Search Results

SmartGuard technology is intended for use with compatible integrated continuous glucose monitors (iCGM), compatible Medtronic continuous glucose monitors (CGMs), and alternate controller enabled (ACE) pumps to automatically adjust the delivery of basal insulin and to automatically deliver correction boluses based on sensor glucose values.
SmartGuard technology is intended for the management of Type 1 diabetes mellitus in persons 7 years of age and older requiring insulin.
SmartGuard technology is intended for single patient use and requires a prescription.

Predictive Low Glucose technology is intended for use with compatible integrated continuous glucose monitors (iCGM), compatible Medtronic continuous glucose monitors (CGMs), and alternate controller enabled (ACE) pumps to automatically suspend delivery of insulin when the sensor glucose value falls below or is predicted to fall below predefined threshold values.
Predictive Low Glucose technology is intended for the management of Type 1 diabetes mellitus in persons 7 years of age and older requiring insulin.
Predictive Low Glucose technology is intended for single patient use and requires a prescription.

SmartGuard Technology, also referred to as Advanced Hybrid Closed Loop (AHCL) algorithm, is a software-only device intended for use by people with Type 1 diabetes for ages 7 years or older. It is an interoperable automated glycemic controller (iAGC) that is intended for use with compatible integrated continuous glucose monitors (iCGM), compatible interoperable Medtronic continuous glucose monitors (CGM) and compatible alternate controller enabled (ACE) pumps to automatically adjust the delivery of basal insulin and to automatically deliver correction boluses based on sensor glucose (SG) values.
The AHCL algorithm resides on the compatible ACE pump, which serves as the host device. It is meant to be integrated in a compatible ACE pump and is an embedded part of the ACE pump firmware.
Inputs to the AHCL algorithm (e.g., SG values, user inputs) are received from the ACE pump (host device), and outputs from the AHCL algorithm (e.g., insulin delivery commands) are sent by the algorithm to the ACE pump. As an embedded part of the firmware, the AHCL algorithm does not connect to or receive data from compatible CGMs; instead, sensor glucose (SG) values or other inputs received by the ACE pump from compatible CGMs via Bluetooth Low Energy (BLE) technology are transmitted to the embedded AHCL algorithm.
The AHCL algorithm works in conjunction with the ACE pump and is responsible for controlling insulin delivery when the ACE pump is in Auto Mode. It includes adaptive control algorithms that autonomously and continually adapt to the ever-changing insulin requirements of each individual.
The AHCL algorithm requires specific therapy settings (target setpoint, insulin-to-carb ratios and active insulin time) that need to be established with the help of a health care provider (HCP) before activation. It also requires five (5) consecutive hours of insulin delivery history, a minimum of two (2) days of total daily dose (TDD) of insulin, a valid sensor glucose (SG) and blood glucose (BG) values to start automated insulin delivery.
When activated, the AHCL algorithm adjusts the insulin dose at five-minute intervals based on CGM data. A basal insulin dose (auto basal) is commanded by the AHCL algorithm to manage glucose levels to the user's target setpoint of 100 mg/dL, 110 mg/dL or 120 mg/dL. The user can also set a temporary target of 150 mg/dL for up to 24 hours. In addition, under certain conditions the algorithm can also automatically command correction boluses (auto correction bolus) without user input.
Meal boluses are the responsibility of the user. The AHCL algorithm includes an integrated bolus calculation feature for user-initiated boluses for meals. When the user inputs their carbohydrate intake, the AHCL algorithm automatically calculates a bolus amount based off available glucose information, entered carbohydrate amount and other patient parameters.
The AHCL algorithm contains several layers of "safeguards" (mitigations) to provide protection against over-delivery or under-delivery of insulin to reduce risk of hypoglycemia and hyperglycemia, respectively.
The AHCL algorithm is a software-only device and does not have a user interface (UI). The compatible ACE pump provides a UI to the user to configure the therapy settings and interact with the algorithm. The AHCL-related alerts/alarms are displayed and managed by the pump.

Predictive Low Glucose Technology, also referred to as the Predictive Low Glucose Management (PLGM) algorithm is a software-only device intended for use by people with Type 1 diabetes for ages 7 years or older. It is an interoperable automated glycemic controller (iAGC) that is intended for use with compatible integrated continuous glucose monitors (iCGM), compatible interoperable Medtronic continuous glucose monitors (CGM) and compatible alternate controller enabled (ACE) pumps to automatically suspend delivery of insulin when the sensor glucose value falls below or is predicted to fall below predefined threshold values.
The PLGM algorithm resides on the compatible ACE Pump, which serves as the host device. It is meant to be integrated in a compatible ACE pump and is an embedded part of the ACE pump firmware.
Inputs to PLGM algorithm (e.g., sensor glucose values, user inputs) are received from the ACE pump (host device), and outputs from PLGM algorithm (e.g., suspend/resume commands) are sent by the algorithm to the ACE pump. As an embedded part of the ACE pump firmware, the PLGM algorithm does not connect to or receive data from compatible CGMs; instead, sensor glucose (SG) values or other inputs are received by the ACE pump from compatible CGMs via Bluetooth Low Energy (BLE) technology are transmitted to the embedded PLGM algorithm.
The PLGM algorithm works in conjunction with the ACE pump. When enabled, the PLGM algorithm is able to suspend insulin delivery for a minimum of 30 minutes and for a maximum of 2 hours based on current or predicted sensor glucose values. It will automatically resume insulin delivery when maximum suspend time of 2 hours has elapsed or when underlying conditions resolve. The user is also able to manually resume insulin at any time.
The PLGM algorithm is a software-only device and does not have a user interface (UI). The compatible ACE pump provides the UI to configure therapy settings and interact with the algorithm. The PLGM-related alerts/alarms are displayed and managed by the pump.

The provided FDA 510(k) clearance letter and supporting documentation detail the acceptance criteria and the studies conducted to prove that Medtronic's SmartGuard Technology and Predictive Low Glucose Technology meet these criteria.

It's important to note that the provided text focuses on demonstrating substantial equivalence to a predicate device, as is typical for 510(k) submissions, rather than establishing de novo acceptance criteria for an entirely novel device. The "acceptance criteria" here refer to the performance benchmarks that demonstrate safety and effectiveness comparable to the predicate and compliance with regulatory special controls.

Here's an analysis of the acceptance criteria and the study that proves the device meets them, based on the provided text:

Acceptance Criteria and Device Performance

The acceptance criteria are generally implied by the comparative data presented against the predicate device (Control-IQ Technology) and the compliance with "iAGC Special Controls requirements defined in 21 CFR 862.1356." The clinical study primarily aims to demonstrate non-inferiority or beneficial outcomes in key glycemic metrics compared to baseline (run-in period).

Table of Acceptance Criteria and Reported Device Performance

Given that this is a 510(k) submission showing substantial equivalence, the "acceptance criteria" are largely derived from the performance of the predicate device and clinical guidelines (e.g., ADA guidelines for Time Below Range). While specific numerical thresholds for acceptance are not explicitly listed as "acceptance criteria" in a table format within the provided text, the results presented serve as the evidence that these implicit criteria are met.

