Pages

Tuesday, 29 September 2015

Government policy on the NHS gets all the attention but it isn't what matters most

When debating how to improve the NHS most of the focus is top-down government policy. It's either the magic solution or the source of all evil, depending on which side of the debate you are on. Both sides are misguided: it's incremental, bottom-up operational improvement that matters most.

I'm going to make an argument about the NHS that will annoy almost everyone from the diehards in the NHS Action Party to the neoliberal blowhards who would put every NHS activity into the private sector. I'm going to argue that the entire debate they are having is irrelevant. It is a complete waste of effort, thinking and newsprint (or electrons, for the younger generation who consume everything digitally).

The debate, whichever side you are on, starts with the assumption that top-down policy is what matters for driving improvement in the effectiveness, efficiency and quality of care. One side argues that things are a mess because of the purchaser provider split and the intersection of competition and austerity. The other argues that the problem is a failure to introduce more competition, more incentives for efficiency and quality and more mechanisms to allow the market to sort things out. And then there are those who argue that the whole framework of NHS regulation is a plot against professional standards of medics and they would do a better job if left to their own devices.

I've taken sides in some of these debates. For example, I'm relatively in favour of provider competition (on quality not price and not in imitation of the corrupt american model where the "market" seems to be rigged in favour of big business to the severe detriment of patients). But the evidence about competition in England (it does exist, see this skeptical review in the BMJ) while it exists, doesn't suggest the effect is large. Also the more integrated NHS systems in Wales, Scotland and Northern Ireland have not demonstrated better long term performance than the system in England in clear contradiction of the idea that the market consumes 14% of the English NHS budget for no good purpose. (The frequently quoted 14% is complete nonsense, by the way, as the rapid responses to this BMJ article which is one of the first to mention the estimate should make clear).

Despite my preference for some provider competition, what the evidence says most clearly is that it doesn't seem to matter much. This argument cuts both ways: the evidence (supplemented by the comparison across the different models of NHS structure in UK provinces) suggests that there are no gains from abolishing it either (unless you count having a system that corresponds to your personal ideology as a source of improvement in your general mental health). The one thing everyone seems to agree on is that major top-down changes to the NHS structure are costly and disruptive in the short term. So, if you are one of the many who think we should abolish the commissioner-provider split, you are going to need much better evidence about the gains to justify the short term cost.

The irrelevance of the top down structure of how the NHS is organised is just one clue that suggests that all those top-down policy arguments are irrelevant. Another is the failure of hospital mergers. The planners in the Department of Health, like central planners everywhere, work with what they know. And what they know is that, in theory, there should be economies of scale. Bigger hospitals should have lower costs than smaller hospitals. The same thinking drives service reconfiguration: bigger A&Es should be easier to manage and cheaper to run than smaller A&Es. But what planners know is only a fraction of the truth. In reality most mergers fail to achieve their goals (see the Kings Fund blog and report on mergers). As far as I can tell this also applies to most major service reconfigurations (there are exceptions for some low volume specialist services such as Stroke and Major Trauma).

Two other clues bolster my hypothesis. One is that the only a small proportion of the cost differences among hospitals seems to relate to size.  Another is that the staffing complement of A&E departments explains essentially none of the variation in performance (this research was done in the mid 2000s and I wrote about it in the BMJ; I think it still holds true). Both point to the idea that management matters more than the scale or resource levels (which is what get the attention because they are easier to observe).

The simple idea I want to propose is that all those top-down strategies can, at best, only make minor improvements to the NHS. It seems obvious that, when we need big improvements to bridge
the £25bn productivity gap, we should aim for big changes in how the system works. But that's not what the evidence says. In reality, top down change has about the same expected success rate as a
paraplegic contestant in an able-bodied arse kicking contest.

Large, sustainable change more often arises from a system that knows how to accumulate many small, achievable changes. This BBC article describes the idea using the way it led to dramatic improvement in Britain's cycling team performance. The article also notes how big the improvements can be in healthcare. Describing the experience of Seattle's Virginia Mason hospital:

But this was just the start. They started to use checklists in the operating theatre, to alter the ergonomic design of surgical equipment, to systematically improve clinical hygiene. Each improvement seemed small, but they rapidly accumulated.

