“Accident and Emergency: Performance update”
Agenda item 3 is on a section 23 report, “Accident and Emergency: Performance update”. The committee will recall that we have previously taken evidence from the Auditor General and Audit Scotland on accident and emergency provision, and this is an interesting update. The Auditor General is here and is accompanied by Catherine Young and Tricia Meldrum. I invite the Auditor General to brief the committee.
Thank you, convener. The report that I bring before the committee today provides an update on how the national health service has been performing against the four-hour waiting time standard in accident and emergency since our last report, in 2010.
A and E departments provide a really important service for patients with serious injuries or illness and it is important that patients are seen quickly. The Government standard is that 98 per cent of A and E patients should be treated and discharged or admitted within four hours of arriving. In April 2013, it also introduced an interim target of 95 per cent, which it expects NHS boards to achieve for the year ending this September.
As the committee is aware from my recent report on NHS financial management, the NHS in Scotland is not currently meeting the four-hour waiting time standard for A and E. Performance against the target has deteriorated since 2010, although it improved during 2013. There is significant variation in performance across A and E departments, with only 14 of the 31 departments meeting the 95 per cent interim target in December 2013.
It is important to say that there is no simple explanation for why more patients are now waiting more than four hours in A and E. A and E departments are part of a much bigger health and social care system and pressures across that system can lead to patients being delayed in A and E. For example, many A and E patients need to be admitted to hospital and delays can be down to a hospital bed not being available right when it is needed. That might be because another patient is waiting to be discharged from hospital later in the day and so is still occupying the bed that the new patient requires.
We know that more patients are now being admitted to hospital from A and E and that there are more delays in A and E because patients are waiting for a hospital bed to become available. There is also some evidence to suggest that patients attending A and E now have more serious health problems than they have done in the past. Across Scotland, just over a quarter of A and E patients were admitted to hospital in 2012-13.
Staffing challenges can also affect how long patients wait. A and E departments need the right number and mix of staff and they need those staff to be available when they are required. Since our last report, the number of A and E consultants has increased by 63 per cent, and there are now about 154 whole-time equivalent consultant posts. However, there are still pressures around medical staffing, including a reduction in some other grades of staff, difficulties in filling vacancies and lower numbers of staff being available at weekends and in the evenings.
In response to the deterioration in waiting time performance, the Scottish Government launched the national unscheduled care action plan in February 2013. One of the aims of the plan is to reduce A and E waiting times and the Scottish Government and NHS boards are taking steps to address some of the causes of delays. The initial work has focused on making improvements in acute hospitals. The next stage is expected to look at the wider health and social care system. It will take time to see the impact of the actions, but there was some improvement in A and E waiting times during 2013.
My report makes a number of recommendations to the Scottish Government. They are mostly about sharing good practice on initiatives that can help to improve A and E departments’ performance and waiting times for patients. My colleagues and I are happy to answer questions.
Thank you, Auditor General. You say that there has been some improvement in 2013 but it is worrying that since you last looked at the issue, there has been a deterioration despite the investment and despite the commitment to improve matters. It is clearly a matter of concern for the public.
Targets are set for a good reason. The politicians and officials who set the targets clearly believe that the targets are realistic and that there is a purpose for setting them. Therefore the fact that the targets have not been met is of great concern. You say that there has been some improvement more recently but I want to raise an issue that is relevant to me locally. Exhibit 7 of the report gives figures for A and E department performance against the target. In the Royal Alexandra hospital in Paisley, performance against the target dipped below 90 per cent in November and December 2013. From a West of Scotland perspective, it is even more worrying that the figures for the Western infirmary are consistently below 90 per cent.
Although hospitals such as the Gilbert Bain hospital and the Royal hospital for sick children in Edinburgh are consistently performing well, others are consistently performing poorly. Have you looked at the specific reasons for those dips in performance at the Royal Alexandra hospital in Paisley and the Western infirmary?
It is important to say that the report is a performance update rather than a full audit, as we carried out in 2010, so, to a great extent, we have relied on the nationally available data. We have investigated that data as far as we can to explore what is associated with better performance or with worse performance.
You are absolutely right to say that A and E waiting times are important, both to all of us as patients and as family members, in that we want people to be seen and treated as quickly as possible, and because there is some evidence that longer delays can compromise clinical effectiveness and the quality of care.
Waiting times matter and that is why the Government has set the targets that it has set. It is also important to keep this in proportion. As at December, 93.5 per cent of patients across Scotland were being seen within four hours.
Equally, there is huge variation across A and E departments, as you have highlighted with the examples from exhibit 7 of the report. One of the most important things that we are trying to draw out in the report is the need for individual A and E departments to really understand the factors that lead to delays in their particular departments, whether it is availability of beds, availability of clinical staff, or availability of alternatives at the right time, and to use that information to put in place solutions for their particular problems.
You will notice, for example, that big, complex, specialist hospitals, such as Ninewells, have very good performance. Case study 1 in the report highlights a range of things that Ninewells has done to achieve that. My recommendations are really about the Government and health boards taking a very similar, tailored approach in order to understand what is causing the bottlenecks in each system and to look beyond A and E for the solutions.
Have you identified problems either with investment or ineffective management? Again, forgive me for being parochial but I will stick with NHS Greater Glasgow and Clyde. If we look again at exhibit 7, we see that in December 2013, Glasgow royal infirmary performance was below 90 per cent; Inverclyde was above it; the RAH in Paisley was below it; the Royal hospital for sick children in Edinburgh was above it, as was the Southern general hospital; and the Victoria infirmary and the Western infirmary were both below 90 per cent. What would cause that type of cluster?
As we say in the report, the factors that affect A and E performance are complex and interrelated. First, the rate of attendance at A and E is affected by deprivation in the local area and by the distance from where people live to the A and E department. That is a starting point. Interestingly, we know that total A and E attendances have dropped very slightly since our last report. There has been a rise in attendance at minor injury units but A and E attendances have gone down. However, the number of older people attending A and E has increased, as has the number of people who are admitted to hospital from A and E, suggesting that people are sicker than they previously were and more seriously in need of attention.
