Jane Smith: The 11th European Forum on Quality Improvement in Health Care

This is my visit to Prague and the first quality form I’ve attended since the first prototype in London, well over 10 years ago. It’s much bigger ‘ over 1000 delegates from 49 countries – and much more professionally run.

The Congress Centre used to host Communist Party congresses and the guide book describes it as ‘supremely ugly’ which is unfair. Its meeting rooms are flexible and comfortable and its circulation spaces generous with stunning views of Prague. Some people have been coming to the European Forum (from next year to be renamed the International Forum on Quality and Safety in Health Care) for years, and they meet in the sunny spaces of the congress hall and are friendly.

No conflict
A Swedish participant commented that there isn’t much conflict or disagreement at the quality forum, everyone is trying to do the same thing, everyone agrees. That makes it friendly, but maybe isolating too. The people who go back from this forum bring back the buzz and excitement of ideas on quality improvement ‘ and tools on how to do it -‘ but they often go into hostile (or perhaps just non-comprehending) environments in their hospitals and clinics and offices. Helen Bevan, who has the wonderful title of director of service improvement at the NHS Institute for Innovation and Improvement, worries over coffee that we don’t get beyond the tools, to actually teach people the how of getting started ? which, she says, is a unique problem in every work place, because the people and the dynamics are all different.

The Swedes may have the right idea to counteract this sense of isolation. There are 249 of them here, and they come in large groups. Stockholm County for example has about 40 people at the forum, including politicians, support people, and clinicians and managers who run the system. I met several of them over lunch and though many of them have been coming for years and have been working on improving their health care system for years, they still say it is hard, and there’s still lots to learn. The politicians in particular want to hear from other health systems, other policymakers.
The need to spread
The urgent issue at this forum seems to be about ‘spread’: how do health organisations move really good improvement work out of a small microsystem – a ward, a clinical service – and spread it throughout the rest of the hospital or health system? Many of the speakers tried to tackle this issue in various ways.

Don Berwick dealt with this in his opening session. Don is the President of the Institute for Health Care Improvement in Boston and one of the gurus of quality improvement in healthcare. Many people come just to hear Don, though most have met and heard him speak elsewhere as well.

I’ve heard quality forums criticised as being a bit like a religious movement, a bit evangelical, with Don as the high priest. But in this talk on Wednesday morning he was less high priest and more ‘marshaller’ and analyst of a vast body of knowledge and experience. He drew on five big system-wide quality improvement enterprises, including Britain’s NHS, J?nk?ping County in Sweden, and the IHI’s 100,000 lives campaign to extract some common lessons (as well as the differences).

The differences were interesting, from England’s top down politically driven system covering a population of 40 million, to the 100,000 lives campaign, which has tried to mobilise activity to reduce deaths in hospital by using the techniques of a political campaign. But despite this, Don drew some general lessons:


that bold aims help

that measurement has to follow the aims (not the other way round)

that all improvement is local – it is necessary to build skills

the best incentives are intrinsic

the voice of the patient can be powerful – professionals will hear things from patients that they might ignore if it comes from a fellow professional

and ‘leadership, leadership, leadership’

Berwick’s other concern was with what he called the ’emerging canon of healthcare policy,? which he claimed is recognisable internationally and driven by ministries of health. The elements of this canon are:


markets and contestability

public reporting (star ratings, league tables etc)

a shift of choice and power to consumers

a shift of responsibility to consumers (perhaps through co-payments)#

pay for performance

He didn?t think this canon would improve health. The ?canon? clearly struck a chord: people were talking about it and recognising it and its sometime negative influence. Quality and Safety in Health hopes to publish Don?s talk in full, and the BMJ will have a shorter version.

Culture, culture, culture
So if the question is spread, what?s the answer? One answer is culture and relationships. Organisational psychology has always been a strong strand in quality improvement, but seems to be less often talked about than the other strands like process mapping, the statistical tools, and team work. Maybe that?s because it?s hard ? ?Soft is hard,? as management guru Tom Peters said.

Paul Plsek, who is a great performer and always enlivens and provokes, argued that the answer was pattern mapping. His organisation, Directed Creativity (www.directedcreativity.com) includes eight heuristics that will get people going with their creative thinking: one is ?Pause and carefully examine ideas that make you laugh the first time you hear them.?

He started by claiming that health care is a complex system (see his series in the BMJ: http://bmj.bmjjournals.com/cgi/search?fulltext=%22complexity+science%22&&journalcode=bmj&&hits=20), which explains why improving the quality and safety of health care is hard, and why it is particularly hard to extend improvements across an organisation or health system: ?Complex systems are like raising a child ? you can do it once but it?s no guarantee of success with the next.?

Plsek?s advice is therefore to study not only structures and process but also patterns. His main claim was that people didn?t pay enough attention to patterns, in particular the key ones of the nature of relationships in an organisation; how decisions are made; how conflict is handled; how power is defined, acquired, and used; and how learning is supported. It may, he said, be the missing link in transformation.

Plsek gave an example of a monthly meeting of IT people in one region in England?s NHS. Each IT person represented his or her organisation and at the meeting they discussed innovations and solutions to problems. But once that group had decided on something they couldn?t go back to their organisation and just implement it, because they weren?t empowered to. Instead they went back and discussed it with their chiefs, who raised issues, so that by the time the IT managers met the next month the issue was discussed again in the light of the issues that had been raised. This went on for several months until something was decided on that would stick and could be implemented. Not surprisingly the IT managers (and their chiefs) were frustrated because decisions took so long to make and implement, but they exposed this pattern and realised that it was caused by the fact that none of them were completely empowered to make the decision. They also recognised that they couldn?t change that, so instead they changed the format of the meeting.

