NHS productivity is critically important but widely misunderstood from the top to the bottom of the system

“Productivity isn’t everything. But in the long run it is almost everything.” Nobel winner Paul Krugman

Improving productivity is the most important challenge facing the NHS. But the leadership of NHSE are in denial that there is a problem, despite abundant evidence otherwise. Widespread misunderstanding of what productivity is abound. This must change or the future looks very dark.

The NHSE leadership’s attempt to defend NHS productivity to the House of Commons Health Select Committee was an astounding illustration of how little grasp they have of what productivity is. Or what the current reality in the NHS is.

But productivity is also misunderstood at every level in the NHS not least because the leadership so often use the word to mean something entirely different.

So what is it and what are the big misunderstandings about it?

What is productivity?

Let's start at the bottom. When I complain that NHS productivity is poor or declining, I frequently get responses from staff saying something like “you are disparaging the effort we are putting into our jobs. We are working as hard as we possibly can. We are not unproductive.”?

This claim exposes two errors about what productivity is:

1. It is about how hard people work
2. It is about local activity (of individuals or teams)

Point 1 is wrong as productivity is not just about inputs, the outputs of the activity also matter. If you are working hard because something else in the system is broken, that work is waste not productivity. Point 2 is irrelevant in any system where the outputs depend on the coordinated work of many individuals or many departments.?

Consider the current state of A&E departments in the NHS.?

In 2010 most people attending an A&E left in <4hr even those who needed to be admitted to a bed for the next stage of their treatment. The final output of the NHS activity for those patients was discharge home once their injury or condition was fixed. Getting to that output required many inputs inside the A&E and also, perhaps, from surgeons, theater nurses and the staff tending to them while they were in a bed. All those inputs contributed to the output: discharging a well patient.

In 2022/23 1.7m of those patients waited more than 12hr to leave the A&E (and, given known patterns from historic data, probably spent an extra day or two in a bed as well). The extra 8hr many spent in A&E required far more inputs from A&E doctors and nurses to maintain the patients while they waited for admission. Extra nursing staff were required to look after them for the extra days they spent in a bed. Everyone involved is now working far harder than they were in 2010; far more work is now being done. But the output–a well patient going home–is exactly the same.?

The extra work is irrelevant; the productivity is now far lower. The same output requires far more inputs. Worse, much of the extra inputs needed are caused by a failure to coordinate the A&E silo of activity with the discharge processes from beds which is probably the responsibility of the staff who design and manage the process of flow through beds. If they design bad processes or operate them badly, doctors end up putting in more work for exactly the same output. If problems in beds cause much longer delays in A&E, the workload in A&E rockets for exactly the same outputs.

If productivity were about the volume of inputs, it would have risen sharply. But we now have a system where far more inputs are required to get exactly the same output.

Productivity is better defined as the total system inputs required to get each unit of output. And, in a complex system with multiple interacting parts, it is the overall output that matters not the output of the individual components or people in the system.

Some have argued that the NHS is naive in how it counts those outputs. An antibiotic prescription is not the same as a replaced cataract which is not the same as a heart transplant. Or that a knee replacement done well is different to one where the op needs to be redone or where the patient picked up an infection. But those factors are taken into account in real productivity calculations. The mix of outputs matters: heart transplants count as more output than antibiotic prescriptions. And there is some adjustment for quality.? If the readmission rate goes down, that also matters in the calculation.

What does the leadership think?

The HSJ reports on what Amanda Pritchard claimed while giving evidence to the Health Select Committee:

“But giving evidence to MPs today, Ms Pritchard said that in fact the NHS was doing more activity, and argued current health productivity measures fail to capture the whole story.”

“There is a misunderstanding at the moment about the state of productivity in the NHS because it’s measured in a way which doesn’t fully reflect either what happens at acute trusts or [investments to improve care quality] or, crucially, it doesn’t reflect what’s happening in community care, it doesn’t count things like virtual wards.”

“And it doesn’t reflect some of the innovation in the ways that we have evolved services to the benefit of patients in using technology.”

