STRIVE: Omnipod 5 SmartAdjust 2.0 versus Omnipod 5

STRIVE is the trial that decides whether the next Omnipod is actually better than the current one. Insulet took 132 people who already wear an Omnipod 5 — children as young as two, adults up to seventy — and had them run the investigational SmartAdjust 2.0 algorithm for four weeks and their existing Omnipod 5 for four weeks, in randomized order. SmartAdjust 2.0 is the software that becomes Omnipod 6: the same tubeless Pod, a new adaptive-learning algorithm, a lower glucose target.¹

Who could take part. You had to already be an Omnipod 5 user on U-100 insulin, and to have spent most of the preceding fortnight at the system's tightest 110 mg/dL (6.1 mmol/L) target — so this is a trial of experienced, well-controlled pod users, not of people starting automation for the first time. Adults could enrol with either type 1 or type 2 diabetes; children were type 1 only.¹

What "success" meant. Notably, STRIVE was designed as a non-inferiority trial on its four co-primary endpoints, all measured at the end of the second four-week period: time below 54 mg/dL (3.0 mmol/L), time below 70 mg/dL (3.9 mmol/L), time in range 70-180 mg/dL (3.9-10.0 mmol/L), and mean glucose. The margins were pre-specified and tight — 0.75 percentage points for time below 54, 3.0 points for time below 70 and for time in range, and 8 mg/dL (0.4 mmol/L) for mean glucose.¹ In plain terms: the bar Insulet had to clear was do no harm, not do better. Any improvement on top of that is a bonus.

The third period is the interesting one

After the crossover, everyone spent a further four to six weeks on SmartAdjust 2.0 with one instruction: take no more than three meal or snack boluses a day.¹ That is not a normal hybrid-closed-loop protocol. It is a deliberate test of how much of the meal work the algorithm can absorb on its own — the same question the CLOSE IT trial asked of the open-source community's algorithm. Other systems already offer different degrees of meal automation; this trial addresses Insulet’s tested configuration.

It is worth being precise about what this is and is not. Reduced bolusing is not no bolusing. Omnipod 6 remains a hybrid closed loop for people with type 1 — you still announce meals — and Insulet describes its fully-closed-loop programme separately, in the context of type 2 diabetes.²³ Period 3 is best read as a stress test of the algorithm's meal-handling headroom, not as the launch of a no-carb-counting Omnipod.

What is known, and what is not

The trial completed on 10 November 2025, and no results are posted on the registry.¹ What exists is company-reported: at the ADA Scientific Sessions in June 2026 Insulet said that in the type 1 arm (ages 14 and over), SmartAdjust 2.0 delivered 77% time in range versus 73% on Omnipod 5, and 54% time in tight range (70-140 mg/dL; 3.9-7.8 mmol/L) versus 47%, with up to 50% more automated insulin and no DKA or severe hypoglycemia.² A type 2 subanalysis reported in July 2026 showed 76% versus 73% time in range and a drop in daily boluses.³

Three honest caveats.

First, these are conference and press-release numbers, not a peer-reviewed paper. They come from the sponsor, they have not been through review, and the full safety and per-age-group breakdowns are not public. Treat the four-point time-in-range gain as promising and provisional.

Second, the headline figures cover the type 1 arm aged 14 and over. The trial enrolled children from age two, but no separate young-children result has been reported — and young children are where AID algorithms are hardest to get right.

Third, no period-3 outcome is reported in the cited sources. The reduced-bolus period addresses how the algorithm handles less frequent meal input; the sponsor’s headline comparison does not settle that question.

The confirmatory study, STRIVE 2, is now recruiting, with primary completion targeted for February 2027.