What happened? Since the new approach was taken, Virginia Mason has overseen an astonishing 74% reduction in liability insurance premiums. It is now regarded as one of the safest hospitals in the world. That is the power of marginal gains.
My point is that effective change comes from the accumulation of small operational improvements in how things are managed on the shop floor. These probably account for 80-90% of all effective improvement (we can let well-designed top-down policy and structural changes have the remainder.)

If I'm right the entire debate about NHS policy is irrelevant. In fact many critics of government policy actively distract attention away from the real sources of improvement by assuming that only more resources can improve anything or only reversing [insert whatever government policy you don't like here] can lead to improvement.

The whole debate needs a new perspective. We need to look for improvement from the bottom-up not top-down. Small marginal changes to operational processes can lead to large sustainable gains in quality and efficiency. And that is what the NHS desperately needs.

Saturday, 26 September 2015

Superior management not superior leadership is what will save the NHS

Stories of the difference made by great leaders are very attractive in the world of organisations and armies. But they grossly exaggerate the role of individuals and understate the role of effective organisational systems. This is another vital lesson the NHS has yet to learn.

There is widely thought to be a crisis around leadership in the NHS with too many vacant senior management posts in hospitals and a general struggle to find good Chief Executives. Richard Vise, for example, say this in his recent Guardian piece on Addenbrooke's:
It is difficult to see why anyone would risk their reputation [as an NHS leader] when the chances of failure are so high.
Recent commentary (e.g. by Polly Toynbee or Roy Lilley) about the situation at Cambridge University Hospitals has tended to blame the regulatory environment, government policy and the general pressures on the NHS for driving out people who seem like good leaders. I've already argued that are learning the wrong lessons from the problems at Addenbrooke's but there is a more general problem with the focus on leadership: leadership isn't what really matters.

You can write great stories about how a visionary leader transforms organisations or wins battles. Stories about management and organisation are harder to write, duller and far less popular. But we should not confuse what makes a gripping narrative with what actually matters. Most narrative on the NHS falls into this trap.

The argument I want to make is that organisational systems matter far more than great leaders. It is vitally, existentially important that this is understood for the future of the NHS. If we focus our attention on leaders and not systems, continued failure will be assured (and hospitals will continue to fail, not just leaders).

It isn't that leaders don't matter, it's just that leadership alone has all the hope of success as a castrated Yorkshire terrier dry humping your leg in an effort to father puppies. And a constant focus on leadership as the silver bullet distracts from the prosaic, but far more important, work of designing and managing effective operations.

The best analysis of the mistake that we might call the leadership fallacy (it afflicts the world of business nearly as much as it afflicts commentary on the NHS) is found in two books by Stephen Bungay. The first, Alamein, is a history and analysis of what happened in North Africa in the second world war. The second, The Art of Action, is a distillation of the lessons for management Bungay learned when reanalysing military history (I've used stories from his analysis of the Battle of Britain to derive lessons about the importance of effective management for the NHS in another blog).

The military battles in North Africa are relevant because the way the story is normally told focusses on the leaders involved and tries to explain what happened by focussing on their characteristics. In telling the stories this way the reader is (implicitly) urged to believe that the selection of the right leader is what makes the difference between success and failure. Their characteristics, their visions, their skills matter and all those other boring details don't.
Bungay makes the point like this:
It is common to identify the actions of armies with their generals, to imagine that the Panzerarmee was quick and clever because Rommel was, and the Eighth Army was slow and cautious because Montgomery was. In fact, those characteristics were rooted in the institutions themselves, in the behaviour of hundreds of middle ranking officers on both sides. The Eighth Army behaved the way it did long before Montgomery arrived, and German troops in every theatre of war behaved much like the Afrika Korps whoever commanded them. Both armies did what they had been told to do in the way they had been trained to do it in the 1930s.
We have been conditioned to believe that the battles in Africa were decided by the characteristics of the generals involved. He argues they were, in reality, determined by the ways the competing armies were organised. These longstanding organisational characteristics, in the British case, made victory far, far harder despite large advantages in the resources available over their better organised German adversaries (the British had enormous advantages in manpower and equipment in North Africa).
The key point is that organisation systems matter far more than leaders (I highly recommend reading Bungay's books not just for the evidence and detail but because they are both very readable and surprisingly insightful).

