Understanding Traffic Lights - The Unwell Child and What to Do Next

When I am driving and see a traffic light ahead the first thought is, "Am I supposed to stop or go?" My next thought is, "what might it change to and what do I do then?"  Assessing the unwell child is like that.  It's not just about the snapshot.  Guidelines look a moment in time but the unwell child is in constant flux making that approach problematic.

The traffic light system for unwell children has been around for a very long time.  It is used across Primary and Secondary care to aid clinicians in their attempts to risk assess febrile and unwell children with all of the non-specific signs and symptoms with which they present.  

I am often asked if I use the traffic light system in my own practice.  The answer is yes and no.  Yes - the system is a useful hierarchy of signs and symptoms.  No - because most childhood illnesses are too dynamic for a snapshot to be completely valid.  Things change constantly.  A risk assessment based on a moment in time is far too simplistic.

That doesn't mean that observation of the child is necessary for decision making.  In most cases it's simply a question of asking how the lights are changing and what I'm going to do with that.

Amber turning green

A 2 year old child presents with a cough, runny nose and a fever.  The parent reports that a couple of hours ago they looked pale and lethargic.  They were shivering, felt hot centrally but had cold hands and feet.  Now they have none of those things happening.  They are walking, talking and cheerfully interactive.

This is a very common scenario.  Parents and carers will often express a certain paradoxical frustration with the apparent wellness of the child.  The child appeared seriously unwell a couple of hours ago and the parent is now feeling that you will think that they have over-reacted.  It is a good thing to acknowledge how unwell the child was and use that as an opportunity to explain why you as a clinician are happy with the child despite how concerning the child's appearance was.  

Giving or signposting to something written is also important.

Green turning amber

A 2 year old child presents with a cough, runny nose and a fever.  When you see them they are miserable but alert and interactive.  They have a temperature of 39.5, heart rate of 160 and are refusing to drink.  They last had any symptomatic treatment 6 hours ago.  The parent reports (you have to ask about this - it won't usually be volunteered) that 2 hrs ago they looked much better and were drinking a bit.

Unlike actual traffic lights, unwell children swing from green to amber and back to green quite normally during uncomplicated self-limiting infections.  There is a reason that we mostly see unwell children between the age of 6 months and six years.  It's not because they are high risk for dangerous infections.  In fact quite the opposite - it is a stage of life characterised by extreme response to simple infections.  The normal physiological response can look bad but usually resolves to reveal a reassuring baseline.  In many ways, a febrile unwell 2 week old is easier from a decision making point of view - that is a very high risk presentation.  A febrile unwell 2 year old is low risk but that presents a different problem - how to recognise the small number that do have a serious illness.

What can be terribly inconvenient is the above situation.  The snapshot we are given is not green but also not red.  Red is also easier from a decision making point of view.  Amber presentations make us have to decide what to do next.  Here are your options:

Every clinician will have a preferred option.  Many working in Primary Care do not feel the need to have a face to face reassessment if the child improves in behaviour and activity.  That is completely valid as such improvement is a good demonstration of physiological change and evidence that the baseline state of the child (active, interactive, good oral intake and no increased work of breathing) is not consistent with sepsis or meningitis for example.

Really good safety-netting advice empowers the parent to make that assessment in a way that is dynamic and continuous.  A reassessment in whatever form (face to face or remote) facilitates documentation of improvement and adds value to the safety-netting advice by giving the opportunity for the parent to further discuss the illness, what to expect and when to worry.

Amber children are a fair bit of work but they are a great opportunity to do what we should consider core business.  We can take a group of children who are reasonably low risk and look for signs (e.g. increased work of breathing, meningism or unexplained tachycardia) that this one is the one with something that needs immediate intervention.  For those that are within what is expected of an uncomplicated infection we can make sure that they have symptomatic treatment in the assumption that they will demonstrate a baseline state of reasonable wellness that effectively rules out serious illness.  Finally we can equip the person caring for that child with the ability to recognise signs of serious illness should those develop later.  That is a lot of great care.

Edward Snelson
Paediatric off-roader
@sailordoctor

Disclaimer - drive on the road when you can, off the road when you have to but always get home safely.  If you need help, call.

Where is Your Focus? Let's Play Spot the Ball.

I think that the hunt for a focus of infection in a child is a lot like a game called “Spot the Ball” in which people looked at a picture from a football game with the ball removed and tried to guess where the ball was.

Finding a focus of infection is a very interesting topic at the moment.  More than ever, primary care clinical assessments are occurring remotely rather than face to face. Febrile children are being assessed without laryngoscopy or auscultation.  This seemingly contradicts the tradition of the need to find a focus of infection.  So what is the deal?  I've been asked a lot of questions about this recently and I thought that the simplest thing to do was to bring together my answers.

Do I need to find a focus of infection?

No.  There is no absolute need to find a focus of infection.  Here are a few facts that disprove the idea that a febrile child needs a physical clinical assessment for focus of infection in every case:

• If finding a focus was mandatory, parents and carers would have to have their febrile child assessed on every occasion.  Self-care without medical assessment would be neglectful and therefore a safeguarding issue.
• Even if you believe that medical assessment should take place, this only provides a snapshot.  Childhood febrile illnesses are dynamic and the focus can change.  A child who has an otitis media now could have mastoiditis before the day is over.  If we always need to know the focus, you would never be able to send the child out of your sight.

Does finding a benign focus rule out serious infection?

No.  That’s not how that works.  To illustrate the point, I'll tell you a story.  A 4 month old child came in with a history of fever and irritability progressing to vomiting and unresponsiveness.  On assessment they had a GCS of 8, heart rate of 200 and capillary refill time of 4 seconds.  I looked in their ear, found that they had otitis media.  We all celebrated the finding of the focus and discharged the patient with oral medication. 

That would never happen of course.  While I give that as an ridiculous example, the principle does apply to the moderately unwell child also.  Finding a focus can be a distraction or a premature conclusion.  If we are misled into thinking that an upper respiratory focus is the end of the decision making, we're missing a bigger picture.  We still need to step back and look at the overall clinical scenario.  So in the need to rule out things like meningitis, sepsis and pneumonia, finding a focus could be diagnostic noise.  Serious infections are ruled out on their own merits, not by finding another focus.

