15 May 2011

The Healthcare Commission Story Part 1

For anyone who has been close to the Stafford Hospital Story over the last three years one of the most significant and contentious events was the decision by the Healthcare Commission to launch an investigation into the hospital in early 2008.
The Midstaffs Public Inquiry has been taking evidence from a number of people who were involved in the HCC at the period that this decision was being made.
This is a brief summary of some of the things that have seemed most important to me. I am aware that I will need to come back and flesh this out much more fully, but I am also aware that the messages I am picking up through the Inquiry raise really serious questions about the assumptions of “a failing NHS” which lie behind the Governments proposed Health reforms. I want to alert more people to this possibility now, at a crucial time for decisions about the health reforms.
The assumption that has been made in the press, and by many of the politicians who rely on the press for their information is that the sequence of events which led to the HCC Stafford Hospital investigation is as follows:
·         The Dr Foster Intelligence produced in mid 2007 a league table of hospital mortality rates which it published in the Daily Telegraph, raising widespread public concern (and also widespread outrage within the NHS about the appropriateness of using the data in his way)
·         Julie Bailey’s mother died in November 2007, sparking a low key local campaign which led to the formation of “Cure the NHS” in Early 2008.
·         Cure the NHS began systematically contacting anyone concerned with the Health service, and build up their contacts with the local press to raise the profile of the concerns about basic nursing care that they felt about Stafford Hospital.
·         Cure the NHS contacted the Healthcare Commission in early 2008 and it was the combination of the Dr Foster mortality rates and the Cure the NHS concerns that triggered the Healthcare Commission investigation of the hospital.
This assumption – according to the testimony given by different members of the HCC is wrong. What actually happened is this:
·         The HCC were aware of the Dr Foster league table, but shared the opinion of other statisticians and regulators within the NHS that the HSMR rates were useful as a management tool, but that the mortality rates produced by this method are insufficiently robust to be used as a league table, and cannot be used as an indicator of quality of care.
·         The HCC did not monitor the local press and so had not picked up on the campaign by Cure the NHS.
·         The HCC received a flood of complaints brought together by the newly formed Cure the NHS in early 2008. They had received no complaints from Stafford before this. They noted the complaints, but would have been unlikely to have triggered an investigation as a response to this.
·         What did trigger the investigation was a new statistical system that the HCC was working on in conjunction with the Dr Foster Unit. (It is important to note that the Dr Foster Unit, which is an academic body attached to imperial college, is entirely separate from Dr Foster Intelligence, which is a commercial company which sells mortality information and consultancy to the health service).
·         This new statistical system was to look at outliers. It was being developed by the Dr Foster Unit as a means of measuring outcomes in the NHS. It is broader that mortality, taking in matters like the development of bedsores, or delays in discharge, or readmissions. Whereas the HSMR system gives a composite figure for a hospital which may for a number of reasons give a very misleading impression, the Outliers system is intended to allow people to focus on the fine detail of performance on a particular ward, or kind of surgery.
·         The reason why this system was being developed is that the Dr Foster Unit and HCC both recognised the limitations of HSMR and needed something that would give managers a much more detailed window into their own performance. This system is, as the evidence of Martin Bardsley makes clear, primarily designed as a tool to assist managers and clinicians drive up quality. It was not conceived as a means of measuring quality of performance.
·         The work on these new outliers was being carried on in a rather clandestine way. The HCC make it clear that they were nervous of word of this new system getting out as they believed that they would be criticised by many people within the health service for this groundbreaking new approach. The Dr Foster unit may also have been aware of the commercial possibilities of the system, and wanted to “keep it under wraps”. The wider Health service was not told about this work which the HCC witnesses have all been at pains to say was a “pilot scheme”. They have said this in reply to questions of why they were not sharing the information they were getting through the alerts generated by the system with the SHA, PCT and Monitor.
·         The very first alert produced by this brand new pilot system came in July 2007 and related to operations on the jejunum carried out at Stafford Hospital. The very cautious approach taken by the HCC statisticians to this new information was that this was not necessarily telling them anything, but it raised questions which they wished to follow up with the Hospital. They expected that the outlier was probably caused by coding problems. The process that they followed was to write a letter to the Hospital asking for further information.
·         Robert Francis speculates on how this would have been received by the hospital. They were being asked to respond to a system that they knew nothing at all about, and they may not have understood what was expected from them. (It is clear that this would have been happening at the time when the hospital was undergoing significant problems as a result of cut backs caused by the financial problems. There were real problems occurring within the hospital and dealing with this extra paperwork exercise may not have seemed like a high priority task.)
·         Between July and November more outlier alerts into a number of different clinical areas all associated with A&E were generated from the hospital. The Mortality Outlier group at the HCC continued to follow these up with more letters seeking further information.
