Thursday, 22 October 2015

It's not the doctors: it's the beds...

People admitted to hospital are more likely to die if the admission happens at the weekend. The government thinks this is because hospitals don't really work 7-days a week. So they are engaged in an attempt to rewrite doctors contracts so they can't opt-out of weekend work. This is the wrong focus.

People don't stop getting sick at the weekend. So hospitals really shouldn't provide a worse service then. But the evidence suggests they do (though it should be admitted that working out the weekend mortality is both hard and controversial). The government thinks this is because consultants can opt-out of weekend working (though how many actually do this is unclear and a subject of significant controversy).

While it is obvious in activity statistics that hospitals function very differently at the weekend, it is a lot less obvious that the doctors are to blame. And focusing on them may be a mistake. A recent study (reported in the BMJ here) casts some new light on the problem that suggests the focus of government policy may be wrong.

The study reported a clear reduction in mortality in a hospital when bed occupancy was reduced. The text below is a version of what I said in a BMJ rapid response and on LinkedIn when I first saw the study.

I'm puzzled that more commentary has not noted the relationship between this study and the current topic of weekend mortality in NHS hospitals.

The government has focussed on trying to force changes to medical contracts to eliminate the ability for doctors to opt-out of weekend work in the hope that this will fix the problem of excess mortality at weekends. But, by focussing on the doctors, they miss the more general point that just having more doctors won't fix broken operational processes at the weekend. This study points to a much broader problem that links mortality to those processes.

The missing link is the fact that the processes for discharging patients and therefore keeping bed occupancy down are widely broken at the weekend. While admissions are lower as few elective patients are admitted, discharges are muchlower. Emergencies, of course, continue to arrive. So the beds fill up as the overall process for discharging patients is usually dysfunctional at the weekend. This leads to very clear patterns (easily observable in activity statistics and for length of stay). In many hospitals beds fill up at the weekend.

The presence of doctors at the weekend isn't (or shouldn't be) the critical factor here. Most of the discharges that should happen are probably routine and could happen automatically without medical supervision. But they often don't because the process for discharge has been poorly designed (this is sometimes because it has an unnecessary requirement for consultant sign off).

We have known for some time that the dysfunctionality of processes associated with the flow through beds is the dominant cause of delays in A&E (which are bad for patients). And we know that A&E delays are bad for mortality and outcomes. Monitor's recent report adds further weight to this hypothesis. This study now reports a direct link to mortality when beds are crowded.

The high bed occupancy is directly bad for patients and is caused by poor operational management of discharge processes. Those processes are much poorer at weekends than they are during the week. This alone may explain the weekend mortality effect.

The weekend mortality problem is not primarily a medical problem, it is a management problem. The lesson for policy is that a focus on medical contracts is a distraction. If we really want to fix weekend mortality we should focus on improving the way hospitals manage the flow through their beds, especially at weekends. As a bonus, this would also lead to major improvements in hospitals' ability to treat patients quickly in A&E. This should be a double win for patient outcomes.

Friday, 9 October 2015

NHS is running out of money, again

NHS hospital finances are in deep deficit (again). But a lack of money isn't the primary problem. While more money is probably needed, the system will still be in trouble if it doesn't address the deep underlying problems it suffers from.

[Note: this is an extended version of a comment I made in response to Alastair McClellan's HSJ editorial on the current financial state of the NHS.]

Over history, the NHS is used to solving every problem with more money but it has little experience with redesigning the way it works to cope with less money. or to do a better job with the same money. Over its life it has, on average, had a budget that increases significantly faster than the economy as a whole. During the Blair years its budget doubled in real terms in less than a decade.

The NHS needs to do better both in quality and efficiency. But every time budgets are tight, the wrong things get cut and the problems get blamed on the budget cuts not the chronic failures to modernise operational practice. 

The trouble is the sort of investments that might lead to significant operational improvements are neither sexy nor popular.

