Few can avoid the escalating hysteria of media headlines declaring that the NHS is yet again in the grip of another winter crisis, seemingly due to the sheer increase in volume of attendances at major Accident and Emergency (A&E) units. However, the facts of the matter are, that there is nothing unprecedented or unexpected about the level of A&E attendances, so what is behind the sharp fall in performance?

While some commentators are acknowledging that the problem might be a little more complex than previously thought, they continue to recycle many of the old explanations, which largely over-look the hard evidence. Perhaps the fact that this crisis generally recurs every year is a sign that we don’t understand the problem and therefore constantly recycle solutions that are either incorrect or poorly implemented.

Current situation in A&E

The main media explanation for the winter crisis, the sheer volume growth, is usually accompanied by a range of suggested root causes: GPs being overwhelmed; young people going straight to A&E instead of their GP; NHS 111 referring too many people to A&E; and increasing numbers of older people with complex health needs, than there used to be. Most commentators subsequently suggest that a key culprit is the lack of out-of-hour’s access to GPs as a result of the change in the GP contract in 2004, which allowed GPs to opt out of providing out-of-hours care.

However, expert analyses by the King’s Fund, the Nuffield Trust, the National Audit Office and others clearly show that the increase in A&E attendance coincided with the opening of many new types of provider units (Type 2 and 3 A&E services such as walk-in centres and minor injury units), which have been included in A&E figures since April 2003. Indeed, between 2003-04 and 2013-14, attendance at Type 3 units increased by 3.7 million (or 114 per cent) while more traditional Type 1 A&E units increased by only 1.6 million or 12 per cent (largely in line with the increases in the size of the population). [i]

This time round, commentators have allowed for the possibility that the problem might be more nuanced. However, many of the old explanations are still receiving more attention than they deserve, causing knee jerk responses that history shows just doesn’t work. An in-depth evaluation of the data provided by NHS organisations can help us understand the facts on the ground and identify the changes that might help improve A&E performance; rather than those that sound plausible to headline writers and policy gurus and which risk diverting resources into actions for which there is no evidence of impact.

Increase in A&E waiting times

The official statistics show that, in the last three months of 2014, the NHS in England failed to meet its A&E waiting time target of 95 per cent of people seen and treated within four hours, with just 92.5 per cent of people being seen within this four-hour deadline. Indeed, more than 90,000 A&E patients waited more than 4 hours and there were 1.4 million emergency admissions. Until now the highest number of emergency admissions was recorded in April to June 2014, when 1.37 million people were admitted as emergencies.[ii]

The first fact of note is that most of the identified problems are occurring in major A&E units (Type 1 A&E) rather than the specialists units or units that deal only with minor injuries and do not open 24 hours a day (Figure 1).

Figure 1. Weekly attendance numbers at different types of A&E Units January 2011-2015


Source: Data source -NHS England Accident and Emergency (A&E) Attendance data[iii] with screen shot generated by Deloitte interactive data analysis tool.

The top chart shows the weekly attendance volume and the bottom half the weekly performance (that is how many patients finish their stay within four hours) of each type of department. This demonstrates that it is only the major A&Es where many patients are now waiting longer than four hours. Too many commentators quote the overall performance of the system while missing the point that the problems are entirely inside the major A&E Type 1 units.

Increase in A&E attendances

One of the biggest and most distracting myths surrounding the crisis is the suggestion that the problem is caused by too many people turning up. Weekly performance data has been published since November 2010. For the week ending 4th January 2015, which a newspaper article described as “the worst performance ever”, three of the identified causes could be summarised as “too many people turning up”. [iv] While it sounds plausible that if too many people turn up, performance will be poor, a closer look at the figures shows there isn’t such a simple correlation.

In reality the numbers attending A&E that week are actually below the average figure for that time of year. Indeed, more granular analysis, using Deloitte’s analysis tool (used to generate the ‘screen shots’ in this article), shows that there is no meaningful statistical relationship between weekly national attendance and national performance. Indeed, the national weekly volume growth has been steady over the last four years (indeed the national long term growth in attendance volume is steady over more than a decade at one-two per cent a year, in line with population growth). While it is a fact that this winter is busy for a winter, it is still quieter than the last few summers when A&E performance in relation to the four hour target was better.

