Sepsis alerts currently embedded in GP software are being over-triggered and the software needs to be improved, a senior NHS England official has admitted.
Dr Celia Ingham Clark, medical director for clinical effectiveness at NHS England and interim national patient safety director at NHS Improvements, said that there was a need to update the software – which has embedded NICE guidance and alerts GPs to potential sepsis cases – after many GPs complained of the alerts being ‘over-triggered’ and switching them off.
GP leaders have warned that GPs are suffering from ‘alert fatigue’ due to the warnings being activated at very low thresholds.
Speaking at the Westminster Health Forum Keynote Seminar on improving sepsis outcomes last week, Dr Clark said that NHS England were talking with GP software providers and that they wanted ‘to ensure that the algorithms that the big software providers use are up to date with NICE guidance and also don’t over-trigger, because we know that if they over-trigger, GPs switch them off and they’re no use at all’.
Dr Clark continued: ‘To be absolutely honest I’m not up to speed with whether the software providers have changed their algorithms or not.
‘When I last interacted with them six months ago, they were saying that they were going to do it and I need to go back to them again. It’s important that what they put on the algorithm fits with the NICE guidance and it’s important that they’re not set at a level that over-triggers, meaning that GPs switch them off.
‘It’s a competitive market between the main software providers and it’s an opportunity for any of them to get a market advantage by doing this well and quickly.’
NHS England conceded in September that the 2016 guidance on sepsis, which includes lists of warning signs that should prompt referral, had been tricky to implement, saying that front-line staff had found it ‘difficult to translate this guidance in to practice.’
BMA GP Committee member Dr Andrew Green described the software alerts as ‘a lesson in how not to use IT’ back in October.
He said: ‘The largest problem with it is the low threshold for sending an alert. The result is that these alerts are ignored. Should that “suspected UTI” patient go on to have sepsis, GPs may find themselves having to justify why they did not take action earlier.
‘Inquests where people have ignored warnings rarely look at the contribution of alert-fatigue.’
