Dr Amanda Gwee is a clinician-scientist fellow in the MCRI Infectious Diseases and Microbiology group. Her area of research interest revolves around the appropriate dosing of antibiotics.
Associate Professor Matt Sabin is the Chief Medical Officer of the Royal Children’s Hospital in Melbourne. It was not in this role that we asked him to speak but rather in his clinical role as a paediatric endocrinologist running the largest tertiary hospital obesity service in Australia.
Professor Harriet Hiscock is a consultant paediatrician and post-doctoral research fellow. Amongst other roles she is co-director of the Unsettled Babies Clinic and it is with this role in mind that we asked her to speak at DFTB18.
Otis is a 3yr old boy presenting to the emergency department with fever, and purulent discharge from his left ear. He otherwise looks well, however, his mother mentions this is his third ear infection since he was born, and he always seems to have a cough and a cold. She asks you if there could be a problem with his immune system?
Some children seem to have constant ear, nose or throat infections during childhood. We know that for a small, but important minority of children this may be the presenting feature of a primary immunodeficiency. Let’s look at how these may present, when to think of it, and what to do about it.
It’s worth stating from the outset, that the majority of children with recurrent ENT infections will not have a primary immunodeficiency. There is a relatively common phenomenon called “physiological immunodeficiency of infancy”, where-by there is a natural nadir in immunoglobulin levels as maternal immunoglobulin fades, and the child’s own immune system has only just become able to produce immunoglobulin for itself. This is at its lowest between 3-6 months and normally resolves by age 1. However, fully developed protection against encapsulated organisms doesn’t reach maturity until between 2-5 years, and IgA production doesn’t reach adult levels until adolescence. It can be completely normal for young children to suffer 4-11 respiratory infections a year (depending on exposure, e.g. siblings, nursery, etc.)
When considering the characteristics of infections that should trigger suspicion for immunodeficiency, we should be thinking about:
Combined immunodeficiency disorders (affecting both cellular and humoral immunity), such as severe combined immunodeficiency (SCID), present in the first 3-6 months with severe, life-threatening infection. Unusually aggressive infections should prompt further investigation
Infections with certain pathogens can point towards specific diagnoses, including respiratory infections with Pseudomonas aeruginosa (think cystic fibrosis or primary ciliary dyskinesia), oral/oesophageal candidiasis (think HIV or chronic granulomatous disease), upper respiratory infections with Pneumocystis carinii (think HIV or other T cell deficiencies) or recurrent otitis/sinusitis with Neisseria meningitidis (think complement deficiency).
Finally, to a lesser extent:
This is the least predictive of immunodeficiency, given the discussion above. Very frequent sinopulmonary infections in younger children with encapsulated bacteria can be the presenting feature of the rare condition X-linked agammaglobulinaemia (XLA: boys who produce no immunoglobulins). In late childhood and adolescence, the same presentation in a milder form may be a sign of combined, variable immunodeficiency (CVID), which is a heterogeneous group of disorders of antibody production.
Other, rare conditions include chronic granulomatous disease (CGD) which may present with deep abscesses of the outer ear or mastoid, or HIV presenting with recurrent otitis media (normally with other associated features)
Some general guidelines have been produced by the Jeffrey Modell foundation for when to consider referral for immunodeficiency workup:
Although having a low specificity, they provide a useful framework when thinking of children with more severe infections than usual.
If considering referral, there are definitely some basic tests are useful to do first (if the child is severely unwell, don’t wait for tests to refer).
Full blood count
This is useful for ANY suspected immunodeficiency. Persistent lymphopaenia in a child <2yrs should prompt screening for SCID.
NB: It can be normal to have transient lymphopaenia or neutropaenia in isolation in young children following a viral illness. Incidental neutropaenia does not need repeat testing if there are no concerns about underlying immunodeficiency.
Immunoglobulins
IgG, IgM and IgA levels are useful to investigate children with recurrent ENT/airway infections.
It is also worth considering an HIV test if symptoms are consistent, but ensure you have a discussion with parents before testing.
If both FBC and immunoglobulins are normal in the setting of recurrent infections, it is perfectly acceptable to wait for 3 -6 months to see if the condition improves before referral.
Further reading: https://www.entmasterclass.com/ENT_Journal_2019_Interactive.pdf page 9
As one of the authors of the Australian Asthma Handbook who better than Professor Adam Jaffe to make things a little clearer for us.
Children have become global gastronomic explorers and are constantly trying unusual foods. It is not these that are generally of concern to the healthcare provider but regular, everyday foods that are found in nearly every larder or store cupboard.