As you touched on in your question, there are real differences in how A and E departments are managed as part of the wider system. In NHS Tayside, they have worked very hard to ensure that they have appropriate specialist medical staff available, not just during the working day but during the evenings and at weekends. As soon as it looks likely that a patient might need to be admitted, they start the process of identifying a bed so that there is not a wait for one. They work very hard at ensuring that general practitioners can refer directly to wards rather than through A and E and they signpost alternatives to A and E departments. That whole approach seems to be very effective at making the system work in NHS Tayside and we think that there is scope for other A and E departments to follow that sort of approach more.
After your recommendations, with the increased investment, and given the fact that areas such as Tayside can achieve the target, did you expect that the target would be met across Scotland?
It is certainly the case that since the Government’s national unscheduled care action plan was introduced in February 2013, we have seen an improvement—as we say in the report. There is scope for that improvement to be more consistent right across Scotland.
Equally, A and E departments can be an important indicator of pressure on the system as a whole. We know that there are financial pressures; that the population is ageing; and that all the pressures that we have discussed before in the committee continue. That is why we think that focusing on that specific indicator is not an end in itself. However, it is an opportunity to look at how the system as a whole works.
Thank you for the report, Auditor General. As well as the statistics on missing the four-hour target, I was slightly worried by a number of other statistics that emerged in your report. In particular, on page 18, in paragraphs 19 to 21, you state quite clearly that
“The number of patients who waited longer than 12 hours in A&E departments has increased”
and that
“The median wait for A&E patients has increased”.
That means that the issue is not just about not meeting the four-hour target; the average experience for patients is getting worse. Very worryingly, a huge number of patients—about 70,000 of them—are now being seen in the last 10 minutes of the four-hour period. That does not paint a very good picture at all, does it?
10:15
When you look at exhibit 8, you get the clear sense of the pressure on A and E departments building up and increasing since our last report in 2010. That is important, because despite all the efforts that the Government and NHS boards are making to meet the target of 95 per cent of people being seen and either treated or discharged within four hours, those pressures are still there. That will not change, because we know that the population is ageing, that older people are more likely to attend A and E and are more likely to need to be admitted and that more people have complex health problems. All that is part of what is going on, which is why we think that A and E departments need to be seen in the context of the whole health and social care system locally, so that the pressures can be properly managed. However, there is no question but that they are real.
Do you agree that it is not only about the one target of four hours and that the fact is that there is a problem across the whole area of A and E?
We say in the report that the increase in the median wait is an indicator of real pressure in A and E and across the health and social care system. The fact that the median wait has increased is not necessarily a bad thing in itself, although it is obviously bad for the people who might previously have been seen in an hour and are now waiting for two or three hours. It might mean that there is more appropriate triaging and that the care that people get is more tailored to their needs. However, it is certainly an indicator that there is pressure in the system, and we believe that a number of boards will find it hard to meet the 95 per cent target by September of this year.
Is that the average experience? You talk about the median. How long are patients waiting now compared with how long they were waiting?
As you highlighted, we report in paragraph 20 that the median wait
“has increased from 99 minutes in 2008/09 to 126 minutes in 2012/13.”
We do not know the figure for December 2013, because it has not been reported nationally, so we used the most up-to-date national data that is available. I would expect there to have been a slight improvement by December 2013, as we have seen across the rest of the A and E performance, but we do not know that yet.
There is a particularly worrying comment about the treatment given to those seen in the last 10 minutes of the four-hour period. Paragraph 33 of the report suggests that
“National data shows that patients who are admitted just before the end of the four-hour period are likely to spend longer in hospital.”
The report also states that
“11 per cent of all admissions to hospital from A&E departments happened within the last ten minutes of the four-hour period”.
That figure has gone up hugely.
Catherine Young might want to add to what I say. First, we would all recognise that one of the inevitable side effects of setting targets is that, for example, if a four-hour target is set, there will be particular attention on patients who are coming towards the end of that four-hour waiting period.
We cannot tell from the available data whether patients are being admitted inappropriately in the last 10 minutes to avoid breaching the target of four hours but, as we say in the report, we tested that by proxy by looking at how long those patients stayed in hospital. Our hypothesis was that, if they were being admitted inappropriately, they would have shorter lengths of stay, as they would be in for a short period and discharged. In fact, we found that the opposite was the case, as patients admitted in the last 10 minutes were likely to have longer lengths of stay. We therefore concluded that it was unlikely that they were being admitted inappropriately. However, we say that the statistic probably highlights real pressures in relation to patient flow through the system.
Does Catherine Young want to add to my comments?
The decision to admit might well have been made earlier in the four hours but, as the Auditor General has said—and as exhibit 14 shows—a bed is not always identified early on in the whole process of patient flow; in some cases and in some departments, it does not happen until late on in the process. That is why we have recommended that the Scottish Government shares good practice on discharge processes to try to start that process earlier for A and E patients.
For a number of reasons, that is very worrying. If the targets that have been set are being modified and are still not being met, that is worrying. If the targets are distorting care, that is also worrying. Is it not worrying in both cases?
We do not conclude that the targets are distorting care. As I said in my answer to your previous question, we tested that. We recognise that, first of all, setting targets and standards can be a good way of focusing the attention of managers and clinicians on things that matter to patients, but they tend to have a distorting effect, because people seek to hit the target rather than allow the way in which patients are worked to run in a more natural flow.
What we are seeing—and what we say in the report—is that there are indications of pressure in the system. Only 14 of the 31 A and E departments met the four-hour target in December of last year, and we think that many boards will struggle to hit the new target by September of this year.
You highlight the availability of beds as being one of the reasons that might be behind that. I believe that the statistics from February this year show that 135,000 beds were lost due to delayed discharge, and more than 1,500 beds have physically been lost over the past seven years in Scottish hospitals. The number of beds that are available in hospitals is down by 1,500 or possibly more. Which of those factors is more important?