For the purpose of this response, I will synthesize the implied performance targets from the "Pivotal Study Observed Results" and "Supplemental Clinical Data" sections and present the device's reported performance against them.

Table 1: Implied Acceptance Criteria (via Predicate Performance/Clinical Guidelines) and Reported Device Performance for SmartGuard Technology (AHCL Algorithm)

Table 2: Implied Acceptance Criteria and Reported Device Performance for Predictive Low Glucose Technology (PLGM Algorithm)

Study Details

1. Sample Sizes and Data Provenance

Test Set (Clinical Studies):

• SmartGuard Technology (AHCL Algorithm) - Pivotal Study:

  • Adults (18-80 years): 110 subjects enrolled (105 completed).
  • Pediatrics (7-17 years): 112 subjects enrolled (107 completed).
  • Total: 222 subjects enrolled (212 completed).
  • Provenance: Multi-center, single-arm study conducted across 25 sites in the U.S. This implies prospective data collection, specifically designed for this regulatory submission. Home-setting study.

• Predictive Low Glucose Technology (PLGM Algorithm) - Clinical Testing:

  • Adults (14-75 years): 71 subjects. In-clinic study.
  • Pediatrics (7-13 years): 105 subjects. In-clinic evaluation.
  • Total: 176 subjects.
  • Provenance: Multi-center, single-arm, in-clinic studies. Location not explicitly stated but part of a US FDA submission, implying US or international sites adhering to FDA standards. Prospective data.

Training Set:

• SmartGuard Technology & Predictive Low Glucose Technology (Virtual Patient Model):
  • Sample Size: Not explicitly stated as a number of "patients" but referred to as "extensive validation of the simulation environment" and "virtual patient (VP) model."
  • Data Provenance: In-silico simulation studies using Medtronic Diabetes' simulation environment. This is synthetic data generated by computational models, validated against real patient data.

2. Number of Experts and Qualifications for Ground Truth (Test Set)

The clinical studies for both SmartGuard and PLGM technologies involved direct measurement of glucose values via CGM and blood samples (YSI for PLGM study, and HbA1c for SmartGuard study). These are objective physiological measures, not subjective interpretations requiring external expert consensus for "ground truth."

• For SmartGuard Technology: Glucose values were measured by the Simplera Sync CGM and HbA1c by laboratory tests. These are considered objective measures of glycemic control.
• For Predictive Low Glucose Technology: Hypoglycemia induction was monitored with frequent sample testing (FST) and frequent blood sampling for glucose measurements (likely laboratory-grade methods like YSI [Yellow Springs Instrument]).
• Expert involvement: While healthcare professionals (investigators, study coordinators, endocrinologists, nurses) were undoubtedly involved in conducting the clinical studies, managing patient care, and interpreting results, their role was not to establish "ground truth" through consensus or adjudication in the sense of image review. The ground truth was physiological measurements.

3. Adjudication Method for the Test Set

Not applicable in the typical sense of subjective clinical assessments (e.g., radiology image interpretation). Ground truth was established by direct physiological measurements (CGM data, HbA1c, YSI/FST blood glucose). The clinical studies were single-arm studies where subject outcomes were measured, not comparative assessments where multiple readers adjudicate on decisions.

4. Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study

No, an MRMC comparative effectiveness study was not described. The clinical studies were single-arm (evaluating the device's performance in a standalone setting against a baseline or predefined safety/efficacy metrics) and did not involve human readers interpreting data from the device to make clinical decisions and then comparing human performance with and without AI assistance. Instead, the AI (algorithm) directly controlled insulin delivery and was evaluated based on patient physiological outcomes.

5. Standalone (Algorithm Only) Performance

Yes, the core of the evaluation is the standalone performance of the algorithms (SmartGuard AHCL and PLGM) in managing glucose, as they are "software-only devices" that reside on the ACE pump firmware. The clinical studies directly measured the physiological impact of the algorithm's actions on glucose levels, HbA1c, and safety parameters in a human population.

6. Type of Ground Truth Used

• Clinical Studies (SmartGuard & PLGM): Objective Physiological Measurements

  • Sensor Glucose (SG) values: From compatible CGMs (Simplera Sync CGM, Guardian 4 CGM)
  • HbA1c: Laboratory measurements of glycosylated hemoglobin.
  • Frequent Sample Testing (FST) / Blood Glucose (BG) values: Clinical laboratory measurements (e.g., YSI) to confirm hypoglycemia during PLGM induction.
  • Adverse Events (AEs): Clinically reported and documented events.
  • These are considered the definitive "ground truth" for evaluating glycemic control and safety.

• In-Silico Simulation Studies: Virtual Patient Model Outputs

  • The "ground truth" for these simulations is the metabolic response of the validated virtual patient models. This computational modeling is used to extend the clinical evidence to various parameter settings and demonstrate equivalence to real-world scenarios.

7. Sample Size for the Training Set

The document does not explicitly state a numerical "sample size" for a distinct "training set" of patients in the traditional machine learning sense for the algorithms themselves. The algorithms are likely developed and refined using a combination of:

• Physiological modeling: Based on established understanding of glucose-insulin dynamics.
• Historical clinical data: From previous similar devices or general diabetes patient populations (though not specified in this document for algorithm training).
• Clinical expertise: Incorporated into the algorithm design.
• The "Virtual Patient Model" itself is a form of simulated data that aids in development and testing. The validation of this model is mentioned as "extensive validation" and establishment of "credibility," implying a robust dataset used to verify its accuracy against real patient responses.

It's typical for complex control algorithms like these to be developed iteratively with physiological models and potentially large historical datasets, but a specific "training set" size for a machine learning model isn't detailed.

8. How the Ground Truth for the Training Set was Established

As noted above, a distinct "training set" with ground truth in the conventional sense of labeling data for a machine learning model isn't described. The development of control algorithms often involves:

• Physiological Simulation: The ground truth for this is the accurate metabolic response as modeled mathematically.
• Clinical Expertise & Design Principles: The ground truth is embedded in the scientific and medical principles guiding the algorithm's control logic.
• Validation of Virtual Patient Model: The "equivalency was demonstrated between Real Patients (RPs) and Virtual Patients (VPs) in terms of predetermined characteristics and clinical outcomes." This suggests that real patient data was used to validate and establish the "ground truth" for the virtual patient model itself, ensuring it accurately mirrors human physiology. This validated virtual patient model then serves as a crucial tool for in-silico testing.

Ask a specific question about this device

Control-IQ+ technology is intended for use with compatible integrated continuous glucose monitors (iCGM) and alternate controller enabled (ACE) pumps to automatically increase, decrease, and suspend delivery of basal insulin based on iCGM readings and predicted glucose values. It can also deliver correction boluses when the glucose value is predicted to exceed a predefined threshold.

Control-IQ+ technology is intended for the management of Type 1 diabetes mellitus in persons 2 years of age and greater and of Type 2 diabetes mellitus in persons 18 years of age and greater.

Control-IQ+ technology is intended for single patient use and requires a prescription.