Much more IT, for example, could lead to the automation of data collection and administration, freeing medics from tedious admin and form filling and thereby leading to better decision making when they treat. Much more usable IT could reduce the burden of fragmented and user-hostile current systems again freeing up useful time for the benefit of patients.

Better management quality especially at the operational level might lead to better performance and a better workplace for nurses and doctors (better matching of capacity and demand is good for patients and better roster design is good for staff). But the naive slogan "more staff to the front line" make better newspaper headlines than better management of the staff on the front line (see my comments here).

Better understanding of costing could lead to improved focus for improvement efforts (if you don't know where the money goes–and many trusts don't according to costing audits–how can you know what needs to improve?) Or how come so many trusts argue they are not being paid enough to operate a service when they have such a poor grasp of how much that service costs?

Sadly none of the key things that are obviously broken in the NHS are things that generate good newspaper headlines for politicians who choose to invest in fixing them. And the internal lobby that calls for more staff and more money to solve the current problems don't have the wit to grasp that more staff in a badly managed system doesn't improve things much.

It is far too tempting, for example, to cut the capital budget for short term gain and far too headline-pleasing to focus on front-line staff numbers than on improving the way staff are organised. 

It is also too tempting to focus on big top-down structural changes rather than the much more effective bottom-up operational changes that actually drive long term improvement (see my comments here).

So the response to a short term crisis makes the long term problems underlying that crisis worse. And the medical lobby that argues that more money fixes everything reinforces the underlying problem by offering a naive analysis of what the problem is.

So, even if the NHS gets a short term injection of money, it will likely be spent on the wrong stuff.

The NHS probably does need a significant injection of cash. But it needs a serious dose of sense about bottom-up productivity and quality improvement even more and we shouldn't let the current debate distract from that.

Friday, 2 October 2015

Good Strategy, Bad Strategy, NHS strategy

Strategizing and planning takes up a humongous amount of management effort in the NHS. Given the quality of what emerges, that is mostly a waste of effort.

There are a lot of business books about strategy. But not many good books about strategy. In fact I’d go so far as to say there may only have been five of any merit in the last 30 years. The most recent in my short list was by Richard Rumelt and is called Good Strategy, Bad Strategy.

Rumelt thinks most strategy is bad and provides a useful diagnostic to help identify it. And he thinks that good strategy has of a kernel consisting of just three things: diagnosis; policy and plan. while his focus is mostly on business strategy the issues apply in the public sector too and I want to look at some examples from the world of the NHS. Not necessarily positively.

But first let’s look at a distilled version of what he says good strategy should be and how to identify when it isn’t. And then I’ll review some NHS strategy stuff against the standards (hint: I won’t be very positive).

He argues that bad strategy tends to dominate:

“Unfortunately, good strategy is the exception, not the rule. And the problem is growing. More and more organizational leaders say they have a strategy, but they do not. Instead, they espouse what I call bad strategy. Bad strategy tends to skip over pesky details such as problems. It ignores the power of choice and focus, trying instead to accommodate a multitude of conflicting demands and interests. Like a quarterback whose only advice to teammates is “Let’s win,” bad strategy covers up its failure to guide by embracing the language of broad goals, ambition, vision, and values. Each of these elements is, of course, an important part of human life. But, by themselves, they are not substitutes for the hard work of strategy.”

If you are starting to feel uncomfortable about your organisations strategy, then welcome to the club.

So what does a good strategy consist of?

“The kernel of a strategy contains three elements: 1. A diagnosis that defines or explains the nature of the challenge. A good diagnosis simplifies the often overwhelming complexity of reality by identifying certain aspects of the situation as critical. 2. A guiding policy for dealing with the challenge. This is an overall approach chosen to cope with or overcome the obstacles identified in the diagnosis. 3. A set of coherent actions that are designed to carry out the guiding policy. These are steps that are coordinated with one another to work together in accomplishing the guiding policy.”