We talk about the issue of bed numbers and bed occupancy in paragraph 40. You are right; the overall number of acute beds has reduced since our previous report, we think by 7 per cent, from 17,374 in 2008-09 to 16,230 in 2012-13. Most of those beds are in acute surgery, reflecting the fact that more surgery is being done on a day-case basis whereas patients were previously admitted at least overnight and possibly for longer. We have reported on that to the committee before.
It is interesting that the average occupancy rate for acute hospital beds has increased over the same period, particularly in acute medicine, which is often where patients who are admitted through A and E need to find a bed. In acute medicine, the average occupancy rate in 2012-13 was 85 per cent. That average can conceal some periods of very high occupancy, and there is some evidence, although it is not conclusive, that above 85 per cent, clinical safety can become more difficult. That is why we think that the focus on bed occupancy is so important here. If bed occupancy is a bit lower, it makes it easier to find a bed, and when we look at the correlation between bed occupancy and performance, there is a clear relationship. I ask Catherine Young whether she wants to comment on that.
In paragraph 41, we mention that we ran a correlation between higher bed occupancy and performance against the target, and we draw on the range between Tayside, where occupancy was 79 per cent—or almost 80 per cent—and Forth Valley, where it was much higher. NHS Forth Valley has weaker performance against the target compared with NHS Tayside. However, the issue is also the use of those beds, and again we highlight the good practice in Tayside, which involves having good discharge processes in place early in the process. It is not just about the numbers; it is also about how the beds are used, timing and the availability of beds.
When we discussed your report in December, we discussed the fact that no health board had met the waiting time target. The committee expressed a lot of concern that, despite the so-called patient guarantee, patients have no recourse whatsoever. Do patients have any recourse in this case if the four-hour target is not met?
The A and E waiting time target is not enshrined in statute in the way that the treatment time guarantee is, so there is no recourse other than through the normal complaints procedures that are taken forward by each health board individually.
Thank you.
Just before I bring in Willie Coffey, I have a question that relates to what Ken Macintosh said. In exhibit 14, you show admissions from A and E in the last 10 minutes of the four-hour target time. Have you looked at what the figures would have been like had those patients not been seen within the last 10 minutes? How much worse would they have been?
We have not done that because those patients were admitted and we did not find any evidence to suggest that people were being admitted inappropriately.
Did you look at all to see whether those patients’ experience was rushed or was less thorough than that of those who were seen at other times in the four-hour period? In other words, is there anything to suggest that hospitals are suddenly rushing people through A and E in the last 10 minutes in order to ensure that targets are met?
No. Unlike in 2010, when we interviewed patients directly and used focus groups to explore their experience, in this update we used just the nationally available performance data. We are planning work on unscheduled care more generally, and as part of that we will want to talk to patients again.
Such work would be useful at some point in the future, because it would be a concern if we found that patients were being suddenly rushed through accident and emergency to meet bureaucratic targets.
I want to consider the numbers in a bit more detail. It is quite correct that the committee examines performance against the target, but it is also appropriate that we acknowledge good performance. To meet the target in 93.5 per cent of cases is pretty good. There are 1,600,000 presentations to A and E every year, and in 1.5 million cases the target is being met. I suggest to committee members that that is not bad, considering the pressures and strains on the NHS, which the Auditor General has mentioned.
Wales is meeting its target in 87.7 per cent of cases, so we are doing significantly better in Scotland, although we want everyone to achieve the target that has been set. How far short of meeting the target are we, in terms of the actual number of patients who present to A and E?
You are right that it is important to keep things in proportion. Straightforward numbers are often easier than percentages to get a grip on.
The most recent figures are from December 2013, at which point it was possible to see an improvement from the low point in January 2013. We say in the report that 8,300 patients across Scotland waited for more than four hours and 118,000 waited for less than four hours. There was good performance for 118,000 people, but it is clear that performance was not as good as any of us would have hoped that it would be for the 8,300 who had a longer wait.
The interim target that Government has set for September is 95 per cent, so I think that we would expect the 8,300 figure to come down slightly to nearer to 7,500. We can give you the exact figure separately, if that would help. Of course, the Government’s 98 per cent standard, which is still in existence, would require a further shift.
In our report, we focused on the variation across Scotland. Fourteen of the 31 A and E departments reached the target in 2012-13 and 17 did not do so. As you can see from exhibit 7 on pages 16 and 17, there is still significant variation between A and E departments. The table shows the departments that are meeting the 98 per cent target in green, the ones that are meeting the 95 per cent target in amber and the ones that are missing both targets in red. It is the variation that should be the focus of attention.
In the context of the total number of presentations to A and E in Scotland in a given year, which is 1.6 million, the number that we need to achieve to reach the 95 per cent target is not significant.
I think that I am at the limit of my ability to do mental arithmetic in front of the committee, so we will give you the figures separately, if we may do so.
I can only agree with you that 93.5 per cent is by no means bad performance. The rate is higher than it has historically been and it is higher than the rate in some other parts of the United Kingdom. However, for each of the 8,300 patients who waited longer than four hours, performance was not ideal in terms of their experience and, potentially, the quality and clinical effectiveness of their care. I think that we are all interested in pushing performance up to the Government’s target and on to the standard in due course. We recognise that there are real challenges in doing that.
I noted in the report that discharges from hospital tend not to occur over the weekend but that presentations to A and E tend to go up on Mondays and Tuesdays, which surprised me. It seems obvious to me that there is an opportunity to make a significant improvement by doing something fairly simple, which is to manage the discharge process earlier in the week so that beds become available when people are expected to present to A and E. Is that something that you made clear in the report? I did not see that among your recommendations.
10:30
It is one of our recommendations. We focus on it in relation to bed availability, where we look at not just days of the week but time of day. Patients who are discharged often leave hospital in the afternoon, which means that if I rock up at the A and E department at 9 o’clock in the morning and the decision is taken to admit me, the bed may not be available until 2 o’clock, which is more than four hours later.
