Medication safety and error prevention in children — branching viva

Opening (candidate)

My first priority is the child. I would support the airway and breathing, give naloxone for the suspected opioid effect, and call for senior and PICU help. I would keep the pump stopped and preserve its settings, the syringe and the original order, and I would check whether any other child is on the same protocol. Once the child is safe I would escalate to the consultant, report the incident, and speak with the family honestly. [1] [13]

Branch A — Definitions and classification

Examiner: Define the terms you would use to describe this event, and classify it. [1]

Candidate: This is a medication error — a preventable process failure — that caused an adverse drug event, the actual apnoea. It is also a near miss for any child who might have received the same erroneous pump setting. I classify it by stage of the medication-use process: the order was prescribed correctly but the administration step failed at pump programming, so this is an administration error. By harm it reached the temporary-harm category requiring intervention. [1]

Branch B — The mechanism

Examiner: Why did this happen? Use a model the examiner would accept. [13]

Candidate: I would use Reason's Swiss-cheese model. Latent conditions — an interruption-prone drug round, a pump library without a hard concentration limit, and no independent double-check for a high-risk opioid infusion — sat in the system. The active failure was a programming slip on a busy shift. When the holes aligned, the tenfold error passed through every defence and reached the infant. The fix is at the layers, not the individual. [13]

Branch C — Tenfold errors

Examiner: Why are tenfold errors so common in children, and how do you prevent them? [8]

Candidate: Weight-based dosing in mg per kg turns one adult dose into a continuum of paediatric values, so a misplaced decimal or unit confusion produces a tenfold error. Kozer showed these are a recurring pattern and that detection method changes the reported rate. Prevention is layered: never write a trailing zero, always use a leading zero, confirm the weight in kilograms, use standard concentrations, and require an independent double-check of high-risk infusions. [8]

Branch D — The double-check

Examiner: You mention an independent double-check. Does it actually work? [14]

Candidate: A double-check helps only when it is genuinely independent, focused and not performed under interruption. Konwinski's human-factors work in a PICU shows that a double-check done as theatre — two people glancing together under pressure — adds little. So I would redesign it: two clinicians check separately against the order, confirm the drug, concentration, rate and patient, and sign independently. [14]

Branch E — The distressed colleague

Examiner: The registrar who wrote the order is distressed. What is your role? [13]

Candidate: That is a second-victim response and it is normal. My role is to ensure the child is safe, relieve the registrar from further high-risk tasks in the short term, and offer peer support and a debrief. I keep the analysis at the system level — this was a layered failure, not one person's fault — so the team keeps reporting and the latent holes get closed. [13]

Close

Confirm the child is safe and monitored, preserve the evidence, report the incident, disclose honestly to the family with a clear plan, and lead a systems analysis that chooses strong actions — a hard pump-library limit, standard concentrations and a genuine independent double-check — over a blame conversation. Check the other exposed children and close the loop with the family and the team. [1] [13]