Search Results
Found 22 results
510(k) Data Aggregation
(67 days)
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:
| Acceptance Criteria (Implied from Clinical Study Focus) | Reported Device Performance (Clinical Study Results) |
|---|---|
| Non-inferiority in change in HbA1c (margin 0.3%) | Mean change in HbA1c was -0.8% (P<0.001 for non-inferiority) |
| No safety concerns associated with device use | No safety concerns identified during the study |
| No unexpected or serious adverse device effects | No unexpected or serious adverse device effects reported |
| No DKA events | No DKA events reported |
| No hospitalizations or emergency visits for severe hypoglycemic events | No hospitalizations or emergency visits for severe hypoglycemic events reported |
| No instances of hyperosmolar hyperglycemic syndrome | No instances of hyperosmolar hyperglycemic syndrome reported |
2. Sample sized used for the test set and the data provenance
- Sample Size for Test Set (Clinical Study): 343 participants
- Data Provenance: The document doesn't explicitly state the country of origin. It indicates it was a "clinical study" performed by Insulet, implying it was prospective within the context of validating the device for T2DM.
3. Number of experts used to establish the ground truth for the test set and the qualifications of those experts
This information is not provided in the document. The clinical study evaluated patient outcomes (HbA1c, adverse events) rather than relying on expert consensus for ground truth on specific assessments.
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, 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 typically relevant for interpretative devices where human experts read cases and the AI provides assistance. The SmartAdjust™ technology is an automated glycemic controller that directly delivers insulin based on algorithms, not an interpretive aid for human readers.
6. If a standalone (i.e. algorithm only without human-in-the-loop performance) was done
Yes, the clinical study essentially assessed the standalone performance of the SmartAdjust™ algorithm (within the Omnipod 5 System) in a human-in-the-loop context. While users interact with the system, the key performance metrics (HbA1c, safety events) are a direct consequence of the algorithm's automated insulin delivery decisions. The document states: "When Automated Mode is turned on, the SmartAdjust™ algorithm (installed on the Pod) controls insulin delivery based on recent CGM values." The clinical study then validated this control for T2DM.
7. The type of ground truth used
The ground truth for the clinical study was based on clinical outcomes data, specifically:
- HbA1c measurements (a standard biochemical marker for long-term glucose control)
- Observation and reporting of adverse events (DKA, severe hypoglycemia, hyperosmolar hyperglycemic syndrome)
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 SmartAdjust™ algorithm. This information is typically not included in a 510(k) summary for a device like an automated insulin delivery system, as the algorithm's development and validation might involve proprietary datasets and methods outside the scope of the premarket notification for an expanded indication.
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 (if any specific to machine learning) was established. Given the nature of a glycemic control algorithm, training might involve simulated data, retrospective patient data with known glucose values and insulin needs, or data from prior clinical trials. However, the document does not specify this. The clinical study described in the summary is for validation of the algorithm's performance with the expanded indication, not for its initial training.
Ask a specific question about this device
(265 days)
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:
| Acceptance Criteria (Implied for this 510(k) update) | Reported Device Performance (as stated in the 510(k) summary) |
|---|---|
| 1. Equivalence in Intended Use and Indications for Use The updated device must have the | The subject device has the same intended use and indications for use as the predicate device. |
| 2. Equivalence in Technological Characteristics No new questions of safety/effectiveness raised by changes. | There are no changes to the design or technology of SmartAdjust™ technology itself, other than the updated iCGM performance specifications. The differences do not raise any different questions about safety and effectiveness. |
| 3. Compliance with Special Controls (21 CFR 862.1356) The device must continue to meet the specific requirements for an Interoperable Automated Glycemic Controller. | The subject device has been shown to meet the special controls for an Interoperable Automated Glycemic Controller. |
| 4. Performance with updated iCGM Specifications The device's performance (safety and effectiveness in automating insulin delivery) must remain acceptable when integrated with iCGMs meeting the updated performance specifications. | Implied by the statement that the "differences in the performance specifications for compatible iCGMs do not raise different questions of safety and effectiveness." The previous clinical data for the original predicate device (K220394) would have supported the device's efficacy with compatible iCGMs. This update likely references test data from the iCGMs themselves or in-silico/bench testing to confirm continued compatibility. |
| 5. Substantial Equivalence to Predicate The overall assessment must confirm substantial equivalence. | SmartAdjust™ technology is substantially equivalent to its predicate. |
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
(119 days)
The SmartBolus Calculator is software intended for the management of diabetes in persons aged 2 and older requiring rapid-acting U-100 insulin. The SmartBolus Calculator calculates a suggested bolus dose based on user-entered carbohydrates, most recent sensor glucose value (or blood glucose reading if using fingerstick), rate of change of the sensor glucose (if applicable), insulin on board (IOB), and programmable correction factor, insulin to carbohydrate ratio, and target glucose value. The SmartBolus Calculator is intended for single patient, home use and requires a prescription.
The SmartBolus Calculator is a software device that is a component of the Omnipod 5 Automated Insulin Delivery System. The SmartBolus Calculator exists on the Omnipod 5 App portion of the Omnipod 5 ACE Pump and relies on the user interface of the App.