The Subject Device, Control-IQ+ technology ("Control-IQ+") is a software-only device intended for the management of type 1 and type 2 diabetes mellitus. The device controls insulin delivery from a compatible alternate controller enabled insulin pump (ACE pump) based on inputs provided by a compatible integrated continuous glucose monitor (iCGM) and inputs provided by the user (e.g., carbohydrate intake, exercise, and sleep schedule). Control-IQ+ technology is meant to be installed on a compatible ACE pump.

Control-IQ+ technology has three different modes: Normal, Sleep, and Exercise. The glucose targets are not individually customizable in these modes but can change based on the mode selected. During Normal mode, Control-IQ+ technology aims to control glucose within a target range of 112.5 – 160 mg/dL, during Sleep mode the target range is 112.5 – 120 mg/dL, and during Exercise mode the target range is 140 – 160 mg/dL.

Control-IQ+ technology includes an integrated feature whereby iCGM values are automatically populated into the glucose field of the integrated bolus calculator when Control-IQ+ technology is active (i.e., the device is operating in closed-loop mode). This feature is disabled when Control-IQ is turned off.

Control-IQ+ technology requires users to input their weight and their total daily insulin requirement, which should be established with the help of a health care provider before using the device.

This FDA 510(k) clearance letter describes the acceptance criteria and study for the Control-IQ+ technology, an interoperable automated glycemic controller.

It's important to note that this device is a software-only device (Control-IQ+ technology), and the primary change described in this 510(k) is the addition of a new compatible insulin (Lyumjev U-100 Insulin) for use with the existing Control-IQ technology. The clearance relies heavily on the substantial equivalence to a predicate device (K243823, Control-IQ+ technology) and a clinical study demonstrating the safety and effectiveness of the new compatible insulin with the existing Control-IQ system.

Here's the breakdown of the acceptance criteria and study proving device performance:

1. Table of Acceptance Criteria and Reported Device Performance

The document does not explicitly present a "table of acceptance criteria" for the Control-IQ+ technology's performance in terms of glucose control metrics (e.g., time in range, hypoglycemia rates) for this specific 510(k) submission. Instead, the "acceptance" for this submission appears to be based on demonstrating non-inferiority or better safety when using the new insulin (Lyumjev) with the already cleared Control-IQ system, compared to established benchmarks.

The primary "performance" mentioned is related to safety, specifically the rates of severe hypoglycemia and DKA.

2. Sample Size Used for the Test Set and Data Provenance

• Sample Size: 179 participants with type 1 diabetes.
  • 70 adults (18-75 years old)
  • 109 pediatric participants (6-17 years old)
• Data Provenance: The study was a "single-arm prospective safety trial." While the exact country of origin isn't specified, FDA clearances typically involve studies conducted in the US or under protocols recognized by the FDA. The T1D Exchange clinic registry is a US-based registry, suggesting a likely US context for the study.
• Retrospective/Prospective: Prospective. The study involved an initial (~16-day) Humalog Lead-in Period and a subsequent (13-week) Lyumjev Treatment Period.

3. Number of Experts Used to Establish Ground Truth for the Test Set and Qualifications

This type of device (automated glycemic controller) does not typically involve expert review for "ground truth" in the same way an imaging or diagnostic AI might. The "ground truth" for glucose control is direct physiological measurements (iCGM readings) and clinical outcomes (hypoglycemia, DKA). Therefore, specific numbers or qualifications of experts for establishing ground truth are not applicable in this context. The study design itself serves to establish the performance and safety against clinical outcomes.

4. Adjudication Method for the Test Set

Adjudication methods (e.g., 2+1, 3+1) are typically used in studies involving subjective expert interpretation of data (e.g., radiologists reviewing images). For a system controlling insulin delivery based on CGM data, adjudication of such a type is not applicable. Clinical events (severe hypoglycemia, DKA) are typically adjudicated by an independent clinical endpoint committee or medical monitors based on predefined criteria, but the document does not specify this level of detail.

5. Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study

No, an MRMC comparative effectiveness study was not done.
MRMC studies are relevant for diagnostic aids where human readers interpret data (e.g., images) with and without AI assistance to measure improvement in reader performance. Control-IQ+ technology directly controls insulin delivery; it is not an assistive diagnostic tool for human interpretation.

6. Standalone (Algorithm Only Without Human-in-the-Loop Performance)

The device is an "Interoperable Automated Glycemic Controller," meaning it operates to "automatically increase, decrease, and suspend delivery of basal insulin based on iCGM readings and predicted glucose values." It also delivers "correction boluses." This indicates that the core function is standalone (algorithm-only) in its closed-loop operation. While users can input data (carbohydrate intake, exercise, sleep schedule) and the device is intended for "single patient use" with a "prescription," the control logic itself functions automatically without continuous human intervention in real-time decision-making for insulin delivery adjustment. The study evaluates the system performance which includes this automated functionality.

7. The Type of Ground Truth Used

The ground truth used for this study was primarily:

• Physiological data: iCGM readings for glucose values.
• Clinical Outcomes Data: Rates of severe hypoglycemia and DKA events. These were compared against "reported frequencies from the T1D Exchange clinic registry" as a benchmark for safety.

8. The Sample Size for the Training Set

The document does not provide information regarding the sample size used for the training set of the Control-IQ+ technology. This 510(k) is for a modification (new compatible insulin) to an already cleared device, implying the core algorithm was trained and validated previously.

9. How the Ground Truth for the Training Set Was Established

The document does not provide information on how the ground truth for the training set of the original Control-IQ algorithm was established. Given the nature of an automated glycemic controller, it would typically involve extensive simulations, in-silico testing, and potentially prior clinical trials where continuous glucose monitoring (CGM) data, insulin delivery data, and corresponding blood glucose measurements were collected and used to train and validate the control algorithms. However, this specific 510(k) submission focuses on the safety and effectiveness of a new component (Lyumjev insulin) with the existing system rather than the initial foundational algorithm development.

Ask a specific question about this device

Control-IQ+ technology is intended for use with compatible integrated continuous glucose monitors (iCGM) and alternate controller enabled (ACE) pumps to automatically increase, decrease, and suspend delivery of basal insulin based on iCGM readings and predicted glucose values. It can also deliver correction boluses when the glucose value is predicted to exceed a predefined threshold.

Control-IQ+ technology is intended for the management of Type 1 diabetes mellitus in persons 2 years of age and greater and of Type 2 diabetes mellitus in persons 18 years of age and greater.

Control-IQ+ technology is intended for single patient use and requires a prescription.

Control-IQ+ technology (Control-IQ+, the device) is a software-only device intended for the management of type 1 and type 2 diabetes mellitus. The device controls insulin delivery from a compatible alternate controller enabled insulin pump (ACE pump) based on inputs provided by a compatible integrated continuous glucose monitor (iCGM) and inputs provided by the user (e.g., carbohydrate intake, exercise, and sleep schedule). Control-IQ+ technology is meant to be installed on a compatible ACE pump.

Control-IQ+ technology has three different modes: Normal, Sleep, and Exercise. The glucose targets are not individually customizable in these modes but can change based on the mode selected. During normal mode, Control-IQ+ technology aims to control glucose within a target range of 112.5 - 160 mg/dL. During sleep mode, this range is changed to 112.5-120 mg/dL. and it is changed to 140-160 mg/dL during exercise mode.