Does a runny nose count as a focus of infection?

Yes, but the question of infection doesn't work like that.  A child with a runny nose and a non-specific cough can be presumed to have an upper respiratory tract infection. The real question is, do they have features of another more significant focus?  If they have difficulty breathing, are seriously unwell or specific features of something else (e.g. reduced conscious level) then the assumption flicks from "presumed uncomplicated URTI" to "presumed complication of URTI".

One of the values of giving credibility to non-specific respiratory symptoms is that it helps to answer the age-old question of "should I get a urine sample from this child?"  That's a whole new can of worms.  In a massively oversimplified answer, if they are well and have a cough and runny nose you don't need one.  If they have no cough or runny nose you do need one, especially if they have abdominal pain, vomiting without diarrhoea or have strong smelling urine.

So if focus can be misleading and can change soon after I've assessed it anyway, what am I actually supposed to be doing? 

Treat the assessment of an unwell child like a game of spot the ball.  There will be clues and distractions.  There will be things that direct and misdirect.  The main task is to look for signs of a focus that needs to be treated.  Uncomplicated upper respiratory tract infection (including tonsillitis and otitis media) are not on the list of things that must be identified and treated.  What is interesting is that uncomplicated lower respiratory tract infection’s (LRTI/ pneumonia) place on the list is increasingly in question.

What is on the list of significant infections and how do I detect these?

• Complications of URTI are rare but significant.  Externally, mastoiditis and lymph node abscess can be detected clinically.  Internally, peritonsillar abscess should be looked for in a child who is unable to swallow despite adequate analgesia.
• Central nervous system (CNS) infection (meningitis, encephalitis, abscess) disproportionately affect function and behaviour.  Reduced or focally abnormal CNS function is a red flag.  Signs and symptoms are often non-specific in children but the non-specific signs such as inability to settle are persistent in a way that is unusual for uncomplicated infections.
• Pneumonia, when accompanied by significant respiratory distress or systemic effects is still going to be on the list.  The fact that there is mounting evidence that pneumonia without significant effects does not mandate treatment simplifies the assessment of focus.  Localised signs on auscultation is a focus but the need for antibiotics in the absence of more significant findings is questionable.
• Urinary tract infection is a paradoxical focus in children.  While UTI can self resolve without significant likelihood of complications, these low grade infections often go undetected.  The UTIs that are clinically apparent in children are the most severe UTIs.  It is therefore normal to treat all UTIs especially in pre-school children.  Urine samples should be tested if there is no clear focus elsewhere.  In the child under three years old, ideally these samples are sent for microscopy and culture.
• Finally, there is the focus that is not actually infection.  Inflammatory diseases such as Kawasaki and PIMS-TS present with 5 or more days of unrelenting fever, mucousitis and a very miserable child.  Consider this possibility when the child is worse or no better on day 5 of a febrile illness.

When you play spot the ball in the traditional way, you don't get to find out if you were correct until the next day.  What could be more like hunting for the serious focus amongst the hundreds of febrile children you will see or assess remotely this year?

Good hunting.

Edward Snelson

@sailordoctor

Everything has changed - Non-blanching rash in children

If you told me that nothing has changed for you this past twelve months, I'd be quite surprised.  This year has been a rollercoaster both in and out of work for every healthcare professional I know.  As a finishing touch to the year that has changed everything, I have one more bit of news for you that will change your practice.  Strap in.  This one is huge.

So how did we get there from where we used to be? - "Fever with non-blanching rash is meningococcal sepsis until proved otherwise."

It has long been recognised that this outdated adage has become obsolete in a population with effective meningococcal vaccination.  When it was first coined, the pre-test probability of meningococcal disease (MD) in a child with fever and non-blanching rash was around 1 in 5.  In an unvaccinated population, the 20% chance of MD is more than enough reason to have a "treat in every case" approach in the absence of a rapid diagnostic test.

Since the introduction of a very successful meningococcal vaccination program, the prevalence of MD has dropped dramatically.  The absolute number of cases of MD (the numerator) became a fraction of the pre-vaccination years.  We continued to see large numbers of children with fever and non-blanching rash (the denominator) but no-one was recording how many.

Enter the PiC study, (1) possibly the most significant academic publication of recent years regarding the management of the unwell child.  This large UK based multicentre prospective study did a couple of very important things.  Firstly, it collected data about the prevalence of meningococcal disease in children with fever and petechial rash.   The number that it found was, as we had all hoped and expected, small.   That number was about 1%.

That was only part of the clinically important information that the study produced.  After all, people might say that a 1/100 risk of MD is enough to justify the continued blind treatment of all such children.

What if you could safely tell who to treat and not to treat though?  The PiC study had enough information about clinical features and outcomes to be able to test the validity of any guideline.  They simulated what would happen if all the children in the PiC study were managed according to a guideline's algorithm.  This allowed them to see how sensitive and specific each guideline is.

The NICE guideline (2) recommends treatment of all febrile children with non-blanching rash.  So no surprises that the PiC study found the NICE approach to have 100% sensitivity but only 1% specificity.  99% of children treated in this way have unnecessary tests, treatment and time in hospital. 

Most major paediatric emergency departments in the UK have been deviating from the "treat every time" approach for many years.  Most centres have guidelines which use a combination of clinical assessment and the use of inflammatory markers to select which children will not be treated.  The PiC study also evaluated the sensitivity and specificity of 6 of these local guidelines.  What this showed was that these guidelines retained 100% sensitivity but improved specificity.  The best guideline (Barts London) achieved a specificity of 36%.  That means that a lot of children are safely avoiding unnecessary treatment and time in hospital.

If you are thinking that this is all very nice but changes nothing for the pre-hospital clinician, the best bit is still to come.  A guideline that wasn't included in the PiC study was the Sheffield Children's Hospital Emergency Department (SCHED) Handbook.  (I believe that the reason that it was not included at the time was that the guideline was being changed.)  The direction of that change was away from using inflammatory markers as part of the decision making process.  The SCHED (3) guideline uses pattern recognition and experienced decision making.  Blood tests are not a recommended part of the process outside of specific circumstances (e.g. diagnosing haematological cause).