The SmartBolus Calculator receives input parameters and settings from other components of the system and calculates a suggested bolus amount of insulin to correct an elevated glucose level (a correction bolus) and/or to cover carbohydrates from a meal (meal bolus). The SmartBolus Calculator allows users to have the option of populating the current estimated glucose value and trend, which is communicated by the connected iCGM. Users may also manually enter the estimated glucose value or a blood glucose (BG) reading from a blood glucose meter. In addition to glucose, the suggested bolus dose is calculated based on the following parameters: user-entered carbohydrates, rate of change of the sensors qlucose (if using a CGM), correction factor, insulin to carbohydrate ratio, target glucose value, and insulin on board (IOB). Once the calculation is complete, the user has the option of delivering the suggested dose of insulin, modifying the amount, or canceling.
The SmartBolus Calculator can be used in the Omnipod 5 Automated Insulin Delivery System with both Manual Mode and Automated Mode.
Here's a summary of the acceptance criteria and study information for the SmartBolus Calculator, based on the provided text:
1. Table of Acceptance Criteria and Reported Device Performance
The provided text does not explicitly list specific acceptance criteria in a quantitative manner for the SmartBolus Calculator's performance. Instead, it states that:
- "Verification activities, as required by the risk analysis, demonstrated that the predetermined acceptance criteria were met and the device is safe for use."
- "Through performance testing, the Subject device has been shown to meet the Special Controls and determined to be substantially equivalent to its predicate."
- "There was no impact to clinical performance of the SmartBolus Calculator for the design change discussed in this submission."
This implies that the assessment for this 510(k) submission focused on demonstrating that the new iOS version of the SmartBolus Calculator (subject device) performs identically to the predicate Android version (K222239) and meets the same safety and effectiveness standards, rather than establishing new performance metrics.
2. Sample size used for the test set and the data provenance
The document does not specify a sample size for a test set in the context of clinical or performance data for the SmartBolus Calculator itself. The testing mentioned is primarily "software verification and validation testing" and "risk management" activities.
The data provenance is also not explicitly stated as retrospective or prospective clinical data. The testing described appears to be internal software development and validation, rather than a clinical trial.
3. Number of experts used to establish the ground truth for the test set and the qualifications of those experts
This information is not provided in the document. The testing described is software-centric, and there's no mention of expert-established ground truth for a test set in a medical diagnostic sense.
4. Adjudication method for the test set
This information is not provided in the document, as it doesn't describe a clinical test set requiring adjudication.
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
An MRMC comparative effectiveness study was not mentioned or described. The device is an "Insulin Therapy Adjustment Device," not a diagnostic imaging device where MRMC studies are typically conducted. The document focuses on demonstrating substantial equivalence of a new software implementation (iOS) to an existing one (Android).
6. If a standalone (i.e. algorithm only without human-in-the-loop performance) was done
The SmartBolus Calculator is described as an "Algorithmic software device." Its function is to "calculate a suggested bolus dose." However, it operates as a component of the Omnipod 5 App, and "Once the calculation is complete, the user has the option of delivering the suggested dose of insulin, modifying the amount, or canceling." This indicates it's a human-in-the-loop system, where the user has ultimate control and decision-making power over the suggested dose. Therefore, a purely standalone clinical performance evaluation without human decision-making is not explicitly implied or discussed in this context. The software's calculation itself is standalone, but its application involves a human.
7. The type of ground truth used (expert consensus, pathology, outcomes data, etc.)
The document does not specify a type of ground truth related to clinical outcomes or expert consensus for its performance evaluation for this particular submission. The "verification activities" and "software verification and validation testing" likely used predefined software requirements, simulated data, and mathematical correctness of calculations as their "ground truth" to ensure the algorithms produced the expected outputs given specific inputs according to the established insulin calculation formulas.
8. The sample size for the training set
The document does not describe a training set in the context of machine learning. The SmartBolus Calculator is an algorithm that computes a bolus dose based on programmable factors and user inputs, not a machine learning model that requires a training set.
9. How the ground truth for the training set was established
As there's no mention of a machine learning model or a training set, this information is not applicable.
Ask a specific question about this device
(119 days)
The Omnipod 5 ACE Pump (Pod) is intended for the subcutaneous delivery of insulin, at set and variable rates, for the management of diabetes mellitus in persons requiring insulin. The Omnipod 5 ACE Pump is able to reliably and securely communicate with compatible, digitally connected devices, including automated insulin dosing software, to receive, execute, and confirm commands from these devices. The Omnipod 5 ACE Pump is intended for single patient, home use and requires a prescription.
The Omnipod 5 ACE Pump is intended to deliver insulin via a tubeless insulin pump (the Pod) that wirelessly connects to and receives insulin delivery commands from the Omnipod 5 App, which is installed on a locked-down Android controller device or a user's personal smartphone device. The predicate device allowed the user to download the Omnipod 5 App to an Android compatible phone. This submission includes an iOS compatible Omnipod 5 App to allow users to download it to a compatible iPhone.
The Omnipod 5 ACE Pump is part of the Omnipod 5 Automated Insulin Delivery System, which also includes SmartAdjust Technology (iAGC), and the SmartBolus Calculator. SmartAdjust Technology and the SmartBolus Calculator functionally independent from the Omnipod 5 ACE Pump. The Omnipod 5 ACE Pump is intended to be digitally connected to a third party iCGM, SmartAdjust Technology, and the SmartBolus Calculator.
