A&E targets distort behaviour

Yesterday the NHS Information Centre published a further analysis of the time that patients wait in Accident and Emergency departments before being seen to. The analysis showed that just before the 4 hour target, there is a huge peak of people being admitted to hospital. See the graph below from the report.

The rising peak just before the 4 hour deadline is very clear.

A and E wait times
A&E wait times

There is an accompanying Excel spreadsheet which lets you see this distribution for each of the trusts that submitted data. Most of them exhibit this kind of pattern.

The report deals mostly with data analysis but little time is given for the reasons why people are treated and admitted in this pattern. I would postulate that the reason is clear. They have a target to deal with people within a 4 hour window and so either people have something relatively minor and they are treated quickly with no follow up or referral to their GP or they are left to wait. Then as they come closer to the 4 hours someone is there to make sure that the target is hit and people are admitted. There is a further chart from the report that is quite revealing.

A&E waits split by type
A&E waits split by type

This is a chart of the destinations of patients who are dealt in each ten minute slot. The bands dealing with ‘referred’ (white on top), ‘others’ (bright green under referred) and ‘discharged – referred to GP’ (dark green, second from bottom) are all fairly flat. The two categories that seem to cause the peaks are ‘discharged – no follow up’ (yellowy green) which peaks around the hour mark, and ‘admitted’ (black, on the bottom) which has a little peak in the first 10 mins and then rises to a very steep peak just before the 4 hour mark.

All the categories drop-off to almost nothing after the 4 hour point.

It is a pretty safe guess to say that the ‘discharged – no follow up’ peak at 1 hour because they have minor issues so they can wait a little while, but then they can be quickly dealt with and go home. (A bit like my cut finger!)

It is the admitted patients that are of the most interest. The 4 hour target is skewing behaviour. Why are so many patients being admitted just before the target? It can’t be coincidence. How many of those patients are being admitted solely to meet the target and not for clinical reasons? The thing that really gives the game away is the massive cliff-like drop-off after the 4 hour point. If there were a smoother drop-off after 4 hours, it would indicate that the system was behaving more normally, but the fact that almost no-one is left to be dealt with after 4 hours, means that patients must be being treated differently as they come up to the 4 hour point.

Now, isn’t it a good thing that people are seen within 4 hours? Well yes and no. It is good if people are seen quickly, of course. Patients with life threatening conditions will surely want to be seen straight away and the rest of us walking wounded don’t want to be hanging around A&E for no reason. But the problem is that this data shows clearly that the 4 hour target is skewing behaviour. Patients are being admitted to meet the target and not to give them the best care. Also if this is happening then it must follow that resources are being used to meet the target. There must be some mechanism that is letting staff know that a patient is about to breach the 4 hour target and then staff are diverted away from other patients to get the near-breach patients admitted. This way of managing resources is taking away from treating people as they arrive which otherwise might move the curve as patients are treated sooner.

The irony is this, by diverting resources to ensure that people don’t wait more than 4 hours you are ensuring that people are more likely to wait longer and thus more likely to breach the 4 hours. And what of the waste of resources in the rest of the trust when people are admitted to meet the target and then take up beds, nursing time etc. when they don’t need it?

This target needs to be dropped and replaced with measures of value to the patient. This will give understanding of demand so the service can be designed to meet the needs of patients, giving maximum value as quickly as possible. If trusts were to be freed from the burden of this target and then given a proper method to improve, they would be able to wipe the floor with the 4 hour target instead of being held back by it.

This data shows that while the Department of Health and central government think that this and similar targets are driving performance, in fact these targets are holding performance back.



If you found this interesting