Places such as Ninewells and Perth royal infirmary are very good at monitoring the time of day at which patients are discharged. They are better at discharge planning and managing the admissions process, all of which means that the system comes together in a way that lets those hospitals consistently achieve strong performance. Other boards and A and E departments still struggle to do that. It is simple but it is not easy.
There is an opportunity there really to make an improvement.
My last point is about the median waiting time for patients who present to A and E in Scotland. The median is the most frequently occurring time; it is not the average. I acknowledge that the median waiting time has gone up but, according to your report, it is two hours and six minutes, which is well within the target time of four hours. We have to put that into context. Would you acknowledge that the median waiting time is still within the target time?
Absolutely. One purpose of this report—and all our reports—is to be as transparent and objective as we can about performance data. It is very clear that the median is two hours and six minutes—well below the four hours. It is also true that it has crept up. Our concern is that that is a sign of growing pressure in the system, which we can clearly see in some of the A and E departments.
I go back to Ken Macintosh’s point, when he referred to paragraph 32 of the report, which states:
“11 per cent of all admissions to hospital from A&E departments happened within the last ten minutes of the four-hour period”.
I think that you said that, in this particular exercise, you did not ask patients directly why that was the case. Can you shed any light on the underlying reasons why that figure is as high as you found in your report?
Our hypothesis is that it is boards doing what they have been asked to do, which is to ensure that as many patients as possible have been seen and either admitted or discharged within four hours. As Mr Coffey said, the median time is two hours and six minutes, so we know that, for most patients, the process is starting well. Many can be discharged very quickly once they have been either treated or referred to a more appropriate place.
For those who cannot be treated or referred, we know that there can be delays in finding a bed, getting a clinical assessment carried out and identifying a specialist to carry out an assessment, when that is needed. Our experience from earlier work is that A and E departments will work very hard to ensure that that is happening within four hours. However, if the system is under pressure, it may well be happening quite close to four hours and there will be an entirely understandable focus on the patients who are approaching four hours, to get them seen, treated and admitted as quickly as possible.
I think that Tricia Meldrum would like to add to that answer.
We talked about Ninewells doing more to identify patients who need a bed and to identify beds for those patients early in the process. Exhibit 14 shows admissions in the last 10 minutes of the four-hour period, and Ninewells is very low down in the table. Ninewells has put in place a number of measures that mean that it is able to admit patients a bit earlier.
I would just say in passing, from personal experience and that of someone in my family, that four hours is a really long time to be hanging around waiting, waiting, waiting.
Presumably if the target was three hours, there would also be a rush in the last 10 minutes. In some ways, it does not matter what the target is. The system is such that, once a target is set, there will be enormous pressure on health professionals to move people out of the waiting room and into the system so that they can tick the box. Is that the case?
That is right. It is clear that targets can be a good device for focusing any service provider on what is being measured and prioritised. As you say, four hours is a long time. It is not an unreasonable target or standard to set but, if we set it, there will be a real focus on getting as many of the patients who are approaching four hours as possible discharged or, in this case, admitted before they hit the four-hour deadline.
That is fair.
On Willie Coffey’s fair point about the median time, why do we have four hours as the target? Why should we put up with four hours as the time that we should expect people to wait and still be within a notional target set by the Government of the day or the NHS? I forget the figure that Willie Coffey gave—let us say that it was two hours and 16 minutes—but, if the median is two hours, why do we not get the target down to two and a half hours?
There are a number of strands to the answer. One is that that is a question for the Government, which sets the target, rather than for us. Equally, any average tends to conceal a wide range of performance. The fact that the median is two hours and six minutes does not mean that we can make the target two and a half hours and have a chance of hitting it. Also, the College of Emergency Medicine has cited evidence that says that, after four hours, there is a risk of quality and clinical effectiveness being compromised. Therefore, as we understand it, there is some basis in the clinical evidence for the four-hour target, even though it can feel like a very long time to wait to be treated or discharged in an A and E department on a busy Saturday night.
Okay. Is it Audit Scotland’s desire to do more work to understand from patients and health professionals why 11 per cent of all admissions are made in the last 10 minutes?
Yes. We are carrying out a wider piece of work on unscheduled care overall, not just A and E. Catherine Young or Tricia Meldrum might want to tell you a bit about our thinking on involving patients in that.
As the Auditor General said, we plan to do some work on broader unscheduled care—not just emergency departments but overall emergency and urgent care. We will also consider primary care, GP out-of-hours services, NHS 24 and ambulance services. We definitely want to do some work with patients.
I am really pleased that you are doing that, but is the NHS doing it too? It should not just be Audit Scotland that has to do such work. Health board chairman, the local boards and clinical professionals, who are paid a lot of money at the top of those organisations, should drive at that kind of stuff, should they not?
Each of the 14 territorial health boards has produced and submitted to Government a local unscheduled care action plan, which we reviewed as part of the work for the report. Most of them are considering the wider system of care and patients’ experiences, but it is too soon for us to see the impact of that and, as we say, variation is a big part of the issue.
It would be fair to say, as you do in paragraph 32, that some hospitals are really good on the measure. The Royal Aberdeen children’s hospital is excellent on it, whereas the Royal infirmary of Edinburgh, which seems to pop up in nearly every report that you produce for us on the NHS, is not good. Presumably, the focus should be on the hospitals that are not delivering.
That is very much the point that we are making: 14 out of the 31 A and E departments are hitting the target and 17 are not. Our evidence suggests that a solution tailored to the particular factors in each A and E department is needed.
You mentioned in your opening remarks that, despite an increase in the number of consultants who are available across the NHS—which is obviously a good thing and welcome—the pressures have increased. Why is that? We are taking on more professionals at consultant level, presumably to target particular problems, but the figures that you have presented to us today are worse. Why would that be?
Exhibit 21 on page 34 tries to summarise the situation.