What I like about this definition is its clarity. In a few sentences it cuts to the heart of what is wrong with a great deal of the sloppy thinking that many organisations call “strategy”. He elaborates a little more:

“The core of strategy work is always the same: discovering the critical factors in a situation and designing a way of coordinating and focusing actions to deal with those factors…

...A good strategy does more than urge us forward toward a goal or vision. A good strategy honestly acknowledges the challenges being faced and provides an approach to overcoming them…

...good strategy includes a set of coherent actions. They are not “implementation” details; they are the punch in the strategy. A strategy that fails to define a variety of plausible and feasible immediate actions is missing a critical component…”

To summarise in my words a good strategy needs the following:
  • A diagnosis of the most critical problem being faced by the organisation
  • A policy to deal with the most critical problems that also acts as guide to rule out actions that will distract from dealing with the most critical problem
  • A realistic, achievable plan of things that can actually be done that deal with the challenge in a way that is consistent with the policy

That doesn’t seem to hard, does it? Apparently, though, it is (and not just for the public sector) as few organisations come close to having strategies that meet those criteria. To be fair it isn’t defining good strategy that is hard, it is generating good strategy which nobody finds easy.

So how do strategies in the NHS stack up?

Here is an example of one way things fail.

Last winter NHS England was keen to encourage CCGs to sort out their perpetual winter crisis. They were happy to spend serious money (I believe more than £500m has been spent in the last two years) on avoiding the bad headlines that come around every year as “winter pressures” appear to overwhelm the emergency care system. I got a phone call from one CCG who had been kicked out of the room for producing an unconvincing recovery plan (I am an expert on emergency care performance having worked on and off in the area since the 4hr A&E target was originally set). I won’t name them to protect the guilty.

I have my own views on what the problem is in emergency care and I’m one of the few people to have produced analysis of the performance data that tries to narrow down where the problem is (the regulator Monitor have recently done an exhaustive analysis that pretty much agrees with my less complete analysis and experience).

So I asked the CCG to tell me about the plan they had presented. I compared it against my list of things I thought were likely to work. There was no overlap. I won’t go into the detail but the key issue was that most actions were focussed on diverting patients from A&E on the assumption that the problem was that the emergency departments were being overwhelmed by demand. This should be one of the easiest hypotheses to refute as the weekly national performance data have never shown any relationship between the attendance volume and performance (You can find some of my analysis of this elsewhere on this blog. Monitor’s assessment agrees with mine.)

So I quizzed the CCG as to why they were focussing on actions that didn’t work and proposed that they should look at some other ideas (for example whether their hospitals had a problem with discharging patients and finding free beds. This is Monitor’s top identified cause of poor A&E performance and the one I’ve been ranting about for years). I suggested some tools (eg this one) they could use to explore whether this issue was significant locally. I assumed they had never thought of this. I was wrong. They had already done this analysis and identified that beds were a serious problem in local providers but had ignored that in their strategy. Why? Because, they said, they didn’t want to annoy the clinicians in the hospitals as they didn’t agree with the analysis (even though it was based on their data).

So they produced a strategy that had essentially no hope of addressing the problem they had. They chose to address headlines and naive explanations about what the problem was rather than face down local political pressures that were driven by denial about the real source of their problem.

This identifies two issues that recur repeatedly in the NHS and lead to much “strategy” being built on a foundation of air.
  • They fail to identify the critical issue in the situation
  • They fail to choose a focus that might upset a strong constituency and prefer multiple actions that make people feel something is being done

Of course, the second factor is often the cause of the first. Even when we know what the problem is we can't address it because of politics.

These failures are pervasive in public sector strategy. Even when one part of the process is done well other pressures undermine the result. In the NHS we sometimes (though far too infrequently) see some decent diagnosis of what the top problem is. But the strategy becomes muddled because of an inability to choose to focus on just the top problem. More frequently, the strategy addresses political issues and never bothers with the objective analysis of what the challenges actually are. This problem is exacerbated because just about every distinct group inside the NHS has some lobby pursuing their interests.

