DBLG1 closed loop kept glucose in range longer in a French trial

HAS found that people using the DBLG1 closed loop spent 68.5% of the time in the target glucose range against 59.4% on pump and sensor alone. It rated the gain as moderate.

  • Not rated a person

    Trial patients spent more time in the target range, but HbA1c did not change, and HAS judged that the long-term gain is still expected.

Depth runs from 1, a convenience or time saved, to 5, a life or a livelihood preserved.

Status
Intended benefit.
Evidence

foundstudiedreportedclaimed

An independent assessment report supports the benefit.

Compared with

Open-loop systems in which an insulin pump and a glucose sensor work independently. Measured.

Without the system
59,4 (± 1,3) % of time in 70-180 mg/dL with pump and sensor in the SP7 trial
With it
68,5 (± 1,2) % with DBLG1, rated amélioration modérée du service attendu (ASA III)
The system’s part
The randomized crossover trial compared the closed loop with the same patients' pump and sensor therapy, which isolates the automated control. HAS describes the module that the maker calls physiological machine learning as predefined equations whose 15 parameters a genetic algorithm adapts to 48 hours of the patient's data. It calls the other modules, including long-term self-learning, deterministic. In the maker's simulation, that module made better dosing decisions than the expert-system module 61.81% of the time.
Not yet checked
LimitsCosts and harmsLearns from dataLater outcome

Download data Cite Suggest a correction

Benefit type
Improved Physical Health
System
DBLG1 System
Made by
Diabeloop
Used by
12 French diabetes centers in the SP7 randomized crossover trial
Where
France (12 trial centers)
Stage
In a pilot
Date
, when the source was published
Kind
Aggregate result
Broad benefit
Controlling Machines and Processes, Estimating the Quantifiable Characteristics of Products, Events, or Information
Collections
Scrutiny and findings
Last revised
31 October 2026

Sources

  1. Assessment report, by authors with no tie to the system.

    Haute Autorité de santé, Commission nationale d'évaluation des dispositifs médicaux et des technologies de santé (CNEDiMTS). Avis de la CNEDiMTS du 28 janvier 2020, DBLG1, système de boucle semi-fermée dédié à la gestion automatisée du diabète de type 1 (dossier 5998, version publiée après occultation). Section 06.2, Niveau(x) d'ASA.

    La Commission considère que le système DBLG1 est susceptible d'apporter un bénéfice sur le contrôle métabolique des patients diabétiques de type 1 sous pompe à insuline et dont le taux d'HbA1c est ≥ 8% depuis plus de 6 mois. Les données disponibles mettent en évidence un effet favorable sur différents paramètres de suivi instantané de l'équilibre glycémique.

Details

AI system

Reported configuration

DBLG1 algorithm v2017.04.20 in the SP7 trial (listed system: software version 1.3.0 of 16 May 2019)

Workflow

A sensor measured interstitial glucose every 5 minutes and sent it to the smartphone. The algorithm drove the insulin pump automatically, and patients confirmed mealtime doses. Trial patients had 48 hours of training in hospital and round-the-clock support from a nurse and a Diabeloop engineer.

Follow-up

Conditions for replication

Patients are adults with type 1 diabetes on pump therapy who receive specific training and education. Patients use the system in closed-loop mode at least 75% of the time, the threshold that HAS set for continuing.

Entry details

Sector

Healthcare (type 1 diabetes).

Purpose

Diabeloop built the DBLG1 System to adjust insulin delivery automatically from continuous glucose readings for adults with type 1 diabetes.

29 more questions not yet checked

model, provider, role, configuration, inputs and outputs, autonomy, human oversight, beneficiaries, reach, distribution and access, duration, every measurement question, every money question, ownership and control, dependencies, cross-border connections, geopolitical implications, replication attempts, availability, institutional decision, grounds, action and result, success criterion, every evidence question, observed or intended, alternative assessments, related entries, other identifiers, category mappings.

Cite this record

Peter H. Vartanian. (2026). DBLG1 closed loop kept glucose in range longer in a French trial (AIBD-0928). Artificial Intelligence Benefits Database, edition of 1 November 2026. https://petervartanian.xyz/aibd/records/AIBD-0928.html (archived at https://doi.org/10.7910/DVN/AADSII, version 7.0)

BibTeX Data Field definitions