Although this guideline is not one of those in the PiC study, it has since been applied to the PiC study dataset of 1300 children with fever and non-blanching rash.  The exciting result of this is that the Sheffield guideline also retains 100% sensitivity (95% CI 82-100%) but achieves an even higher specificity at 69% (95% CI 66-72%).  (3)

The exciting thing about this approach is that it is a decision that can be made anywhere.  What the decision is made up of is the following

• Continuing default treatment in a few cases (rare but important)
• Fever and purpuric rash
• Fever and petechial rash and clinically probable sepsis
• Identifying other possible causes (such as mechanical cause from vomiting) using pattern recognition to identify those at low risk
• For those children who do not have another diagnosis or deemed to need default treatment, allowing an experienced decision maker to choose whether to treat or discharge.

Here is an adapted flowchart from the Sheffield Children's Hospital Emergency Department Handbook:

The next question you might be asking yourself is, "Am I an experienced decision maker?" when it comes to the child with fever and petechial rash.  While there is no simple answer to this, the likey answer is yes if you have 5-10 years of postgraduate experience in a role that includes decision making about unwell children.

Everything has changed in the management of the well child with petechial rash and fever.  Thanks to vaccination and high quality research, we can take a very different approach and avoid overtreatment of what is now known to be a low risk clinical presentation.

Edward Snelson
@sailordoctor

References

1. Waterfield T, Maney J-A, Fairley D, Lyttle MD, McKenna JP, Roland D, Corr M, McFetridge L, Mitchell H, Woolfall K, Lynn F, Patenall B, Shields MD, Validating clinical practice guidelines for the management of children with non-blanching rashes in the UK (PiC): a prospective, multicentre cohort study, The Lancet, November 2020
2. NICE. Meningitis (bacterial) and meningococcal septicaemia in under 16s: recognition,diagnosis and management | Guidance and guidelines | NICE. 2015 [cited 2017 Oct 10]
3. Snelson E, Waterfield T, Testing the limits of pragmatism in children with fever and non-blanching rash, Correspondence, The Lancet March 2021

There, I've Said It: There is No Such Thing as Early Sepsis

Words are really important when it comes to communication.  Certainly, non-verbal communication counts for a lot, but words are very powerful ways of getting a message across.  We should be responsible about how we use them.

Diagnostic error in paediatrics is a very emotive issue.  If a clinician is wrong (always in retrospect) and a child is involved, it can be difficult to be objective.  No-one is right all the time though, and diagnoses  that are later apparent are sometimes easy to miss or even too elusive to reasonably detect in their early stages.

There is one important exception to this: sepsis, because there is no such thing as early sepsis.  This is really important because clinicians are both criticised and self-critical when a child has been diagnosed with sepsis and there was an earlier clinical contact.

You're going to want me to back this one up no doubt.  I haven't got any academic references, but I don't need them since it's a case of simple logic.  I'm going to Spock the heck ot of the idea of early sepsis.

Starting with the terminology that we are working with, there is already difficulty with the definition of sepsis.  Ever since the word ceased to mean infection in any form and came to be used for an effect of the infection, sepsis has remained impossible to decisively define.

The best definition that we have of sepsis "a life-threatening organ dysfunction caused by a dysregulated host response to infection," comes from the 2016 Third International Consensus Definitions for Sepsis and Septic Shock.  This definition is inherently subjective.  There is no formula and no test.  Note also the lack of attempt to define early sepsis at this event.

Why do I care if people use the term early sepsis?  Because words.

The term early sepsis implies that there is a clearly definable and therefore recognisable entity which has therefore earned its own name.  Implications are more dangerous than overt statements because they go unchallenged and become part of the profession's assumption that because a term is used, it must be valid.

The harm of the term being given validity is that (again with the logic) if it is a definable entity, whenever a child is diagnosed with sepsis and they had an earlier contact with a clinician, that person has failed to recognise early sepsis.  In reality, they will have seen a child with features that are attributable to sepsis.  These features are also commonly found in children who are febrile but not sepstic.

How damaging is that?  Apart from the medicolegal implications, the negative impact on a clinician's confidence and reputation is potentially huge.  Using the term early sepsis risks leaving a string of second victims in its wake.

Just to be clear, I am 100% behind the idea of recognising sepsis early.  Swapping those words around is all it takes to make them functional again.  I'm also all for anything that improves the early recognition and treatment of sepsis.  So far, no strategy has proved successful in achieving earlier diagnosis of sepsis in children.  Awareness, careful clinical assessment are key, as is treating every illness as dynamic.  That is why appropriate observation and good safety-netting are key interventions.  Whether the child is managed at home or in a health-care setting, no illness is 100% safe until the child is better.

The focus on sepsis over the past decade or so has improved the timely treatment when sepsis is diagnosed, but making that decision in the first place remains a complex business.  Here's a nice oversimplifiaction of that process:

The red and green patients are relatively easy in terms of decision making.  The amber patients represent a small uncertainty which needs to be managed expertly.  In the small proportion that later become red and therefore relatively easy to define as having sepsis, retrospectively calling the preceding illness "early sepsis" defies logic and undervalues the difficulties of managing a large volume of moderately unwell children.

Next time you hear someone talk about recognition of early sepsis, politely challenge them and explore whether they mean early recognition of sepsis.  Sepsis is a thing.  The point when an illness goes from not sepsis to sepsis is not sudden and therefore easily missed.  Implying that there is a clearly definable and recognisable thing called early sepsis risks the vilification of front line clinicians in both primary and secondary care.

Edward Snelson
@sailordoctor

What am I missing? The child with fever but no obvious cause

The fear that a significant illness wil go unreconised in a child is one that is always present for the clinician who sees acutely unwell children.  The stories that we hear of infections and other illnesses being "missed" fuels that anxiety.  The common sense side of us tells us that significant illness should manifest itself in an obvious way, but that doesn't stop us from asking the question, "What am I missing?" when we see a child with a febrile illness and no apparent cause?