The Omnipod 5 ACE Pump can operate in Manual Mode, delivering insulin based on userprogrammed basal rates, or in Automated Mode, where insulin is automatically delivered based on the calculations and command of a compatible iAGC. Currently, the Omnipod 5 ACE Pump is compatible with SmartAdjust Technology, the software which is pre-installed on the Pod and the App. Future alternate controllers may be established for use with the Pod, in which case the software modules of the SmartAdjust Technology would be disabled. The Pod is a bodywearable insulin pump that affixes to the user on the back of the arm, the lower back, the abdomen, the thigh area, or any site that has a layer of fatty tissue available. It is held in place by an adhesive pad and provides up to three days of insulin before it is removed and replaced with a new Pod. The Omnipod 5 App is a software application installed on a handheld touchscreen device (Android and iOS) that connects to the Pod via Bluetooth Low Energy (BLE) and serves as the user interface of the system. In addition to programmed basal delivery and automated insulin delivery, the Omnipod 5 ACE Pump allows users to deliver bolus doses at values that are either inputted manually or calculated by the SmartBolus Calculator based on the user's settings and user-entered parameters. The Pod has the ability to connect to a compatible iCGM through BLE and receive data for use with SmartAdiust Technology and the SmartBolus Calculator.
The Omnipod 5 App has the ability to wirelessly connect to the Insulet Cloud which it utilizes for registering new devices, authenticating users, ensuring hardware devices and host operating systems are compatible, and completing over the air software (OTA) and firmware (FOTA) updates.
The provided text is a 510(k) summary for the Omnipod 5 ACE Pump, which describes the addition of an iOS compatible mobile application. The document focuses on demonstrating substantial equivalence to a previously cleared predicate device (K203768).
Based on the provided text, the acceptance criteria and study details are primarily related to software verification and validation, interoperability, cybersecurity, electrical safety and EMC, and human factors validation for the new iOS application, rather than clinical performance of the insulin delivery itself (as the Pod itself was not modified).
Here's an attempt to structure the information based on your request, highlighting what is and isn't explicitly stated in the document:
Device: Omnipod 5 ACE Pump (with added iOS compatible Omnipod 5 App)
Purpose of Submission: To add a new mobile application compatible with iOS mobile devices. The Omnipod 5 App (iOS) is the new configuration being added to the previously cleared Omnipod 5 ACE Pump.
1. Table of Acceptance Criteria and Reported Device Performance
The document describes several types of performance testing and their adherence to standards and regulations. It doesn't present specific quantitative acceptance criteria or detailed performance metrics in a tabular format that is typically seen for accuracy, sensitivity, or specificity in AI/ML medical devices. Instead, it states that tests were performed and demonstrated that predetermined acceptance criteria were met and the device is safe and effective for use.
| Category | Acceptance Criteria / Standard Compliance | Reported Device Performance |
|---|---|---|
| Risk Management | Compliance with ISO 14971:2019. | "Verification activities, as required by the risk analysis, demonstrated that the predetermined acceptance criteria were met and the device is safe for use." |
| Software Validation | Compliance with IEC 62304:2015-06 and FDA’s guidance “General Principles of Software Validation – Issued January 11, 2002,” and “Content of Premarket Submissions for Device Software Functions - Issued June 2023.” | "Software verification and validation testing were performed..." "Software documentation was provided..." "Software verification testing has demonstrated the device records timestamped critical events, including information related to its state, user inputs, and device settings, as required by the ACE Pump special controls." (for Data Logging) |
| Interoperability | Adherence to FDA Guidance “Design Considerations and Pre-market Submission Recommendations for Interoperable Medical Devices - Guidance for Industry and Food and Drug Administration Staff - Issued September 6, 2017.” Specifies validated interface specifications, partnership agreements, and post-market reporting. | "Interoperability documentation was provided as is relates to changes due to the Omnipod 5 App (iOS) according to the FDA Guidance..." |
| Cybersecurity | Compliance with various FDA cybersecurity guidances (2014, 2016, March 2023, April 2022 draft). | "a cybersecurity analysis was performed for the OP5 ACE Pump with the Omnipod 5 App (iOS)... In addition, Insulet has provided a software bill of materials and penetration testing." |
| Electrical Safety & EMC | Compliance with IEC 60601-1:2020-08 and IEC 60601-1-2:2020-9, and FDA guidances on Electromagnetic Compatibility and Radio Frequency Wireless Technology. | "testing was performed to verify that the Omnipod 5 ACE Pump with the Omnipod 5 App (iOS) meets its requirement to comply with IEC 60601-1:2020-08... and IEC 60601-1-2:2020-9..." |
| Human Factors Validation | Compliance with IEC 62366:2015-06, HE75:2009(R)2018, and FDA’s guidance “Applying Human Factors and Usability Engineering to Medical Devices - Issued February 3, 2016.” | "A robust validation evaluation was performed to demonstrate safe and effective use of the Omnipod 5 App (iOS) with intended users in the expected use environments, including associated training and accompanying documentation. The results of the validation demonstrate that the device has been found to be safe and effective for the intended users, uses, and use environments." |
| Special Controls (21 CFR 880.5730) | Evaluation supports safety and effectiveness. | "evaluation of the Special Controls for this device (regulation 21 CFR 880.5730) supports the safety and effectiveness of the device." "Through performance testing, the Subject device has been shown to meet the Alternate Controller Enabled Insulin Pump special controls..." |
Note on Insulin Delivery Accuracy: The document states that the "Pod Delivery Accuracy (tested per IEC 60601-2-24)" for basal and bolus rates is a characteristic of the predicate device and is unchanged for the subject device. It is listed as:
- Basal: ± 5% at rates ≥ 0.05 U/hr
- Bolus: ± 5% for all set values ≥ 1.0 unit, ± 0.05 unit for set values < 1.0 unit
However, the current submission focuses on the new mobile app, not the Pod's fundamental insulin delivery mechanism.