Control-IQ+ technology includes an integrated feature whereby iCGM values are automatically populated into the glucose field of the integrated bolus calculator when Control-IQ + technology is active (i.e., the device is operating in closed-loop mode). This feature is disabled when Control-IQ is turned off.

Control-IQ+ technology requires users to input their weight and their total daily insulin requirement, which should be established with the help of a health care provider before using the device.

Here's an analysis of the provided text to extract information about acceptance criteria and the study proving the device meets them, presented in the requested format.

It's important to note that the provided text is an FDA 510(k) clearance letter and a 510(k) summary. These documents primarily focus on demonstrating substantial equivalence to a predicate device, rather than providing a detailed technical breakdown of novel acceptance criteria for algorithm performance in the same way one might expect for a new AI/ML device where performance metrics (e.g., sensitivity, specificity, accuracy) are the primary basis of clearance.

For Control-IQ+ technology, the key change is the expansion of the intended use to include Type 2 diabetes patients, supported by a pivotal clinical study. The algorithm itself is stated to be "unchanged," implying that its core performance characteristics (e.g., glucose control targets, basal rate adjustments) are those already established for Type 1 diabetes with the predicate device. Therefore, the "acceptance criteria" here largely pertain to the clinical safety and effectiveness in the new population, rather than the internal algorithmic performance metrics.

Acceptance Criteria and Device Performance Study for Control-IQ+ Technology

Given that the device is a refined version of an already cleared product ("Control-IQ technology" K232382) and the core algorithm is "unchanged," the acceptance criteria for this 510(k) submission (K243823) are primarily focused on demonstrating the safety and effectiveness of expanding its indications to include Type 2 diabetes patients. The primary evidence for this is a clinical study.

1. Table of Acceptance Criteria and Reported Device Performance

As this is a 510(k) focusing on an expanded user population for an existing algorithm rather than a de novo submission for a novel AI algorithm with specific performance cutoffs, the acceptance criteria are framed in terms of clinical outcomes rather than typical AI performance metrics like sensitivity/specificity.

2. Sample Size Used for the Test Set and Data Provenance

• Test Set (Clinical Study):
  • Sample Size: 319 subjects
  • Data Provenance: Not explicitly stated (e.g., country of origin), but implies a multi-site prospective clinical trial. The study was a "randomized controlled trial."
  • Retrospective/Prospective: Prospective

3. Number of Experts Used to Establish Ground Truth for the Test Set and Qualifications of Those Experts

This information is not provided in the document. For a clinical trial of an automated glycemic controller, the "ground truth" for evaluating effectiveness is typically objective clinical measurements (e.g., HbA1C tests, CGM data, adverse events) rather than expert consensus on diagnostic images. Clinical outcomes are inherently the "ground truth." Expert physicians would design the trial, manage patients, and interpret results, but they aren't establishing a "ground truth" per se in the way radiologists might do for disease detection.

4. Adjudication Method for the Test Set

This information is not provided. For a randomized controlled trial measuring objective metrics like HbA1C, formal adjudication in the sense of a diagnostic agreement (e.g., 2+1 radiology consensus) is generally not applicable, as the outcomes are quantitative and measured directly.

5. If a Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study Was Done, What was the Effect Size of How Much Human Readers Improve with AI vs. Without AI Assistance

• MRMC Study: No, an MRMC study was not done. This type of study is relevant for diagnostic imaging AI, where human readers (e.g., radiologists) interpret images with or without AI assistance.
• Effect Size of Human Improvement with AI: Not applicable, as this device is an automated glycemic controller, not an AI assisting human interpretation of medical images. The clinical study compares the device's performance directly against a control group (CGM with continued insulin injections), rather than evaluating human reader performance with and without AI assistance.

6. If a Standalone (i.e. algorithm only without human-in-the-loop performance) Was Done

• The "Clinical Testing" section describes a pivotal clinical study where Control-IQ+ technology's performance (an automated insulin delivery system) was evaluated in patients. While it's implied that the system operates autonomously based on CGM inputs and user settings, the study measures the system's performance in a real-world setting with human users interacting with the device, not a purely standalone algorithmic evaluation in a simulated environment. The study is effectively evaluating the device's performance as intended for clinical use.

7. The Type of Ground Truth Used

• The primary ground truth for the clinical study was outcomes data, specifically the change in HbA1C (Glycated Hemoglobin), which is a key clinical endpoint for diabetes management. Other clinical measures from the iCGM (e.g., glucose values, time in range) and safety outcomes would also serve as ground truth for evaluating the device's performance.

8. The Sample Size for the Training Set

• This information is not provided in the document. The document states that "The design of the Control-IQ+ algorithm is unchanged" from its predicate. This implies that the training of the core algorithm was done prior to the predicate device's clearance and is not being re-evaluated for this submission. The 510(k) is primarily about demonstrating safety and effectiveness in the expanded user population, not re-training or re-validating the core algorithm's initial development.

9. How the Ground Truth for the Training Set Was Established

• This information is not provided as the algorithm itself is stated to be unchanged from the predicate device. Details about the original algorithm's development, including its training data and how its ground truth was established, would have been part of the predicate device's (K232382) submission.

Ask a specific question about this device

SmartAdjust™ technology is intended for use with compatible integrated continuous glucose monitors (iCGM) and alternate controller enabled (ACE) pumps to automatically increase, decrease, and pause delivery of insulin based on current and predicted glucose values. SmartAdjust™ technology is intended for the management of type 1 diabetes mellitus in persons 2 years of age and older and type 2 diabetes mellitus in persons 18 years of age and older. SmartAdjust™ technology is intended for single patient use and requires a prescription.

SmartAdjust™ technology is a software-only device that enables automated insulin delivery by taking in glucose inputs from a connected iCGM and calculating insulin micro-bolus outputs for delivery by a connected ACE Pump.

SmartAdjust™ technology is part of the Omnipod 5 Automated Insulin Delivery System, which also includes the Omnipod 5 ACE Pump and the SmartBolus Calculator regulated devices. The Omnipod 5 ACE Pump and the SmartBolus Calculator are functionally independent from SmartAdjust™ technology. SmartAdjust™ technology is intended to be digitally connected to a third party iCGM, the Omnipod 5 ACE Pump, and the SmartBolus Calculator.

The SmartAdjust™ technology software is installed on both the Omnipod 5 Pod and Omnipod 5 Controller (which contains the Omnipod 5 App), the 2 physical components that make up the Omnipod 5 System.

The Omnipod 5 System is a hybrid closed loop system and therefore can operate in either open loop (Manual Mode; SmartAdjust™ technology disabled) or closed loop (Automated Mode; SmartAdjust™ technology enabled). When Automated Mode is turned on, the SmartAdjust™ algorithm (installed on the Pod) controls insulin delivery based on recent CGM values.

Here's a breakdown of the acceptance criteria and the study that proves the device meets them, based on the provided text:

This document describes the 510(k) premarket notification for the SmartAdjust™ Technology, which is an interoperable automated glycemic controller. The submission aims to expand the indications for use to include individuals with Type 2 Diabetes Mellitus aged 18 years and older, in addition to the existing indication for Type 1 Diabetes Mellitus.