You are right: there has been something like a 63 per cent increase in the number of whole-time-equivalent consultants in post, which is a good thing because the senior decision making seems to make a real difference. There has been a reduction in the number of doctors in training working in A and E partly because of changes to the way that doctors are trained in general, which aims to give them more of a generalist and less of a specialist training so that they are better able to meet the needs of an ageing population.
Some A and E departments clearly have difficulties in filling vacancies. The posts can be very pressured and there are all sorts of difficulties in recruiting and retaining staff. However, the other aspect is making sure that staff are there at the right time—that staffing is not just Monday to Friday, 9 to 5, but that specialists are available in the evening, at weekends and at peak times. Again, we are going to sound like cheerleaders for Tayside, but Ninewells does that well. The hospital ensures that it has specialist staffing till midnight and some cover in the quiet period after midnight, and it really matches staffing to when it knows that patients arrive.
It is not only about the numbers; it is about how people are used as well.
That is fair.
Finally, I want to understand the point that you made in answer to the convener’s question about targets. The 95 per cent target is an interim target or something—I forget exactly how you described it—and then there is a standard target of 98 per cent. If we are not hitting 95 per cent, the 98 per cent is neither here nor there, is it? That takes me to the point that targets are irrelevant to the experience of real patients and real people.
The target of having 98 per cent of people seen within four hours has been the Government standard for a while, and it is what the Government is aiming for. When it became clear during 2012-13 that there were real pressures and that performance was deteriorating, the Government introduced the 95 per cent target, with the aim that it would be hit by September of this year, as an interim step on the way to 98 per cent. As we say in the report, some A and E departments are already hitting that and some will struggle to do so. That is why we are focusing on the system as a whole.
Will we know in September whether the 95 per cent target has been hit across the health service?
The team will keep me straight about when the data will be available.
It will be by the year ending September 2014, so we will know by then.
In response to Tavish Scott, you pointed out that there is a tendency for a last-minute surge of admissions near the end of the four-hour target period, but the point of the report, which I thought you were highlighting, is that that surge has dramatically increased. There was a little peak before, but the figure has gone up from 45,000 patients admitted in the last 10 minutes to 70,000 patients. That is a huge surge that cannot simply be caused by the four-hour target, because the target was always there. It is a huge increase at the end of the period.
The point that we make throughout the report is that, in spite of great efforts by people across the NHS and especially those working in A and E departments, it is tough to meet the four-hour target or standard that has been set.
To a great extent, that is for reasons that are outside the control of A and E departments. That is why we think that it is important to really understand the problems in each A and E department and to tailor the solutions around the availability of beds and specialists and signposting people to alternatives. The pressure is there, even though a great deal of effort is going into ensuring that as many patients as possible are treated within four hours.
I thank the Auditor General for her helpful report, which has helped us to begin to have a much clearer understanding of what is going on, although, obviously, there is much work still to be done.
On the issues that Tavish Scott and Ken Macintosh raised, I think that Catherine Young began to explain to us that although people might find a bed in the last 10 minutes of the four hours, that is not the beginning of their journey, because their A and E experience starts a lot earlier. Their diagnosis and the process of finding out what is wrong, discovering what the issue is and coming to a conclusion is on-going throughout that four hours, although the bed might be found at the end of the period. Will you expand a bit on that, please?
As the Auditor General mentioned, since our previous report on the issue, there has been evidence that A and E departments are seeing more serious cases. In 2012-13, 50 per cent of patients were categorised as flow 1—as one of the exhibits shows, that is minor injury and illness—compared to 55 per cent in 2008-09. That indicates that more serious cases are being seen. That is combined with the fact that more patients are being admitted from A and E into hospital, which also links to the idea of complexity of care. In looking at the breakdown by 10-minute intervals, we need to consider that more sick patients will need more tests and more blood work, which could be a factor in the longer wait for some of those patients.
On those who are more sick, I was intrigued by a comparison of exhibits 9 and 5 in the performance update report. Exhibit 9 shows attendances at A and E and performance against the four-hour standard, while column 2 in exhibit 5 shows the proportion of referrals from 999 services. Apart from at Ninewells, which seems to stick out in all areas in terms of good practice, there seems to be quite a strong correlation between a high incidence of referrals from 999 services and departments that perform less well on the A and E standard. I think that that is the point that Catherine Young made.
The highest figure is shown against Z, or the Royal infirmary of Edinburgh, which, as we see from exhibit 5, gets 28.1 per cent of its referrals from 999 services. A is University hospital Ayr, where 18.4 per cent of people who are referred come from 999 services. G is Forth Valley, which is also in the bottom quartile and gets 29 per cent of referrals from 999 services. I could go on. There is a whole series of examples at the bottom end of the range—apart from Ninewells and the Western infirmary, for some reason. There seems to be a correlation between referrals from 999 services and A and E departments that perform less well against the target. What are your reflections on that?
10:45
With reference to exhibit 5, on referral sources, we highlight in endnote 10 that there are some inconsistencies in how A and E departments record self-referrals. Some record self-referrals who come in by ambulance as self-referrals, whereas others record them as 999s. There was actually no strong correlation between 999 calls and performance.
It is likely that, if departments recorded more consistently, there would be a link. In our previous A and E report, we found that patients who were referred by GPs or through 999 calls were more likely to stay longer. They were the more complex cases, and dealing with them would have an impact on performance. We found a link between more complex cases and performance.
There are some issues around consistency and those codes. It might look as if some hospitals have higher figures than others, but that is because they record 999s as self-referrals.
It seems to me that a bit more work needs to be done in that area to examine what is going on. Other parts of your report reflect that: they describe the complexity and the interrelated nature of different data. The data is collected differently in some places, and there are a variety of practices.
For instance, paragraph 22 on page 20 says:
“it is difficult to draw clear conclusions about the relative performance of A&E departments because the services provided vary across the country.”
Paragraph 25 on page 21 says:
“the methodology A&E departments use to define patient flows differs”.
