ARTIFICIAL INTELLIGENCE

Artificial intelligence (AI)  is “intelligence” exhibited by machines.

Intelligence is exhibited by any device that perceives its environment and acts upon the information to work towards achieving a goal.

So machines mimic cognitive functions that humans use to learn and problem solve.

In health, the potential to assist nurses and other health professionals make decisions is clear.

One artificial intelligence function you may be familiar with is optical character recognition (OCR).

OCR is where a computer is able to “read” a document or any other piece of typed, handwritten or printed text and recognise the words.

Using words can be easily searched to extract data.

But that is just one example. Other examples are where the computer understands human speech, play strategic games at a high level, or drive autonomous vehicles.

DeepMind is a UK company that was founded in 2010. 

Google acquired it in 2014 but it operates independently from them . 

Their aim is to build AI technologies and prove that the technology can have a positive social impact.

They have worked with and in the UK National Health Service (NHS). 

DeepMind recognised that nurses and doctors do not have the tools to instantly analyse every test result.

The flow on was that if test results were being analysed, then the right treatment could not be commenced.

And then every patient that needed urgent or complex care was referred to experienced practitioners.

When patients do not receivethe right care at the right time because of a failure to recognise and respond to patient deterioration in time is known as “failure to rescue”.

DeepMind developed an instant alert application called Streams to address this issue.

In the current system, there are seven additional steps that occur when the results become available.

Each step has a built-in delay from notification to action.

Streams reduced the steps to two.

The first step after the results are available is that the Streams app checks the result.

If Streams detects an abnormality using AI, the results are send via mobile to the appropriate person who can act upon the information.

According to DeepMind, they cite a report that says that thousands of people in UK hospitals die preventably from time critical conditions such as sepsis and acute kidney injury.

This is because the warning signs were not picked up and acted upon in time.

The Streams app is currently in use at two hospitals in the UK.

The app quickly reviews the test result for serious issues.

If one is found then the system sends an urgent secure smartphone alert to the right clinical staff to help. 

Also included with the alert is information about pervious conditions so that the clinicians can make an immediate diagnosis.

One hospital claims that nurses are saving over two hours each day using the app.

Sarah Stanley, a consultant nurse who leads a “patients at risk and resuscitation team” said: 

    “Streams is saving us a substantial amount of time every day. The instant alerts about some of our most vulnerable patients mean we can get the right care to the right patients much more quickly. Instead of spending time checking a number of different systems for information about patients, all the information is contained in the app, which frees us to spend more time delivering face-to-face care to patients.”

So if this system works out, future patient care will not depend upon the manual analysis of test results to identify potential issues, followed by an endless round of phones calls, pagers and desktop systems.

Instead the future will be where AI tools learn how to analyse test results and scans to immediately identify when a patient might be at risk.

And the system will continually improve to get even better with machine learning.

As always, patient data and privacy needs to be carefully protected with the highest standards of security.

So far Streams is working out on a small scale.

The trick will be scaling it up to analyse many more results.

To learn more about artificial intelligence: 

DeepMind Website 

Royal Free Hospital web page about Streams

YouTube video about DeepMInd Streams Application

Hogan, Helen, et al. “Preventable deaths due to problems in care in English acute hospitals: a retrospective case record review study.” BMJ Qual Saf (2012): bmjqs-2012.

Johnston, Maximilian J., et al. “A systematic review to identify the factors that affect failure to rescue and escalation of care in surgery.” Surgery 157.4 (2015): 752-763.