1. A table of acceptance criteria and the reported device performance

The provided text doesn't explicitly list a table of "acceptance criteria" with numerical targets in a typical performance study format. However, it does report on a primary safety endpoint from a clinical study which effectively serves as a key performance metric for the expanded indication.

Here's a table based on the information provided, focusing on the clinical study's primary safety endpoint:

Ask a specific question about this device

SmartAdjust™ technology is intended for use with compatible integrated continuous glucose monitors (iCGM) and alternate controller enabled (ACE) pumps to automatically increase, decrease, and pause delivery of insulin based on current and predicted glucose values. SmartAdjust™ technology is intended for the management of type 1 diabetes mellitus in persons 2 years of age and older. SmartAdjust™ technology is intended for single patient use and requires a prescription.

SmartAdjust™ technology is a software-only device that enables automated insulin delivery by taking in glucose inputs from a connected iCGM and calculating insulin micro-bolus outputs for delivery by a connected ACE Pump.

SmartAdjust™ technology is part of the Omnipod 5 Automated Insulin Delivery System, which also includes the Omnipod 5 ACE Pump and the SmartBolus Calculator regulated devices. The Omnipod 5 ACE Pump and the SmartBolus Calculator functions are functionally independent from SmartAdjust™ technology. SmartAdjust™ technology is intended to be digitally connected to a third party iCGM, the Omnipod 5 ACE Pump, and the SmartBolus Calculator.

The SmartAdjust™ technology software is installed on both the OP5 Pod and OP5 Controller (which contains the OP5 App), the 2 physical components that make up the Omnipod 5 System.

The Omnipod 5 System is a hybrid closed loop system and therefore can operate in either open loop (Manual Mode; SmartAdjust™ technology disabled) or closed loop (Automated Mode; SmartAdjust™ technology enabled). When Automated Mode is turned on, the SmartAdjust™ alqorithm (installed on the Pod) controls insulin delivery based on recent CGM values.

This document is an FDA 510(k) clearance letter and summary for the Insulet SmartAdjust™ technology. It states that the device is substantially equivalent to a previously cleared predicate device (K220394), with the only change being an update to the specifications for glucose sensor performance for compatible integrated continuous glucose monitors (iCGMs). This means the core algorithm and its performance were already established with the original clearance.

Therefore, the document does not contain the detailed study information typically found in a new device submission or a clinical trial report that directly proves the device meets specific acceptance criteria in a new clinical study. Instead, it relies on the substantial equivalence argument, implying that the previous studies for the predicate device, or studies demonstrating the updated iCGM performance, are sufficient.

Based on the provided text, here's what can be extracted and what cannot:

Information Available:

• Device Name: SmartAdjust™ technology
• Predicate Device: SmartAdjust™ technology (K220394)
• Change: Updated iCGM Performance Specifications per 21 CFR 862.1356(1)(iv).
• Conclusion: The subject device is substantially equivalent to its predicate. The differences do not raise new questions of safety and effectiveness.

Information NOT Available (because this is a substantial equivalence submission for a minor update, not a new clinical trial submission document):

• A table of acceptance criteria and the reported device performance: This document does not detail specific performance metrics or acceptance criteria for a new study, as it's an update to an already cleared device. The performance requirements were presumably met by the predicate device and the updated iCGM.
• Sample sizes used for the test set and the data provenance: Not provided.
• Number of experts used to establish the ground truth for the test set and the qualifications of those experts: Not provided.
• Adjudication method: Not applicable/provided for this type of submission.
• If a multi-reader multi-case (MRMC) comparative effectiveness study was done: Not applicable/provided.
• If a standalone (i.e., algorithm only without human-in-the-loop performance) was done: The device is an algorithm/software, but the clinical performance data related to its efficacy (how well it manages glucose) would typically be from human-with-device studies from the original predicate submission. No new standalone study details are here.
• The type of ground truth used: Not specified, but for a glycemic controller, this would typically involve actual blood glucose measurements.
• The sample size for the training set: Not provided.
• How the ground truth for the training set was established: Not provided.

Reconstruction of "Acceptance Criteria" based on the document's type:

Since this is a 510(k) for an update, the "acceptance criteria" revolve around demonstrating that the change (updated iCGM specifications) does not negatively impact safety or effectiveness, and that the device still meets the requirements for its classification.

Table of "Acceptance Criteria" and "Reported Device Performance" as implied by a 510(k) for an update:

Study Information (based on the context of a 510(k) for an update):

• Sample size used for the test set and data provenance: No specific sample size for new clinical testing is mentioned. The approval hinges on the existing data for the predicate device and the updated iCGM meeting its own specifications. The data provenance for the original predicate would have included clinical trial data (likely prospective).
• Number of experts and qualifications, Adjudication method, MRMC study, Standalone performance: These details are not relevant to this specific 510(k) summary, as it's establishing equivalence based on a minor technical update, not presenting new clinical efficacy data for the core algorithm from scratch. Such studies would have been part of the original K220394 submission.
• Type of ground truth used: For the underlying technology (which is unchanged here), the ground truth for an automated glycemic controller would be actual blood glucose measurements, measured by a reference method (e.g., lab venous plasma glucose, or accurate point-of-care devices).
• Training Set Sample Size and Ground Truth Establishment: Not mentioned, as this is an established, already-trained algorithm.

In summary, this document is a regulatory approval for a minor update to an already-cleared medical device software, rather than a detailed report of a new clinical study. The "proof" lies in the argument of substantial equivalence to the predicate device, with the specific change (iCGM specifications) not introducing new safety or effectiveness concerns.

Ask a specific question about this device

CamAPS FX is a mobile app intended for managing glucose levels in people with type 1 diabetes, aged 2 years and over, using a hybrid closed loop approach (automated basal insulin delivery with manual bolusing for meals). Additional age and other restrictions may apply depending on the chosen continuous glucose monitor and insulin pump.

CamAPS FX requires an insulin pump and a continuous glucose monitor (CGM) to fulfil its intended use. The list of supported insulin pumps and CGMs is provided in the User Manual.

CamAPS FX is indicated for use in pregnancy complicated by type 1 diabetes provided that the linked continuous glucose monitoring system is suitable for use in pregnancy.

CamAPS FX is for prescription use only.

CamAPS FX is a mobile app.

This FDA 510(k) clearance letter (K232603) for the CamAPS FX device does not contain the detailed acceptance criteria and study information typically found in a clinical study report or a more comprehensive summary of safety and effectiveness.

The document primarily focuses on the regulatory clearance of the device, stating that it is substantially equivalent to legally marketed predicate devices and outlining the indications for use. It details regulatory compliance aspects such as general controls, special controls, the Quality System regulation, and reporting requirements.

Therefore, I cannot extract the requested information from the provided text. To answer your questions, one would need access to the full 510(k) submission, including the performance data and clinical study summaries.

Based on the provided text, the answer to all your questions about acceptance criteria and the study that proves the device meets them would be: "Information not available in the provided document."