We see lots of children with suspected infections. Modern microbiology techniques have opened up a huge array of tests: some new and expensive, but we are often still reliant on good old fashion microscopy and culture.
With so many tests so readily available, we need to think hard about diagnostic stewardship. This means testing the right patients for the right reasons. We must be wary of over-diagnosis, preventing confusion, anxiety or unnecessary treatment, and making choices that represent good value. Many tests can be expensive and are often not necessary to make management decisions.
With that in mind, let’s take a look at some of the most common diagnostic tests for infections, and when we should (or shouldn’t!) be deploying them.
Urinary tract infections (UTIs) are the most common serious bacterial infection in high-income countries. There are many departments where it is routine to set up every febrile child to get a “clean catch” urine as soon as they arrive. This is unwise, because it is VERY EASY to contaminate a urine sample from a clean catch. We have all seen children or parents putting their hands/feet/face in the bowl, and let’s be honest – if the child is sitting on the container, it’s basically directly under the body’s primary waste pipe.
Accepting a decent risk of false positives, we must aim to test only those who need the test. So when should we do it?
This is the primary indication for doing a urine dip, and it is a sensible one. However, still not every child with fever and no source needs a urine dip. Older children can report urinary symptoms, and the absence of these makes a UTI much less likely. In addition, by school age, UTIs in males with normal renal tracts become very rare, so urine testing also becomes less useful.
As a framework, urine dips should be performed in the following groups with fever and no source (assuming they have no risk factors for UTIs and have no red flags):
Outside of these groups, use your clinical discretion to decide if the pre-test probability justifies the risk of a false positive – take into consideration the child’s age, gender, duration of symptoms, how unwell they appear, and obviously if they have known risk factors such as renal abnormalities or previous UTIs.
This seems obvious – but it’s worth stating that once urinary symptoms are present (increased frequency, dysuria) you should dip the urine to check for infection, and it may be worth sending samples for MC&S even if they are dip negative in this scenario (you can withhold treatment pending results).
It is worth taking more care for children with non-urinary symptoms, such as abdominal pain or vomiting (which is probably not predictive of UTI). Once at school age (particularly in boys) these symptoms are unlikely to be a symptom of a UTI so a higher threshold for testing should be adopted.
Some people say that all children with rigors require urine testing. Rigors are not evidenced to have any influence on the risk of UTI (or any significant risk of bacterial infection). If there is another source of the fever, urine dip is certainly not indicated on the basis of a rigor alone.
For more information on relative risks for UTIs in younger children, the supplementary materials to the UTI risk calculator study make an interesting read.
This becomes a slightly more controversial topic, and decisions require risk stratification based on the age of the child. For example, a febrile neonate with bronchiolitis might be lucky to escape the full shebang of a septic screen anyway – and a quick in/out catheter is unlikely to yield a false positive.
The literature on this topic is a bit confusing because of varying definitions of UTI and bronchiolitis (some studies including any child with RSV detected in their nose). The most recent meta-analysis with more stringent criteria for diagnosing UTI found a rate of concomitant UTI with bronchiolitis of 0.8% – low enough that testing is not advised.
Bottom line: if an infant has a fever and a clinical diagnosis of bronchiolitis, then urine dip is not necessary in most instances – however this should be given strong consideration in infants <60d and should be performed in neonates.
For a full myth busting exercise in blood cultures, please read the recent DFTB post on this topic. Some things to bear in mind if you’re thinking of taking a blood culture:
When it comes to swabbing for microscopy, culture and sensitivity (MC&S), there is a golden rule*:
Do not swab any non-sterile site that you have not already clinically diagnosed as being infected.
A skin swab, throat swab, eye swab etc. will grow bacteria 100% of the time, because these places are non-sterile. They will often grow pathogens, because many pathogens are quite happy just being colonisers a lot of the time, and actually some of them are more often found as bystanders than as trouble-makers (Pseudomonas aeruginosa is a prime example – it is very rarely pathogenic in non-sterile sites). A positive swab does not diagnose infection.
YOU have to diagnose infection; a swab will just tell you what bacteria is causing it.
I would like to give a special shout out to gastrostomies at this point – just because they are “mucky” is not a good reason to swab. If you do swab it, you will find good old Pseudomonas (it loves playing in wet stuff). Skin and soft tissue infections are red, hot and inflamed +/- a bit of pus. Yellowish clearish greenish stuff is normally just serous fluid, so don’t worry about it and don’t swab it!
The same goes for babies sticky eyes. If you swab it, it will grow bacteria, but this tells you nothing about whether they are infected. Look for inflammation, if you find it then diagnose infection, treat empirically and send a swab if you are concerned about resistant bacteria.
