Episode 43, Tandem Control-IQ+ and Mobi

Why this episode exists

Most conversations about automated insulin delivery start with the algorithm and end with the target. Control-IQ does not let you change its target at all, it is fixed, and yet real-world outcomes on it vary enormously between users on the same hardware. The registry data explains why: the setting most people never hear named, correction factor, has a bigger effect on time in range than basal rate, insulin-to-carb ratio, or anything a target dial could offer.

Laurel Messer has practised in paediatric diabetes for over two decades and now leads Global Medical Affairs at Tandem. She is not a neutral observer of her own product, and this episode names that plainly. What she brings is the real-world registry analysis behind the claim, and a working knowledge of every lever in Control-IQ and the newer Control-IQ+ update, including the ones aimed at very young children, very high insulin users, sick days, and pregnancy.

In this episode

John and Laurel work through the lever hierarchy behind Control-IQ, the real-world registry study that put correction factor at the top of it, and everything that changes with the Control-IQ+ update: a much wider total daily dose range, a correction factor ceiling that no longer caps out at the level built for six-year-olds, a temp basal that no longer requires leaving automation on a sick day, and an extended bolus built for the meals that outlast a normal bolus curve.

They also cover the FDA’s fresh clearance for Control-IQ in pregnancy, what the CIRCUIT trial found for time in pregnancy range, and close on the Tandem Mobi, a matchbox-sized pump running the same algorithm without the dangling tubing.

Watch or listen

Key themes

Correction factor, the lever that gets overlooked

Control-IQ’s glucose target is fixed at 6.1 mmol/L (110 mg/dL) and cannot be changed by the user or clinician. Every other adjustable-target AID system offers a dial that Control-IQ simply does not have. What it offers instead is a set of three tunable settings, correction factor, basal rate, and insulin-to-carb ratio, and a real-world registry analysis of over 20,000 users (Messer 2023, a cross-sectional association, not a randomised trial) found a strong association between correction factor and outcome: the strongest correction factor quartile ran roughly 14 percentage points higher time in range than the weakest. Laurel Messer, who led the analysis, describes it as a lever most clinicians barely think about, “there’s that too”, when the association she found points to it as a major driver of the difference.

The GNL teaching framework built from this evidence, reviewed with Laurel Messer in April 2026, ranks the three settings in strict order: correction factor first, basal percentage second (a calculation, not a target to chase), and insulin-to-carb ratio third. A companion 2026 practice-guidance paper (Shah, Choudhary, Halperin, Rabbone, Zaharieva, Messer, Diabetes, Obesity and Metabolism) sets out the same hierarchy and gives clinicians a specific threshold: correction factor rules below roughly 88 divided by total daily dose (in mmol/L terms) associate with the highest time in range without a meaningful rise in time below range. Laurel Messer co-authors both papers, so this is not an independent replication of the correction-factor finding; a genuinely independent registry or trial confirming the hierarchy has not yet been published.

Control-IQ+, a much wider dose range

The Control-IQ+ update extends the usable total daily dose range from 10 to 100 units up to 5 to 200 units, according to Laurel Messer. That change opens the algorithm to people who need very little insulin, mainly very young children, and to people who need a great deal, without the earlier range acting as a ceiling on either end. As reported in this episode, the same update lowers the FDA-cleared age indication to two years and above, alongside Control-IQ+’s dedicated correction factor and insulin-to-carb ratio ranges for that age band.

Control-IQ+, the correction factor ceiling raised

On the original Control-IQ, the correction factor is capped at a level equivalent to roughly 11 mmol/L (200 mg/dL) regardless of what is programmed, a constraint set for the six-and-older population it was licensed for. Control-IQ+ raises that ceiling to roughly 33 mmol/L (600 mg/dL), reported in this episode as a change aimed squarely at very young children, whose insulin sensitivity can shift by what Laurel Messer describes as an “eightfold” swing between the dusk hours and the middle of the night. A correction factor strong enough for 5pm to 10pm can be dangerously strong at 2am; the wider range lets clinicians programme that swing directly into time-blocked settings rather than compromising on a single number, or resorting to workarounds such as artificially logging carbohydrates to keep the algorithm active overnight.