The RCGP (the lobby for GPs) thinks the key problem for the whole NHS is that the government hasn’t done enough to mitigate demand on them and hasn’t recruited enough new GPs to make the workload bearable. They even produce numbers to back these ideas up. But the numbers are bogus as the NHS doesn’t collect reliable data about how many GP appointment there are. Moreover there is significant evidence that GPs who apply some operational insights to how they book and deal with appointments can reduce their workload substantially while improving patient satisfaction. The correct diagnosis of the problem might be a poorly designed and inflexible appointment process and not a tsunami of demand.

Or, in recognition of the NHS’s inability to operate properly at weekends, Jeremy Hunt demands changes to doctors contracts. Nobody seems to have done any analysis of the real problem.  Why bother when the headline solution seems to be so attractive. But, while it is clear the NHS doesn’t operate well at weekends, it is far from obvious that the fault is driven by lack of medical cover. Piss poor processes for managing discharges from beds seems to me far more likely, and that isn’t fixed by having more doctors around.

And there are big areas of NHS strategy where we could apply those criticisms. Especially where anything involving those newfangled things, computers. The National Information Board (NIB) has a framework for the future of NHS IT (which is describes as a framework, wisely since it isn’t much of a strategy by Rumelt’s definition). The NIB framework says a lot of plausible things about what the NHS should do with computers and IT. But it lacks a compelling diagnosis of what is broken or why the NHS hasn’t done obviously beneficial things in the past. And it flunks the second test as well by failing to choose. It defines a wide range of actions rather than focussing on the most important action. In an attempt to please many, it fails to make a compelling case to drive a focus on anything that might make a difference.

And many of the other strategies in the NHS suffer exactly the same problem.

A hospital is in trouble. Let’s have a merger! We should get some economies of scale at least. We have compelling evidence this strategy doesn’t work. And one of the key reasons is that scale isn’t usually the problem and mergers distract from rather than fixing the underlying operational problems that cause the trouble in the first place.

Emergency performance is poor in the region. Let’s have a major service reconfiguration! Again, these rarely work as the diagnosis is wrong at the start. Scale isn’t usually the problem and scale doesn’t lead to better operational performance not least because larger units are harder to organise than smaller units.

Demand is growing faster than we can cope. Let’s pursue integration of social care, primary care and secondary care! I’m not even sure what the logic for this is. We know that few NHS organisations are any good at internal coordination among their departments even before we bring others into the mix. And if coordination is what we need we don’t need organisational integration to achieve it. Coordination is an information problem not an organisation structure problem. The drive for integration is not based on a coherent diagnosis or a coherent policy approach. It is pure wishful thinking based on a goal we’d like to pursue which it isn't even clear would make any difference to the underlying problem.

Even when good ideas emerge in the NHS they are often undermined because the other essential elements of effective strategy are missing. The recent saga of Cambridge University Hospitals is an example. I’ve written in more detail about this here. The now departed boss had a reasonably plausible vision for where he wanted the hospital to go and it was based on a reasonable diagnosis of a key organisational problem: the lack of good quality information about patients and hospital activity. His visionary solution: a new hospital-wide eHospital system from respected supplier Epic. Nothing wrong with that. But compelling vision is merely a wish fulfillment fantasy in the absence of the other elements of an effective strategy. It seems like his strategy lacked both focus and the hard, detailed operational plan that would turn it into reality. As a result he has departed and the hospital is in a deep, deep hole.

I could go on, but I’m not sure it would help.

Given the humongous amount of time and effort the NHS devotes to planning and strategizing you might hope that the system would do a better job. For those of you who have to contribute to this futility, save yourself some effort by reading Rumelt’s book. Make sure your strategy has a kernel that works: a diagnosis, a focus and a plan. If it doesn’t, then don’t even bother writing it down or printing it. In fact, run for the hills: you are on the road to nowhere.