“If you take into account of all of those things, actually the start point is an NHS which is doing far more work and differently than it was pre-covid.”

Steven Powis is reported as saying much the same:

“it is perfectly possible to make productivity gains ..sometimes where we make productivity gains is not where we measure productivity”.

“We would all recognise that things like virtual wards, the transactional costs that are removed from using digital for instance to interact with your GP or the hospital, are productivity gains but they don’t sum up in the way that we would want them to sum up, into the figures that we produce.”

These claims argue that we are not measuring productivity correctly. But the examples they use to illustrate the point make it look as though they don’t understand what productivity is. Virtual wards, for example, sound like a good idea. But the point is to reduce hospital bed occupancy and thereby increase hospital throughput. But we are not seeing that increased throughput so the effort put into virtual wards cannot be considered as productive activity. Maybe I’m wrong on this, but neither Powis nor Pritchard showed any actual evidence to persuade me otherwise.?

Arguing that the productivity is really increasing but we need a better way to manage it is an incredible claim given the current calculations of NHS productivity are the most sophisticated in the whole public sector.?

If they had claimed that the NHS was now doing the same number of things but had switched to doing bigger more significant things (even more knee replacements, fewer antibiotic prescriptions), the argument might have made sense. But that isn’t the argument they made and nor would it be true if they had claimed it. Rather they seem to say more effort is going in and there is therefore more productive activity.?

What is really happening?

Several independent bodies have analysed the state of NHS productivity recently. The IfS have weighed in as has the IfG. The ifG analysis is the clearest and most devastating.

“Between December 2019 and December 2022 the number of full-time equivalent (FTE) junior doctors, and nurses and health visitors, increased by 16.4% and 10.9% respectively. But despite this the number of people being treated in hospitals is only marginally higher than it was before the pandemic. On some important metrics it’s lower.”

They estimated that productivity was 20-25% lower than it was before covid. Only GPs are “producing” more and not because there are more GPs (numbers have fallen) or because they have used technology to improve the patient experience (some have, most have not).

Important areas of public concern are not being tackled. Waiting lists for elective activity are growing as the hospital activity needed to reduce them has not increased. The already lengthy waits for emergency care are getting longer. Access to GPs doesn’t appear to be improving, though there does appear to be more activity.?

But Powis and Pritchard seem to be arguing that productivity is improving. But just not anywhere that affects the most important areas of NHS activity. They only appear to be able to quote areas where more effort is being expended, not any where the quality or volume of outputs is increasing. The areas where the NHS has clear goals for improvement–and where productivity is most important–are not seeing any impact.

In response to the NHSE claim, the IfS felt compelled to respond with new analysis which started with this startling chart (red annotation by me):

They said:?

“in our view, the available evidence strongly points to the NHS – or, at the very least, NHS hospitals – having an ongoing productivity problem”

The IfS analysis also pointed out that the problem is recent and has not always existed. This chart compares key growth rates of inputs and outputs before and after 2019:

NHS output (and productivity) was rising before the covid funding boost and large staff increases but output has stalled since resulting in a big hit to productivity.?

The NHSE leadership could have argued that the NHS is focussed on the wrong outputs. Perhaps the system should be be focussing on quality not volume. They didn’t make this argument. And would have been doubly wrong if they had since both are declining.

They could have argued that the NHS is focussing on the wrong type of activity. Perhaps more activity in prevention reduces the need for hospital treatment. More activity in the community might reduce the need for hospital treatment. That would be an improvement in productivity (and would probably count as one even given the current way it is measured). But more activity doesn’t translate into more hospital admissions avoided. Anyway, they provided no evidence for any such shift.

In some future utopia the NHS might not be doing so much hospital activity. But, right now, most of the biggest visible problems and the biggest problems the public care about will only be fixed with more hospital activity. Despite a lot more money and a lot more staff, that activity isn’t happening. Pretending that some other area of NHS activity has grown and this makes the system more productive is an act of denial not a subtle claim for better measurement. NHSE should be renamed NHS Excuses.