Scenario

You have just seen Billy, a 1 year old child with a fever that started today.  They have no cough or runny nose.  They appear well and have a heart rate of 120.  Chest is clear, heart sounds are normal and abdomen is soft.  Tympanic membranes are not inflamed.  You may or may not have looked at their throat but if you did, there is nothing obvious to see.

What do you do?

The answer to this is to understand what the possible causes of fever are, know how to exclude them and have an idea of how likely they are.  The latter brings us onto an important question:

How likely is significant or dangerous infection in a child?

That depends on the child.

The most common risk factor that we encounter is the infant.  The likelihood of an unwell newborn having a significant infection is high.  This is further compounded by their non-specific symptoms and lack of physiological response in the first few weeks of life.  The risk of serious infection multiplied by the risk of underestimating the illness makes a baby under the age of 60 days a high risk patient.

After those first few weeks, the infant becomes less reliant on maternal antibodies and begins to produce a more vigorous response to infection, most of which are now viral.  As a result, the risk diminishes inversely.

Other risky patients are those with ongoing reasons to either have more significant infections or less obvious signs of serious illness.  These include children with neurodisability, immunodeficiency or chronic illness.

For the usually healthy child beyond early infancy, the very great probability is that an illness will be benign and that those infections that are dangerous will manifest themselves in some significant way.  This itself brings a challenge: complacency.  We become so used to good outcomes and fruitless investigations that we start to think that everything is an uncomplicated viral infection.
It usually is, but what if it isn't?  That brings us on to the next question:

What are the less common causes of fever in a child?

One way to think about the causes is within categories:

When there are so many possibilities, it is often best to consider the least common first.  Let's start with the non-infective.  These illnesses cause inflammation without active infection.  They are all very uncommon compared to other things on the list but that makes them easy to forget and therefore miss.

Kawasaki Disease - This is a vasculitis which can look like a prolonged viral illness.  The cause has not yet been identified but it is presumed to be a post-infective phenomenon.  If a child has had a fever for five or more days without a clear cause, we should check if the child fulfils the criteria for Kawasaki disease.  You can check the criteria in a book or use an online tool such as this one linked here.

Leukaemia - Haematological malignancy in children occasionally presents as an unexplained and prolonged pyrexia.  More often there are other symptoms and signs such as increasing lethargy, weight loss, pallor, bruising, bleeding and unexplained pains.

Systemic Onset Juvenile Idiopathic Arthritis (JIA) - this subtype of JIA is rare but is one possible cause of unexplained fever.  Often the fever is accompanied by a typical salmon pink rash and joint pains, even if there is no clinically obvious joint swelling yet.

These non-infective causes of fever should not be over-thought.  The presenting symptoms of childhood illness are often so non-specific that it can be all to easy to imagine zebras instead of horses.  The key to not missing these is to be aware of them as entities and to look for features of these in the child with unexplained fever, especially when that fever is prolonged.

Next, the sepsis question.  Every febrile child should be assessed for sepsis, whether a focus the fever is found or not.  That decision can be made easy for you in one of two ways.  Either the child is very well to the extent that sepsis can be ruled out, or the child is so unwell that sepsis is presumed.  Everything in between is a case of careful assessment, including risk factors and the trajectory of the illness.

So, if we have considered the very rare and the sepsis question, what we should be left with is a child who we think does not have sepsis and yet has a fever without an immediately obvious focus.  At this point we return to the list of possibilities.

The task in a child with fever and no clear focus is to rule these possibilities out, which is usually based on clinical assessment.  Start with the complications of upper respiratory tract infection (URTI) as these are the most common significant infections in children.

Mastoiditis - infection of the mastoid is usually a complication of otitis media infection there should be evidence of that.  Mastoiditis is excluded clinically if there is no erythema, swelling or significant tenderness of the mastoid process.

Peritonsillar abscess - A collection of pus in the peritonsillar tissues is manifested by swelling which displaces the tonsil.  At the time of writing this, throat examination is not routinely performed due to the COVID-19 pandemic.  However, peritonsillar abscess is highly unlikely in a child who is willing to drink or eat.  If the child is refusing all oral intake, it may be necessary to use droplet PPE, including eye protection, to visually exclude peritonsillar abscess.

Lymph node abscess - Inflamed or enlarged lymph nodes are a common finding in children with URTI. Occasionally, the lymph node becomes bacterially infected.  When this happens, the lymph node is more enlarged and painful.  The overlying skin is often erythematous.  Another common feature is that the child becomes reluctant to turn their neck due to the pain from inflammation of the surrounding tissues.  These infected lymph nodes may respond to high dose oral antibiotics, however they may require incision and drainage.  Discussion with or referral to ENT is therefore advisable.

Osteomyelitis and septic arthritis - This is a good example of something that is rare but also often missed when it is a cause of unexplained fever.  Infection in a bone or joint can be visible or hidden.  If a parent has noticed a swollen, red or hot area or that the child has localising signs in a limb, that can lead to early diagnosis.  It is also the case that in a significant number of cases, the infection is not identified early on.  It is no surprise when a febrile child is miserable and moves less.  It is not common practice for clinicians to examine every bone and joint in a febrile child.  However, this is something that needs to be done if a child has an unexplained fever.  If limbs have not been examined for swelling, hot spots or erythema at first presentation, I would suggest that this should be done at the second assessment should fever persist and remains unexplained.

Urinary tract infection (UTI) - UTI is probably the most common cause of fever without a clinically obvious focus in children.  The younger the child, the less likely they are to present with specific symptoms.  Fever without obvious cause is an indication to screen the child for UTI.  Blind treatment with antibiotics is not recommended.  Urine should ideally be sent for culture so that treatment is based on the most robust result - a significant bacterial growth.

Meningitis and encephalitis - Central nervous system (CNS) infection is the most feared of the causes of fever without focus.  In the younger child, symptoms are less specific.  Infants may be irritable, jittery and not feeding well.  Vomiting and excessive sleeping are also common features but again, non-specific.  In an infant with an open fontanelle, this should be examined.  A bulging fontanelle (when not crying) is a red flag sign.  Older children may exhibit classical signs of neck stiffness, headache and photophobia.  Younger children are more likely to stand out because they just won't settle or have an abnormal tone or conscious level.  CNS infection is usually ruled out by the child demonstrating normal interaction or behaviour, often after adequate analgesia has been provided.

