The evenings are drawing in, and NHS minds are once again focusing on the impending winter pressures.

Over the last few years, even the best run hospitals have faced increasing problems achieving the national target to treat people within four hours of arriving at A&E. The number of people attending A&E has continued to grow, as has the proportion admitted as an emergency. But as funding fails to keep pace with demand, the prospect of hitting the four-hour target seems to slip further away every year. Hospitals are now struggling to meet the target all year round, leading the Chief Executive of the NHS Confederation to say ‘the winter crisis is now an all year round crisis’.

The government’s ambition extends beyond just getting through this winter. Its annual mandate to the NHS sets an expectation that overall 95% of people will be seen within four hours by March 2018, in accordance with the national target. In previous years, a winter relief fund was set up to help manage seasonal pressures, before being incorporated into CCG budgets in 2015/16.

Of late, there have been substantial efforts to prepare for this winter. Recent measures taken to help reduce pressures include an increase in people having the flu vaccine, more 111 calls being handled by clinicians and £100m of capital investment.

However the true impact of these measures is not yet known, and there have been calls for additional dedicated funding to help providers improve access to emergency care for patients. But how much would it cost to meet the 95% target this winter? And if the money was available, is it already too late?

We estimate it would cost at least an extra £360m to add sufficient capacity to meet the national four hour standard for A&E this winter. But it may already be too late to wisely invest this in the most effective way.

With the help of three charts, here is how we’ve worked this out:

  1. Firstly, we need to look at how many people are likely to attend A&E this winter.
  2. Secondly, how many of those are likely to wait longer than four hours.
  3. Then, how many of these are severe enough to require an emergency admission to hospital?

In each case, the trends follow a clear seasonal pattern, which we’ve accounted for to produce projections for the likely activity exceeding the four hours target between January and March 2018. (Note: these are projections, not predictions. The true value will be affected by many factors including the weather or severity of any outbreak of flu.)

Finally, we then need to estimate the extra capacity that would be needed in A&E and inpatient departments to get back up to 95%.

Chart 1 – how many people are likely to attend A&E this winter?

Number of attendances at A and E in England, Q3 2010/11 to Q4, 2017/18

Download the chart (PNG)

You will notice that there are actually fewer attendances at A&E in winter than at other times of the year. 5.9m people attended A&E in the latest quarter (July to September), and we actually project around 5.8m people to attend this winter (January to March). This is 600,000 more attendances than in winter 2010/11 (a 12% increase). However, the issue is not the number of people coming through the door. It is that more of the people who do come through the door take longer to treat, and are more likely to need to be admitted, adding pressure on available beds.

Chart 2 – how many attendees are likely to wait longer than four hours?

Number of people waiting more than 4 hours in A&E in England, Q3 2010/11 to Q4 2017/18

Download the chart (PNG)

While the number of A&E attendances is lower in winter, more of them are severe enough to require an emergency admission. (We project around 1.1m admissions this winter, 19% of projected attendances). This means that the number waiting more than four hours is consistently higher. If this trend continues, our projections suggest around 735,000 people will wait longer than four hours in Q4 2017/18. That’s 150,000 more people than between July and September, and 311% higher than in the winter of 2010/11.

These projections together suggest that just 87% of people will be seen and treated within four hours in Q4 2017/18, which would potentially be the worst quarterly performance on record. Meeting the 95% target set in the mandate would require an additional 445,000 people to be seen and treated within four hours.

The average A&E attendance in 2015/16 cost £138. Cost inflation since then was estimated by NHS Improvement at 3.1% in 2016/17 and 2.1% in 2017/18, with an assumed efficiency growth of 2% each year. Using this, we have estimated a minimum cost of around £60m to help add capacity.

However, with lower numbers of attendances in winter, it’s not just about A&E capacity. The inpatient capacity needed to accommodate the additional admissions must also be addressed.

Chart 3 – how many of these are severe enough to require an emergency admission to hospital, and wait longer than four hours to be admitted?

Number of people waiting more than 4 hours in A&E in England, Q3 2010/11 to Q4 2017/18

Download the chart (PNG)

To estimate the additional capacity for admissions, we can look at the number of people who are waiting longer than four hours on a trolley. These are people who definitely need to be admitted, but for whom there isn’t a bed available yet, so they wait longer than four hours between the decision to admit and a bed being available. While this doesn’t account for all admissions over the four-hour A&E target (which will be due to a combination of A&E capacity and bed availability), it provides a minimum estimate for the most severe cases. This group has grown by an average 40% a year since 2011/12, a total increase of 414% (albeit from a relatively low base). Again, you can see there’s a clear seasonal pattern. This winter, if this pattern continues, we would expect this to be around 173,000 people.