What misunderstandings feed the problem?

The basic failure to understand that productivity is a property of the system and not the individual components causes many errors in thinking.

Doctors working harder is not them being more productive if the work is only required because other parts of the systems are broken.

Cost cutting is not productivity improvement if the cost cutting affects some other part of the machinery that keeps the system working. Cutting other staff groups to achieve “more resources to the front line” is a persistently cursed goal that has led to huge damage in the effectiveness of the overall system. Cutting operational management cuts the coordination needed to keep different people and departments synchronised so the overall output is sustained. Cutting admin support leaves admin in the hands of doctors and nurses, taking time away from patient facing tasks and leaves them with less output for more workload. Underinvesting in IT has the same effect. So does cutting capital spend on better designed buildings. Those cuts create short term cost savings but hurt productivity.

Sometimes the NHS behaves like a car owner who cuts maintenance spending on his vehicle and doesn’t replace her engine oil. These actions save money today (productivity looks to be up as the input costs go down but the miles driven stays the same). The ultimate result is obvious: eventually the engine seizes, the transmission fails or the wheels fall off because basic maintenance wasn’t done. The productivity gain was purely superficial and short term and miles driven falls to zero until a vast maintenance bill is paid for or a whole new car is purchased (assuming the owner didn’t perish in a bad accident making productivity zero forever).?

A car is a complicated system that needs a balance of inputs to achieve its goal of getting you from A to B. Too much oil will destroy the engine as surely as too little.?

The wrong mix of inputs matters too. There is no point in keeping the engine lubricated at the cost of skimping on lubricating the transmission and gearbox.

NHS decision makers often have the wrong idea about improving productivity because of similar types of failures. It is often assumed that keeping senior doctors busy makes the system more productive. GPs often assume that putting their experienced doctors on a task like triaging requests is damaging to productivity. But this depends on how the GP practice is organised. If the mix of demand is wide (which it usually is in GPs), then putting highly qualified GPs at the front door may lead to much better allocation of patients to other staff members in the practice with minimal risk. Some patients need a short time with a pharmacist; some a physio appointment; some have simple problems an AHP can handle; some need some administrative problem sorted; some have a problem needing 20 mins with a mental health specialist; some do need examination by a GP. Efficient quick allocation to those people can give a far more productive GP practice than pushing them all to a 10 min session with a GP. Only by analysing the whole practice as a system can the productivity of triage first be understood. Yet many GPs resist because such an approach seems to reduce the productivity of the individual professionals doing the triage. This twitter thread from a GP counters that argument pretty well. But the majority of GPs still focus on the individual productivity instead of the system productivity. A similar issue occurs in A&E departments where senior doctors at the front door improve the flow by making better calls about where to direct each patient. Such productivity-enhancing process changes are resisted by both the staff and their leadership.

The general problem of paying attention to just parts of the system and not the whole system is systemic in thinking about how to improve the NHS. The right mix of staff on the team is neglected in favour of more staff. The role of managers in designing better systems or coordinating existing systems is neglected in favour of having more people who do the clinical activity. The role of support staff in removing administrative burdens allowing clinicians to spend more time with patients is neglected. The coordination of different departments to create a productive hospital is neglected in favour of throwing more resources at the symptoms of failure. A&E staffing has grown sharply as performance has declined even though the cause of delays is rooted in the flow through beds which A&E staff can’t fix. More local activity does not contribute to system productivity. More surgeons are recruited to increase surgical output when the bottleneck is a lack of theatre nurse or theatre scheduling is about as uncoordinated as the limbs of a patient with cerebral palsy.

Improving productivity is the biggest problem the NHS has and it is an existential one

What can be done?

According to standard economic thinking productivity growth is critical for modern economies.?

Economists tend to argue that there are three key factors driving increased productivity:

1. Improved skill of the workforce
2. Capital investment
3. Innovation

These are not independent. And how they materialise in an organisation like the NHS needs some work to look at specific “levers” that NHS policy could influence.

Some are more obvious than others. Capital investment is relatively clear and the persistent lack of it has been a huge drag on NHS productivity.