Three or four time blocks across the day are generally enough to capture this. Laurel Messer’s clinical rule of thumb, echoed by John, is to simplify rather than multiply, sessions with ten or more time segments are a sign of overcomplication, not precision.

The sick day hack, temp basal without leaving automation

Before Control-IQ+, a dramatic short-term shift in insulin need, most often an acute illness, meant leaving automation altogether to run a manual temporary basal. As reported in this episode, Control-IQ+ allows a temporary basal rate to run while Control-IQ automation stays active, for up to three days at a time. Laurel Messer frames this as a sick-day tool first: the person keeps the hypoglycaemia protection and time-in-range benefit of automation while telling the system, for a defined window, that more insulin is coming. The same mechanism doubles as a low-stakes way to trial a basal increase, for example ahead of an expected change in insulin sensitivity, before committing to a permanent setting change; any such change is a conversation with the diabetes care team.

The extended bolus, built for the meals that outlast it

As reported in this episode, Control-IQ previously allowed an extended bolus of up to two hours while automation was active, with anything longer meaning exiting automation; Control-IQ+ extends this to up to eight hours. Laurel Messer highlights two populations this serves: young children, whose eating pace and quantity across a meal or an event is genuinely unpredictable, and adults with delayed gastric emptying (gastroparesis), where insulin needs to be spread over a longer absorption window. The extended bolus can be cancelled part-way through if the situation changes; spreading a larger dose over several hours also spreads the window in which a lower-than-expected glucose can develop, so checking glucose partway through a long extended bolus, especially overnight, is part of using it safely.

The pizza test, why high-fat meals need more insulin

John poses a version of a well-known finding to Laurel Messer as a quiz: how much extra insulin, on average, does a high-fat, high-protein meal (the pizza-with-toppings scenario studied by Bell and colleagues) need compared with an equivalent low-fat meal. Her estimate lands short. The teaching point used in GNL clinical practice starts at 25 per cent extra insulin, stepping up to 50, 75, or 100 per cent if that is not enough, delivered as an extended bolus because the effect plays out over several hours, not the usual two; a larger total dose spread this way carries its own delayed-hypoglycaemia risk if the fat-and-protein effect is smaller than expected on a given occasion, which is why glucose is worth checking partway through rather than only at the end. Mechanistically, excess dietary fat is broken down in the liver into diacylglycerols, molecules that interfere with insulin’s action at the muscle and liver cell level, which is why even a well-functioning automated system struggles to keep pace with a very high-fat meal without a deliberately extended, larger bolus.

FDA clearance for pregnancy, and what the CIRCUIT trial found

As reported in this episode, Control-IQ received FDA clearance for use in pregnancy in the fortnight before recording. This is a United States regulatory clearance only. Control-IQ is not licensed for pregnancy in the UK or Europe; the trial evidence behind it was generated under a research protocol using the system off-label. Anyone considering an AID system in pregnancy in the UK should discuss current licensed options with their diabetes-in-pregnancy team.

The evidence behind the US clearance is the CIRCUIT trial (Donovan et al. 2025, JAMA), which randomised pregnant women with type 1 diabetes to Control-IQ or their standard therapy with CGM. Over the 16-to-34-week gestation window, Control-IQ users spent approximately 3 hours per day more time in the pregnancy-specific target range (3.5 to 7.8 mmol/L, 63 to 140 mg/dL) than the standard care group, an adjusted difference of around 12.5 percentage points, without a comparable adjustable-target system to benchmark against since Control-IQ’s target cannot be changed. A separately licensed system, CamAPS FX, has shown a comparable adjusted benefit of 10.5 percentage points over standard care in the UK AiDAPT trial (Lee et al. 2023, NEJM). Time in pregnancy range is a surrogate; it is used because it tracks the outcomes that matter (large-for-gestational-age birth weight, neonatal hypoglycaemia, and other neonatal complications) rather than being the outcome itself, and CIRCUIT was not powered to show a direct difference in those endpoints. As Laurel Messer puts it, the lever that gets you there does not have to be a target: “settings are settings, outcomes are outcomes.”