We estimate the average cost for an emergency admission from reference costs at just over £1,700. Applying the same adjustment for cost and efficiency as above, we estimate the extra cost of admitting this group of people would be around £300m.

So this gives our minimum estimate of £360m if we’re to increase both A&E and inpatient capacity to meet the four hour target this winter. This is similar to estimates from providers themselves.

How might the £360m be spent?

The more complicated question is how might this be spent. The first challenge is that it is not easy to predict where and when this additional capacity might be needed, so an oversupply of beds would almost certainly be required. The second is that there is already a shortage of nursing staff, so there would need to be a higher use of bank and agency staff to open these extra beds. This is a key cost area that NHS Improvement are trying to reduce. There is a similar argument for physical capacity. The cost of opening temporary beds or wards may be higher than maintaining current ones. This suggest that the true figure might be much higher.

The counterargument is that there has been £100m of capital investment to help ease winter pressures, more people have received the flu vaccine, and more 111 calls are handled by clinicians. If these measures are successful then the number of people waiting longer than four hours may buck the trend and be lower than we expect. Equally, we must acknowledge that people waiting are seen eventually, so the capacity does currently exist to treat them, but just not within the target time. Conversely the risk is that a major outbreak of flu or norovirus, or a severe drop in temperatures (which we have largely avoided in recent years), may buck the trend the other way.

This money might not be best spent in hospitals. Another way to increase hospital capacity is to reduce the number of delayed transfers of care (DTOCs), where people are medically fit to leave hospital but can’t due to delays arranging packages to support their return home. DTOCs were 65% higher in 2016/17 than in 2011/12. In Q4 2016/17, DTOCs accounted for 583,000 days, 5.2% of all hospital bed days.

Care packages to help people out of hospital are generally cheaper than treating them in acute care settings, so investing the money this way may be more effective. There is a target to reduce DTOCs to 3.5% of all hospital bed days. This target is backed by extra social care spending, announced in the last Budget, and would free up 215,000 bed days. But the drive to free up hospital capacity before winter has made only limited headway and, given how close we are to winter, achieving radical reductions in delays seems unlikely in the short-term.

With the temperatures already dropping, it is almost certainly too late to invest wisely in schemes to improve NHS performance this winter. But without at least some level of additional investment, it is hard to imagine how additional capacity can be found to improve access for patients who need emergency care this winter, next winter and beyond. Any additional investment would be welcome for this year, but a long-term plan needs to be developed as soon as possible for next year and those after. Winter is not going away.

Share this blog on: Facebook | Twitter | LinkedIn

Adam Roberts is Head of Economics at the Health Foundation, @ADRoberts777


stephen black

This is an interesting approach to estimating the potential cost of "fixing" poor A&E performance over winter.

I wonder, though, whether the cost base used for the calculations is the right one. The costs of the average A&E attendance and admission are the average costs across all admissions. But, surely, the real costs we need to estimate how much is required to improve flow are the marginal costs of extra activity (and these are likely to be significantly different to the overall average). Perhaps, though, there are no good estimates of the marginal costs so it will prove hard to do better.

A more serious objection is that trying to assess to cost of fixing A&E in a way that doesn't account for the bottlenecks in the process is problematic. It has been known since the early days of e target that the biggest problem causing delays is finding empty beds for admissions. This problem can't be fixed by spending more inside A&E, though the existence of a block in flow outside A&E probably increases the workload inside A&E. If this is correct, then speeding inside A&E makes little difference to performance. But it also means that focussed spending the right place might make a large impact.

Also, it is arguable that the major problem with beds isn't overall resourcing, but coordination of arrival and discharge. An estimate for DH in the mid 2000s suggested that the single easiest way to generate more free beds for emergency admissions would be to ensure most simple discharges were done in the morning. This would ensure perhaps 15-20% of all beds were empty by the point of the day where emergency demand was highest. This suggests that while putting more resources into beds might help, putting more resources into coordination might help a lot more. And failing to fix coordination might drastically limit the gains from more beds. And, if discharges are problematic because of factors outside the hospital (eg no beds in social care homes) then the key investment are not even in the hospitals.

So, while it is interesting to have back-of-an-envelope estimates for how much money is required to fix A&E, it might be better to keep pushing the message that understanding why A&E performance is poor is more important. There are too many zombie hypotheses about why A&E performance is worse. Killing those and directing attention to the real root causes will yield better results than any amount of extra money.

Add new comment

* indicates a required field

Your email address will not be published on the site and will only be used if we need to contact you about your comment.

View our comments policy