The document mentions:

• Device Name: CamAPS FX
• Regulation Number/Name: 21 CFR 862.1356, Interoperable automated glycemic controller
• Regulatory Class: Class II
• Indications for Use: Managing glucose levels in people with type 1 diabetes, aged 2 years and over, using a hybrid closed loop approach (automated basal insulin delivery with manual bolusing for meals). Also indicated for use in pregnancy complicated by type 1 diabetes, provided the linked CGM is suitable. Requires an insulin pump and CGM. Prescription use only.

However, it does not provide any details regarding:

1. A table of acceptance criteria and reported device performance.
2. Sample sizes or data provenance for test sets.
3. Number and qualifications of experts for ground truth.
4. Adjudication method.
5. Multi-reader multi-case comparative effectiveness study results or effect sizes.
6. Standalone algorithm performance.
7. Type of ground truth used.
8. Training set sample size.
9. How training set ground truth was established.

Ask a specific question about this device

DEKA Loop is intended for use with compatible integrated continuous glucose monitors (iCGM) and the DEKA alternate controller enabled (ACE) insulin infusion pump to automatically increase, and suspend delivery of basal insulin based on iCGM readings and predicted glucose values. It can also recommend, and with the user's confirmation, control the delivery of correction boluses when glucose values are predicted to exceed user configurable thresholds.

DEKA Loop is intended for the management of Type 1 diabetes mellitus in persons six years of age and greater.

DEKA Loop is intended for single patient use and requires a prescription.

DEKA Loop is an interoperable Alternate Glycemic Controller (iAGC) and works to control an ACE (Alternate Controller Enabled) insulin pump to automatically increase, decrease, and suspend delivery of basal insulin based on readings from an iCGM (integrated continuous glucose monitor) and glucose values predicted by DEKA Loop. DEKA Loop can also recommend, and with the user's confirmation, control the delivery of correction boluses when glucose values are predicted to exceed user configurable thresholds. It is controlled by an iOS app that is downloaded to a user's iPhone.

The provided document, an FDA 510(k) K234055 clearance letter for the DEKA Loop device, focuses on demonstrating substantial equivalence to a predicate device (Tidepool Loop K203689). It details the device's technological characteristics and mentions performance data primarily through in silico testing for clinical equivalence. However, the document does not directly provide specific acceptance criteria or detailed results of a study designed to prove the device meets those criteria in the format requested (e.g., a table with numerical acceptance values and reported performance). Nor does it describe patient-level ground truth establishment, expert adjudication, or MRMC studies.

The information below is extracted and inferred from the provided text, highlighting what is available and what is explicitly not mentioned or detailed in relation to your specific questions.

Here's a breakdown based on the provided text:

Device Performance Acceptance Criteria and Study Details (Based on available information)

The document primarily relies on demonstration of substantial equivalence to a predicate device (Tidepool Loop) based on technological, functional, and performance characteristics, rather than establishing de novo performance criteria. The "Performance Data" section specifically states: "Additionally, in-silico software challenge testing demonstrated clinical equivalence to the predicate device."

1. Table of Acceptance Criteria and Reported Device Performance:

The document does not provide a table with specific numerical acceptance criteria and corresponding reported device performance values. Instead, it states that "in-silico testing proves that the DEKA Loop algorithm is clinically equivalent to the Tidepool Loop Algorithm." This implies that the 'acceptance' for clinical performance was demonstrating equivalence through in silico methods.

2. Sample Size Used for the Test Set and Data Provenance:

• Sample Size (Test Set): Not explicitly stated. The document mentions "in-silico software challenge testing." This implies a simulated patient cohort, but the size of this cohort is not provided in terms of "samples."
• Data Provenance: The nature of in silico testing means it's not based on ex vivo or in vivo patient data in the traditional sense for the test set. It's a computational simulation.
  • For the predicate device's clinical performance (which the subject device aims to be equivalent to), the document states: "Tidepool Loop clinical performance is supported by representative 1,250 participants in a 15 months duration real-world, observational, single arm study of DIY Loop including both pediatric and adult participants." This refers to the predicate's data, not the subject device's in silico test set.

3. Number of Experts Used to Establish Ground Truth and Qualifications:

• Not applicable / Not stated. Ground truth, in the context of an in silico study for a glycemic controller, would likely refer to the accuracy of the simulated physiological model against known physiological principles or real-world data characteristics, rather than expert annotation of medical images or diagnoses. No human experts are mentioned for establishing ground truth for the in silico test set.

4. Adjudication Method for the Test Set:

• Not applicable / Not stated. Given the in silico nature and lack of human expert involvement in "ground truth" establishment as typically understood in AI imaging, no adjudication method is described.

5. If a Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study was done:

• No. An MRMC study is relevant for human-in-the-loop performance studies, particularly in medical imaging where radiologists or clinicians interpret cases. The DEKA Loop is an automated glycemic controller. The document does not describe any study where human readers (e.g., clinicians) used the DEKA Loop (or a simulated version) to assess its comparative effectiveness against a standard of care or the predicate with outcomes like improved blood glucose control. The stated clinical performance evaluation was in silico device-to-predicate algorithm equivalence.

6. If a Standalone (i.e., algorithm only without human-in-the-loop performance) was done:

• Yes, implicitly. The "in-silico software challenge testing" is an algorithm-only (standalone) performance evaluation. It assessed the DEKA Loop algorithm's performance against the predicate's algorithm in a simulated environment, without direct human intervention in the loop of glucose regulation within the test.

7. The Type of Ground Truth Used:

• For the in-silico testing, the ground truth would be based on the simulated physiological model's behavior, which is designed to accurately represent human glucose metabolism and insulin action under various conditions. This is a form of simulated data / model-based ground truth rather than expert consensus, pathology, or direct patient outcomes data from a clinical trial for the subject device itself. The goal was to prove "clinical equivalence to the predicate device," meaning the in silico performance mirrored what the predicate device achieved in its real-world clinical study.

8. The Sample Size for the Training Set:

• Not stated. As this is a 510(k) submission for an existing algorithm (the "Loop" algorithm, which DEKA Loop is shown to be equivalent to), and not a novel AI/ML algorithm requiring de novo training, details about its original training set (if any, as an "algorithm" might be more deterministic control logic than a learned AI model in some cases) are not provided in this regulatory document. The focus is on the validation that the DEKA Loop implementation of the algorithm is equivalent to the predicate's.

9. How the Ground Truth for the Training Set was Established:

• Not stated. Refer to point 8. If the algorithm involved machine learning, its original training (if any) would have required a separate dataset and ground truth establishment method, which is not detailed in this 510(k) summary. Given the description ("predicts glucose levels... based on prior iCGM readings, insulin delivery history, and user input... uses that prediction to adjust insulin delivery"), it sounds more like a model-based predictive control algorithm rather than a deep learning model trained on a large dataset with ground truth labels.

Ask a specific question about this device

Control-IQ technology is intended for use with compatible integrated continuous glucose monitors (iCGM) and alternate controller enabled (ACE) pumps to automatically increase, decrease, and suspend delivery of basal insulin based on iCGM readings and predicted glucose values. It can also deliver correction boluses when the glucose value is predicted to exceed a predefined threshold.

Control-IQ technology is intended for the management of Type 1 diabetes mellitus in persons 2 years of age and greater.