*there are some exceptions to the golden rule, including burns and chronic wounds in immunosuppressed patients.
Before starting – let’s remember that you cannot diagnose a bacterial throat infection with a swab alone. If you are considering swabbing a throat for MC&S, you must have already clinically diagnosed infection.
Guidelines vary quite widely in their recommendations to swab or not swab when diagnosing tonsillitis. It is worth considering that a throat swab has a reasonable sensitivity for group A Strep, if performed correctly. Sadly – we are all dreadful at performing throat swabs in children (who are usually very good at not wanting a throat swab), and often get a good dose of tongue and palate. Not good.
A further thing to consider is that approximately half of all throat swabs positive for group A Strep just indicate carriage – you’ve found the bug, but it’s just a bystander.
This means that if you swab and haven’t found the bacteria, it might be there but you’ve missed it, and if you have found it, there’s a 50% chance it’s not causing the illness anyway…
If it’s extremely important you detect the presence of group A Strep (for example in populations high risk for rheumatic fever) then I would definitely do a swab. If it’s not (and it usually is not), then make your decision to treat or not on clinical grounds alone.
Also, remember that in children <4yrs group A Strep tonsillitis is rare and almost never causes complications, so if you’re thinking of doing a throat swab for a child in this age group you need to have a very good reason.
Respiratory tract infections are extremely common in children. There is a fair amount of controversy and disagreement about the role for respiratory virus testing. It can have several roles:
A group of children you should not test for respiratory viruses is anyone with cough and coryza. They do not need a test – they can be safely diagnosed clinically, and the presence or absence of a virus on testing does not change anything.
What about in lower respiratory tract infections? We can imagine that the discovery of a virus would prevent unnecessary antibiotics. However, respiratory viruses are common (even among non-hospitalised populations) and co-infection with bacteria is also common in viral infections. The presence of a virus does not preclude a bacterial infection. As such, their use in this context is contentious, and they do not appear to reduce antibiotic use.
For a thorough look at the principles and evidence of respiratory virus testing in children, I would recommend this excellent review paper.
One afternoon my team broke the news to three different families that their children had a non survivable condition. That same week I was involved with a patient transitioning to a palliative pathway focused on comfort. I returned home to utter the words, “She is so sweet, I hope she dies soon.”
For many of us, days like these, occur commonly.
Being a doctor is a privilege, an honour, a calling. Our jobs are stressful, diagnostically challenging, involve managing team members, and effectively communicating and engaging with different families whom have different needs. We are reliant on our knowledge and skills. What sets our job apart from other high stress environments is that any given day can involve death and dying. We see distressing conditions. Our day includes the uncommon, the unlucky and the unfortunate events of life. To the public these events occur few and far between, but for us it may be a daily occurrence -a relentless barrage of traumatic events, poor outcomes and sad stories.
The intensive care environment is difficult to navigate. The rates of burnout, mental health issues and self medication are high amongst our peers. 70% of junior doctors feel burnt out following a neonatal rotation. Strikingly, their (our) rates of suicide are twice that of the general population. Most of us have heard the words compassion fatigue. Some of us may even be familiar with vicarious trauma – the negative experience of working directly with traumatised populations. Compassion fatigue and vicarious trauma are on a spectrum. We initially may feel overwhelmed by our interaction but this can develop into symptoms of post traumatic stress.
At DFTB18, I spoke about some of the things we can do to reduce this happening to us, and the events above reinforced that message;
And remember, it’s ok not to be ok
For more on this topic of the difficulties of dealing with death and burn out hit up DFTB at:
A short story about death by Andy Tagg
Nimmo A, Huggard, P. A systematic review of the measurement of compassion fatigue, vicarious trauma and secondary traumatic stress in physicians. Australian Journal of Disaster and Trauma Studies. 2013;1:37-44.
Stress, burnout and vicarious trauma: looking after yourself. RACGP Webinar Series.
When a child is picked up by paramedics or brought into an emergency department, their weight is not always known and cannot always be formally measured. Many research teams across the globe are trying to find the best method to estimate a child’s weight, so medication can be dosed and equipment sized appropriately. Traditionally, age-based formulae have been used, but these are known to be very inaccurate. More reliable methods are available, however all require input of more information than just age, whether that be height, mid-arm circumference, a parent estimate or a smartphone image. You can find a summary of weight estimation techniques in this post from Andy Tagg. The question is, what method is sufficiently accurate and will work best in practice?
