How NHS Digital is using data science to cut down on A&E visits

NHS Digital hopes to ease the burden on local accident and emergency (A&E) units by analysing call data into its 111 emergency telephone service.

Speaking at the Tech for Britain conference in London last week Daniel Ray, director of data science at NHS Digital said: "Where we have started to generate new big data and apply data science is in areas like NHS Pathways. So if you phone the 111 service, up until recently no one had analysed that data."

The NHS received 15 million 111 calls per month, two million of which end up with a patient being referred to a local A&E. The NHS 111 telephone helpline has actually increased the number of people turning up at emergency departments and calling ambulances since being launched, the opposite of its intent to ease the burden on A&E units.

Ray explained that NHS Digital has always collected 111 call data as "when a call comes in the handler sits in front of a clinical system and they capture information", he explained. "What sits behind that is a clinical algorithm decision tree."

Ray noticed that the data from these calls was just "sitting there" when he arrived at the organisation around a year ago, so he kicked off the project to try to get more insight into the referral process and to hopefully improve the clinical algorithm to reduce the number of people being sent to A&E where possible.

With a small team of two data scientists the organisation started looking at what the patient actually does after a 111 call and secondly, if they should have referred that patient to A&E at all. By linking up the 111 call data and local A&E inpatient records he believed "we could potentially stop hundreds of A&E attendances simply by tweaking the clinical algorithm that the nurse fills out at the other end of the phone", he said.

Predicting A&E demand

Ray, who joined NHS Digital last year as part of the new Centre of Excellence for Big Data and Data Science, is tasked with finding new ways to use data science to deliver better care across the health service. Another project he is working on is to try and improve the predictions the NHS can make around A&E demand.

By analysing streaming data from sources like sporting event databases and the Met Office for weather Ray claims to be able to increase its demand prediction models for A&E attendance by up to 25 percent. This is "enough for someone who is in charge of a hospital to redirect how many resource they need to cope", he said.

Something as innocuous as weather data is significant because "the temperature outside makes a massive difference to both the total volume of A&E attendance and the type of patients that come through the door", Ray said. "When the sun is out everyone is out breaking themselves."


Copyright © 2017 IDG Communications, Inc.

Shop Tech Products at Amazon