2. Sample Sizes Used for the Test Set and Data Provenance
The document does not specify quantitative sample sizes for the various testing protocols (e.g., software validation, human factors validation). It broadly refers to "performance testing" being completed.
- Data Provenance: Not explicitly stated as retrospective or prospective data in the context of clinical studies for direct performance evaluation. The testing described is primarily non-clinical, involving software validation, cybersecurity analysis, and human factors validation. The device is a pump with an app, not an AI/ML algorithm that analyzes patient data to provide a diagnosis or risk assessment.
3. Number of Experts Used to Establish Ground Truth for the Test Set and Qualifications
N/A. This submission is for an Alternate Controller Enabled Pump (ACE Pump) with a new mobile application. It's not an AI/ML diagnostic or prognostic device where expert ground truth interpretation of images or other clinical data would be required. The "ground truth" for this device's validation is adherence to engineering and usability standards and ensuring the software functions as intended and securely.
4. Adjudication Method for the Test Set
N/A. Not applicable for this type of device submission. Adjudication methods like 2+1 or 3+1 typically apply to the establishment of ground truth in image-based diagnostic AI/ML models.
5. If a Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study was done
No, an MRMC comparative effectiveness study was not done. This type of study focuses on how AI assistance impacts human reader performance, typically in diagnostic tasks. The Omnipod 5 ACE Pump with its iOS app is an insulin delivery system and its controlling interface, not a diagnostic AI.
6. If a Standalone (i.e., algorithm only without human-in-the-loop performance) was done
The "algorithm only" performance for an insulin pump would refer to its ability to accurately deliver insulin based on programmed rates or commands. The document states that the Omnipod 5 ACE Pump's Pod (insulin pump itself) has not been modified, and therefore, performance testing applicable only to the Pod was not completed for this submission. The original validation of the Pod's insulin delivery accuracy (referenced in the table) would be considered its standalone performance. The new submission focuses on the software application for iOS as the controller. The software's performance is validated through various non-clinical tests (software validation, interoperability, cybersecurity, human factors) to ensure it reliably sends commands to the Pod.
7. The Type of Ground Truth Used
The "ground truth" for this submission are the established technical standards, regulatory guidelines, and functional requirements for medical device software, cybersecurity, electrical safety, and human factors. For instance:
- Software Validation: Ground truth is defined by the functional specifications and requirements of the software.
- Cybersecurity: Ground truth is defined by cybersecurity best practices and regulatory guidance.
- Human Factors Validation: Ground truth is defined by usability engineering principles and standards, demonstrated by the device being "safe and effective for the intended users, uses, and use environments" after testing.
8. The Sample Size for the Training Set
N/A. The Omnipod 5 ACE Pump and its associated app are not explicitly described as an AI/ML device that uses a "training set" in the context of supervised learning for a diagnostic or prognostic task. The software development process would involve iterative testing and debugging, but this is distinct from training a machine learning model on a labeled dataset.
9. How the Ground Truth for the Training Set Was Established
N/A, as there is no mention of an AI/ML training set in the context of this device and its submission.
Ask a specific question about this device
(171 days)
The Omnipod GO Insulin Delivery Device is intended for the subcutaneous infusion of insulin at a preset basal rate in one 24-hour time period for 3 days (72 hours) in adults with type 2 diabetes.
The Omnipod GO Pod helps manage diabetes by providing continuous subcutaneous insulin delivery. To facilitate insulin dose titration and provide appropriate options across a wide range of daily insulin needs, the Omnipod GO Device will come in seven different models: 10, 15, 20, 25, 30, 35, and 40 units per day, with each model delivering insulin at a fixed rate over the 72hour life of the device. There is no ability to deliver a bolus dose of insulin using the Omnipod GO device. Like the proposed predicate device (V-GO Insulin Delivery System, K103825), Omnipod GO will be entirely self-contained in an on-body device that is single-use, sterile, and disposable. It is small, lightweight, and designed to the body via an adhesive pad. The adhesive backing keeps the device securely in place for up to 3 days (72 hours).
The provided text describes the Omnipod GO Insulin Delivery Device, but it does not contain specific acceptance criteria or the study details to prove the device meets those criteria, especially in the format of a table with reported device performance.
Here's what can be extracted from the document regarding performance and evaluation, and where information is missing:
Missing Information:
- Specific Acceptance Criteria Table: The document doesn't provide a table explicitly listing acceptance criteria for performance metrics (e.g., specific accuracy ranges they aimed for). It mentions "predetermined acceptance criteria were met" but doesn't detail what they were.
- Reported Device Performance: While it states the device delivers insulin accurately and detects occlusions, specific numerical performance results (e.g., X% accuracy, occlusion detection time) are not provided in the summary.
- Test Set Sample Size: The document refers to "human factors validation" and "performance testing" but does not specify the sample size used for these tests.