That volume isn’t the problem is even clearer if you delve into the detail of the numbers for individual trusts. Indeed, about 30 per cent of trusts have actually seen volumes decline over the last four years while others have seen much larger increases than the national average.[v] However, our analysis shows there doesn’t appear to be any systematic relationship between these changes in attendance and local performance. This calls into question the time and energy being devoted to diverting people away from A&E when it seems highly unlikely that this will make a measurable or systematic difference to A&E performance.

Increase in emergency admissions

Meanwhile, our analysis of the weekly data reveals what looks like a much better explanation of why performance is so poor - the increase in the number of emergency admissions. Unlike attendance volume there is a notable relationship between the number of emergency admissions and performance. Moreover, emergency admissions are close to record levels and have risen sharply in the last year. There have been around 4,000 more admissions a week compared to 2013-14 or emergency admissions were at 1,013,307 in quarter three, a 5.6 per cent increase on same quarter last year. [vi] 

Once again, however, the story is not as simple as it might appear. While it is reasonable to think that the reason more admissions causes slower treatment is that sicker people need more time in A&E, the reality is more complex. Sometimes they do, but the real reason admissions cause delays is because finding a free beds is a serious challenge for many hospitals, who are increasingly operating at record levels of bed occupancy. A great deal can be gleaned by looking at the pattern of waits for different groups of patients and at different times of day. Patients requiring admission to a bed constitute between 15 and 35 per cent of all attendances but they often suffer long waits. This isn’t usually because they need a lot of treatment in A&E (some do, but most don’t), usually it is because the A&E department can’t find a bed to admit them to.

An analogy to help explain why - imagine you arrive at a hotel where you have booked a room but the hotel doesn’t know which rooms are free so has to send a squad of staff to knock on each door to find free rooms. This will clearly take time but is remarkably similar to how most hospitals operate with few hospitals having access to real time bed management information. While readers are likely to be thinking that the analogy is flawed (hotel bookings systems unlike hospitals are simple and reliable to operate and people turn up according to a schedule and the hotel knows when people will arrive and depart); this is true for many hospital admissions and departures—elective patients—who are booked just like a hotel stay and hospitals know reliably when they will arrive and depart. Moreover, emergency patients also arrive in very predictable patterns (the volume peaking in the morning and remaining much the same from day to day), and even the reasons for attending conform to a discernible pattern.

Two things can make a difference, one is how well the A&E department is organised; the other is how well the whole hospital is organised, including how effectively it co-ordinates and manages its discharge planning. While the above could be perceived as a capacity problem, again, it isn’t. The problem is that many hospitals do their discharges in the afternoon despite demand for emergency admissions peaking in the mornings. As a result, many beds are filled with patients who are fit to leave but haven’t been discharged yet while others, who really need the beds, wait in the A&E queue. Patients waiting for a bed sometimes make up more than half the queue in A&E though they only constitute around 15-35 per cent of the attendance. This suggests it is more a coordination problem not a capacity problem.

Managing the flow and stock of beds is therefore a whole hospital problem not an A&E problem. So if availability of beds is the cause and A&E performance suffers, the focus is likely to be on the wrong people. This also implies that if the NHS fails to address the way hospitals manage their beds there is little hope that the A&E problem will be fixed. There is independent evidence about this which too many parts of the system appear to ignore. Audits of the clinical notes of patients’ currently occupying beds, to determine whether their medical state requires them to stay in a bed, consistently show 30-40 per cent of people have no medical need to be in a bed. Some may be fit to leave, but need follow up care at home and some may need to return to or be in a care home, which may well cause delays as the necessary arrangements are agreed and put in place.