Appendicitis - In an older child, recognising appendicitis is done in a similar way to adult practice.  Appendicitis is rare in younger children but when it does occur, it can easily be missed.  Guarding tends to be a later sign in the pre-school child because their abdominal wall muscles are not very strong.  Children often cry or otherwise appear distress when their abdomen is examined, leaving the clinician uncertain.  Analgesia and reassessment is a good way of clinically ruling out appendicitis if the initial assessment is ambiguous.

Pneumonia and empyema - Lower respiratory tract infection (LRTI) is common in children.  Cough and fever are non-specific symptoms and are not grounds for diagnosing LRTI on their own.  Hearing crepitations on auscultation is also a common finding that should not be given too much weight.  Many LRTIs in children are viral and self-limiting.  Important discriminators are how unwell the child is, their work of breathing and more specific focal signs such as localised reduced air entry or a dull percussion note.

Tropical diseases - If a child has an unexplained fever and has recently returned from an area with e.g. malaria, they need to be referred to secondary care for investigation.

And finally...

So if Billy looks well and behaves in a way that effectively rules out sepsis and meningitis, his fever without clear focus means that we should look just a bit harder.  A urine sample should be taken to exclude UTI; blind treatment with antibiotics is not recommended without good evidence of UTI.

If there are reasons to suspect one of the less common (than uncomplicated viral infection) causes of fever, referral to secondary is likely to be the way forward.  If there is no evidence of a significant cause and what you are left with is a reasonably well child with an unexplained fever, the final question is, "should I refer this child or send them home with safety-netting advice?"

Both options are valid and the choice should be made in the best interest of the child.  In secondary care, the assessment of the child should be clinical in the majority of cases.  As such, referral may simply add a further clinical history and examination.  If a second opinion or physical period of observation is felt to in the child's interest, that is fine.  If not, it may be best to keep the child away from hospital and the risks associated with a secondary healthcare setting/

Before a final decision is made, risk should be considered.  For the majority of healthy children presenting to Primary Care (including the Emergency Department) with no specific risk factors, the likelihood of any febrile illness being a serious bacterial illness is very low.  That makes it perfectly reasonable for a child who has had a careful clinical assessment to be managed conservatively and with good safety-netting advice.

There are children who have a significantly higher risk.  As mentioned above, the most commonly encountered risk factor is the baby.  If your patient is a baby, especially if not yet started on their primary vaccinations, fever without focus warrants a referral to paediatrics.

Fever in a child who does not have an immediately obvious focus is a clinical conundrum for all of us.  Many children can be managed with a thorough history and examination.  If there are significant risk factors or specific findings then appropriate referral is likely to be the next step.

Edward Snelson
99% Type 1 decision maker
@sailordoctor

Disclaimer: If it's the clinician who has no focus, there's nothing I can do for you.

Paediatric Warning Scores - why they are always right but never the answer

This decade has seen a massive rise in the use of early warning scores.  For many clinicians early warning scores are a thing with which they are well familiar.  For those who were in practice before early warning scores became commonplace, they can confusing. For those wondering what the place of these scores is, it is important to be clear about what they are and what they do.

What is a Paediatric Early Warning Score and which is the best?

Paediatric early warning scores are ways of flagging patients up for further assessment and guiding the urgency of that assessment.

The original paediatric early warning score is called PEWS and exists in many forms.  It tends to be chart based and was designed for monitoring of inpatients.  The idea behind it was to take a traditional observation chart and give thresholds for concern and action to the numbers being measured.  The reason for doing this was an awareness that sometimes observations were recorded but that deterioration or severity of illness was not acted upon.

PEWS charts have colours to show how deranged the numbers are.  Anything not in the green is used either as a threshold in its own right (i.e. a red parameter for one physiological variable has a required action) or is part of a score (i.e. added to the values of other parameters) or both.

Because children have differing physiology* at different ages and because each clinical environment is different, it is essential to use a PEWS chart that is appropriate to the child and for the actions to be appropriate to the place where it is used.

Although PEWS was designed for ongoing inpatient monitoring, it has been used in other environments including primary care community settings such as out-of-hours care providers.

A relatively new kid on the block is the Paediatric Observation Priority (POPS) score.  This scoring system includes a score for nurse/clinician gut feeling and a somewhat subjective score for breathing, making it an interesting blend of measured parameters and clinical judgement.

The other big difference with POPS is that it was designed for use in the front end of an Emergency Department.  Essentially, it was crafted to identify those patients who were more likely to need early senior decision making and early investigation or treatment.

Because it is not an ongoing observation chart, it can all cleverly fit on one page which has the numerical values* for each age.  For all of these reasons, POPS is increasingly used in community settings as a front door assessment tool.

*Note the careful avoidance of the word normal when referring to physiological parameters in children.  There is no normal, just numbers.

I wouldn't say that any one is better than another.  They have different usability and different emphasis.  What works best will vary by people and place of use.

Where a snapshot assessment is needed for high volumes of patients, I find that POPS works well.

What does the score mean?

The obtuse answer to that is, "A warning score in itself is not an answer."  The score was never designed to be an answer.  Instead it is part of a process that asks questions.  A score is simply a way of assigning a numerical value to a set number of variables.  It standardises a snapshot but in no way replaces appropriate clinical judgement.

The questions that are being asked are:

• Is this child seriously unwell?
• Who should be assessing this child?
• Is the assessment time critical?

Of course these are questions that have always existed before scores were invented.  So if the score just asks the obvious questions, what is the point in them?

If the score is high but I am happy with the patient, does the score prove I am wrong?

No.  Let's prove that with an example.