- Data Provenance (Test Set): Not explicitly stated whether the test data was retrospective or prospective, or its country of origin.
- Number of Experts & Qualifications (for Ground Truth): Not mentioned for any ground truth establishment.
- Adjudication Method: Not mentioned.
- MRMC Comparative Effectiveness Study: Not mentioned as being performed.
- Effect Size of Human Readers Improvement with AI: Not applicable as no MRMC study with AI assistance is mentioned.
- Standalone (Algorithm Only) Performance: The device is an insulin delivery device, not an AI diagnostic/screening algorithm, so this concept (algorithm only performance) doesn't directly apply in the typical sense for medical imaging or similar devices. Its performance is inherent to its physical operation.
- Type of Ground Truth Used: For mechanical performance, "ground truth" would implicitly be highly accurate measurements against known standards. For human factors, it would be based on participant observations and feedback. However, it's not explicitly detailed.
- Training Set Sample Size: Not applicable/not mentioned, as there is no mention of a machine learning model being "trained" in the typical sense for this device.
- How Ground Truth for Training Set was Established: Not applicable/not mentioned.
What is provided in the document:
The document outlines a series of non-clinical performance data categories and general statements about compliance and effectiveness.
Summary of Non-Clinical Performance Data (What was done, but not detailed results or acceptance criteria):
- Risk Management: Completed in accordance with ISO14971:2019. "Verification activities... demonstrated that the predetermined acceptance criteria were met." (Details of criteria and results are missing).
- Safety Assurance: Device utilizes an insulin pump design marketed with other Omnipod devices. Risk mitigations from previous devices are similarly implemented to provide safety against overdose and underdose (occlusion detection, delivery accuracy).
- Human Factors Validation: Followed IEC 62366:2015-1, HE75:2009, and FDA guidance. "A robust validation evaluation was performed to demonstrate safe and effective use... Results... demonstrate that the Omnipod GO has been found to be safe and effective." (Specific data, sample size, or acceptance criteria are missing).
- Software Validation: Performed in accordance with IEC 62304:2015 and FDA guidance.
- Cybersecurity: Analysis performed following FDA guidance.
- Performance Testing: "Demonstrated that the device delivers insulin accurately at various flow rates and that it can effectively detect when an occlusion occurs and promptly notify the user." (Specific metrics, acceptance criteria, and results are missing).
- Biocompatibility: Evaluation performed per ISO 10993-1:2018 (patient-contacting materials are the same as currently marketed Omnipod Pod devices).
- Sterilization: Product adoption completed into the family of devices under sterilization validation.
- Electrical Safety and EMC Testing: Performed to verify compliance with IEC 60601-1:2005 and IEC 60601-1-2.
Key Performance Claim related to Predicate Device Comparison (Table 1.01):
| Element of Comparison | Subject Device: Omnipod GO Insulin Delivery Device | Performance Claim/Difference |
|---|---|---|
Insulin Delivery Accuracy | +/- 5% | Improved from predicate's +/- 10% |
This is the only specific quantitative performance metric with a value mentioned in the entire document. The document implies this "improved" accuracy value meets the new device's acceptance criterion for insulin delivery accuracy.
Ask a specific question about this device
(189 days)
Ask a specific question about this device
(24 days)
The SmartBolus Calculator is software intended for the management of diabetes in persons aged 2 and older requiring rapid-acting U-100 insulin. The SmartBolus Calculator calculates a suggested bolus dose based on user-entered carbohydrates, most recent sensor glucose value (or blood glucose reading if using fingerstick), rate of change of the sensor glucose (if applicable), insulin on board (IOB), and programmable correction factor, insulin to carbohydrate ratio, and target glucose value. The SmartBolus Calculator is intended for single patient, home use and requires a prescription.
The SmartBolus Calculator is a software device that is a component of the Omnipod 5 Automated Insulin Delivery System. The SmartBolus Calculator exists on the Omnipod 5 App portion of the Omnipod 5 ACE Pump and relies on the user interface of the App. The SmartBolus Calculator receives input parameters and settings from the other components of the system and calculates a suggested bolus amount of insulin to correct an elevated glucose level (a correction bolus) and/or to cover carbohydrates from a meal (meal bolus). The SmartBolus Calculator allows users to have the option of populating the current estimated glucose value and trend, which is communicated by the connected iCGM. Users may also manually enter the estimated glucose value or a blood glucose (BG) reading from a blood glucose meter. In addition to glucose, the suggested bolus dose is calculated based on the following parameters: user-entered carbohydrates, rate of change of the sensor glucose (if using a CGM), correction factor, insulin to carbohydrate ratio, target qlucose value, and insulin on board (IOB). Once the calculation is complete, the user has the option of delivering the suggested dose of insulin, modifying the amount, or canceling. The SmartBolus Calculator can be used in the Omnipod 5 Automated Insulin Delivery System with both Manual Mode and Automated Mode.
The information provided focuses on the substantial equivalence of the SmartBolus Calculator to a predicate device, specifically regarding the expansion of its age range of intended users. The document details the clinical study performed to support this expanded indication, rather than a general acceptance criteria study for the device's core functionality.