In the NHS there is some interaction between innovation and workforce skill. But current thinking tends to ignore the interaction and confuses innovation with “bright shiny new technology or medicines” when those are fairly unimportant.

Some more concrete categories of productivity enhancing activities might make the position of the NHS clearer. So here are where some might fit in the big economics classification:

Workforce:

1. The right mix of skills in front-line? teams?
2. The right mix of different skill sets across the whole workforce
3. The right level of training for front-line workers
4. The right level of training for managers and leaders
5. Higher skilled analysts

Innovation:

1. Better diagnostic equipment
2. Better process design of clinical work
3. Better processes for coordinating flow across clinical and departmental boundaries
4. New, more effective, treatments
5. Better ways to stop activities that don’t work

Part of the problem at driving productivity is that the multiple factors don’t add up, they multiply. More skilled workers may have no influence on overall productivity if capital is short cna they can’t apply their news skills effectively; better designs or the processes of work may need more effective investment in IT without which they fail. A big push to improve one factor can be completely negated by another factor being missing. So a serious attempt to improve productivity cannot be made without a good understanding of the whole system and how each of the productivity improving factors interact.

The NHS has ignored this point for decades. It has focussed on increasing the front-line workforce but has starved the system of capital and failed to innovate in the design of work processes.

In some other ways policy has actively hurt some of the critical factors. Headline numbers of front line staff have been prioritised over ensuring the right mix of skills. Management has become a dirty word and something to be minimised; analytical skill has been neglected (there are too few analysts and they are paid too poorly compared to their non-NHS peers to get a sufficiently skilled workforce); training has been perfunctory and almost non-existent for managers or leaders. The system finds it almost impossible to stop paying for new ideas that turn out not to work; it has starved itself of the capital required to build stocks of better, newer equipment; information technology has also been starved of capital and the necessary budgets to implement it well, undermining the possibility of enabling the improved communication and coordination needed for better designed ways of working.

In effect, the NHS has set some of the levers that could improve productivity at zero despite nullifying the effect of other levers pulled as hard as possible (the factors multiply, remember, so a zero anywhere leaves the whole system at zero). The accelerator pedal is pressed hard but the gearbox is set to neutral and the car doesn’t move.

NHS strategy has tended to try to improve productivity one lever at a time without understanding the interacting effect of all the levers on overall productivity.

The implications of ignoring productivity are dire

Productivity isn’t everything but, as Krugman says, in the long run it is almost everything.

An NHS where productivity is increasing is one where the service can do more good while not becoming unaffordable. One where productivity remains at current disastrous levels will become very unaffordable very quickly.

Short term decisions to recover the system’s financial balance (largely a consequence of ignoring the importance of productivity when trying to get higher output by throwing money at the system) are actively harming the very factors that could drive up long term productivity. Capital spending has already been cut in general and IT spend in particular; efforts to improve system coordination and to design better processes have long been neglected in favour of more front line staff (and fewer of the operational managers needed to design and implement better coordinated processes).The NHS has a severe shortage of good analysts to identify where productivity could most improve and won’t pay them the market salaries needed to retain them.?

The opportunity to improve productivity is huge. But it requires doing many things that feel politically very difficult. Hiring more and better managers has long been anathema. Disciplined long term investment in better information systems seems impossible and what investment there is seems to be misdirected to a search for “silver bullets” that offer the promise of magical solutions but will turn out to be snake oil (maybe the FDP, certainly undisciplined hope for artificial intelligence). Even silver bullets don’t work if you neglected to buy the gun needed to fire them. Capital spending on better buildings and equipment is unlikely to happen without a complete rethink of the spending processes to enhance long term planning and provide a bulwark against short term capital to revenue transfers.

This would be less catastrophic if productivity-enhancing spending had not been neglected for most of the last two decades. But too much NHS decision making has seen the long term as anything beyond the next 12 months.?

Productivity and the failure to invest in it is the biggest problem the NHS has. NHSE’s denial that there is a problem is a harbinger of doom for the NHS.