Bungay's book on management makes this point (my emphasis):
We need not worry about how to make the structure perfect. However, unless the structure of the organization broadly reflects the structure of the tasks implied by executing the strategy, the strategy will not be executed. Every organizational structure makes doing some things easy and doing other things difficult. If the structure makes doing some things so difficult that there is a conflict between structure and strategy, the structure will win.
The lesson generalises. I've argued before (using Bungay's work) that the RAF won the Battle of Britain not because it had better pilots, better technology or better aircraft but because it had a better system to organise the technology, aircraft and pilots.

This is where the NHS needs to pay attention. If the focus is on appointing great leaders with great visions then the focus is wrong. The best leader and the best vision will be undermined if the organisational processes in their hospital are broken. Vision is a fine thing to have, but only attention to operational detail will translate it into effective action. When we tell stories about great NHS leaders and their transformational ideas we distract attention from the simple fact that those leaders matter little compared to the organisational systems in their hospitals. If those boring operational management details are not addressed then leadership is irrelevant.

This is where the specific failures at Addenbrooke's tell us something important for the whole NHS. Keith McNeil was by most accounts a great leader and had an important vision for how the hospital would run (a large part of which involved a new eHospital system that would put all administration and record keeping into a single online system). There is nothing wrong with that vision. If it worked, it would transform the quality and efficiency of day to day activities at the hospital. But McNeil, by his own admission, had no grip on operational detail or finances. In other words, he had no grip on the organisational processes required to translate his vision into effective action. And the hospital failed to appoint anyone else to the board who could fill in that gap. So, instead of leading to to a better and more effective hospital, the vision has led to chaos on the ground where basic processes no longer work reliably.

The leadership at Addenbrooke's (not just McNeil) focussed on the vision and neglected the boring management detail around the basic operational processes that have to work to translate the vision into action. In doing so they epitomise a pervasive neglect in the NHS. You can tinker with top level organisational policy and structure as much as you want but you won't see effective improvement unless you get the operational processes and management right from bottom up. Operational management matters and, if you don't get it right, you will cripple the best vision, the most carefully crafted top down strategy or the most highly respected leader.

I don't want to go so far as to say that leadership doesn't matter at all. But the greatest vision and most attractive and popular personality matter little if they have no grip on the operational detail. Until the NHS and the leaders it appoints get to grips with the prosaic detail of effective operational management, there is little hope for improvement and leaders will continue to fall.

Tuesday, 22 September 2015

What really went wrong at Addenbrooke’s


The commentariat have issued a torrent of bullshit on the decision to put Cambridge University Hospitals into special measures. Most explanations are misleading nonsense that don’t match what we know. Seeking the right explanation is an essential first step to fixing anything in the NHS.

The decision to put the hospital into special measures has been controversial. Senior people have resigned. And far too many commentators have responded by rolling out clichéd explanations that pander to their prejudices and don't fit the specific facts of the case. Roy Lilley blames the regulatory environment. Polly Toynbee agrees and adds another explanation: the funding crisis (which, despite the flat NHS budget, apparently consists of “savage cuts” though that might just be a cliché obsessed sub editor). Another Gruaniad story claims problems in social care keeping people stuck in beds. Others have blamed unsustainable surges in emergency demand. Many bemoan the loss of a highly regarded CEO and blame the government, the CQC, Monitor and just about every other central body in the NHS for making hospital leadership roles impossible.

While many of these supposed explanations are real problems in the NHS, none explain what went wrong at Cambridge or why it happened so quickly. In leaping to conclusions without checking the facts commentators are creating a fog of distraction that seriously inhibits identifying the real issues and fixing them. And the real issues are pretty important for the whole NHS so we really should be paying attention and trying to learn something that will help in the future rather than contributing to the miasma of ignorance.

The thing that seems to cause confusion among the commentariat is that a hospital with motivated clinicians which recently had a top rating for care quality has fallen over. The thing they don’t seem to understand is that motivation, vision and general worthiness of intent are not adequate to run a hospital: you also need good operational management. All the good will in the world won’t compensate for dysfunctional operation processes and data.