Here's an analysis of the provided information based on your requested points:
1. Table of Acceptance Criteria and Reported Device Performance
The document does not explicitly state formal acceptance criteria in a tabular format for the SmartBolus Calculator's overall performance. Instead, the "acceptance criteria" can be inferred from the clinical study's primary objective, which was to evaluate safety based on glucose metrics. The study aimed to show that the use of the SmartBolus Calculator in the expanded age group (2.0-5.9 years) did not significantly worsen these safety metrics compared to manual entry of blood glucose values.
| Acceptance Criteria (Inferred from Study Objective) | Reported Device Performance |
|---|---|
| Safety Metrics (4-hour post bolus period): | |
| Mean % time Blood Glucose < 70 mg/dL (Phase 2 vs Phase 1) | Mean decrease of 1.13% in Phase 2 (4.03%) vs. Phase 1 (5.16%). (P = 0.6250, not statistically significant) |
| Mean % time Blood Glucose > 180 mg/dL (Phase 2 vs Phase 1) | Mean decrease of 2.03% in Phase 2 (33.2%) vs. Phase 1 (35.2%). (P = 1.0000, not statistically significant) |
| Number of deaths | 0 |
| Number of serious adverse events | 0 |
| Number of unanticipated adverse device effects (UADE) | 0 |
| Number of non-serious adverse events | 1 (prolonged hyperglycemia) |
2. Sample Size Used for the Test Set and Data Provenance
- Sample Size: 5 subjects in the preschool cohort (aged 2.0-5.9 years).
- Data Provenance: The study was a "single-arm, multi-center, prospective clinical study" conducted across "2 US clinical sites." Therefore, the data is prospective and from the USA.
3. Number of Experts Used to Establish the Ground Truth for the Test Set and Qualifications of Those Experts
This document does not provide details on the number or qualifications of experts used to establish ground truth for the test set. The clinical study collected glucose metrics directly from the subjects, which served as the primary data for evaluation.
4. Adjudication Method for the Test Set
The document does not mention an explicit adjudication method for the clinical study's data. Clinical study data, particularly objective metrics like blood glucose levels, typically do not require adjudication in the same way imaging or subjective diagnostic interpretations might. Safety events would be reviewed by the study investigators and reported.
5. If a Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study was done
No, an MRMC comparative effectiveness study was not done. This study was a clinical trial evaluating the safety and effectiveness of the device in a specific patient population, not a comparative study of human readers with and without AI assistance.
6. If a Standalone (i.e., algorithm only without human-in-the-loop performance) was done
This study was a human-in-the-loop study. The SmartBolus Calculator is a software device that recommends a bolus dose, but the user "has the option of delivering the suggested dose of insulin, modifying the amount, or canceling." The clinical study evaluated the device's performance when used by patients/caregivers in Manual Mode.
7. The Type of Ground Truth Used
The ground truth for the clinical study was based on objective clinical outcomes data (glucose metrics and adverse events) directly measured from the study participants during the two phases of the study (Manual Mode with manual BG entry vs. Manual Mode with SmartBolus Calculator use).
8. The Sample Size for the Training Set
The document does not provide information on the sample size for the training set used to develop the SmartBolus Calculator. The submission focuses on the clinical validation of the device's expanded indication, not its initial development.
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
(400 days)
The Omnipod 5 ACE Pump (Pod) is intended for the subcutaneous delivery of insulin, at set and variable rates, for the management of diabetes mellitus in persons requiring insulin. The Omnipod 5 ACE Pump is able to reliably and securely communicate with compatible, digitally connected devices, including automated insulin dosing software, to receive, execute, and confirm commands from these devices. The Omnipod 5 ACE Pump is intended for single patient, home use and requires a prescription.
The Omnipod 5 alternate controller enabled (ACE) Pump is intended to deliver insulin via a tubeless insulin pump (the Pod) that wirelessly connects to and receives insulin delivery commands from the Omnipod 5 Application (App), which is installed on a locked-down controller device or a user's personal compatible smartphone device.
The Omnipod 5 ACE Pump is part of the Omnipod 5 Automated Insulin Delivery System, which also includes the Omnipod 5 Interoperable Automated Glycemic Controller (iAGC), Omnipod 5 Bolus Calculator, and the third-party Dexcom G6 iCGM. Omnipod 5 iAGC and Bolus Calculator functions are functionally independent from the Omnipod 5 ACE Pump. The Omnipod 5 ACE Pump is intended to be digitally connected to the iCGM, the iAGC, and the Bolus Calculator.
The Omnipod 5 ACE Pump can operate in Manual Mode, delivering insulin based on userprogrammed basal rates, or in Automated Mode, where insulin is automatically delivered based on the calculations and command of a compatible iAGC. Currently, the Omnipod 5 ACE Pump is compatible with the Omnipod 5 iAGC, whose software is pre-installed on the Pod and the App. Future alternate controllers (iAGCs) may be established for use with the Pod, in which case the software modules of the Omnipod 5 iAGC would be disabled.
The Pod is a body-wearable insulin pump that affixes to the user on the back of the arm, the lower back, the abdomen, the thigh area, or any site that has a layer of fatty tissue available. It is held in place by an adhesive pad and provides up to three days of insulin before it is removed and replaced with a new Pod. The Omnipod 5 App is an Android software application installed on a handheld touchscreen device that connects to the Pod via Bluetooth Low Energy (BLE) and serves as the user interface of the system.