Scenario 1 -
You go to see a child who has a POPS score of 5...
When you look at the breakdown of how they got  that score, they are a wheezy 3 year old with slight increased work of breathing (+1), a heart rate of 160 (+1), respiratory rate of 24 (+1), and Temp of 38 (+1). They were born prematurely at 32 weeks (+1) but have no ongoing complications of this.
When you go to see them, they run past you shouting, "I'm Spiderman!"
You complete your assessment and conclude that they have an uncomplicated viral URTI with a moderate viral wheeze.  It responds really well to salbutamol.
Four hour later they are tucked in bed at home with no respiratory distress.

Scenario 2 -
You go to see a child who has a POPS score of 5...
When you look at the breakdown of how this unwell 6 month old got  that score, they have a temperature of 37.9, heart rate is 176 (+1), a respiratory rate of 40 and they are grunting (+1).  They are subdued but if someone makes a noise the child produces a weak cry (+1).  They look seriously ill and the person completing the initial POPS score gave them 2 points for gut feel (+2).
Completing you assessment you decide that they are probably septic and act accordingly. Within 4 hours they are on a paediatric critical care unit, ventilated and on inotropes.  Two weeks later they leave hospital and make a full recovery.

The score indicated a need for an early and expert decision to be made.  In the case of scenario 1, the decision that the child was essentially well was entirely valid.  The score was valid also.

A score which is made up of these parameters cannot be wrong.  It is just a score.

Should I be using POPS scores in my clinical practice?

That depends on what you are using the score for.  If the score is to identify which of several children need to be put in front of a senior decision maker and how quickly that needs to happen, then that is a good use of a POPS score.  If the score is intended to identify a possible deterioration of a patient who is being monitored and observed, then this is what PEWS was designed for, though some will use POPS in that scenario.

If you have already made a clinical decision about the best course of action, then what do you want a score for?  After all, the score is simply a volume control for a question.  If you have already decided what the answer is, what is the score for?

Edward Snelson
Qualitative quantifier
@sailordoctor

Comment from Dr Damian Roland - Consultant in PEM at Leicester Royal Infirmary and one of the original developers of the POPS score:

Many thanks Edward - great piece and love the volume analogy! I use a very similar example of two different patients with the same score and what this means. It's an important example as children with higher scores (if admitted) have longer length of stay (based on the initial assessment POPS). A child with a score of 5 and above if admitted is likely to be in hospital for over 24 hours and also has a slightly higher risk of return if discharged (but the latter association actually quite weak). But lots of children with a POPS of 5 at presentation are safely and correctly discharged. So POPS helps highlight a patient who needs an intervention (which may be something as simple as senior review) but shouldn't tell you what to do.

I try to emphasise that the patients presenting to an Emergency Department are usually in a pre-treatment phase of their illness (they've not had medications) and therefore the range of possible acuities is very wide (which is why POPS is 0-16). This is different from an inpatient PEWS which is looking to identify children in a post treatment phase who are deteriorating. The bandwidth of PEWS (typically 0-7/8) is therefore smaller.

Finally POPS was designed to also help support decisions on 'wellness'. A POPS of 0 has a very high positive predictive value for being discharged without subsequent return and admission. It's the volume analogy in reverse. A POPS of 0 screams: why are you admitting me? And in some cases (DSH, Risk from specific past medical history, Social Concerns) you will do but it's the cognitive prompt that's the important thing.

Thanks Edward for shining a light on this important topic. I am happy to share copies of my paper explaining POPS "Scoring systems in paediatric emergency care: Panacea or paper exercise?" https://onlinelibrary.wiley.com/doi/abs/10.1111/jpc.13123 on request.

How do we diagnose sepsis in children? The Sepsis Jigsaw

Sepsis in children is something that we all fear.  It is difficult to define and  difficult to diagnose early.  This millennium has seen a huge rise in the presence of sepsis in education, campaigns and guidelines.  I believe that one of the reasons that we're talking about it so much is that we're still trying to understand what we mean.  Within that, we are trying to find ways to explain some of the things that we know.  That is because a lot of what we know about recognising sepsis is tied up in tacit knowledge.

Tacit knowledge refers to the things that we know but are not easily explained.  For example, it is  difficult to explain all the elements involved in driving a car.  Much of what we do in our lives relies on tacit knowledge.  How do you find things?  How do you figure things out?  These are far easier to do than to explain.

The very nature of the recognition of sepsis makes it something that needs completely taking apart and putting back together.

Sepsis is not easily definable in the first place.  2016 saw the Third International Consensus Definitions for Sepsis and Septic Shock (1).  This came from a process that involved two previous attempts to find consensus definitions, a recognition that none of the previous definitions were perfect, and a third brave attempt to find a definition for something that is somewhat amorphous.

The resulting definition:  "life-threatening organ dysfunction caused by a dysregulated host response to infection" is a good one and I would agree with it.  However, it does little to help us diagnose sepsis in children.  Recognising severe sepsis is not a great challenge.  Recognising early sepsis in children is very difficult because of the way that children respond to illness.

There is a bit of a misunderstanding that could result from many of the recent guidelines and publications about recognising sepsis in children: that fever plus tachycardia equals sepsis.  Since febrile children are routinely tachycardic, this does not make sense.  The misunderstanding comes from a retrospective approach to guideline definitions of sepsis.  If you look at all the children who were diagnosed as septic, what were the common features at presentation?  Abnormal temperature (high or low) and tachycardia come up a lot.

There are two sides to this coin.  Sepsis in children is not simple.  It is difficult to recognise and thwarted by many biases.  Yet it is deadly and anything that we can do to improve our recognition of sepsis is going to save lives. So complexity is no reason for complacency.

Since we don’t have a retrospectoscope when we see our next patient, we need to have a good way of recognising possible sepsis and serious bacterial infection (SBI) amongst the large numbers of children with uncomplicated illnesses.  If fever and tachycardia are not specific, what can we rely on?  Despite hopes to the contrary, routine near patient testing (e.g. CRP) in a primary care or emergency department setting will not give us the answer.

If neither numbers nor tests can sort the few out from the many, what is left?  Simply put, a global assessment made by an experienced clinician is what really brings the magic to the decision making.  So what is it that helps them to make a decision?  The answer is complicated but essentially, they put together a jigsaw of features and come up with enough of a picture so that the puzzle makes sense.  Some of the jigsaw pieces are fairly obvious but some of them are less well known or involve that tacit element of the process.  It is worth being aware of the various factors that influence this crucial decision.