And this is the real explanation for what went wrong. The hospital tried a big-bang implementation of a visionary new eHospital system and (i’m guessing a little) botched the implementation process or seriously underinvested in it. The result was that the basic information operational managers needed to actually run the hospital was missing or corrupted. So basic processes no longer happened the way they should. Patients in A&E couldn’t be tracked so the 4hr target wasn’t met; surgical activity wasn’t correctly recorded so the PbR payments due from commissioners were wrong; appropriate matching of staff with activity couldn’t be done…

I can’t be sure of all the details because I haven’t visited the hospital to check. But I know that the clinicians thought the implementation was a catastrophe and were concerned that top management didn’t want to know their visionary system wasn’t working. I know that A&E performance fell off a cliff the week the system was started up. I know that the financial situation deteriorated very suddenly at the same time.

Roy Lilley blames the regulators. But they were just recognising the emerging catastrophe and did nothing to cause it.

Polly Toynbee blames cuts to social care and general government policy. But they affect everyone and impact only slowly. They don’t explain why things happened so fast in Cambridge.

A&E performance didn’t suddenly fall because of a surge in attendance. There was no surge. And attendance has no effect on performance anyway.

200 beds were not suddenly blocked by social care problems outside the hospital. Chances are that number is the same this year as last when the hospital was performing well.

Digitalhealth.net reports the following:
The introduction of the eHospital programme, which included a major infrastructure upgrade by HP, and the first implementation of the Epic electronic patient record system in the UK, may have contributed to the deficit, but the CQC report shows it may have impacted patient care as well.

In its key findings, the CQC says that the introduction of Epic, which is in common use in the US, has “affected the trust’s ability to report, highlight and take action on data collected on the system.”

This includes the ability to access information from diagnostics tests such as electrocardiographs, while agency staff are not always able to access information about patients they are supporting.
In other words basic operational processes couldn’t be done properly and this started to affect the quality of care. This explanation explains the facts, the timing and the speed with which the problems arose. Other explanations don’t.

Too many commentators have focussed on general distractions and pressures rather than the specific issues at Cambridge. Roy Lilley, for example, highlighted a radio interview given by Keith McNeil just after his resignation as chief executive. Early in the interview McNeil blames general system pressures (which is what Lilley wants us to see as the cause for losing a good chief executive). But immediately afterwards McNeil argues that he was a “vision and strategy” person lacking the “granular detail” about management and the hospital now needed someone with operational “grip”. This is pretty much an admission that the problems were about a failure to get to grips with the detail of how the operations would be affected by his grand vision.

There is a really important lesson here. Operational management matters. Hospitals don’t spontaneously organise themselves to give quality care, they have to be organised. You have to know what is going on inside your hospital or the efficiency and quality of care will suffer badly. Grand visions of how new IT will make things better are fine but if you neglect the detail the vision won’t happen. If you don’t spend enough effort to implement the new system and make it useful for and usable by the doctors and nurses who will have to work with it every day, your grand plan will fail and the clinical care will be worse.

Too many commentators believe the foolish idea that management consists of a bunch of bureaucrats who just get the way (I’ve ranted about this before). The idea that management doesn’t matter is a pervasive and damaging idea widely believed (though implicitly and unquestioningly) by commentators about the NHS and many of the staff inside the system.

Effective operational management needs good information about the activity in the hospital. And doctors need good information to deliver the right care to their patients. An effective hospital needs good information, good operational managers and good clinicians to do a good job for its patients.

Cambridge is a perfect illustration that good clinicians and a grand vision do not lead to great care: you also need good information and good operational management. The torrent of misleading commentary on the situation in Cambridge has distracted from this critical lesson. We need to ignore the commentary and pay attention or we won’t be able to make the NHS better.