In addition to programmed basal delivery and automated insulin delivery, the Omnipod 5 ACE Pump allows users to deliver bolus doses at values that are either inputted manually or calculated by the Omnipod 5 Bolus Calculator based on the user's settings and userentered parameters. The Pod has the ability to connect to a compatible iCGM through BLE and receive data for use with the Omnipod 5 iAGC and Omnipod 5 Bolus Calculator.
The Omnipod 5 App has the ability to wirelessly connect to the Insulet Cloud which it utilizes for registering new devices, authenticating users, ensuring hardware devices and host operating systems are compatible, and completing over the air software (OTA) and firmware (FOTA) updates. In addition, data from the App uploads regularly to the Insulet Cloud for data management purposes.
The Omnipod 5 ACE Pump is intended for the subcutaneous delivery of insulin and is stated to be substantially equivalent to the Omnipod DASH Insulin Management System with Interoperable Technology.
Acceptance Criteria and Reported Device Performance
The provided text primarily focuses on regulatory compliance and substantial equivalence to a predicate device rather than precise acceptance criteria and their corresponding empirical results in a clear tabular format. However, the document does mention performance aspects, particularly regarding Delivery Accuracy and Occlusion Detection.
Based on the available information, the following can be inferred:
| Acceptance Criteria | Reported Device Performance | Study Type/Context |
|---|---|---|
| Delivery Accuracy | Basal: ± 5% at rates ≥ 0.05 U/hr | Performance Testing (tested per IEC 60601-2-24) |
| Bolus: ± 5% for all set values ≥ 1.0 unit, ± 0.05 unit for set values < 1.0 unit | Performance Testing (tested per IEC 60601-2-24) | |
| Occlusion Detection | Detects occlusion at 5.0 units. | Performance Testing |
The document also mentions compliance with various standards, which implicitly sets acceptance criteria for aspects like software validation, risk management, human factors, and cybersecurity. However, specific quantitative acceptance values for these broader categories are not detailed in the provided text.
Study Information
-
Sample size used for the test set and the data provenance:
The document does not specify the sample sizes (number of devices, test conditions, etc.) used for the performance tests (e.g., delivery accuracy, occlusion detection). It also does not provide details on the data provenance (e.g., country of origin, retrospective or prospective). The testing appears to be primarily lab-based performance verification rather than clinical data from human subjects. -
Number of experts used to establish the ground truth for the test set and the qualifications of those experts:
This information is not provided. The performance data presented (e.g., delivery accuracy, occlusion detection) would typically be derived from objective measurements against known standards rather than expert-established ground truth in a diagnostic context. -
Adjudication method for the test set:
Not applicable, as the performance tests are quantitative measurements against defined specifications. -
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:
Not applicable. The Omnipod 5 ACE Pump is an insulin infusion pump, not an AI-assisted diagnostic imaging device that would typically involve a multi-reader multi-case study. The focus is on the device's functional performance and safety. -
If a standalone (i.e., algorithm only without human-in-the-loop performance) was done:
The device itself (the Pod) performs autonomously in its defined functions (e.g., insulin delivery, occlusion detection) based on its pre-programmed software and commands. Therefore, the "Performance Testing" data can be considered standalone algorithm performance for specific functionalities. However, the system is designed to be used with a human-in-the-loop (the user managing their diabetes) and interfaces with an automated insulin dosing software (iAGC). The document indicates "software verification and validation testing" was performed, which would cover the device's algorithmic performance. -
The type of ground truth used (expert consensus, pathology, outcomes data, etc.):
ForDelivery AccuracyandOcclusion Detection, the ground truth would be based on objective physical measurements against established engineering and medical device standards (e.g., volumetric measurements for insulin delivery, controlled pressure or flow scenarios for occlusion detection). This is not expert consensus, pathology, or outcomes data. -
The sample size for the training set:
Not applicable. The document describes a medical device, an insulin pump, which is not an AI/ML model that learns from a training set in the typical sense. Its software is developed and validated through traditional software engineering processes, not machine learning training. -
How the ground truth for the training set was established:
Not applicable, as there is no mention of a machine learning training set.
Ask a specific question about this device
(400 days)
The Omnipod 5 SmartBolus Calculator is software intended for the management of diabetes in persons aged 6 and older requiring rapid-acting U-100 insulin. The Omnipod 5 SmartBolus Calculator calculates a suggested bolus dose based on user-entered carbohydrates, most recent sensor glucose reading (or blood glucose reading if using fingerstick), rate of change of the sensor glucose (if applicable), insulin on board (IOB), and programmable correction factor, insulin to carbohydrate ratio, and target glucose value. The Omnipod 5 SmartBolus Calculator is intended for single patient, home use and requires a prescription.
The Omnipod 5 SmartBolus Calculator is a software device that is a component of the Omnipod 5 Automated Insulin Delivery System. The SmartBolus Calculator exists on the Omnipod 5 App portion of the Omnipod 5 ACE Pump and relies on the user interface of the App.
The Omnipod 5 SmartBolus Calculator receives input parameters and settings from the other components of the system and calculates a suggested bolus amount of insulin to correct an elevated glucose level (a correction bolus) and/or to cover carbohydrates from a meal (meal bolus). The Omnipod 5 SmartBolus Calculator allows users to have the option of populating the current estimated glucose value and trend, which is communicated by the connected iCGM. Users may also manually enter the estimated glucose value or a blood glucose (BG) reading from a blood glucose meter. In addition to glucose, the suggested bolus dose is calculated based on the following parameters: user-entered carbohydrates, rate of change of the sensor glucose (if using a CGM), correction factor, insulin to carbohydrate ratio, target glucose value, and insulin on board (IOB). Once the calculation is complete, the user has the option of delivering the suggested dose of insulin, modifying the amount, or canceling.