·         By the end of 2007 the HCC mortality outlier group were becoming concerned by the number of the outliers, and by the lack of a full response from the hospital. They had still not drawn any conclusion about what this might actually mean, they were perfectly aware that the results could easily be explained by either the case load, or by idiosyncrasies in the way that the Hospital was coding cases. (In practice I think that Martin Yeates and his team would have considered that this matter would be covered by the work being carried out by Professor Mohammed of Birmingham University, at the instigation of the SHA to look at the High HSMR figures in 6 of the hospitals in the West midlands region.) The HCC mortality outliers group were however sufficiently concerned to refer their findings to the HCC Investigation team who then prepared to carry out a preliminary visit to the hospital.
·         The Preliminary unannounced visit took place in early 2008. Heather Wood describes that they were concerned about the EAU where they found an elderly lady out of the sight of a nursing station, in a dark corner of the ward, and at risk of falling out of bed. The lay out and Staffing levels that they found on this visit to the EAU was sufficient to persuade them that a full inspection was necessary.
·         Once an inspection was announced then the HCC rapidly became aware of the Cure the NHS campaign and their complaints fed into the investigation process.
·         The high level of attention given by the press to the investigation and to the Cure the NHS campaign ensured that more people came forward to tell the HCC team about their stories. Many of these people have never been directly connected to the Cure the NHS campaign, but felt that they had information which might be relevant.
How did the wider Health service see the investigation?
·         People at all levels within the health service have shown how concerned they have been about the role played by the press. Whilst no one has directly criticised individual papers yet, there is a very high level of awareness of that stories need to be managed in order to try and prevent the damage that can be done by misleading and sensationalised stories in the press.
·         There has been considerable disquiet about the length and the conduct of the investigation, and the damage that has been done to the hospital and public confidence in both the individual hospital and the health service as a whole both by the investigation and the consequences that flowed from it.
·         The investigation has clearly been extremely divisive within the health service with many direct criticisms having been made both of the process and some individuals within the team. Heather Wood made it clear that her relationship with the rest of the health service has been deeply damaged by this. The animosity that she feels towards a number of individuals was clearly displayed within her evidence. The key members of the Investigations team interviewed by the Inquiry are no longer working for the CQC.
What did the Investigation tell us?
·         When Ben Bradshhaw as Secretary of State for health had discussions about the investigation in May 2008 he asked the question “is this another Maidstone and Tunbridge Wells.” He was given the assurance – some four months after the investigation had begun that this was not the case. So this means that even after four months actively looking for problems that the HCC investigations team were not seeing anything that led them to expect that large numbers of deaths had occurred.  
·         With my personal reading of the Healthcare Commission report I saw that a very thorough trawl of information about the hospital had been done and that there were a number of real problems that had been identified. I also read with interest the sections on the statistical material which did make it clear that there were significant problems with coding which would have made the Dr Foster material unreliable.
·         A reading of the report showed plenty of areas where improvement needed to be made. The body of the report does this with a certain detachment. The Summary is much more dramatic than the report, and many people within the health service were concerned about the “sensational” way in which the report was presented.
·         Though it is completely clear that there were many things going on that will have made peoples stay in hospital undignified, unpleasant, and perhaps at times unsafe I personally found it difficult to see anything in the report that might have explained death on a widespread scale.
·         The “excess death figures” that appeared everywhere in the press and have continued to be used ever since were particularly puzzling because they simply do not appear in the report at all.
Why don’t the Excess death figures appear?
·         The figures which have formed the central assumption in every piece of press coverage from March 17th 2009 onwards is that there were 400-1200 excess deaths as a result of poor care. These figures apparently appeared as part of an appendix in a draft version of the report
·         Sir Ian Kennedy explains that the decision not to use the excess death figures was taken by him.
·         A very limited number of people have seen this material. I am not one of them, so I cannot give any reliable account of what the document said or implied.
·         The assumption made by the press is that the figures were removed as a result of pressure from the DOH. Sir Ian makes it very plain that he does not bow to that kind of pressure and that the DOH did not ask him to do this. Bill Moyes from Monitor did say to him that the figures were insufficiently robust and should not be used.
·         Ian Kennedy says that his decision to leave the figures out is based on his experience of Bristol, where there were a number of “unnecessary deaths” identified, but there was no way of connecting these to individual deaths. His concern was that if a figure of numbers of deaths were to be used by the press that this would cause a great deal of pain for many people who had lost relatives in this period. They would wonder if there was more that they could or should have done.
·         This fear is entirely borne out by the evidence given by many of the bereaved who have given evidence to the Inquiry. They believe that their loss was “the tip of the iceberg” and they feel driven to pursue to the matter so that “no one else has to suffer”. As Ian Kennedy gave his evidence I was sat directly behind a lady who has said to me in so many words “My (relative) was one of the 400”.
·         It is clear from the evidence of Heather Wood and Nigel Ellis that they felt that the figures should have been included in the report. At least in the case of Nigel Ellis he felt that the figures should be there for completeness but with strong provisos that these were theoretical figures which related to probabilities, and did not tell us that any number of people were known to have died as a result of poor care.