Tuesday, 8 September 2015

Top-down changes won't bridge the NHS funding gap

It is all too tempting for the leadership of the NHS to assume that the top-down levers they can easily control will be the things that will solve the big problems facing the NHS. They won’t.
Few disagree that the NHS has to generate something like £25bn of extra activity in the next 5-10 years without getting the £25bn of extra cash required to pay for it if the costs of that activity remain the same. In other words: there is a productivity gap.[1]
Both politicians and the leaders of the NHS have produced plans to avoid this future crunch. And those plans look a lot like other plans proposed or implemented over the last three decades that have arrived with the promise they would improve the NHS. They all share some characteristics that make it unlikely they will actually help.
This is why I think they won’t and what the alternatives are.
Governments in particular suffer from a belief that the things they know about–which being made up largely of professional politicians and lawyers and having very few engineers, scientists or practical managers consists of legislation and top-down organisational changes–are that only things that matter much in driving change. They are like one club golfers or perhaps footballers who only know how to tackle and not how to kick the ball.
There are two reasons for this narrowness of vision. One is simply that that is all they know. Legislators know how to write laws, they don’t know how to manage an organisation. The other reinforces that ignorance by using legislation as a signalling mechanism with no intention of it actually doing any good in the first place (legislation limiting government freedom can always be abolished if it doesn’t work out unless it can be written into a rigid constitution limiting future government freedom to amend it; the UK has no such framework). So in the UK we have had laws to limit the size of the government deficit and laws mandating a reduction in child poverty. Neither has any credibility in driving action yet both wasted valuable government time that might better have been spent understanding the problems.
One consequence for the NHS has been a major structural and legislative change every 5-7 years since Margaret Thatcher won the 1979 general election. Given that many of the changes take perhaps 3 years to enact and implement and the system takes perhaps another 3-4 years to get used to how the new structures work, there has probably been no point in my adult lifetime when we could judge whether the long term effect of any particular organisational form for the NHS was likely to lead to sustained improvement.
The NHS faces many major challenges and many of these have existed for decades. The leadership, however, is perpetually tempted by one major immediate challenge that often squeezes out its ability to deal with the big long term challenges: the need to be seen to do something right now. This reinforces the temptation of all those top-down changes. “We know” the political and organisational leadership seem to argue “how to do structural and legislative changes and at least we will look like we are doing something.”
The core problem faced by the NHS as a whole is the mismatch between the demand and the capacity of the system. However much money we spend, it seems that demand will always outstrip capacity. But it is worth clarifying what this means.
Some argue that growing demand on the service could be curbed by better effort devoted to prevention. even if true, though, this plan might not pay back for 30-40 years. Even if we could make a significant difference by doing the right thing now, the benefit will fall to our children and the current NHS will still face the damning consequences of mistakes made four decades ago.
On the other hand there is plenty of current activity that isn’t that beneficial. Many very expensive cancer treatments lead to only small increases in life expectancy or only work on a small proportion of patients. Shockingly, many patients currently receiving aggressive treatment for cancer would live longer if we stopped aggressively trying to cure them and started palliative care so they could have a more pleasant death.[2] And it isn’t just cancer. Many routine NHS treatments seem on contribute little to the wellbeing of the patients being treated. Whether improved data (more PROMS[3], for example) combined with improved decision making would lead to a lower NHS activity without harming patients is not proved, but it looks promising. If it worked that would be an immediate contribution to the big challenges.
More important and less controversial is the idea that the NHS needs to get more productive. More treatments for a given spend. Nobody disagrees that would be a good idea (though some think it is impossible to achieve). But there is a very lopsided argument about the best way to achieve it. And we don’t always have a sensible definition of what productivity is. It isn’t just about more treatment; it is also about better treatment and fewer errors in treatment.
Major organisational changes are always justified by the effect they are supposed to have on productivity. But, while organisational structures can blunt the incentives to be productive, they don't have a big impact and they do disrupt the system for years every time they are implemented so, if done too frequently, they permanently distract the management capacity of the NHS away from a focus on what matters. This applies at multiple levels to reinforce some of the temptations of top-down management. For example, there are some economies of scale in many hospital services. This should lead to larger hospitals having lower costs than smaller ones. And this appeals to top-down planners (who only seem to understand top-down economics) and hospital managements (who always prefer a bigger empire to a smaller one). But the evidence that scale matters a lot for whole hospitals is scarce (to be fair there is good evidence for some highly specialised services like stroke or major trauma but not for many routine services). But hospital mergers have a poor track record of delivering improvement in the NHS (and NHS hospitals are already on average bigger than almost any others in Europe). One possible reason is that, according to one unpublished study I’ve seen, size only accounts for about 10%-20% of the cost differences in NHS hospitals. The rest of the difference is about local operational, organisational or geographic factors.
