Medical devices, digital health and remote monitoring — OSCE

Station brief (candidate)

• Counsel the family on the validity of the smartwatch rhythm flag as a screening signal rather than a diagnosis, and outline the device-independent confirmatory pathway (the 12-lead ECG and clinician overread). [9]
• Counsel the family on the CGM physiology and the plasma-interstitial lag, review the ISPAD 2024 ambulatory glucose profile targets, and reach a shared decision to tune the thresholds rather than abandon the device. [4]
• Counsel the family on the demonstrated inaccuracy of pulse oximetry in dark skin pigmentation and low perfusion, and agree a written action plan that names the equity limitation and the threshold to escalate to the emergency department. [5] [6] [7]

Setting

A general paediatric outpatient clinic. The parents (a dark-skinned couple) attend with their three children. The candidate is the general paediatric registrar. Nine minutes for the encounter, one minute for the examiner's marking. [8]

Encounter script

Parent: "Doctor, our smartwatch flagged an irregular heart rhythm on our eight-year-old at the weekend. We're frightened — is it atrial fibrillation? And separately, our four-year-old with diabetes has been having alarms on her monitor every night at three, and we want to take it off her. And the baby's home oxygen monitor — we never know when to trust the number." Candidate (smartwatch): I would begin by acknowledging the family's concern, then explain that a smartwatch rhythm flag is a screening signal rather than a diagnosis. I would take a structured history — palpitations, syncope, chest pain, exercise intolerance, family history of sudden death or arrhythmia — and examine the child's cardiovascular system for a structural murmur. I would obtain a 12-lead ECG and arrange a clinician overread of the device tracing, because the algorithmic label is not the diagnosis. I would name the common paediatric causes of a smartwatch flag: sinus arrhythmia, the normal physiological variation of heart rate with respiration, premature atrial contractions, supraventricular tachycardia, and the rare true atrial fibrillation. I would not act on the algorithmic label without the confirmatory test. [9] Candidate (CGM): On the four-year-old, I would not stop the CGM. The device measures interstitial glucose, which lags the plasma by five to fifteen minutes, so the night-time low alarm at three o'clock is a real signal that the basal insulin is driving the child low overnight. I would review the ambulatory glucose profile with the family — the ISPAD 2024 targets are time in range over 70 per cent, time below 70 mg/dL under 4 per cent, time below 54 mg/dL under 1 per cent — and tune the low alert threshold to silence the non-critical alarms while keeping the critical alarm audible. I would confirm the immediate hypoglycaemia protocol is in place: any reading under 54 mg/dL is confirmed with a fingerstick, treated with 10 to 15 grams of fast-acting oral carbohydrate if the child is conscious, or glucagon 0.5 milligrams intramuscularly for a child under twelve years or 1 milligram at twelve years and over if the child is unconscious, and rechecked in fifteen minutes. Abandoning the device without a safety-net would expose the child to unrecognised hypoglycaemia. The Wadwa trial moved CGM and hybrid closed-loop systems into the standard of care for the very young. [1] [4] [8]

Candidate (pulse oximetry): On the baby, I would name the equity limitation directly. The pulse oximeter is biased by dark skin pigmentation and by low perfusion, so a normal SpO2 reading in a dark-skinned infant does not exclude hypoxaemia — the Sjoding cohort showed occult hypoxaemia was nearly three times more common in Black patients than in White patients, and the Gudelunas study extended the finding to low-perfusion states. My mitigations are three: I train the parents to read the baby before the device — the colour, the work of breathing, the feeding, the alertness — and to use the device's number in that context; I give them a written action plan that names the equity limitation and the threshold to escalate to the emergency department; and I keep a low threshold to obtain an arterial blood gas with co-oximetry when the family is concerned, even if the SpO2 is in the normal range. [5] [6] [7]

Examiner marking domains

• Clinical knowledge (3): Distinguishes a consumer-grade screening signal from a medical-grade diagnosis; states the CGM plasma-interstitial lag and the ISPAD 2024 AGP targets; states the pulse oximetry equity limitation and the device-independent confirmatory tests (12-lead ECG, arterial blood gas with co-oximetry). [4] [5]
• Communication (3): Avoids jargon, acknowledges the family's concern, names the equity issue directly and without judgement, and checks the family's understanding at each step.
• Shared decision-making (2): Reaches a documented shared decision to continue the CGM with threshold tuning and a written plan; to investigate the smartwatch flag with a confirmatory ECG before any action; and to continue the pulse oximetry with an equity-aware written action plan.
• Safety-netting (2): Provides a written action plan with explicit thresholds and a 24-hour escalation contact for each device, and confirms the family can execute it at any hour.