Another example is at paragraph 31 on page 26:
“Our previous audit highlighted that opening hours and levels of staffing vary across the country. Up-to-date national information about how hospitals use these units and how they operate is limited.”
There is a sense of that throughout the performance update report.
I raise the point because it has become clear to me that, to get a full understanding of where we need to make the improvements that will be required in future years, we need a much more serious, in-depth analysis or investigation—although I am not suggesting that the report before us is not a serious analysis—about where we can make improvements in a system in which interrelationships exist. If we address one factor in isolation, we might end up disturbing another bit of the system and making it worse. I know that you will do further work on that, Auditor General.
Do you think that it would be appropriate and useful if, as Tavish Scott suggested, as well as having the health boards themselves and the Government consider the matter, a parliamentary committee undertook a fuller investigation of A and E in the context of the whole exercise? That would allow a good examination to be carried out.
It is certainly true that, with the report before you, we have done an update of a fuller report that we carried out in 2010. We focused on the data that is available nationally about A and E. We have used that data as fully and rigorously as we could. Lucy Jones, who is not at the table, has correlated every possible set of factors to see what might be interesting patterns and explanations for them. We have pulled out that information as far as we can. We have also tried to highlight where we think that there are consistencies or where the data to draw conclusions is simply not available.
The intention behind the work that the Government has asked each of the 14 health boards to do as part of their local action plans is to address that, and we will consider that work as part of our next, wider piece of work on unscheduled care as a whole. There may well be aspects that a parliamentary committee would wish to explore. The trick is to use the data to ask the questions, which we have done as well as can be done, and then to explore what the answers to the questions are locally, and what that means with respect to solutions.
That was very helpful.
Your comprehensive report shows an improvement in waiting times over the past 12 months. If performance continues to follow that trend, do you expect the number of people who wait more than four hours to fall in the coming years?
The conclusion that we have reported is that some health boards will find it difficult to meet the four-hour target by September this year. Given that challenge, whether the NHS as a whole will meet the 95 per cent target is not a call that I would like to make, but I think that it is very unlikely that all 31 A and E departments will hit the four-hour target by September. We all hope that they will do, and a lot of good work is going on, but a number of indicators behind the four-hour figure suggest that there is real pressure in many A and E departments.
We have talked about bed numbers. Am I right to say that in 2012-13 there was an increase in bed numbers of about 183?
I do not have the figures available to confirm that just now. We are clear that the reduction was an appropriate reduction that reflected a move to day surgery, but there are signs of pressure around acute medicine beds, where the occupancy level across Scotland is around 85 per cent. That is another area where a better understanding of what is happening would be very helpful, both nationally and, more importantly, in each of the 14 health boards.
Could you look at that and let us know?
Sure.
I will ask a general question that arises from paragraph 3. Paragraph 3 refers to your “Emergency departments” report of August 2010, which included the Scottish Ambulance Service and NHS 24. You made some clear recommendations that set out
“a clearer strategic direction for emergency care services”.
However, you now say:
“Since then ... performance against the standard deteriorated”.
Were your recommendations, guidance and strategic direction taken on board, and did the performance against the standard deteriorate as a result, or were your recommendations ignored?
Some of them were accepted and implemented, and some of them were accepted and implemented a little bit later. I ask Catherine Young to highlight what happened.
In part 2 of the report, we comment on progress on the recommendations that we made in our 2010 report. Overall, through the work of the national unscheduled care action plan, a lot of recommendations are now being progressed.
As the Auditor General mentioned, the issue is quite complex. We have highlighted again in the latest report some of the quick-win solutions, and we are seeing some evidence of implementation of the longer-term, strategic recommendations in the new unscheduled care action plan. Recommendations on staffing, benchmarking information, use of assessment units and so on are now being progressed and there is a lot more evidence of outcomes.
I presume that if all the recommendations had been adhered to, we would have seen not a deterioration but greater progress.
In your comments at the beginning of the meeting, Auditor General, you said that more over-65s were presenting to A and E. We know that we have an ageing population, and you have mentioned a couple of times that over-65s have more complex needs.
My question is really a supplementary to those from Ken Macintosh and Tavish Scott. More than 18 per cent of people at Hairmyres and the Royal infirmary of Edinburgh are being seen in the last 10 minutes of the four-hour period. Almost 20 per cent are being seen in the last 10 minutes, and if that happens they are more likely to be admitted to hospital, and to stay longer.
From the evidence that you have given today, can we conclude that patients with less complex needs are being seen quicker because the target, rather than clinical need, is being prioritised? Are the targets distorting what is happening? It seems that there is almost sufficient evidence in what you have said today to suggest that that is the case.
We did not find evidence of that, and we tested for it specifically in relation to the patients who were admitted in the last 10 minutes of the four-hour period. We started with the assumption that, if those patients were being admitted to avoid breaching the four-hour limit, they would probably be discharged more quickly than other patients. In fact, we found that they were staying in hospital for longer than other patients.
Clearly, accident and emergency departments need to manage the flow of patients who arrive at their front door to ensure that the most seriously ill and injured patients get priority and receive the range of assessment and treatment that they need. We found no evidence that that is not happening. What we found was an example in Tayside of the reception process for a new patient being very clear about whether someone has a relatively minor condition, which might mean that they could be treated and discharged without too much complexity of care, or whether they have a condition that is likely to need more complex assessment and, potentially, admission. That made a real difference to overall performance and also to the quality of care that those groups of patients got. The hospital described that as “streaming”.
From the data, we cannot tell whether that is happening consistently across Scotland. However, we did not find evidence of gaming the target; we found evidence only of the fact that the target inevitably has an impact on where the attention of the managers and clinicians tends to focus as the four-hour limit nears.
There is no doubt that Tayside is a model of best practice. However, to be fair, that best practice is not replicated in the 30 other A and E units.
