon-Woods, M. (2026) How can improvement and innovation save the NHS? THIS Institute (The Healthcare Improvement Studies Institute), 24 June. Available at: https://www.thisinstitute.cam.ac.uk/blog/how-can-improvement-and-innovation-save-the-nhs/ (Accessed: 31 August 2026).

East London NHS Foundation Trust (ELFT) (2024) Chief Quality Officer, Dr Amar Shah, is First National Clinical Director for Improvement. Available at: https://www.elft.nhs.uk/news/chief-quality-officer-dr-amar-shah-first-national-clinical-director-improvement (Accessed: 31 August 2026).

ELFT Quality Improvement (n.d.) Prof Amar Shah. Available at: https://qi.elft.nhs.uk/team/dr-amar-shah (Accessed: 31 August 2026).