Control IQ technology is intended for single patient use and requires a prescription.

Control-IQ technology (Control-IQ, the device) is a software-only device intended for the management of type 1 diabetes mellitus. The device controls insulin delivery from a compatible alternate controller enabled insulin pump (ACE pump) based on inputs provided by a compatible integrated continuous glucose monitor (iCGM) and inputs provided the user (e.g., carbohydrate intake, exercise, and sleep schedule). Control-IQ technology is meant to be installed on a compatible ACE pump.

Control-IQ Technology has three different modes: Normal, Sleep, and Exercise. The glucose targets are not customizable in these modes but can change based on the mode selected. During normal mode, Control-IQ Technology aims to control glucose within a target range of 112.5 – 160 mg/dL. During sleep mode, this range is changed to 112.5-120 mg/dL, and it is changed to 140-160 mg/dL during exercise mode.

Control-IQ technology includes an integrated feature whereby iCGM values are automatically populated into the glucose field of the integrated bolus calculator when the Control-IQ technology is active (i.e., the device is operating in closed-loop mode). This feature is disabled when Control-IQ is turned off.

Control-IQ technology requires users to input their total daily insulin requirement, which should be established with the help of a health care provider before using the device.

The provided text describes the acceptance criteria and the study conducted for the Tandem Diabetes Care Control-IQ Technology (K232382).

Here's the breakdown of the information requested:

1. Table of Acceptance Criteria and Reported Device Performance

The document does not explicitly present a table of specific numerical acceptance criteria for the Control-IQ technology. However, it states the primary outcome of the pivotal clinical study. The device's performance is reported in relation to this outcome.

2. Sample Size Used for the Test Set and Data Provenance

• Sample Size: 102 subjects were enrolled in the pivotal clinical study.
• Data Provenance: The document does not specify the country of origin. It indicates the study was a prospective, randomized controlled trial.

3. Number of Experts Used to Establish Ground Truth for the Test Set and Their Qualifications

This information is not provided in the document. The study's "ground truth" for glucose values would inherently come from the continuous glucose monitors (CGM) used by the subjects in the trial, but the involvement of independent experts to establish a "ground truth" for the test set is not mentioned.

4. Adjudication Method for the Test Set

This information is not provided in the document.

5. If a Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study Was Done, and the Effect Size of Human Readers Improve With AI vs. Without AI Assistance

A MRMC comparative effectiveness study involving human readers and AI assistance was not done. This device is an automated glycemic controller, where the AI (Control-IQ technology) directly controls insulin delivery, rather than assisting human interpretation of data for medical decision making.

6. If a Standalone (i.e., algorithm only without human-in-the-loop performance) Was Done

Yes, the pivotal clinical study directly evaluated the performance of the Control-IQ technology (algorithm only without human-in-the-loop for basal insulin adjustments and correction boluses based on predictions). Subjects were randomized to either Control-IQ or Standard Care, directly comparing the automated system's performance.

7. The Type of Ground Truth Used

The ground truth for the primary outcome (percent time in range 70-180 mg/dL) was based on CGM measured glucose values.

8. The Sample Size for the Training Set

The document does not provide information on the sample size used for the training set of the Control-IQ algorithm. The clinical study described is a pivotal trial for evaluation of the device, not necessarily for its training.

9. How the Ground Truth for the Training Set Was Established

The document does not provide information on how the ground truth for the training set was established.

Ask a specific question about this device

The iLet Dosing Decision Software is intended for use with compatible integrated continuous glucose monitors (iCGM) and alternate controller enabled (ACE) pumps. A self-monitoring of blood glucose (SMBG) meter may also be used for manual input of blood glucose values to continue insulin dosing for a limited period of time when input from the iCGM is temporarily not available. The iLet Dosing Decision Software autonomously determines and commands an increase, decrease, maintenance, or suspension of all basal doses of insulin and autonomously determines and commands correction doses of insulin based on input from an iCGM, and it autonomously determines and commands meal doses of insulin based on meal announcements. iLet Dosing Decision Software is intended for the management of type 1 diabetes mellitus in people 6 years of age or older. iLet Dosing Decision Software is intended for single patient use and requires a prescription.

The iLet Dosing Decision Software is an iAGC indicated for the management of type 1 diabetes mellitus. It autonomously determines and commands an increase, decrease, maintenance, or suspension of all basal doses of insulin and autonomously determines and commands correction doses of insulin based on input from an iCGM, and it autonomously determines and commands meal doses of insulin based on meal announcements. The iLet Dosing Decision Software is intended for the management of type 1 diabetes in people 6 years of age or older.

The iLet Dosing Decision Software works in conjunction with a compatible alternate controller enabled (ACE) pump. The iLet Dosing Decision Software only requires initialization with the user's body mass (body weight).

The iLet Dosing Decision Software does not require carbohydrate counting by the user or the use of carbohydrate- to-insulin ratios. Although the iLet system does not require a user to enter an exact carb amount to calculate and administer a meal bolus, it does require that the user announce the meal (e.g., breakfast, lunch, dinner) AND provide an estimated carb content as "Usual", "More", or "Less" than is routine for that meal type.

The iLet Dosing Decision Software does not require any information about the user's total daily dose of insulin, basal or long-acting insulin requirements, or insulin correction factors. It is an insulin titration system that requires no insulin-dose determinations by the user or provider. During normal operation, the iLet bionic pancreas (iLet ACE Pump with the iLet Dosing Decision Software installed) autonomously responds every five minutes to a glucose signal, from an iCGM that is worn by the user, by computing a control signal that translates to a dose of insulin, which is delivered to the user through the subcutaneous (SC) route. The iLet dosing decision software has three insulin controllers (algorithms) running in parallel: an adaptive basal insulin controller, which continually adapts to each individual's basal metabolic need for insulin, an adaptive bolus controller which provides doses that are required above and beyond the basal metabolic needs, and an adaptive meal dose controller which provides insulin in response to a meal announcement.

The iLet is intended to dose insulin based on CGM data. In the events where CGM stops providing glucose data to the iLet Dosing Decision Software BG-run mode feature will serve to temporarily continue insulin delivery. BG-run mode will determine and command basal insulin based on past requirements and will allow announcement of meals and entry of fingerstick BG measurements, which will be treated as iCGM data and may result in commanding administration of insulin or temporary suspension of basal insulin. BG-run mode use should always be for the shortest duration possible with the goal to resume CGM.

Here's a breakdown of the acceptance criteria and the study details for the iLet Dosing Decision Software, based on the provided document:

1. Table of Acceptance Criteria and Reported Device Performance

The document does not explicitly state formal "acceptance criteria" in a quantitative manner (e.g., "HbA1c must decrease by X%"). Instead, it presents key outcomes from the clinical study and concludes that the modified device's performance regarding safety and effectiveness is comparable to the predicate device.