As a result it has become fashionable to blame cuts in social care for many of the “bed-blockers” but again bed audits often tell a different story:

  • one trust who did such an audit at the same time as a catastrophic decline in A&E performance found that while a fair numbers of patients who were fit to be discharged were delayed by problems with social care; a much larger number were delayed entirely due to the hospital (for example failure to coordinate discharge rounds, delays providing medicines and basic problems with paperwork)
  • the NHS‘s own figures on delayed discharge show that in November 2014, 66.5 per cent of delays were attributable to the NHS (the largest proportion due to patients awaiting non-acute NHS care) and 26.5 per cent were attributed to social care (the most common being patient’s awaiting a care package in own home)[vii] 
  • analysis by the Association of Adult Social Care suggest that over the last four years, while the number of delays attributable to social care has gone up, the proportion of delays has decreased from 33 percent to 26 percent.[viii] 

The above analyses mean that fixing the social care system, which all experts agree is needed, would still not fix a majority of A&E delays in hospitals. Though the weekly performance data gives us hints about the problems with beds, it also tells us that the problem is far more complex than a simplistic view would imply. Digging into the data and patterns in individual trusts provides further insight. Such analyses show that not every trust with rapidly rising admissions has had problems with performance. For example, Figure 2 generated by the Deloitte analysis tool illustrates how one home counties district general hospital (A) has maintained an exemplary performance record in 2014 despite rising attendance and 50 per cent more admissions than in 2010 (a very high rise by national standards); and Figure 3 shows how a similar size hospital (B) serving a similar population, but with slower attendance growth and similar growth in admissions has seen their performance deteriorate sharply.

Figure 2: Weekly A&E performance of hospital trust A January 2011-2015


Figure 3: Weekly A&E performance of hospital trust B January 2011 to 2015


Indeed, a number of trusts appear to have maintained or performed better at higher admission levels. Again this contradicts any simple model of what the problem is and undermines the chance that reducing attendances or admissions is the only answer. Many discussions assume that the national pattern is replicated in most hospitals. So, for example, admissions are rising everywhere. In fact, if you spend time looking through the trust-level data you will see every possible pattern in the relationship between admissions and performance which strongly argues against a simple national cause or solution.

Understanding reasons behind higher admissions

The assumption is that higher admission levels are an inevitable consequence of increasing numbers of older people with more complex health needs and that there is little that can be done to tackle this as they will often need admission. Again, our analysis suggests that while the population is aging it isn’t aging fast enough to explain the rapid growth in admissions seen in 2014. Notwithstanding the current challenges in primary and social care in providing care and support to increasing numbers of frail older people as demonstrated in our report Better care for frail older people . The thresholds for admission are not very clear cut and there is, in a number of hospitals, a relationship between the proportion of admissions and the busyness of the hospital (some hospitals seem to admit more when busy, perhaps as a way of meeting the target; others admit fewer perhaps because they use stricter criteria to avoid filling up beds; others have consistent thresholds however busy they are (all this is evident in
the hospital level dashboards in Deloitte’s analysis tool).

Another insight, that higher admissions are driven by need (by Quality Watch), based on analysis of detailed patient-level Hospital Episode Statistics data, found that recorded morbidity while increasing, was again nothing like fast enough to explain the scale of increase in admissions. [ix]

Tackling the A&E challenge

One simple lesson: diverting people away from A&E is probably futile and is unlikely to improve the speed of A&E. Projects focussed on this problem are likely to be a waste of time and effort as they are dealing with a problem that doesn’t exist. While many patients could, in principle, be treated elsewhere, those patients can be dealt with very quickly by a well-organised A&E department so they don’t affect performance much unless performance is already poor.

It is also unrealistic or indeed unsafe to divert people from hospital beds, at least not the people who really need to be in a bed. If hospitals are admitting people who don’t need to be admitted (this happens because admission thresholds are fuzzy and are sometimes systematically changed to help meet one target or another) this may only postpone an A&E problem and will create knock on problems elsewhere. If hospitals fail to admit those in genuine need, they may be hurting patients unnecessarily. Again, the pattern of actual behaviour is wide. The proportion of arrivals being admitted varies a lot across hospitals as seen in Deloitte’s analysis tool and Figure 4.