The pieces of a sepsis jigsaw puzzle:

Temperature

Abnormally low or high, infection will affect temperature in some way.  This is an oversimplification which fails to address some of the subtleties of temperature and its relationship to bacterial infection and sepsis.

Factors to consider are:

• Low temperature in the context of an unwell child is more indicative of sepsis
• The relationship between height of temperature and sepsis/SBI is loose.  Although there is a correlation between very high temperatures and SBI, it is a weak one.  Children with viral infections may well get temperatures over 40˚C.
• Temperatures that are more persistent or fail to come down with antipyretics are often seen as more concerning.  Again, this is a poor discriminator as this can be seen in viral illnesses.  However, it is also true that a child with a persistent temperature may not get the opportunity to demonstrate their wellness by having a little run around.
• A normal temperature at the time of assessment does not rule out sepsis.

Circulation: Heart rate, central capillary refill and peripheral perfusion

The normality of these factors is quite rightly reassuring.  If outside of a reference range, these features may or may not be significant.  Each of these factors can be affected by pain, fear, pyrexia and environment.  Again, the extremeness of the abnormality is a consideration as is the persistence of deranged markers of circulation.

Respiration: Respiratory rate and work of breathing

Abnormal respiration is more discriminatory for SBI and sepsis, assuming that there is no other reason for being unwell and breathing abnormally (e.g. viral wheeze).  The reason for this is that respiration is less prone to the physiological changes that affect circulation.  Abnormal breathing may be caused by acidosis or hypoxia but is less likely to be due to a simple illness.  This ties in nicely with the definition of sepsis that relates to organ dysfunction.  While circulation changes may be a reaction to an uncomplicated viral illness, respiratory changes are more likely to be due to organ dysfunction.

Significant episodes

Since we might only see the child for a few minutes, it is important to take seriously any significant events that have occurred recently.  Pale, floppy or blue episodes are all notable.  Shivering and shaking are also worth taking into account.  They are not in themselves proof of serious infection.  Any of these things can occur during a temperature spike in an uncomplicated viral illness.  Remember that each of these is only a piece of a jigsaw.  You need to look at the whole picture and if the child is now running around pretending to be Spiderman, they’re probably OK despite the thing that happened.

Fluid balance

A well hydrated child (wet mucosa etc) who is drinking well and has good urine output is what you are looking for here.  Where these things are not adequate, sometimes all that is required is analgesia and a fresh start.  It all depends on how the rest of the pieces of the jigsaw are coming together as to whether it is time to go down a particular path.  Dehydration and poor urine output combined with other features is more significant.

Activity, behaviour and interaction

Now we are truly into the area of tacit knowledge.  (I wondered when he was getting around to that...)   Very little is published about the relationship between a child’s ability to smile, play, run or do anything for that matter and their risk of having SBI or sepsis.  However, it is reasonably intuitive that a child who runs in, smiles and talks the hind leg off of you is less likely to have sepsis than a child who is carried in, interacts little and looks miserable.   These factors rarely feature meaningfully because they are impossible to quantify.  Each appraisal is as different as each child is unique.  I couldn’t tell you what my threshold for ‘active’ or ‘interactive’ is because it will be specific to the child and depends on factors that I could not explain easily.  That is tacit knowledge in a nutshell.  While no-one can tell you what you are looking for in this category, it is an important piece of the jigsaw and should be give the weight it deserves.  Your instinct here is vital.
If you use these things in your decision making then that is completely normal.  An article in Archives of Disease in Childhood this year (2) published a consensus of which behaviours are seen to indicate that a child does not have sepsis.

Parental anxiety

More tacit knowledge here folks.  We will ask about symptoms and are looking to get some fairly specific answers.  Much of what we want to know will feed into the features already mentioned.  However, there may be things going on that a parent will struggle to articulate.  It is our job to distinguish between unwarranted anxiety (“I saw that news story about the child who died of sepsis…”) and the anxiety that comes from  a parent knowing that something is deeply wrong and being unable to articulate the reason why they know that.  The latter is the parent’s own tacit knowledge being given to you in the form of a person who cannot be reassured.

The trajectory of the illness

I believe that this may be one of the most important yet least discussed pieces of the jigsaw.  No one has told me about it and it may be that no one has ever told you, but when I say it, your own tacit knowledge about assessing unwell children will hopefully agree with the following statement:  An illness that has extreme fluctuation in symptoms (i.e. very unwell followed by surprisingly well) is almost certainly an uncomplicated viral illness.  I am talking about the “you wouldn’t believe how unwell they looked” kind of illness.  Sepsis and SBI don’t give you time off.  Viral illnesses, it seems, do.  So much so that a child who was floppy and lethargic can within the hour be smiling, playing drinking and complaining that they don’t want to go home because they want to play with the toys that you have.  It’s not in the guidelines but it is very important because the opposite is also true.  Two children can have the same heart rate, temperature, hydration and appearance, but the one who hasn’t had a return to normal in the past few hours is the one to really worry about in my opinion.

Many of these jigsaw pieces are the more quantifiable and traditional features that guidelines rely heavily on.  The rest are more woolly and difficult to define, let alone describe.  These are the pieces of the jigsaw that only you, the experienced clinician, can piece together.  If you would like to do a bit more reading about decision making in paediatrics, here is an article published in ADC (open access) (3) which further explores that issue.

Interestingly, there is a paediatric decision tool that takes into account some of the tacit knowledge features described here.  The POPS (Paediatric Observation Priority Score) includes features such as gut feel alongside physiological values (4).  This scoring system is both simple and over-simple in equal measure.  While it is quick, easy to do and validated, it only gives you a number at the end, not an answer or a diagnosis.  That number tells you to look at the jigsaw and see what the numbers mean.  The higher the number, the harder and longer you need to look and the better the explanation you need in order to be happy.

The other thing about POPS is that it doesn’t include my much neglected feature: the trajectory of the illness.  I think I’ll make a modified version of POPS which includes this.  I’ll call it POPcycleS.