Ask a specific question about this device

The iLet Dosing Decision Software is intended for use with compatible integrated continuous glucose monitors (iCGM) and alternate controller enabled (ACE) pumps. A self-monitoring of blood glucose (SMBG) meter may also be used for manual input of blood glucose values to continue insulin dosing for a limited period of time when input from the iCGM is temporarily not available. The iLet Dosing Decision Software autonomously determines and commands an increase, decrease, maintenance, or suspension of all basal doses of insulin and autonomously determines and commands correction doses of insulin based on input from an iCGM, and it autonomously determines and commands meal doses of insulin based on meal announcements. iLet Dosing Decision Software is intended for the management of type 1 diabetes mellitus in people 6 years of age or older. iLet Dosing Decision Software is intended for single patient use and requires a prescription.

The iLet Dosing Decision Software is an iAGC indicated for the management of type 1 diabetes mellitus. It autonomously determines and commands an increase, decrease, maintenance, or suspension of all basal doses of insulin and autonomously determines and commands correction doses of insulin based on input from an iCGM, and it autonomously determines and commands meal doses of insulin based on meal announcements. The iLet Dosing Decision Software is intended for the management of type 1 diabetes in people 6 years of age or older. The iLet Dosing Decision Software works in conjunction with a compatible alternate controller enabled (ACE) pump. The dosing decision software includes adaptive control algorithms that autonomously and continually adapt to the ever-changing insulin requirements of each individual to enable lifelong adaptive learning. The iLet Dosing Decision Software only requires initialization with the user's body mass (body weight). The iLet Dosing Decision Software does not require carbohydrate counting by the user or the use of carbohydrate- to-insulin ratios. Although the iLet system does not require a user to enter an exact carb amount to calculate and administer a meal bolus, it does require that the user announce the meal (e.g., breakfast, lunch, dinner) AND provide an estimated carb content as "Usual", "More", or "Less" than is routine for that meal type. The iLet Dosing Decision Software does not require any information about the user's total daily dose of insulin, basal or long-acting insulin requirements, or insulin correction factors. It is an insulin titration system that requires no insulin-dose determinations by the user or provider. During normal operation, the iLet bionic pancreas (iLet ACE Pump with the iLet Dosing Decision Software installed) autonomously responds every five minutes to a glucose signal, from an iCGM that is worn by the user, by computing a control signal that translates to a dose of insulin, which is intended to be delivered to the user through the subcutaneous (SC) route. The iLet dosing decision software has three insulin controllers (algorithms) running in parallel: an adaptive basal insulin controller, which continually adapts to each individual's basal metabolic need for insulin, an adaptive bolus controller which provides doses that are required above and beyond the basal metabolic needs, and an adaptive meal dose controller which provides insulin in response to a meal announcement. The iLet is intended to dose insulin based on CGM data. In the events where CGM stops providing glucose data to the iLet Dosing Decision Software BG-run mode feature will serve to temporarily continue insulin delivery. BG-run mode will determine and command basal insulin based on past requirements and will allow announcement of meals and entry of fingerstick BG measurements, which will be treated as iCGM data and may result in commanding administration of insulin or temporary suspension of basal insulin. BG-run mode use should always be for the shortest duration possible with the goal to resume CGM.

The provided text describes the iLet® Dosing Decision Software, an interoperable automated glycemic controller (iAGC), and the study conducted to demonstrate its performance.

Here's an analysis of the acceptance criteria and study as requested:

1. A table of acceptance criteria and the reported device performance

The document doesn't explicitly list "acceptance criteria" in a bulleted or numbered format with corresponding performance metrics like a typical FDA performance table. However, the "Endpoints" section in the Clinical Performance summary serves as the de facto acceptance criteria for the clinical study outcomes. The "Conclusions" section then describes how the device performed against these.

2. Sample size used for the test set and the data provenance (e.g. country of origin of the data, retrospective or prospective)

• Sample Size for the Clinical Study (RCT): 440 adult and child participants.
• Country of Origin: United States (16 clinical sites).
• Study Design: Prospective, multi-center, randomized controlled trial (RCT).
• Ancillary Study (BG-run feature): Participants in the BP Groups had the option of participating in this ancillary study, but a specific sample size for this ancillary study is not provided, only that it followed the RCT.

3. Number of experts used to establish the ground truth for the test set and the qualifications of those experts (e.g. radiologist with 10 years of experience)

This information is not provided in the document. The ground truth for the clinical study was established by the actual physiological responses and clinical outcomes of the participants with Type 1 Diabetes, measured by standard medical metrics (HbA1c, CGM data, adverse events). There is no mention of external experts establishing a "ground truth" for the device's dosing decisions themselves, as the device is designed to operate autonomously. The study evaluated the effectiveness and safety of the device's autonomous decisions.

4. Adjudication method (e.g. 2+1, 3+1, none) for the test set

This is not applicable to this type of study. Adjudication methods like 2+1 or 3+1 refer to expert consensus processes for evaluating medical images or diagnoses, typically used when establishing ground truth for AI algorithms in diagnostic imaging. For this device, which makes automated dosing decisions for diabetes management, the "ground truth" is physiological response, not expert interpretation. Adverse events would typically be adjudicated by a Clinical Events Committee (CEC), but the specific method (e.g., how many members reviewed each event) is not detailed.

5. If a multi reader multi case (MRMC) comparative effectiveness study was done, If so, what was the effect size of how much human readers improve with AI vs without AI assistance

• A Multi-Reader Multi-Case (MRMC) comparative effectiveness study was not done.
• This type of study is primarily relevant for diagnostic imaging AI where human readers interpret medical images. The iLet Dosing Decision Software is an automated glycemic controller, not an imaging interpretation aid.
• The study was a randomized controlled trial comparing the iLet system (which is the AI, managing insulin autonomously) to "standard care" (human-managed insulin delivery, either by pump or injections, though with CGM monitoring). It assesses the device's performance versus standard human-led care, not how human readers improve with AI assistance.

6. If a standalone (i.e. algorithm only without human-in-the-loop performance) was done

• Yes, a standalone study was done in the sense that the iLet Dosing Decision Software operates autonomously, commanding insulin doses without real-time human intervention in its decision-making process. The clinical trial directly evaluated this autonomous "algorithm only" performance within the iLet Bionic Pancreas System.
• The comparison was between the iLet system (operating autonomously) and standard human-managed care.

7. The type of ground truth used (expert consensus, pathology, outcomes data, etc)

The ground truth for evaluating the iLet Dosing Decision Software's performance in the clinical study was primarily outcomes data and physiological measurements:

• HbA1c (a measure of average blood glucose over time).
• Continuous Glucose Monitoring (CGM) metrics (e.g., time in target range, time spent in hypo/hyperglycemia, mean glucose, standard deviation).
• Safety outcomes (severe hypoglycemia, diabetic ketoacidosis, other serious adverse events).

8. The sample size for the training set

The document does not provide information regarding the sample size used for the training set of the iLet Dosing Decision Software algorithm. It only details the clinical study for validation of the device.

9. How the ground truth for the training set was established

The document does not provide information on how the ground truth for the training set was established. The iLet Dosing Decision Software employs "adaptive control algorithms that autonomously and continually adapt to the ever-changing insulin requirements of each individual to enable lifelong adaptive learning." This suggests a machine learning or adaptive control approach, which would have been trained on or developed using a dataset, but the specifics of that training data and ground truth establishment are not disclosed in this summary.

Ask a specific question about this device