I turn to exhibit 5, which Bruce Crawford talked about. I have a difficulty with the figures. This time round, you examined accident and emergency figures in A and E units. However, I believe that, the last time that you considered this issue—before I was on the committee—you included the figures for NHS 24, the Ambulance Service and so on. The fact that we are now looking at only one part of quite a large model causes me some difficulty. I think that, in your own words, the accident and emergency service needs to be seen as an overall part of the health and social care system.
Exhibit 5 was more than interesting to me. For example, it shows that 82.5 per cent of patients at the Belford hospital in Fort William are self-referred and that no patients are referred from 999 services or GPs. When I dug slightly deeper into that, I discovered that 19 of the 31 units had zero referrals from GPs. Are people just bypassing GPs? Perhaps I am reading too much into this, but I have to say that I was shocked by the figures.
I was delighted to see that there were extremely few referrals from minor injuries units. That tells me that they are doing an excellent job and are freeing up resources in A and E. However, NHS 24 was responsible for 0.7 per cent of referrals to the Southern General hospital but 8.6 per cent of the referrals to St John’s hospital at Howden. Bruce Crawford has mentioned the figures for the 999 services, so I will not go into them. However, the disparity between different areas of the country makes it look almost as if different healthcare models are being used. As I said, the Belford hospital has a figure of 82.5 per cent for self-referrals, but the Royal infirmary in Edinburgh has a figure of 46 per cent.
We are getting only one part of the picture here, but if there is any further drilling down to be done, it would seem to be needed in this area.
11:00
I will highlight two points, and Catherine Young might want to expand on them.
First, as you say, the model of care varies a lot across Scotland. Partly, that is entirely to be expected, as conditions in the islands and remote parts of the Highlands are different from those in the major cities in the central belt, and what good care looks like is likely to vary as well.
In part 3 of the report, we comment that there is still room for guidance from the Government about different models of care and the way that they work. The relationship between minor injury units, assessment units, admissions units and so on is still variable across Scotland, and that will have an effect that cannot be understood through the national data.
Secondly, as Catherine Young said in response to Mr Crawford’s question, the data in exhibit 5 is recorded inconsistently, especially in relation to 999 arrivals. It is hard to envisage any accident and emergency department not having at least some 999 referrals. We understand that, for some hospitals, if the patient or family dials for the ambulance themselves, that is called self-referral, and if the ambulance is called by a GP, that is a GP referral rather than a 999 referral. There are some inconsistencies in the data that need to be better understood.
As an MSP for the Highlands and Islands, I am always quick to look at the data for the islands and remote areas. I used the example of the Belford hospital in Fort William, which is close to our highest mountain and has a very busy accident and emergency department. What I should have said was that Hairmyres and the Southern general are at 80 per cent and 79.8 per cent self-referrals, respectively. I also want to point out the difference between Scotland’s two biggest cities. Edinburgh has 46 per cent self-referrals, and Glasgow has 80 per cent. We cannot take into account any rural or remote factors with those figures. My question is really this: are all parts of accident and emergency working well together?
I am very impressed with the fantastic figures for the minor injury units; they have made 0.2 per cent of the referrals to A and E and are obviously dealing with what needs to be done. Do we need more minor injury units? Are there more problems in accident and emergency departments in areas in which there are fewer minor injury units? I do not know. We need more information about the figures. Are we making best use of the Ambulance Service and paramedics, who do a fabulous job and prevent many people from going to accident and emergency?
I am frustrated today, convener, because we are looking at one part of the service, but there are different patterns in how everything works together throughout Scotland. The shocking figure for me is that 19 accident and emergency departments out of 31 had no referrals from a GP. Will the additional work that you mentioned to Bruce Crawford drill down into those figures? Is there no recommendation for best practice nationally that would improve the figures in the longer term?
Mary Scanlon is absolutely right. The solution to ensuring that everybody who needs to go to accident and emergency is treated quickly and effectively is not about accident and emergency departments; it is about the whole system. The data help us to pose questions and to answer some, but not all, of them. We will take that a bit further in our work; the Government is trying to do that through its unscheduled care action plan.
Earlier, Mary Scanlon asked about the recommendations in the 2010 report. One of the important ones that has not been responded to fully is about providing guidance on the most effective models of care. Exhibit 4 shows that the distribution of activity between accident and emergency departments and minor injury units, for example, varies a great deal across Scotland, and not in ways that can easily be explained by geography, deprivation or anything else. The question is not just about minor injury units; it is also about admissions units, assessment units and the links with GPs, the Ambulance Service, and NHS 24. Getting those models right in each part of Scotland will go a long way towards relieving the pressure that we know exists.
Finally, there also seems to be an issue about culture. I was shocked to see that in Glasgow, self referrals are at 80 per cent, and 0.7 per cent come through NHS 24. However, at St John’s hospital at Howden, referrals from NHS 24 are at 8.6 per cent. Is there a culture of people just turning up at hospitals? Is NHS 24 being underused in some areas, compared with others?
We did not look specifically at NHS 24, although in the 2010 report we looked at referrals from NHS 24 and found that they were mostly appropriate because the patients were quite sick and ended up being admitted to hospital.
I would like to pick up on the point that was made about GP referrals and GP referrals for admission. Our report highlights huge inconsistencies in those two codes. GP referral for admission is a new code to the datamart over the past 18 months or so, but that is why we made a recommendation about the Scottish Government sharing good practice on that process and on the way in which patients are referred to A and E. For example, some patients bypass the A and E department and go directly to a ward or to an especially acute receiving unit, and some go via A and E.
In order to understand the impact that the current models have on performance, ISD Scotland is having discussions with boards about completing those codes correctly. We know that NHS Lothian records high GP referrals for admission. The split between GP referrals for admission and GP referrals was not quite right, and we discovered in our fact check with Lothian NHS Board that there needs to be a better split between those two codes. There is on-going work with boards on the new code for GP referrals for admission, and we expect to see that coming through the datamart over the next few months.
Mary Scanlon raised a point about the Southern general, which emphasised the complexity of the issue and the dependence on other services that are available in the city. Because the Royal hospital for sick children exists in Glasgow, more NHS 24 referrals for children may be made there than to other hospitals. That emphasises the complexity and interrelationships even more.