This is where I get to speculate. My hypothesis is that the majority of differences in productivity are determined by operational management. In other words, how you organise the work[4]. A well organised A&E department will treat more patients faster than a poorly organised one even if the poor one has far more staff. And the quality of treatment will be better. But the quality of operational management is not directly visible to top-down planners. Worse, top-down planners like campaigns to reduce bureaucrat numbers but lack the ability to distinguish between wasteful pen-pushers and good operational managers so place staffing limits on all management. In an organisation like the NHS where the management capacity is remarkably low to start with this is a catastrophe.
The thing is we know that when operational management is done right, productivity and quality improve. It isn’t all about the amount of resources available. The medical (or nursing) life isn’t all about being overwhelmed by uncontrolled demand. If you take control of how the work is organised, work can be made less pressurised, resources can be freed up, patients can be happier and quality can be transformed.
One of my favourite examples comes from here and it is about transforming the patient and GP experience in primary care by applying simple operational principles.[5] Those of you who follow @HarryLongman on Twitter will know the evidence already. You can search his website for the detailed evidence that his process works, but my simplified summary is below. Instead of assuming that all requests for a GP appointment are equal (and fit in a 10min slot) practices should assess all requests by phone or face to face. Only then should any appointment be booked and the appointments should be adjusted to the nature of the demand. What happens when this is done is that much of the demand goes away or is dealt with very quickly and often over the phone. This frees up so much practice time that it usually becomes possible to meet all demand for an immediate appointment on the same day and leaves some over leaving less pressurised staff and happier patients.
I’d go so far as to say that the success of his approach shows that the actual problem in primary care isn’t, as headlines and RCGP campaigns suggest, overwhelming demand and underfunded primary care budgets, but a chronic lack of the operational skills to design GP services that match the needs of patients. And a naive adherence to an inflexible model of how GPs should operate. Maybe the RCGP should offer operational training to GP practices before starting any more campaigns for a bigger share of the NHS budget.
Another example from the acute sector reinforces the belief that much of the NHS lacks the operational expertise to run its services effectively. In a recent HSJ article, Rob Findlay (@Gooroohealth on twitter) points out that elective waiting lists don’t demonstrate a fundamental lack of capacity in hospitals. If we were really short of capacity, waiting lists would rapidly grow without respite. They don’t. They rise at some parts of the year and fall in others with very small long term drift. He argues that the problem is they way hospitals respond to the pressure. Instead of smoothing the capacity to match the demand, many hospitals only adjust the capacity in a panic when things get out of hand in the short term. This increases the burden on staff and costs much more as much of the “extra” capacity is bought with expensive overtime and out-of-hours payments. He argues that much of this could be avoided (reducing costs and staff disruption and burnout) by fine-tuning capacity well in advance of the panic resulting in far better use of available resources (of staff and equipment). He argues:
...we could manage the main working week with the resources we already have, instead of continually incurring the high marginal costs, inefficiencies, and risks of relying on ‘extra’.”
Again, the point is that good operational management can greatly improve the productivity of a service in a way that is good for both patients and staff.
My main point is that real change in the NHS will come bottom up from operational improvement and not from more top-down changes. And they should be easier to sell to staff as they see the benefits immediately rather than having to wait years to see how new structures, incentives or tariffs work their way through the system.
I hope that the Vanguard programme launched with the 5-year Forward View will sneak some more of this sort of thinking into NHS strategy. But I know from some recent conversations that a great deal of thinking at the top is being focussed on the top down stuff that likely won’t have much effect like major changes to the way tariff works.
The NHS can meet its productivity gap, but bottom-up operational improvement is the key, not big top-down initiatives. Let’s hope that’s what we get.

[1] Some have described this as a budget cut. It isn’t. The NHS will--even under conservative plans--have more money in real terms in 5 years than it does today.
[2] See this article that argues: “A new study finds palliative care doesn't put patients out of their misery; it puts the misery out of the patients.” And there is a surprising disparity between how doctors choose to die and how they choose to treat their patients faced with the same conditions. See this article about how they often choose low-intervention rather than aggressive treatment and how their deaths are more pleasant as a result.
[3] Patient Reported Outcome Metrics: a systematic way of assessing whether some common surgical interventions actually lead to improvements that matter to patients.
[4] The bottom-level work in each ward and each specialty, not the top-down structure of the whole system.
[5] And, incidentally, not blaming the government for insatiable demand, a shortage of GPs, reduced share of the NHS budget etc.. Maybe there are things the government should do differently, but if GPs take no responsibility for managing their own workloads things are not going to get better soon.