At the moment, Australian guidelines still use age-based formulae (namely the original APLS formula, weight = 2 x (age+4)). Even though they are very inaccurate, they have a number of advantages:
They are very quick. Most prescribers use these formulae in conjunction with resuscitation aids, emergency manuals or clinical practice guidelines which mean they do not need to remember the formula or do the calculation themselves as they are given a table with corresponding weight to age. Given age-to-weight conversions are often provided, staff do not need to be trained on how to gather the estimate. They do not require any additional equipment, which may be hard to find if an ambulance or emergency department rarely sees paediatric critical cases. An emergency department can predict the weight of the child that is about to arrive by ambulance if they have the child’s age, and can therefore start drawing up medications in advance.
Stress and cognitive load have been shown to be the key precipitating factors of human error in paediatric critical events. Human errors in these scenarios include significant medication errors, such as ten-fold errors (where 10x the medication is prescribed or administered because the decimal point is moved or the concentration incorrectly calculated). These have been shown to cause significant patient morbidity and mortality. Efforts to gather the further information needed to make the weight estimate more accurate (e.g. measuring the child, taking a sufficient quality image, finding a parent) increase the complexity of the weight estimation phase. Increased complexity is likely to increase cognitive load, and thus increase the risk of human error at all phases in the dosing process.
We need to find a weight estimation tool that can be used by anyone who might need to manage a paediatric critical event. This includes paramedics, junior medical staff and adult emergency department personnel that may need to manage patients before they reach a tertiary children’s hospital or paediatric emergency department. This means we need clear, easy-to-follow guidelines and associated training that can be rolled out broadly. It also emphasises the need to ensure we keep the cognitive burden as low as possible, as many prescribers will be in an unfamiliar, stressful situation, both of which further precipitate human error. Future protocols may also differ based on the paediatric emergency expertise and training available in that setting, for example, a paediatric emergency department may choose a more accurate method with higher cognitive load than an ambulance service.
Another important consideration is the time delay involved in each weight estimation strategy. Most events requiring weight estimation are time-critical in nature. It is important to not only consider the time involved in getting the estimate, but also the time needed to find the appropriate equipment, make subsequent dose calculations and prepare the dose for administration. This highlights the significant advantage of emergency departments being able to draw up medications prior to the child’s arrival, as having doses pre-calculated and pre-prepared would significantly reduce the time delay in drug administration.
Given rising rates of childhood obesity, we need to find a weight estimation strategy that will work for all body types and medication types. Some drugs should be dosed based on ideal body weight (IBW) whilst others should be dosed based on total body weight (TBW), depending on their pharmacokinetic properties. Similarly, dosing medication by TBW in obese children can lead to overdose. Sydney Children’s Hospital has given a nice overview to some of the adjustments which should be made for specific medications. However, adjusting weights for specific medications in a paediatric emergency may further add to the cognitive load.
Overall, the pros and cons of each group of techniques can be summarised in a table:
So, how important is it that we have an accurate weight estimate? And how important are other considerations such as reducing cognitive load and practicality (eg. speed, equipment and staff training requirements)?
Unfortunately, there is very limited data on patient outcomes available to help guide us. The small number of studies into the impact of weight errors look at incorrectly documented weights, such as where the wrong weight unit was recorded (pounds instead of kilograms) or where a decimal point was moved (6). No study has looked specifically at the harms caused by weight estimation error in paediatric emergencies. There is no suggestion that using the original APLS formula in Australia is currently causing harm to patients, but there is also no evidence that proves that it is not. Reducing error should always be the goal, however increasing the complexity of generating a weight estimate could increase the cognitive load, and thus increase the risk of more significant errors. When deciding on which weight estimation technique to use, we need to find a middle ground between accuracy and practicality with an emphasis on reducing overall cognitive load.
Wells M, Goldstein LN, Bentley A. The accuracy of emergency weight estimation systems in children – a systematic review and meta-analysis. Int J Emerg Med. 2017;1:1. Available from: https://intjem.biomedcentral.com/articles/10.1186/s12245-017-0156-5
Sutherland A, Ashcroft DM, Phipps DL. Exploring the human factors of prescribing errors in paediatric intensive care units. Arch Dis Child. 2019;0:1-8. Available from: https://adc.bmj.com/content/104/6/588.long
Khoo TB, Tan JW, Ng HP, Choo CM, bt Abdul Shukor INC, Teh SH. Paediatric in-patient prescribing errors in Malaysia: a cross-sectional multicentre study. Int J Clin Pharm. 2017;39(3):551-9.
Doherty C, McDonnell C. Tenfold medication errors: 5 years’ experience at a university-affiliated pediatric hospital. Pediatrics. 2012;129(5):916-24. Available from: https://pediatrics.aappublications.org/content/129/5/916.long
Foster M, Tagg A, Klim S, Kelly AM. Accuracy of parental estimate of child’s weight in a paediatric emergency department. Emerg Med Australas. 2019; in press.
