Urgency AI

What is Urgency AI?

When a request is submitted by a patient, Urgency AI automatically flags the request as either:

  • Emergency – Red flag. Should be resolved on the same day and may need emergency services input. Goes to the top of the inbox.
  • Urgent – Orange flag. Should be resolved within 48 hours. Goes to the top of the inbox but below red flag emergencies.

It will also pre-populate the 'How clinically urgent is this request?' triage decision for you in the patient request. You can read more about triage decisions in PATCHS here

Urgency AI's suggestions are meant to assist, not replace clinical judgment. You should therefore still triage and reply to all PATCHS requests ASAP.

 

Screenshot_2021-07-15_at_11.25.53.png

What are the benefits of Urgency AI?

Without Urgency AI, initial triage decisions are usually made by receptionists. Advantages of PATCHS doing this instead are that it:

  • Improves patient safety by speeding up triage decisions, helping urgent requests get reviewed quicker.
  • Reduces workload by suggesting triage decisions instead of having to come up with them yourself.
  • Improves triage accuracy because it's learned how to triage from hundreds of GPs using PATCHS.
  • Reduces clinical risk placed on receptionists by supporting them to triage patients.

How accurate is Urgency AI?

Interviews and workshops with PATCHS users have shown the most important thing for Urgency AI is its ability to detect urgent and emergency patient requests. The best way to do this is to measure the percentage of urgent and emergency requests it gets right, which is called sensitivity. (This is different to specificity, which is the percentage of not urgent or emergency requests Urgency AI gets right.)

Urgency AI is primarily designed to be safe, and has therefore been built to make sure it doesn't miss any urgent requests i.e. it has a high sensitivity. This means it will often suggest requests are urgent or emergency just to be on the safe side.

In a study of 19,805 requests submitted to 15 GP practices using PATCHS, Urgency AI correctly detected 94% of the 3,899 urgent and emergency requests. Research suggests that human physicians have a sensitivity of 79% to detect urgent and emergency requests.

What should I do if I disagree with Urgency AI?

If you disagree with the urgency suggested by Urgency AI - or you think it's missed an urgent or emergency request - you should change the urgency yourself when making your triage decision. Because PATCHS is always learning, this will help it correct mistakes in future.

 

Screenshot_2021-07-15_at_11.26.48.png