Manufacturer: Roche Diabetes Care GmbH, Mannheim. Family: Accu-Chek SmartGuide CGM device, SmartGuide app, SmartGuide Predict app, AC Care HCP platform. CE marked: July 2024, MDR class IIa. UKCA non-adjunctive: MHRA, May 2025, 18 years and older. Wear duration: up to 14 days. Application site: back of the upper arm. Sensor profile: 5.9 mm height with adhesive, 5 g, 33.3 mm diameter on skin (Glatzer 2024). Reading interval: every 5 minutes via Bluetooth Low Energy. Run-in: one hour. Calibration: two paired finger-prick calibrations on day one (around 12 and 14 hours after insertion); factory-style operation thereafter. Display: Accu-Chek SmartGuide app on iOS and Android (no dedicated reader). Indication: non-adjunctive (no confirmatory finger-prick required for insulin dosing once calibration is complete). Age indication: 18 years and older. Hospital use: not intended.

The Predict app, separately. A second app, the Accu-Chek SmartGuide Predict app, is CE-marked MDR class IIa medical software in its own right. It carries the three prediction layers (30-minute, two-hour, overnight) and runs as cloud-supported inference on iOS and Android. The CGM app and the Predict app are distinct downloads with shared data flow.

What it does. Classifies the next 30 minutes as high-risk or not for hypoglycaemia. When high-risk, an alarm fires on the SmartGuide app.

What it does not do. It does not predict the size or depth of the low; it does not catch every event. It is reactive, not preventive.

When to teach it. First. This is the active safety net the wearer relies on while learning the longer-horizon layers.

What to do when it fires. Finger prick if uncertain, treat per usual hypo plan, plan the next 30 minutes (driving, meeting, sleep), re-check after the response window.

Pitfall. Wearers who silenced “predictive low” alerts on previous CGMs because they fired too often. The SmartGuide LGP fires meaningfully less; ask them to keep it on for two weeks before deciding.

What it does. Projects a glucose trace 2 hours forward, with a 50% confidence band drawn around it. Refreshes every 5 minutes.

What it does not do. It does not see meals or insulin doses the wearer has not entered. It does not account for alcohol, exercise, or illness. A wide band means low confidence.

When to teach it. Second. Frame it as “the curve you plan around for the next hour, the curve you act on for the next forty-five minutes”. The split is the teach. Parkes A+B agreement is 99.3% at 45 min and 96.3% at 2 hours (Herrero 2024).

What to do when it lights up. 0 to 45 min, act on the curve direction (eat ahead of a low, defer a correction if a fall is shown). 45 min to 2 h, plan around it (meeting timing, drive, exam, run).

Pitfall. Selling the 2-hour mark as reliable. It is not. Wearers who treat the 2-hour endpoint as actionable will be disappointed quickly, and that disappointment is the form factor of a sensor falling out of routine wear.

What it does. Issues a Red, Amber, or Green prediction of overnight low risk between 9pm and 2am, refreshable every 20 minutes. Personalises over around 28 days of input data.

What it does not do. It is a population-level confidence band, not a personal probability. It misses approximately 45% of overnight hypoglycaemia events at the published threshold; LGP remains the active safety net during the night.

When to teach it. Third. The wearer should already trust the LGP alarm and the GP curve. The night plan is what unlocks confidence to sleep.

What to do when it lights up. Red, first half of the night (model ROC AUC 0.902): fast carbohydrates before bed are the option most aligned with the risk; discuss with the care team. Red, second half (ROC AUC 0.730, lower confidence): a conversation about protein, slow-acting carbohydrates, or a temporary basal reduction (if on a pump). Amber: re-check in 20 minutes, watch for trend.

Pitfall. Treating both halves of the night identically. The ROC AUC asymmetry means a red light at 1am and a red light at 4am are not the same signal. Land that at onboarding.

The SmartGuide rewards a 28-day onboarding pattern that current CGM marketing does not always reflect.

• Night Low Predict needs around 28 days of input data before it reaches its full personalised accuracy. The first few weeks may be less reliable; do not draw conclusions from the first night.
• If the prediction shows red in the first half of the night, fast-acting carbohydrates are the option most aligned with that risk; discuss with your care team. If red in the second half, the model’s confidence is lower; protein, slow-acting carbohydrates, or a temporary basal reduction (if on a pump) are the conversations to have, again with the care team.
• You can re-request a Night Low Predict every 20 minutes between 9pm and 2am. If you take action, check whether the risk rating changes.
• The 45-minute Glucose Predict window is more reliable than the two-hour window (Parkes A+B 99.3% at 45 min vs 96.3% at 2 hours; Herrero 2024). Treat 45 minutes as an action window and the two hours as awareness and context.
• The predictions do not account for alcohol, exercise, or illness. Adjust your interpretation if any of these applies.
• If the worm is wide at two hours, confidence is lower. Do not treat it as a reliable forecast.
• The two day-one finger pricks (at approximately 12 and 14 hours after insertion) take a few minutes each and activate the full predictive features. Good hand washing and technique matter.

The teaching shape of the SmartGuide differs from a standard CGM in ways that matter at the first appointment.

• Teach the first-half versus second-half distinction for Night Low Predict from the outset. The model’s ROC AUC is meaningfully higher for first-half risk than second-half (0.902 vs 0.730; Herrero 2024); the action menu and the confidence band differ accordingly.
• Group onboarding sessions generate richer learning than one-to-one. The calibration conversation lands differently when people hear from peers who still finger prick routinely.
• The mandatory day-one calibration is an opportunity to re-establish the value of finger prick testing, not a device drawback.
• Send people away to use it and come back to report. The most useful education comes from early adopters describing what they actually did, not from the manual.
• Current indication is adults 18 and older only. No paediatric licence and no AID compatibility in current form.
• The DTN Competency Assessment Framework is being uploaded to the DTN website shortly. It maps team skill mix against a four-tier national standard and links to annual appraisal documentation.