Medical devices, digital health and remote monitoring — formative SAQs

SAQ 1 (10)

A four-year-old with type 1 diabetes for eighteen months is brought to clinic. The family uses a continuous glucose monitor (CGM) and report recurrent night-time alarms and a recent reading of 48 mg/dL (2.7 mmol/L) at three o'clock in the morning, treated with juice. The family ask whether they should stop the device. [4]

1. Explain how a CGM works at sensor level, including the plasma-interstitial lag and its clinical consequence. (3) [4]
2. State the ISPAD 2024 ambulatory glucose profile targets for time in range, time below 70 mg/dL and time below 54 mg/dL, and outline how you would read the AGP at this visit. (4) [4]
3. Outline your immediate management of a CGM reading below 54 mg/dL (3.0 mmol/L), and your shared decision with the family about whether to continue, modify or stop the device. (3) [8]

Model answer

Sensor physiology and the lag. A CGM measures interstitial fluid glucose, not plasma glucose. A flexible sensor is placed subcutaneously — usually the upper arm in a child — and a glucose-oxidase enzyme electrode generates a current proportional to the interstitial glucose. Because glucose must move from plasma to interstitial fluid before the sensor reads it, the interstitial value lags the plasma value by five to fifteen minutes, longer during rapid change. The clinical consequence is the lag: a falling glucose can be missed at the trough, and a post-meal rise can lag the capillary value. For this reason any critical CGM reading — particularly a low — must be confirmed with a fingerstick before action. [4]

AGP targets and how to read the report. ISPAD 2024 and the Battelino international consensus set the operational targets: time in range — glucose between 70 and 180 mg/dL (3.9 and 10.0 mmol/L) — should exceed 70 per cent of the day; time below 70 mg/dL (3.9 mmol/L) should be under 4 per cent; and time below 54 mg/dL (3.0 mmol/L) should be under 1 per cent. At this visit I would open the AGP report, look first at the time in range percentage against the 70 per cent target, then at the time below range against the 4 per cent and 1 per cent targets, then at the median trace to see whether the hypoglycaemias cluster at night (suggesting basal insulin or alarm-threshold tuning), and finally at the sensor wear to confirm the data are complete enough to interpret. [4]

Immediate management and the shared decision. For a CGM reading below 54 mg/dL (3.0 mmol/L) I confirm with a fingerstick, give 10 to 15 grams of fast-acting oral carbohydrate if the child is conscious and able to swallow (glucose tablets, juice, glucose gel), give glucagon 0.5 milligrams intramuscularly for a child under twelve years or 1 milligram at twelve years and over if the child is unconscious or unable to swallow safely, and recheck the glucose in fifteen minutes. I counsel the family to keep glucagon available at home and to use it. On the shared decision, I would not stop the device — the recurrent night-time alarms are a threshold-tuning problem, not a device failure. I would raise the low alert threshold to limit non-critical alarms, schedule a period of protected sleep with only the critical alarm audible, review the basal insulin pattern that is driving the night-time lows, and arrange a follow-up review of the AGP in two to four weeks. Abandoning the device without a safety-net would expose the child to unrecognised hypoglycaemia. [8]

SAQ 2 (10)

A six-month-old infant with bronchopulmonary dysplasia is discharged home on home oxygen with a finger pulse oximeter. The family are first-time parents and the infant has darkly pigmented skin. [10]

1. Outline your home pulse oximetry programme for this family — selection, calibration, family education, the written action plan and the review pathway. (4) [8]
2. Explain the demonstrated inaccuracy of pulse oximetry in dark skin pigmentation and in low perfusion, citing the relevant studies, and state the bedside mitigations. (4) [5] [6] [7]
3. State why a single night of home pulse oximetry is an inadequate basis for a clinical decision, and how you would use the device over the longer term. (2) [10]

Model answer

Home pulse oximetry programme. The right child is the child with a defined clinical question the device can answer — here, the titration of home oxygen in bronchopulmonary dysplasia. I would select a medical-grade, TGA-registered pulse oximeter with a probe sized to the infant; calibrate it against a ward device before discharge; train the family on what is and is not an alarm (a saturation reading, a trend, a colour change, a work-of-breathing change); agree a written action plan with explicit thresholds (when to increase oxygen, when to call the team, when to come to the emergency department); and arrange a review point at which the device is either continued, modified or stopped. The Foster AAP policy on remote monitoring of patient- and family-generated health data frames the deployment, and the written plan is the safety-net the family executes at any hour. [8]

Pulse oximetry inaccuracy and the bedside mitigations. A pulse oximeter estimates arterial saturation by passing two wavelengths of light — 660 nanometres red and 940 nanometres infrared — through a pulsatile vascular bed and ratioing the absorbance of oxyhaemoglobin to deoxyhaemoglobin. The Sjoding cohort of over ten thousand paired measurements showed occult hypoxaemia — a SpO2 in the normal range when the arterial saturation was low — was nearly three times more common in Black patients than in White patients, because dark skin pigmentation increases the red-wavelength absorbance and biases the saturation upward. The Gudelunas prospective study extended the finding into low-perfusion states, where darkly pigmented skin was again associated with missed hypoxaemia. The Rathod collaborative recommendations framed the issue as a patient-safety rather than a tolerance problem. The bedside mitigations are clinical correlation (treat the child, not the number), a low threshold to escalate to an arterial blood gas with co-oximetry when the reading and the child disagree, and equity-aware thresholds that account for the bias. A normal SpO2 in an unwell infant with dark skin does not exclude hypoxaemia. [5] [6] [7]

Single-night inadequacy and the longer-term use. The Hoppenbrouwer study of night-to-night variability of pulse oximetry features in children showed that a single night under- or over-estimates the burden of desaturation, so clinical decisions — oxygen titration, the wean, the readiness for discharge — must rest on a trend across several nights rather than one recording. Over the longer term I would review the multi-night trend at each clinic visit, look for the pattern of desaturations in the context of the oxygen flow rate, and use the trend to guide the wean. I would also review the family's experience of alarms and distress at each visit, because alarm fatigue is the most common reason for non-adherence and abandonment of a home device. [10]