Now at each monthly meeting they meet, make a decision, break off to telephone their home base, re-discuss it in the light of new issues, break again, and so this as many times as necessary until they have a workable decision. The difference is that the decision is made within the one meeting rather than over several.

It?s not a perfect solution, agreed Plsek: ?They speeded up the decision without challenging power gradient,? but the point was that they had identified what was causing the problem and come up with a better solution. ?Identify the thing that?s stupid and change it.?

What high performing organisations do
What Paul Plsek conveyed in a barnstorming presentation matched very closely the more sober message of Paul Bate of University College London and Peter Mendel of the RAND Corporation. Their session, on the secrets of high performing healthcare organisations in Europe and the US, was the result of research project to try to get beyond what works to why it works.?

They examined their 6 organisations (in the US, the UK, and the Netherlands) by studying (questioning, observing) top management, middle managers, and people interacting with patients.

They found the familiar things: good IT systems, measurements and data, PDSA (plan, do, study, act) cycles, process design, and the input of IHI. Their surprise was that most of the stories were about human and cultural and organisational factors. Their findings are complex and detailed (and about to be described in a book) but essentially they found that in these successful organisations the patient level and top management were working in synchrony: the organisation?s leaders were letting the micro level get on with improving patient care but protecting it from politics and constraints. They commented that they would really liked to have studied 6 unsuccessful organisations but that they hadn?t the time or the funds.

These insights into organisational realities were nicely illustrated by a short presentation by Lorelei Lingard on how a hospital in Toronto had introduced a checklist into its operating room procedures to improve reliability and communication. The checklist is used just before the operation by the surgeon to brief and gather information from the anaesthetist and operating room nurse. She commented that they had decided that the surgeons had to run the checklist briefing because they?d judged it was not politically acceptable for the anaesthetists or nurses to run it ? if they did the surgeons might not take it seriously.

Because the checklist project had started with some research funding, Lingard?s team had had to get the informed consent of all the original teams who had designed and piloted the checklist (three staff had declined). Although that fact didn?t prompt any comment in that session, the issue of whether quality improvement work needs informed consent is a hot one amongst some quality workers. The BMJ is thinking of running a controversy on it.

Whenever that question pops up I often think about non-health organisations handle this sort of issue with their staff, and one analogy is mystery shoppers in the retail sector (who pretend to be ordinary shoppers but are actually inspecting the service retail staff provide). So I was delighted to see a poster on ?Mystery shoppers as a tool for sexual health service evaluation.? It wasn?t the winner of the poster conference (though that too was on sexual health services), but it got my vote.

?Plot the bloody dots?
To another barnstorming session, this time by Davis Balestracci, who describes himself as a ?right-brained statistician.?

His subject was the sort of data analysis you need to understand the problems you?ve got. He was scathing about externally directed goals, traffic light systems, comparisons of this month with last month and to the same month last year, etc. Rather tellingly he used a real example of data from an English hospital on four hour waits in the accident and emergency department to show that the figures that management highlighted as months when ?something went wrong? were nothing of the sort, but well within the range of random variation of that process.

The key problem is the old one identified by W Edwards Deming (American quality guru, whose thinking in post-war Japan is credited with leading to Toyota?s famous lean production) ? muddling up special cause and common cause variation. Common cause variation is that caused by the normal random variation of your system, and a key problem is treating something that is random variation as if it was a real outlier ? ie special. So the trick is to identify the random variation in your system, see if there are any genuine outliers (and tackle them), and if there are not, think about what might improve the current system and shift the band of variation.

Balestracci?s message was: ?Crude measures of the right thing are better than precise measures of the wrong things? and his advice: ?plot the bloody dots?: look at the actual data. He had lots of fun examples for us to try (he suggested doing this in boring meetings when people are looking at histograms and testing whether the data are normally distributed). We plotted the number of drug errors each month over four years and saw immediately that something strange was going on every January, something that was completely obscured by the summary statistics we had initially been given. Plotting the data won?t tell you what to do, but it may help you ask the right questions.

The conference ended on a high note with a bravura performance from John Halamka, the chief information officer at Harvard?s Caregroup health system. Using an internet connection he logged on to the web based information system available to hospital clinicians and related primary care providers. CareWeb links information sources, 9 million patients? records, laboratory results, and imaging in an uncluttered clear interface ? and was impressively fast.

He also illustrated a new way that patients can help their doctors. Instead of leaving her body to medical research Martha Ford had left her data ? so he pulled up her records, clicked on her laboratory results, and would have shown an image or two if the conference laptop had had Acrobat installed. The system has voice recognition for 3000 clinicians, with 98% accuracy, so they don?t have to type. Patients can access the site and email their doctors. The doctors were worried they?d be swamped with emails ? but they haven?t been. Patients average 1.2 messages a month and the emails have replaced phone calls.

The system had impressive safeguards for patient confidentiality and allowed patients to opt into the system or out of it.

The NHS person I was sitting next to said that the NHS?s information system should ultimately be as fast as CareWeb. If it?s anything like as impressive it could be excellent. But, as Halamka said, summing up both is session and the forum, people and process are always more important than technology.

Quotes heard
?Beware the tranquillising drug of gradualism? (quoted by Don Berwick in his talk)

?Understanding involves 3 phases: simplistic, complex, and the profoundly simple? (A Schultz)

?Soft is hard? (Tom Peters)

?If we are actually trying to do the wrong thing the only reason we may be saved from disaster is because we are doing it badly? (David Kerridge)

Competing interest: The BMJ Publishing Group (for whom I work) runs the European Forum along with IHI. I?m part of the group working on the next and future forums, but apart from chairing a session I had little input into the Prague Forum
Rapid Responses:

Read all Rapid Responses

One can observe a lot just by watching-Learning from experience is also research.
BM Hegde
bmj.com, 8 May 2006 [Full text]