Researchers have warned GPs not to use paediatric risk scores to assess sick children after showing they do not work in general practice.
The team at Cardiff University have now assessed three risk tools – including one recommended by NICE – and found they classify too many children as seriously unwell and fail to spot those who go on to need admission.
Developed in hospital settings, they do not translate well into primary care populations, the researchers said.
In the latest study, they tested the Liverpool quick Sequential Organ Failure Assessment (LqSOFA) and Paediatric Early Warning System (PEWS) using records from 6,703 children under the age of five presenting to their GP in England and Wales with an acute illness.
Reporting in the British Journal of General Practice, they looked at children admitted to hospital within two days of seeing their GP and assessed whether the risk score accurately predicted how sick they were.
It showed that if the risk scores had been followed almost two-thirds of children would have been rated as needing hospital care. When it came to identifying serious illness, they were “no better than chance”.
An earlier study from the same researchers testing the ‘traffic light’ system recommended by NICE for use in general practice, also found it ‘performed poorly’ because it categorised almost all children as moderate or high risk of serious illness.
There has been much debate about use of risk scores developed in hospital in primary care.
The Royal College of GPs recommends the NEWS2 score to assess clinical deterioration in adults should not be used as a replacement for clinical judgement because it has not been validated in general practice.
This latest study shows the same is true of risk tools developed to identify very poorly children, the researchers said.
‘Unconsidered use by GPs could result in unsustainable referrals,’ the team concluded. ‘Further research is needed to derive and validate an accurate scoring system in general practice that is both easy to use and accurate,’ they said.
Study leader Dr Kathryn Hughes, a GP and senior clinical lecturer of primary care at the University of Cardiff said the population of children seen in emergency departments and in general practice were not comparable.
They often present earlier and the same tests cannot be done due to lack of equipment.
‘I feel very passionately that this is not acceptable. The prevalence of serious illness is so different, it is crazy to think that something developed in hospitals will automatically apply to general practice.’
Any risk score – which would be helpful – needed to be developed in and tested on GP populations, she added.
One had been developed in general practice in Belgium that showed promise but needed further validation.
‘If a child has got an abnormal oxygen SATs, heart rate, or whatever they probably do need to be in hospital, it’s just that you might miss children if you rely only on that symptom or sign. So, I think it is still important that we do check these but the clinical gut feeling, in the absence of a good tool, is probably what we do need to rely on.’
There is no substitute for experience and gut feeling in General Practice. Somebody (either A&E or you the GP) should actually see the patient. The GP will only take a snapshot of what is going on. In the case of sepsis this might be before any real signs show themselves, therefore you need to give clear worsening instructions which may include taking the patient to A&E within minutes of leaving your surgery.
Children deteriorate and improve so quickly. It is important to listen to the parents carefully. Normally their concern is right. Sometimes it is not appropriate to delay A&E assessment and send them straight to A&E. This could always apply to the very young. You must always record what you find in a clear and detailed way. If you think “Could it be sepsis?” You will normally be OK. I do think that one should try to record all the appropriate vital signs because they can help track deterioration but relying on a score and only intervening according to that is a recipe for disaster. I am a retired medical director of an OOH service.
I saw a child once on a Saturday morning, just before closing for lunch, who looked quite well, behaving normally to me, speaking coherently and answering fine, no vomiting, no fever, no rash, no signs, but ‘not feeling quite right’. I had a feeling, and could not find any explanation for him being not right, so sent him to children’s ward. (NB, I do not send EVERYONE to childrens’ ward !!) He was clerked in and set for observation overnight with no diagnosis or serious concern (and they were probably thinking of something critical to write about my having sent him for admission)
During the evening a friend of the family, who happened to be a GP, was visiting and sat on the edge of the bed chatting, when the child started to develop the typical rash of Meningococcal meningitis.
He would have failed all paediatric illness screening tools prior to that point.
Think it’s a no brainer that anyone with any confidence in their clinical skills will only use tools as way of documentation. Maybe a tool could advise, but rarely direct.