Friday 21 May 2021

Joining The Dots - How to recognise the seriously unwell child

In the previous post, I outlined how to tell the difference between abnormal signs that are part of a functional and fully compensated response and those that are part of an illness that is having a more significant clinical effect.

The child who is completely well and the child who is visibly seriously unwell and decompensating are both fairly straightforward scenarios.  The child who is very well causes no clinical concern.  The child who is decompensating causes unambiguous and immediate clinical concern.  If only decision making was always this easy.

The child who is febrile with tachycardia but a reassuring level of activity and interaction is also relatively straightforward.  The trouble is that we end up having to make decisions about those children who have a reduced activity level or whose interaction is not completely reassuring.  Often there is a disconnect between what guidelines tell us to worry about and how worried we actually are.  I believe that this often occurs when we intuitively include something in our decision making that is rarely featured in guidelines: the pattern of the illness.

Everything written tends to be geared towards the snapshot:

  • What is the heart rate?
  • What is the temperature?
  • How active and interactive is the child?

This is problematic in any acute specialty assessing febrile children.  Catch the child at the wrong time and they seem to trigger multiple red flags.  Base your assessment on the snapshot alone and you may be falsely reassured.

As covered in the previous post, physiology in young children (not so much babies and older children) responds to illness with what can be dramatic changes even in uncomplicated low-risk infections.  Unfortunately the same abnormalities can be seen in more clinically significant infections.  No one thing is particularly sensitive or specific when deciding whether to be worried.

So if a snapshot can be misleading and there is significant overlap between low-risk and high risk scenarios, how can we decide when a clinical presentation is high risk for sepsis or serious bacterial infection?

The pattern of symptom progression is probably the answer.  In the past, research has concentrated on the snapshot, over-emphasising the assessment of various parameters at a single point in time.  Human intelligence allows us to incorporate the more complex business of considering three very important factors that previous research has not often considered or emphasised:

  • The most recent worst state of the child
  • The most recent best state of the child
  • The pattern and progression of symptoms over time

The evaluation of the best and worst states and the pattern may be include retrospective (history) and prospective (a period of observation) information.  The likelihood is that you already put a great deal of weight on these factors in your decision making.  It is also likely that you have recognised that there are two main patterns of illness in children.  I have represented these in the following diagram, without labelling which is the low risk scenario and which is high risk.

Intuitively, I would assume that you recognise the dotted line as the high risk pattern and the solid line as the typical pattern of a low risk clinical picture of childhood febrile illness.

I suspect that the explanation for this is that the child represented by the solid line is exibiting a physiological reponse to an infection that does not cause organ dysfunction.  Therefore as their immunological response fluctuates, they swing from one extreme to the other.  The child represented by the dotted line is sufferng from an infection which is causing significant physiological dysfunction.  The result is an inability to return to normal.

If you buy that, have another look at the diagram and look at the lines in the first half from a time point of view.  In the left hand part of the diagram, severity of symptoms is often worse for the child who swings from being more unwell to a return to baseline.  I believe that this is the reason that formulaic risk-assessment for serious illness that is based on a snapshot is impractical and problematic.

The end results of this over-emphasis on clinical information taken at a single point in time are twofold:
  • Using a snapshot will over-diagnose serious illness, simply due to the pre-test probability of sepsis and serious bacterial infection (SBI) in the low-risk (which is most children) child.
  • Over time, the repeated realisation that children usually have an uncomplicated, self-limiting illness despite the severity of symptoms and abnormal physiological parameters risks de-sensitises the clinician to the possibility of sepsis/SBI.  Abnormality becomes normalised and we learn to ignore things that have poor specificity for a serious outcome.
I would strongly encourge stepping back from the snapshot to see the big picture over time.  Doing so, either through the history, observation or both might help you to recognise which children are best managed symptomatically and with good safety netting advice.  It migh also allow earlier recognition of the less common scenario of the child who persistently fails to return to (or close to) baseline and is more likely to have a more significant illness.

Telling the difference between self-liniting childhood illness and serious infection is complex and requires us to process an awful lot of information.  When it comes to features like best and worst states and the pattern of symptoms over time, your intuition is probably already joining these dots for you.  I think you should trust that intuition.

Edward Snelson
Dot-joiner but likes to colour outside the lines
@sailordoctor