Sleep Activity, a tighter target used around the clock in pregnancy

Sleep Activity is not a target-adjustment feature in the way some other systems offer; it is an additional layer that tightens Control-IQ’s every-five-minute insulin delivery to a narrower range (6.25 to 6.7 mmol/L, 112.5 to 120 mg/dL) while disabling the predictive Autobolus. In the CIRCUIT trial protocol, participants ran Sleep Activity around the clock rather than overnight only, on the basis that pregnant participants were bolusing consistently for meals and did not need the Autobolus as a backstop. This is a trial-protocol detail, not a general recommendation; whether and when to run Sleep Activity continuously is a decision made with a diabetes care team, and running it around the clock without consistent meal bolusing is associated with more hyperglycaemia, not less, because the Autobolus safety net is switched off while it runs.

The Tandem Mobi, the same algorithm without the tubing

Tandem Mobi runs the same Control-IQ algorithm family as the t:slim X2, on a tubeless, matchbox-sized on-body pump with no screen; all interaction happens through a phone app, plus a single bolus button on the pump itself for when the phone is not to hand. As reported in this episode, the automation itself runs on the pump communicating directly with the CGM, so it continues even if the phone is in another room. The minimum insulin fill is reported in this episode at 30 units, which Laurel Messer notes matters for people with very low total daily doses who would otherwise waste insulin filling a larger reservoir. Tandem Source now accepts direct uploads from the app, removing an earlier dependency on a separate upload platform.

The correction-factor-first framework, in detail

The lever hierarchy behind this episode, reviewed with Laurel Messer (Tandem Global Medical Affairs) in April 2026, sets correction factor first, basal percentage second, and insulin-to-carb ratio third, with Sleep Activity as an additional consideration rather than a level of its own. This is a population-average teaching framework, reviewed by Tandem Medical Affairs but not a Tandem-endorsed configuration recommendation; every setting change sits within Tandem’s own adjustable range and is a conversation with your diabetes care team.

Practical exploration

For people living with type 1 diabetes and their families

None of this is a substitute for a conversation with your diabetes care team; it is context for that conversation.

• If you use Control-IQ and are unsure which of correction factor, basal rate, or insulin-to-carb ratio is doing the most work in your settings, that is a specific, answerable question for your next clinic appointment.
• Control-IQ’s target cannot be changed, on either version or either pump. If a clinic conversation implies otherwise, that is worth clarifying.
• A high-fat, high-protein meal (a large pizza, a takeaway) can need substantially more insulin than usual, delivered over a longer window than a standard bolus. This is a known physiological effect, not a dosing mistake.
• If you are pregnant or planning pregnancy and using or considering an AID system, ask your diabetes-in-pregnancy team which systems are currently licensed for pregnancy in your country; licensing differs between the UK, Europe, and the United States.
• If you are considering Tandem Mobi, ask specifically about current UK MHRA and NHS availability; US FDA clearance does not mean UK availability.

For clinicians and educators

The teaching shape that comes out of the registry data is specific enough to apply directly.

• Lead with correction factor, not basal rate, when optimising Control-IQ or Control-IQ+; the real-world effect size is the largest of the three adjustable settings (Messer 2023; Shah 2026).
• Keep time-block schedules to three or four blocks across the day; ten or more is a sign to simplify, not personalise further.
• For very young children with a pronounced dusk-to-night insulin sensitivity swing, Control-IQ+’s wider correction factor range (up to roughly 33 mmol/L, 600 mg/dL, as described in this episode) is built for exactly that pattern.
• Use the Control-IQ+ temp basal (reported in this episode as available for up to three days) as a structured, time-limited trial before committing to a permanent setting change, not only as a sick-day tool.
• Any correction factor, basal, or insulin-to-carb figure discussed with a patient is a population-average starting point; their actual settings are set and adjusted with their own diabetes care team.
• Sleep Activity run continuously without consistent meal bolusing increases hyperglycaemia risk by disabling the Autobolus; match the Sleep Activity schedule to actual bolusing behaviour.

About the guest

Laurel Messer is VP of Global Medical Affairs at Tandem Diabetes Care, following a clinical and research career of over 20 years in paediatric diabetes technology, including leading the real-world registry analysis (Messer et al. 2023, Diabetes Technology and Therapeutics, n=20,764) that identified correction factor as the strongest predictor of time in range on Control-IQ. She previously joined the GNL Podcast to discuss diabetes technology and skincare, and returns here to open the GNL AID series.

Related reading on GNL