Problem-First Procurement: What Buying a House Taught Me About MedTech Innovation

When I bought my house, I had to commission a survey.

This is standard practice in England and Wales. The buyer pays, the surveyor inspects, and the report belongs to the transaction. If the sale collapses, the report is largely worthless to the next buyer, who commissions their own. Same property. Different surveyor. Fresh cost. No institutional memory.

Scotland operates differently. Under the Scottish system, the seller commissions a Home Report before the property goes to market. Every prospective buyer receives the same document. The condition of the asset is established before the market engages with it.

This structural difference lodged itself somewhere useful, because it maps with uncomfortable precision onto something I observe repeatedly in NHS MedTech procurement.

When a MedTech company approaches an NHS provider, it arrives with a product and an estimated benefit case. That estimate is not a proven value proposition; it is a reasonable projection based on published literature, analogous deployments, and the company's understanding of the clinical problem. It is the starting point for a conversation, not the conclusion of one.

The pilot exists to test that estimate with real data. In principle, this is the correct design. A pilot should capture the health economic evidence needed to determine whether the product adds sufficient value to justify procurement. If it does, the innovation moves forward. If it does not, the pilot ends, and the provider looks elsewhere. These are both legitimate outcomes.

The problem is not the pilot. The problem is what happens before it.

A pilot can only capture the health economic data it was designed to capture. If the provider and company have not agreed, in advance, on what the relevant metrics are, what the cost baseline looks like, and what threshold of value would justify procurement, the pilot will generate data. It will simply generate the wrong data.

This is the origin of pilotitis, and it is misdiagnosed almost every time it occurs.

Pilotitis is commonly attributed to NHS risk aversion, commercial impatience, or a cultural reluctance to move from pilot to scale. These factors exist, but they are secondary. The primary cause is simpler. Pilots that are not designed around a pre-agreed health economic framework cannot produce procurement-ready evidence, regardless of whether the technology works. The company has not necessarily failed. The pilot has proved the wrong things. So the company moves on to another provider, another design, another estimated benefit case, and the cycle restarts.

There are only three honest outcomes from a well-designed pilot. The health economic data shows the product adds meaningful value, and procurement follows. The data shows it does not, and the pilot correctly ends; the product is not the answer to this problem. Or the pilot was never designed to capture the right data, and nobody learns anything useful.

The first two outcomes are acceptable. The second is even, in its way, a success: the system has correctly filtered out a product that did not meet the threshold. The third outcome is a system design failure. It wastes resource on both sides, leaves the underlying problem unsolved, and adds another data point to the narrative that the NHS cannot adopt innovation at pace. It is also, in practice, the most common of the three.

In property transactions, the survey answers a defined question before the market engages: what is the condition of this asset, and what are the material risks a buyer should know? Buyers and vendors then negotiate around a shared, independently established understanding of the facts.

NHS procurement rarely starts here. It starts with the market arriving at the door, estimated benefit case in hand.

The consequence is that the pilot, which should be a rigorous evidence-generation exercise, is instead asked to do two jobs simultaneously: define what good looks like and then prove it has been achieved. These are not the same job, and conflating them is where the system design failure originates.

The model has three stages, each dependent on the one before it.

The first stage is the Problem Brief. Before engaging the market, the provider defines the problem in health economic terms. Not a strategic aspiration, but a specific, costed articulation: this is the patient pathway, these are the friction points, this is what those friction points cost in clinical time, bed days, readmissions, or community resource, and this is the threshold of value at which a solution becomes worth commissioning. The Problem Brief is available to every prospective supplier from the outset. It functions like the Scottish Home Report: an independently established baseline that precedes market engagement.

The second stage is the Proof-Ready Pilot. Companies responding to the Problem Brief enter the pilot already knowing what they must demonstrate. The estimated benefit case the company brings is now tested against a pre-agreed framework rather than a loosely defined aspiration. The data collected is specified in advance. The success threshold is defined before anyone commits resource to implementation. A pilot that generates data showing the product does not meet the threshold is not a failure of the system; it is the system working correctly. A pilot that cannot generate relevant data at all has no place in this model.

The third stage is Procurement. Because the success criteria were pre-specified and the health economic data was collected against a shared framework, the transition from pilot to procurement becomes a decision rather than a negotiation. The provider is not asking "was this valuable?" after the fact. It is asking "did this meet the threshold we defined?" The evidence either supports that conclusion or it does not.

This is Problem-First Procurement. The value proposition belongs to the problem, not the product.

There is a fair challenge here and it deserves a direct answer.

Defining a problem in rigorous health economic terms requires expertise, time, and analytical resource. Many NHS providers do not have a health economist embedded in their procurement or innovation function. The clinical team knows the problem intimately but may lack the infrastructure to cost it precisely. Asking providers to construct a Problem Brief before engaging the market is asking them to invest in a capability they have historically outsourced to the companies pitching to them.

This is real. But it is a reason to invest in that capability, not a reason to continue a model that systematically produces the third pilot outcome.

The Scottish Home Report works because there is a regulated profession, a standardised methodology, and a legal framework that mandates the process. Problem-First Procurement would benefit from analogous infrastructure: shared methodologies for problem costing developed at system or Integrated Care Board level, and access to independent health economic support that does not sit inside a commercial organisation with a stake in the conclusion. The building blocks exist. NICE reference cases, Health Innovation Network analytical functions, and ICB population health teams all represent relevant capacity. What is largely absent is the norm: the expectation that the problem is defined, costed, and shared before the market is invited in.

If Problem-First Procurement became standard practice, the consequences would compound over time.

Providers would develop genuine institutional knowledge of their own cost architecture. The Problem Brief is analytically valuable independent of what the market subsequently offers.

Companies would enter pilots knowing exactly what they need to prove. The category of pilot that fails not because the technology does not work but because it was never designed to generate procurement-ready evidence would cease to exist.

Pilotitis would become a diagnostic rather than a chronic condition. When a pilot ends without reaching procurement, the question is no longer "why did this not work?" It is "did the product meet the pre-defined threshold?" If yes, procurement follows. If no, the problem remains open and the next company builds on the same foundation.

That last point matters more than it might appear. In the current model, when a pilot does not proceed, much of the analytical work done around it is lost. The next company starts again. In the Problem-First model, the Problem Brief persists. The investment in defining the problem is not tied to the fate of any individual product. The survey has already been done.