That is exactly right. There is a need to understand the whole system and to ensure that we know which models of care work well. It is partly about what services are available and partly about how well the health boards signpost people towards them, so that they know that they exist and know what is appropriate. It is also partly about developing those services further so that, as Catherine Young said, if it is more appropriate for GPs to refer patients directly to a ward, rather than their going through A and E, there is a route for them to do that. It can all make a difference.
Thank you for your report. I found it interesting, given comments from sources including Jason Long, who is the chair of the College of Emergency Medicine and who has been quite supportive, and Ian Ritchie, who is the president of the Royal College of Surgeons, who was pleased to see the on-going work that is being done. It seems that we are moving in the right direction, after coming through a period when we were particularly affected by norovirus and stuff like that, to judge by your report and by what others are saying.
A report on beds was produced in March by ISD, in relation to the quarter ending December 2013. It states:
“The number of available staffed beds in acute specialties was recorded as 16,223 in the quarter ending December 2013. This is an increase of 1.1% from 16,041 beds in December 2012.”
I do not think that you have taken that into consideration in your report, have you?
What we quote in paragraph 40 is the shift between our baseline from our last report in 2008-09 and the latest available figures when we were preparing the report, which were for 2012-13. The figures are consistent with yours. We have 16,230 as the number of beds for March 2013. It sounds as though your figures have gone down slightly from that, from 16,230 to 16,223, but they are very close.
The publication date for the report was 25 March.
Your figure was 16,223. Is that right?
That is what the figure was in the quarter ending December 2013.
That is seven beds lower than the figure that we have for March 2013, but it is very close.
The figure represents an increase of 1.1 per cent. I know that the Royal infirmary of Edinburgh has opened a new ward, so there are obviously pressures there as a result of the building having been built too small; half the problems come from that itself. I am thinking about what the professionals are saying outside, and about where we have come from, and it seems that we are heading in the right direction. I associate myself with Bruce Crawford’s comments about an in-depth look at things by the Health and Sport Committee being appropriate as we head forward.
In terms of the overall direction of travel, we have been very clear in the report and in all our comments about it that performance had deteriorated slightly since our previous report in 2010, but improved during 2013. We think that that is a result of the Scottish Government’s national unscheduled care action plan and the action that is being taken. We are not raising a specific concern about bed numbers—the data do not support that—but the bed occupancy rates in acute medicine are at a level that is starting to cause clinicians concern. There is a shared agreement that going a step further to understand fully the interplay among the different factors—at national level and, more important, locally—is the key to helping A and E departments to manage the pressures that they face, which are real and include demographic change, overall financial pressures and all the other pressures with which we are familiar.
Okay. Thank you.
A couple of my questions have been asked. I would like to return to the issue of patients being admitted in the last 10 minutes of the four-hour period. Is there evidence to show that, as Catherine Young said, the decisions to admit those people are not being left to the last 10 minutes? Mary Scanlon suggested that the most complex cases are being left to the end. However, is it possible that the more complex cases are being looked at earlier, as you suggested in your response, but that because of the complexities of those people’s situations, it is taking that length of time to find appropriate beds for them?
The short answer is yes. It is no surprise that there is a peak of activity just before the four-hour deadline—that is, understandably, what targets produce. We are all human, and the staff will do their best to hit the target. We found no evidence that that was being done inappropriately. However, through case studies such as the Tayside study, we have found evidence that if the planning is started as early as possible within the four hours, that peak can be smoothed down.
As Catherine Young said, at Ninewells hospital, only about 4 per cent of admissions from A and E occur in the last 10 minutes. We think that that is because the staff are identifying very early which patients are likely to be admitted, and are able to start the process of finding beds for them. They are not taking three hours to decide that the patient needs to be admitted and then rushing for the last hour to find a bed; it is happening right through the four hours, which is better for everyone involved.
I think that there is a general consensus that other hospitals should follow that best practice. Thank you.
I have a final question for the Auditor General. You and others have mentioned the number of people who are presenting at accident and emergency units. In the wider context, a lot of excellent work is being done as staff cope with that level of demand. Did you examine whether people are going to accident and emergency departments predominantly through self-referral, rather than using GP out-of-hours services?
We explored that as far as was possible through the data. As I said, this is an update, rather than a full audit such as we did in 2010 and will do again next year. As Catherine Young said, there is evidence that the people who attend A and E departments are getting sicker, if I can put it in crude terms. They are in the higher-flow categories, and fewer people with minor illnesses or injuries are attending. We also know that more older people are attending. They tend to be sicker and have more complex needs and so are more likely to be admitted. Both those developments are adding to the pressure.
Overall, attendances at A and E departments, as opposed to minor injuries units, have fallen slightly, and attendances at minor injuries units have gone up markedly since our previous report. That suggests that there is a move in the right direction. However, you can see from various exhibits throughout the report that the situation is not consistent throughout the country.
Do you have statistics that show the demand on GP out-of-hours services and what the trend has been?
We have not used them in the report. Catherine Young may know what information is available.
We looked at that issue in the previous A and E report. In fact, we carried out a survey in which we asked patients why they chose to attend an A and E department. Overall, we found that they felt that that had been the most appropriate place to attend. It is quite difficult to get behind the reasons for attendance.
11:15If there was a big increase in attendance in a particular A and E department, we would expect the board to try to get behind the reasons for that, as part of its local unscheduled care action plan. We talked about signposting people away from A and E to more appropriate services; boards should be looking at, for example, capacity in GP in-hours and out-of-hours services in order to ensure that there is somewhere for patients to go.
Thank you.
For the benefit of the committee, I have been trying to work out what the shortfall is between the current rate of 93.5 per cent and the 95 per cent target: it is just over 24,000. I hope and expect that it is not beyond us to meet the 95 per cent target, given the discussions that we have had.
We look forward to the next report. I thank everyone for their contributions to the discussion.
Next
Section 22 Report