New NHS Waiting Time App Data Capped At Four Hours

The data behind a new app giving waiting time information at Wrexham Maelor and other hospitals has been limited in the raw information it provided, as it was ‘never intended’ for publication.

Earlier this month an app was released which was described as allowing people to ‘get up to the minute information on waiting times for Emergency Departments via their smartphone or tablet’.

Wrexham.com accessed to raw data above four hours however that has been capped, with also a possible difference stated in what information the app displays. Currently a definitive ‘Longest Wait’ label is shown in the app, however we have been told that figure is based off calculations from the average total time patients are waiting in the department.

At the time Liz Rix, Chief Nurse at the University Hospitals of North Midlands NHS Trust, said: “The app gives real-time information on waiting times and show how busy the departments are at any time during the day. This is about empowering patients to make the most appropriate choice for them.”

As readers will be aware we sniffed out the data sources via a very simple process, which showed the app requested data from waittimes.uhnm.nhs.uk . Looking at the url further it allowed raw information to be pulled publicly – you can see what is called JSON output here – http://waittimes.uhnm.nhs.uk/APIv2/GetSitesGeo?lat=53.043040&lng=-2.992494&radius=5.

An example of this is below, with the relevant figures shown below for the wait time and queue.

We created a very simple piece of code that grabbed those two numbers every thirty minutes and tweeted them out via our new @WrexMaelorBot. The wait time did not appear capped, with data such as the below showing on the raw feed while the app showed ‘over four hours’.

Current longest waiting time at the Maelor is stated as 5 hrs 40 mins. There are 30 people in the queue.

— Maelor Wait/QTwitBot (@WrexMaelorBot) January 12, 2017

On Thursday the bot went haywire, and tweeted that the wait time was at that point over a year.

Current longest waiting time at the Maelor is stated as 9999 hrs 99 mins. There are 30 people in the queue.

— Maelor Wait/QTwitBot (@WrexMaelorBot) January 12, 2017

We checked the code and the raw information, which was showing as 9999 for the wait time:

After some testing it appeared that now anything over four hours will show 9999 rather than a precise number.

We incorrectly enquired with BCUHB on why the waiting time app had been capped, and was correctly told, “There haven’t been any changes to the app – it has never shown detailed times above four hours.”

“The app is intended to give the public a helpful guide as to the maximum time they are likely to spend in one of our minor injuries or emergency departments if they attend with a non-urgent condition, to assist in their choice of where best to go. Obviously the time people spend in the departments varies depending on how busy the department is, the clinical urgency of a patient’s injury or illness, the time needed for tests and scans to be carried out and the nature of any treatment that needs to be provided. The figure displayed in the app is calculated from the average total time that patients currently in the department have been there.”

“We don’t show detailed times for beyond four hours because of the uncertainty in making valid predictions about what the waiting time situation might be like the further ahead you look – we’d be claiming a level of accuracy that we couldn’t justify.”

We returned with a query on the data rather than the app, as it was the raw data that had changed rather than the app (although our initial query did mention ‘app data’) and asked has the data provided to the app been capped at four hours, and if so, why.

We were told, “The app has only ever been designed to show forecast wait times up to 4 hours because we don’t believe we can make detailed predictions of likely waiting times further ahead than that.”

“Yes, an adjustment apparently has been made to how the data is provided to the app to ensure that this position is maintained.”

We then provided examples of the app data showing what appears to be a natural rise and fall in the data we logged of one mornings wait times pointing out that the information over the 4 hour cap appeared at similar granularity to under 4 hour information.

Any information over 4 hours below would show ‘over 4 hours’ in the app, whereas our twitter bot would show the raw information as below.

BCUHB described this as ‘output of the initial calculations that feeds into the app’, adding, “It was not being provided to anywhere else and was never intended for publication in that format.”

Currently the app and raw data is still capped at four hours.

The current output from our bot is below:
Tweets by WrexMaelorBot