Will linked data finally show us what happens after the ambulance arrives?

Sarah Scobie takes a closer look at new ambulance response standards, explores the variation in how providers are performing, and argues the case for a single national ambulance dataset.

Blog post

Published: 28/08/2019

Ambulances are called out more than 20,000 times each day in England, and the recent NHS Long Term Plan said they are “at the heart of the urgent and emergency care system”. Yet we know very little about how the care they provide affects patients’ overall outcomes and experiences. That leaves unanswered a series of important questions about the timeliness and quality of care, and what drives them.

What has been the impact of the revised ambulance response standards?

The 2017/18 Ambulance Response Programme brought in radical changes to ambulance targets and how services in England prioritise calls. This is the first time that response standards have changed since ambulance services became part of the NHS in 1974, and has required major shifts in how the service works. 

The new standards cover all calls, from Category 1, the most urgent, to Category 5, the least urgent calls that only require a response over the phone. 

Category 1 is for “life-threatening” conditions needing immediate intervention, such as cardiac arrest or airway obstruction. The target is now an average response time of seven minutes. We can’t compare this directly with the previous eight-minute target for the most urgent response, because this was based on a cut-off, not an average. 

Performance for these most urgent calls has improved since April 2018, and the target has been met since March 2019. But while this is an important achievement, it is important to remember that a high proportion of patients in this category will not survive – for example, only 10% of patients who had a cardiac arrest survived (based on data from April 2018-January 2019).

So what is happening to patients who do not quite meet this criteria, but will often have their life hanging in the balance? This includes people suffering from a probable heart attack, serious injury, stroke, sepsis or fits. Under the old system these patients should also have received a response within eight minutes, although this was only met around two-thirds of the time.

Now they are categorised as Category 2 “emergency” calls, with a target of 18 minutes (average time) for a response. Since 2017/18, services have struggled to meet the new target: in May 2019 the average response time was 21 minutes.

But because ambulance data doesn’t link up to the record of what happens to patients next in hospitals, there is no information available about the outcomes for them, and whether these are related to how long it took for an ambulance to arrive.

How concerned should we be about ongoing and wide variation between ambulance providers?

It’s hardly news that some NHS organisations perform better than others, and this is also the case for ambulance services. However, each of the ambulance services – with the exception of the Isle of Wight – serve large populations and deal with hundreds of thousands of incidents each year. If an ambulance service is not meeting standards, this affects a large part of England. 

In May 2019 only one ambulance trust – London – met standards for all four categories of response. Looking more closely at Category 2 incidents, only four trusts met the standards. The average response time was 25 minutes or longer for four trusts, and in the south west 10% of “emergency” patients waited longer than an hour.    

Average and 90th percentile response times for “emergency” (Category 2) patients, by ambulance trust, May 2019 27/08/2019



Nuffield Trust analysis of NHS England Ambulance System Indicators

Read more

What impact does this have on patients in those areas? Again, the answer is currently largely beyond our reach because we cannot see what happens to them once they leave the ambulance.

How will changes in A&E performance targets impact ambulance responses?

NHS England is currently piloting new A&E performance measures in 14 acute trusts, aiming to replace the four-hour target with measures based on average waits. Exactly what would constitute success is not yet clear, but part of it must be the impact on ambulance services and the wider urgent care system.

There is a strong relationship between A&E performance and ambulance response times, particularly for Category 2 calls. More patients waiting longer than four hours in major A&E departments is associated with longer waiting times for an ambulance (see chart). 

The strong association between the two measures reflects similar patterns of need impacting on demand for services. At the same time, worsening performance in A&E leads to longer delays when handing over patients – meaning ambulances and staff aren’t available to respond to the next call.

We don’t yet know how the new targets will affect what happens in A&Es. But given the close relationship to ambulance performance, we can expect there to be a knock-on effect that needs to be monitored.

Ambulance waits for “emergency” (Category 2) patients compared with patients waiting more than 4 hours in major (Type 1) A&E departments 27/08/2019


Read more

Looking for the answers

The planned development of a single ambulance dataset could offer us some answers to these questions. This would let us analyse response times for different groups of patients, and be a first step to enabling ambulance response data to be linked to hospital episodes. 

But linking data in the way we need will remain difficult, because ambulances are not routinely able to capture patient details or NHS numbers. Working with London Ambulance and other partners, the Nuffield Trust has found that it was feasible to link data by using the dispatch number and crew details, fields that are now part of the new emergency care dataset

Once this data is flowing, we’ll finally be able to address how ambulance response times impact the chances of being admitted, length of stay, and outcome for patients. We could look at how the response categories are actually dividing up patients with different conditions, how long patients wait in total from an ambulance call to a hospital bed, and much more. Linked data will also enable the current set of quality measures for ambulances to be extended.

In the coming years, meanwhile, the Long Term Plan heralds a battery of further changes, including joining up triage systems for ambulances, 111 and other services.

If we are going to understand the impact of the last set of changes, let alone the next ones, a single national ambulance dataset linked to hospital information will be vital.

Suggested citation

Scobie S (2019) “Will linked data finally show us what happens after the ambulance arrives?” Nuffield Trust comment.