Shaw KN, Lillis KA, Ruddy RM, Mahajan PV, Lichenstein R, Olsen CS, et al. Reported medication events in a paediatric emergency research network: sharing to improve patient safety. Emerg Med J. 2013;30(10):815-9. Available from: https://emj.bmj.com/content/30/10/815.long
James was 13 the first time I treated him in A&E. He was rushed in after being hit on the head with a metal bar, but he wouldn’t tell us what had happened. Some others had seen him, rushed him, started beating him with bats and bars – that was all he said he remembered. It was clear that James was already deeply involved in the world of gang violence, and it was equally clear that if he wasn’t saved from it soon this world would destroy him. We tried to get James to stay in A&E long enough to hear the results of his scan, but as soon as he got a chance he slipped out and back onto the streets. My fear was that before long the streets would deliver him back to us, only this time he wouldn’t be able to walk out again.
“Pick a card…any card”, as the saying goes.
I remember being the recipient of this common trick when a medical student, at a friend’s for dinner where a magician was a fellow guest. The scientist in me wanted to know how it was done, so I prepared to follow my card studiously. I chose carefully and deliberately, feinting to pick one card from the fan offered to me, before choosing – in my eyes – a less obvious card. However, after returning the card to the deck, the magician threw his arms up with a flourish, the pack hitting th eceiling with a thud, causing cards to scatter across the room. Amazingly, my chosen card was stuck to the ceiling, fixed in place by a drawing pin.
I made peace with the fact I was always going to choose that card, and applauded in awe.
A few years ago, during a lecture, I was introduced to the concept of the‘magician’s choice’ by Daniel Sokol (a medical ethicist and magician). He used it as an innovative analogy to illustrate how, in obtaining consent, the presentation of the different options is nothing like the free choice it is intended to be. Rather, the possibilities are presented, much like a deck of cards, only this time the ‘trick’ is that the clinician– whether consciously or not – weights information, making a certain choice more likely. I was reminded of this concept recently, after reading about a new app due to be rolled out in the English National Health System. This new platform, it is claimed, will enable patients to see waiting times at different emergency departments, the implication being that they will choose to go to the one with the shortest wait. Unsurprisingly, this sparked Twitter debate about whether or not this actually was a useful tool to help patient flow, with strong opinions on either side.
To my surprise, however, the assumption from everyone seemed to be that this app would offer choice. To which my first thought was: ‘choice, what choice?’
For many years, the NHS has offered explicit patient choice for specialist outpatient care, offering a range of different hospitals at the point of referral. In recent years, the system has evolved to display waiting times. Sounds good so far? To an extent, I agree. A transparent system that allows patients, in conjunction with their GP, to book in to an appointment – even obtaining a date and time – is a pretty simple but effective use of technology. My objection is that the offered choice is nothing but an illusion.
The patient is given little idea of the constraints through which their choice is being made. As an example, many of the children I see have a long-term condition (epilepsy). Effective, joined-up care is much more than a quick hospital consultation, achieved through a combination of having a point of contact via their specialist nurse, close liaison with school, tracking of development progress, and surveillance of mental health and wellbeing. This is best achieved through local, networked care, a system which has been carefully developed over many years.When I see a patient outside this framework, the care is not just difficult to deliver – in what is far from a delicious irony, it actually gets delayed. When I have to inform families that ‘I can’t access that information’, ‘they won’t accept that referral from me’, or ‘our nurse doesn’t cover that area’ (all recurring themes from my practice over the last year), I empathise with the fact that they took a decision that they thought would get quicker access to care for their child, only to now discover that delay was just shifted further down the road. There are potential ways around this – not least patient record systems that speak to each other. Yet that sort of change doesn’t happen overnight so, until it does, surely patients deserve better? Until then, this resembles the ‘three cup and ball’ trick. The patient believes they have options, when actually the system is too constrained to offer the truly free choice that is advertised. No matter how much you try to pick the cup with the ball under it, the pieces move and the magician ensures you never can.
Blowing the whistle: Kim Holt at DFTB19
Team DFTB. Blowing the whistle: Kim Holt at DFTB19, Don't Forget the Bubbles, 2019. Available at:
https://doi.org/10.31440/DFTB.20736
You may recall the headlines surrounding the case of Baby P. Back in 2007 a 17 month old boy died as a result of injuries suffered over months of abuse. During that ordeal he had been seen by the London Borough of Haringey Children’s services and multiple concerns were raised. But nothing happened. Not until it was too late. Eight years earlier the same council had failed to intervene possibly leading to the death of eight year old Victoria Climbie.