How do we disgnose sepsis in children?  It remains a clinical diagnosis, best made by someone who has all the pieces of the sepsis jigsaw.

Edward Snelson
Perpetually puzzled physician
@sailordoctor

Disclaimer - If there is a piece of the jigsaw missing, go back and reassess the child.  They have probably eaten it.

References

1. Singer M, Deutschman CS, Seymour CW, et al. The Third International Consensus Definitions for Sepsis and Septic Shock (Sepsis-3). JAMA. 2016;315(8):801–810. doi:10.1001/jama.2016.0287
2. Snelson E, Ramlakhan S Which observed behaviours may reassure physicians that a child is not septic? An international Delphi study Archives of Disease in Childhood 2018;103:864-867
3. Roland D, Snelson E ‘So why didn’t you think this baby was ill?’ Decision-making in acute paediatrics Archives of Disease in Childhood - Education and Practice Published Online First: 01 March 2018. doi: 10.1136/archdischild-2017-313199
4. Roland D , Lewis G , Fielding P , et al . The paediatric observation priority score: a system to aid detection of serious illness and assist in safe discharge. Open J Emerg Med 2016;4:38–44.doi:10.4236/ojem.2016.42006

You Better Think! - A three dimensional guideline for recognising the unusual diagnosis in the ill child (including Kawasaki disease)

When assessing ill children, it is easy to presume that the problem is an uncomplicated viral infection.  Most of the time it is.  The odds are severely stacked against a more significant diagnosis to the extent that it is easy to become overly presumptive.  This, combined with the fact that a simple and benign illness will share many features with a rare or dangerous illness means that spotting the unusual or harmful diagnosis is very challenging indeed.

Much of the work done on congitive and diagnostic error takes the errors and then works backwards.  For a long time there have been reports on the number of deaths in healthcare that are related to error. (1)  These are reverse-engineered and start from the point of the problem.  People died - what is the evidence that there was any flaws in the care/ diagnosis/ treatment?  This is very different from the alternative approach of:  People had a healthcare episode- what happened next?

Outcome based stats are dangerous in that respect.  If you have 10% more adverse events than your colleagues but see 50% more patients (because you're awesome at your job) then please come and work with me.  You might flag up as a dangerous clinician if someone looks purely at incidents rather than the big picture.

I think that the most effective clinicians are those capable of recognising well children and capable of changing gear when something is unusual.  This is sometimes referred to as type 1 and type 2 thinking as per the model descibed by Croskerry. (2)

Using this model, we are most efficient when we are thinking inuitively and making gut feel decisions (type 1 thinking) and most effective at making the more complex diagnoses and managing the most dangerous scenarios when we are more considered and thorough (type 2 thinking).

Let's use this example to consider a child with non specific symptoms such as fever, rash, lymphademopathy and pharyngitis.  The reasonable but also dangerous assumption is that the child has an uncomplicated viral illness.  The possibility of another outcome is small but the consequences of missing an alternative diagnosis are great.  So, we need to use type 1 thinking to be efficient and be prepared to go into type 2 thinking when needed.

The obvious questions are then, what am I looking for and when do I look for it?  Guidelines on the subjects of febrile children, URTI in children and recognising complications such as sepsis tend to be written as if the problem was one dimensional or that the same guideline could be used in every circumstance.  This is one of the reasons that guidelines can sometimes frustrate.  Clinicians don't think that way, so it jars when a "fits all sizes" guideline over-simplifies such a complex process.

Here's an example of something that is useful but fairly simplistic.

This tells us what normal and abnormal look like.  It does very little to tellus what it all means.  Stopping here would be fine if we are just going to tell people when to refer or not.  To do that safely, such guidance will invitably err on the side of caution.

What it fails to do is to address what may be causing the red flags or atypical findings.  While a diagnosis is not necessesary in order to make a decision to refer, having a suspected diagnosis helps us to get the right child to the right place at the right time.

Lets take two of the possible complex and dangerous diagnoses as examples.  A child has a febrile illness with conjunctivitis, phayngitis, swollen lymph nodes, a rash and is pretty miserable.  Good to know.  If I told you that the onset of symptoms was within the past 24hrs, would you consider Kawasaki disease? No.  If I told you that it was day 6 of the illness and that for the past 3 days the child was neither better nor worse would you think that the diagnosis was likely to be acute sepsis? No, but can we get a guideline to help us get there?

Since it is a factor in our decision making, we could add in the dimension of time and disease progression to our guideline.  If we did that I think that it could look something like this:

Even adding this dimension doesn't fulfil our need for something which maps to our way of thinking.  We now have the bit that focuses on the child in front of us and the bit that takes into account the real world where patients present in different ways, but many guidelines fail to take into account the fact that different diseases behave differently.  Worse than that, the differences can be subtle.

Guidelines often struggle to deal with the fact that medicine is a complicated subject.  Do you write a guideline for a clinical scenario (e.g. febrile child)?  If so, you need to include every possible cause and when to think of it.  Do you write your guideline about a specific disease (e.g. Kawasaki disease)?  If so, how will people know when to use the guideline?  If they have looked it up, they are 90% of the way there and the guideline is going to be more useful as confirmation and treament advice.

For these reasons, guidelines will never be a substitute for the need for clinical knowledge and understanding.  Our child with non-specific symptoms guideline needs to have another layer - specific diagnoses, what they look like and when to consider them.

We need guidelines to be both simple in order to be practical and complex because nothing is simple.  We need them to be based on real-world clinical practice and to be honest about the uncertainties inherrent to that.

The short answer to the child with non-specific symptoms?  Anything is possible, including Kawasaki disease.  Early recognition of Kawasaki disease is important as treatment will reduce complications.  So, you better think.  In fact, because type 1 thinking will do very nicely most of the time, but not all of the time, you better think think.

Edward Snelson
Occasional overthinker
@sailordoctor

Disclaimer: Over and under-thinking are both perfectly acceptable in the right circumstances.

References

1. Kohn LT, Corrigan JM, Donaldson MS. To err is human: building a safer health system.National Academies Press, 1999.
2. Croskerry P. A universal model of diagnostic reasoning, Acad Med. 2009 Aug