Same Child, Different Room, More Risk? #WILTW

This is the 152nd #WILTW

In just under 2 weeks time we will move into a brand new Emergency Department (ED).

Picture via @LeicChildHosp

Although daunting we’re really looking forward to it.  It’s an exciting prospect for our patients as we are optimising the ‘front door‘ of the patient pathway to improve clinical input and patient experience. We will further improve access of children and young people to an appropriate healthcare professional following an initial assessment in a common triage area. The co-location of services is a recommendation of the Royal College of Emergency Medicine and in the new build we will have General Practitioners, working along side ED staff (many who initially trained as Paediatricians), both having easy access to the Children’s Hospital admitting teams.

One of the consequences will be the need for senior staff to recognise the challenge of being able to stratify risk appropriately between patients assigned to to different clinical staff. Why? Well, whether they realise it or not, all clinical staff who make diagnostic decisions apply a form of Bayesian statistics (Great summaries by St.Emlyn’s and Casey Parker) in nearly everything they do.

Patients arrive with a ‘pre-test‘ probability of disease. This is essentially the total  number of patients who have a specific diagnosis over a given time period i.e. it might be that 10% of children who present to the ED have a chest infection. After taking a history and examining the child, and maybe preforming some investigations, a clinician will make a decision on whether they think a chest infection is present or not. Those with a high ‘post-test‘ probability of disease will be more likely to get treated (‘test‘ in Bayesian Statistics doesn’t necessarily mean a blood test but could be any number of interventions including simply what the person’s gut feeling is about the presence of disease).

For example if you work somewhere where very few children have urinary tract infection (let’s say 2%) and you see a child with no relevant history or symptoms  at all it’s really unlikely this child has an urinary infection. Their post-test probability will be even less than 2%. Conversely if a urine dipstick comes back positive this will increase their post-test probability of having an infection. But remember this is just probability. The mistake made by many is that a positive test means a positive diagnosis. No! It just increases the probability of having a disease – there are few absolutes in medicine.

Pre-test probability varies between clinical settings. The risk of sepsis in children who go to a General Practitioner is very very low. There are 11 million children in the UK and in 2012/13 only 1000 were admitted to intensive care as a result of severe sepsis (100 died as a result). This clearly makes it vital we have systems to recognise sepsis promptly and avoid the tragedies that have occurred in the last 5 years. But the challenge is that a General Practitioner may go through their entire career and never see a child with sepsis. The risk of having sepsis increases if you are seen in an ED (this isn’t because an ED makes you ill but because parents are a good judge of their children’s health so are more likely take them straight to the ED when they perceive them to be very unwell). Finally because the ED screens and discharges a number of patients with more simple illness your pre-test probability of having sepsis will be highest if admitted into hospital.

So back to our new department. The same child, with the same signs and symptoms seen by a General Practitioner, ED doctor or Paediatricians will be perceived as having a potentially different risk of illness by the 3 doctors. This is because they are consciously, or unconsciously, aware of what the pre-test probability is for their normal working environment. If the GP asks a question of the Emergency Department Consultant (as is likely to happen in a co-located department) that Consultant will need to acknowledge their different frame of reference of risk. While the phenomenon of differential risk assessment is not new (GP colleagues often phone for advice) this is going to be at at a very different scale and pace. Something everyone is going to need to be mindful of to protect patients from both over- and under- investigation.

What have you learnt this week? #WILTW

You can now follow WILTW on Facebook by liking this page . Browse previous posts here or insert your e-mail address in the box on the right hand side to receive future posts

One thought on “Same Child, Different Room, More Risk? #WILTW”

  1. Great blog (again!), Damian!
    As so often, it makes me reflect on geriatric practice. Many parallels, and differences.
    ?UTI is a ‘working diagnosis’ (I’m being charitable here) we geriatricians see often, and at least 50% are not that. By that time, the frail patient has been given broad spectrum antibiotics, often without blood cultures having been sent, rendering the patient at risk from C diff.
    We need to get those priors – and the (poor – ~50%) positive predictive value of urinalysis – out there, and/or more folk spending training time in geriatric assessment units!

Leave a Reply