Figure 4: Range of admission rates and numbers of trust with that rate


This suggests that solving the problem of availability of beds is not an A&E problem. It is a whole hospital problem that will only respond to whole hospital solutions. If the behaviour of the clinical community in the hospital is not coordinated with the demand from A&E (which it mostly isn’t) then there will be problems with finding the beds required for A&E patients (notwithstanding that most hospitals are making valiant efforts to try and resolve this challenge). Nevertheless, this last point is critical to attempting to resolve the A&E crisis. The biggest benefit might come from investments that don’t have anything to do with A&E. For example, better coordination of and more capacity in community care. But an even bigger gain should be possible with investment in real time bed management and better coordination of the flow through beds in the hospital, helping hospitals to improve control over the timing of discharges and avoid internal delays in the discharge process.

So what next?

In responding to the current crisis, without access to the evidence and understanding, the NHS risks wasting a great deal of effort on interventions that will do little to improve the current problems in A&E. It is vital that future interventions and investments are based on a detailed analysis of the data rather than being knee-jerk responses to “obvious” problems or the demands of lobbyists. The majority of proposed interventions are not compatible with the evidence from our data analysis nor are the more nuanced interventions based on the more detailed patient-level evidence.

The single most effective intervention—improving the flow through beds via better coordination inside the hospital—cannot be achieved by health economy restructuring or by patient education. More investment in social care would certainly help, but not as much as fixing the internal management processes for patient discharge. Meanwhile schemes to divert patients from A&E will have little, if any, effect. A&E performance should be seen as a symptom of a different problem – with more success coming from treating the disease.

Karen TaylorKaren Taylor
Research Director, Deloitte UK Centre for Health Solutions




Stephen Black0002

Dr Stephen Black
Head of the Deloitte Health Analytics Insight and Knowledge Unit




[i] All A&E performance figures are from the NHS monthly or weekly A&E attendances and admissions data sets. See also:

[ii]   ibid

[iii]   ibid



[vi] ibid

[vii] Statistical press notice. Monthly delayed transfers of care data, England November 2014. See also:

[viii] A&E pressures, social care and the Better Care Fund. David Pearson, President of the Association of Directors of Adult Social Services personal view 15 January 2015. See also:

[ix] Health Foundation and Nuffield Trust Quality Watch Focus on A&E Attendances, July 2014:


  • Hi - emergency admissions have increased as providers manage their more complex A&E attendances in units called
    Emergency Assessment Units
    or Clinical Decision Units
    or Medical Assessment Units etc

    They are for patients who are waiting for tests or specialised inputs. Subsequently patients are discharged or transferred to a hospital ward.
    These units have beds but are not usually part of the provider bed stock. Significantly patients in these units are admitted on the hospital system and so get counted in your analysis. In the past, before providers had these units, patients remained on trolleys in A&E waiting for the same specialist inputs but were not formally admitted. So demand hasn't changed but the number of admissions has gone up.

    A&E 4 hour targets and PBR have incentivised this management practice

    Posted by: Paul Cavendish on 27/07/2016

  • Interesting analysis...but it doesn't consider workforce. If you also looked at the number of A+E consultants in 1997 through to present and measured the indexed rate of change in these you would see that Nationally there has been no increase in numbers of A+E consultants from then to now. So whilst bed utilisation is obviously important, there is still a consideration of more demand, less supply (consultants) and so a point when no more can be done unless supply is increased...It would also be helpful to see the admission rate over time to see the pattern in this, as if admission thresholds have fallen it may be consequential of change in practice that has resulted in higher rates of admission

    Posted by: Contrarian on 27/07/2016

Verify your Comment

Previewing your Comment

This is only a preview. Your comment has not yet been posted.

Your comment could not be posted. Error type:
Your comment has been saved. Comments are moderated and will not appear until approved by the author. Post another comment

The letters and numbers you entered did not match the image. Please try again.

As a final step before posting your comment, enter the letters and numbers you see in the image below. This prevents automated programs from posting comments.

Having trouble reading this image? View an alternate.


Post a comment

Comments are moderated, and will not appear until the author has approved them.