Every paediatric doctor has a fear of sending home a young, septic child. We constantly review our practice and guidelines on how to identify the septic child amongst the thousands of febrile two-year-olds that present to ED. But what are the consequences if we do miss one?
This study looked at children with meningitis or septicaemia to see if they had presented to hospital in the days preceding their eventual diagnosis.
Vaillancourt S, Guttmann A, Li Q, Chan IYM, Vermeulen MJ, Schull MJ. Repeated Emergency Department Visits Among Children Admitted With Meningitis or Septicemia: A Population-Based Study. Ann Emerg Med. 2014.
This was a retrospective cohort study looking at all non-elective hospital admissions in Ontario, Canada over a five year period (following the implementation of the pneumococcal vaccine programme).
Included patients were:
- aged 30 days to 5 years
- with a diagnosis of meningitis or septicaemia
- had a length of hospital stay of at least 4 days (or died in hospital)
Further patients were excluded if:
- they had been discharged from hospital in the preceding 14 days
There were 521 children included in total.
There were two groups – those who had previous ED presentations (in the preceding five days), and those who were admitted on first presentation to ED.
The primary outcomes were: length of stay; critical care admission; mortality.
Secondary outcomes measured included: type of ED for previous presentation; time of registration in ED; and triage acuity.
125 of the 521 children (24%) admitted with meningitis or septicaemia had an ED presentation in the 5 days prior to admission.
Over 2/3 of the repeat presentation group were given one of the following diagnoses initially: fever; otitis media; upper respiratory tract infection; viral infection; gastroenteritis; urinary disorder, and seizures.
96.4% of patients returned within 72 hours.
Both groups had similar lengths of stay, similar rates of critical care admission, and similar mortality rates.
Those who were treated in a community ED without review by a paediatric specialist were more likely to be in the repeat presentation group.
29.8% of the repeat presentation group had their repeat presentations in different EDs. This emphasises that it is hard for us as clinicians to receive feedback on our patients and improve our practice. It’s important to have good communication between hospitals in similar geographic areas.
1. Did the study address a clearly focused issue?
Sort of – at the start, the study aim is set out as ‘How often do children visit the emergency department in the 5 days before diagnosis of meningitis or septicaemia?’. Actually what the authors are asking is ‘Does having repeat presentations prior to admission for meningitis or septicaemia affect outcome?’.
The authors do not directly answer the first question, although it appears from the data that all 114 children were admitted on second presentation to hospital.
2. Was the cohort recruited in an acceptable way?
Yes – retrospectively via the patient database using diagnosis at admission.
3. Were the outcomes accurately measured to minimise bias?
Yes – retrospectively collected and independently verified.
4. Have the authors identified and taken into account confounding factors?
The authors identify that prior presentations to Primary Care services could affect the groups but this is not able to be accounted for.
5. Was the follow-up long enough?
Yes – patients were followed up until discharge or death.
6. Do you believe the results?
The authors identify some factors that may skew the results – for example, patients admitted on first presentation are likely to be more acutely unwell and therefore it may make that group appear to have worse outcomes. Also, information about previous treatment/antibiotics could have an impact the outcomes too.
7. Can the results be applied in our local population?
8. Do the results fit in with the other available evidence?
It is reassuring to think that discharging a child with meningitis or septicaemia does not affect that child’s overall outcome. However, some more information about the initial clinical encounter would have been helpful – for example, the patient’s initial clinical symptoms, observations, and investigations/results.