Degrees of certainty about the Excess death figures.
At the moment there is still huge uncertainty about these figures. This is something that the inquiry still needs to actively explore.
I do not know the following:
·         Who wrote the deleted material?
·         What is the statistical source of the excess death figures?
·         How the deleted material is phrased – is the word “If” included?
·         If the figures are based on the Dr Foster figures then has it been made clear that the NHS as a whole does not accept these as a robust indicator of quality of care?
·         If it is based on the mortality outliers then has this material been subject to peer review?
·         What do other statisticians think of it?
What I do know is that the degree of certainty with which people talk about the Excess death figures is in inverse proportion to the understanding of their statistical basis.
We can trace this through the evidence given to the Inquiry.
·         Martin Bardsley who was the person within the HCC who understood the statistical material best is very cautious about the conclusions that can be drawn from the material. He sees the Mortality outliers as management tools, there to assist in improving quality. They are not to be seen as an indicator of quality.
·         Nigel Ellis, is almost as cautious about the use of the figures. He is an investigator who will use statistics, rather than a statistician. He does see the excess death figures as telling us something about what happened at Stafford. He wanted the figures to be used with qualifying statements about what they meant.
·         Heather Wood who was the head of the investigation team says clearly of herself that she is not a statistician. She has clearly accepted from her more expert colleagues that the figures are telling her something, and she is then falling back on her own preference for dealing with what other people pejoratively call “anecdotal evidence”. It is Heather Wood who was directly questioned by the press at the press release of the HCC report, following the leak to the press of the excess death figures on the previous day, and her response to their questioning gave the press the go ahead to attribute these figures to the HCC.
·         Sir Ian Kennedy is at a further remove from this, he accepts the information that is given to him by others working within the HCC and his comment is that he thinks their methodology is sound, but that it would be unhelpful to use the figures because the press would misinterpret them and they would be into a major dispute about statistics rather than being able to focus on the real problems associated with poor care and the impact that had on individual families.
The Impact of the figures:
·         Sir Ian was completely right about the way the press would misread the information.
·         The fact that the material was not in the report and we could not see where it came from means that we have taken from March 2009 to May 2011 to see what actually happened, and that even now we do not know the detail of what was said in the appendix.
·         What perhaps could not have been anticipated is the degree of enthusiasm with which politicians then used these figures.
·         We have already seen from the evidence of Bill Cash that he had no idea where these figures came from, he simply accepted them at face value, as something he had picked up from the newspapers. He certainly continued to use these figures in his election leaflets for 2010 Even after the publication of the Robert Francis Independent Inquiry.
·         Other local politicians followed suit, and the figures were widely used in the Local election campaign of 2009.
·         We do not know what David Cameron and Andrew Lansley understood about these figures. It is possible that if they had their information from Bill Cash, or even from Ian Kennedy or Heather Wood that they may have read more into them than the statisticians themselves would have wished. What we do know is that Andrew Lansley and David Cameron have used these figures on a number of occasions, and have made Stafford a central “image” in their general election campaign.
·         I am not certain that they have used the figures since the statement on the publication of the Robert Francis Inquiry, with its clear indication that the figures are unsafe, Though David Cameron did use the figures in PMQs immediately before the statement. What I can say is that they have never attempted to correct the impression that they have repeatedly given that Stafford is proof of the failure of the health service, and is a justification for major reforms of the health service.
Is anyone to blame for the confusion?
I came to the evidence of the HCC sharing the anger felt by many people within the Health service about the way in which the Stafford hospital story has been sensationalised and the way in which misleading material has been used. I was prepared to blame individuals within the HCC for what had happened.
Having listen to and read the evidence so far I am able to see understand the motivation of the individuals concerned much more clearly.
Everyone who I have seen from the Health service at the inquiry has been genuinely concerned to try and deal with the real challenges that the health service is experiencing. They want to find ways to help the health service deliver the best possible standard of care. They each bring different skills and different approaches to this task.
The task that the PCT, The SHA, the HCC, and now the Care Commission are all seeking to perform is finding ways of seeing how good the quality of care is and how it can be improved. This is an ongoing task. We are not talking about some perfect time in the past when we could clearly see this information, the tools for doing this job have been under active development for the last decade or so.
Targets were a step on the way, The Dr Foster HSMR figures were a step on the way, The Clinical Dashboard developed by the SHA is a step on the way, The Mortality outliers system being developed by the Dr Foster Unit and HCC are a step on the way.
All of these things are tools designed to assist managers to manage. They are not and should not be used as hard and fast indicators of quality of care, in a way that the Press and public can use as a single headline indicator of how good a hospital is.
These are good systems, being driven by good people for a good reason. They are also highly confusing.
The problem only arises when they are used for a purpose that they are not intended by the press and politicians.