The Omnipod 5 SmartBolus Calculator can be used in the Omnipod 5 Automated Insulin Delivery System with both Manual Mode and Automated Mode. When the Omnipod 5 SmartBolus Calculator is used with manually-entered BG readings, it suggests a bolus dose based on the same calculations as the currently cleared Omnipod DASH Insulin Management System (K180045, most recently cleared in K192659).
The document describes the Omnipod 5 SmartBolus Calculator, a software device for diabetes management. Here's a breakdown of the acceptance criteria and the study proving it meets these criteria:
1. Table of Acceptance Criteria and Reported Device Performance
The acceptance criteria are implied through the statement that the device meets "Insulin Therapy Adjustment Device special controls and to be safe and effective." The clinical study focused on glycemic control within 4 hours of bolusing. While explicit acceptance criteria for these percentages are not given as strict pass/fail thresholds, the performance observed is presented as evidence of safety and effectiveness, particularly regarding hypoglycemia.
| Glycemic Measure (as measured by CGM) | Acceptance Criteria (Implied: Safety & Effectiveness) | Reported Performance (CGM-Informed SmartBolus Calculator) |
|---|---|---|
| Time in range (70-180 mg/dL) | Expected to maintain time in range within acceptable levels given bolus calculations | 63.8% (SD 15.7) |
| Time spent < 70 mg/dL (Hypoglycemia) | Expected to minimize hypoglycemia, especially post-bolus | 2.1% (SD 2.0) Lower than Standard SmartBolus Calculator |
| Time spent < 54 mg/dL (Severe Hypoglycemia) | Expected to minimize severe hypoglycemia | 0.3% (SD 0.7) |
| Time spent > 180 mg/dL (Hyperglycemia) | Expected to manage hyperglycemia | 34.0% (SD 16.0) |
| Time spent ≥ 250 mg/dL | Expected to manage high hyperglycemia | 9.7% (SD 10.3) |
| Time spent ≥ 300 mg/dL | Expected to manage very high hyperglycemia | 2.6% (SD 3.7) |
Note: The primary analysis focused on comparing the CGM-informed SmartBolus Calculator with the standard one, and a statistically significant difference was observed for time spent < 70 mg/dL, indicating improvement in minimizing hypoglycemia.
2. Sample Size Used for the Test Set and Data Provenance
- Sample Size: 25 participants
- Data Provenance: The document does not explicitly state the country of origin. The study appears to be a prospective clinical study, as participants actively used the device in two distinct phases of device usage.
3. Number of Experts Used to Establish the Ground Truth for the Test Set and Qualifications of Those Experts
This information is not provided in the document. The "ground truth" for the clinical study was measured directly by CGM data. There is no mention of experts establishing ground truth for the bolus calculations themselves or for the CGM data.
4. Adjudication Method for the Test Set
This information is not provided in the document. Given that the clinical study involved direct physiological measurements (CGM data), it is unlikely that an adjudication method for "ground truth" labels was necessary in the traditional sense.
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
An MRMC study was not conducted, and therefore, there is no information on the effect size of human readers improving with AI assistance. This device is an automated bolus calculator, not an AI assisting human interpretation of images or data in an MRMC setting.
6. If a Standalone (i.e., algorithm only without human-in-the-loop performance) was done
The presented clinical study is a comparison of two versions of the SmartBolus Calculator (standard vs. CGM-informed) in a human-in-the-loop setting, as users still had the option to deliver, modify, or cancel the suggested dose. While the algorithm calculates the dose standalone, the clinical study evaluates its performance within the user's decision-making process. The document also mentions "Software Validation" as a non-clinical performance test, which would likely include standalone algorithm validation, but specific details of such a study are not provided in terms of performance metrics or acceptance criteria.
7. The Type of Ground Truth Used
The "ground truth" for the clinical performance data was continuous glucose monitoring (CGM) data. This is objective physiological measurement data collected from the participants.
8. The Sample Size for the Training Set
This information is not provided in the document. The document describes a clinical study as "Summary of Clinical Performance Data" which would be considered the test set for validating the device's clinical impact. There is no mention of a separate training set for the algorithm "training" itself within the scope of this regulatory submission summary. The bolus calculator is based on established algorithms for insulin dose calculation.
9. How the Ground Truth for the Training Set was Established
This information is not provided in the document. As mentioned in point 8, the document describes a clinical validation study rather than a de novo algorithm training process. The bolus calculation algorithm is based on programmable settings (correction factor, insulin to carbohydrate ratio, target glucose value) and real-time data (glucose readings, carbs, IOB), which are clinical parameters established through diabetes management principles rather than a "ground truth" derived from a specific training dataset in the AI/ML sense.
Ask a specific question about this device
(400 days)
SmartAdjust technology is intended for use with compatible integrated continuous glucose monitors (iCGM) and alternate controller enabled (ACE) pumps to automatically increase, 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 6 years of age and older. SmartAdjust technology is intended for single patient use and requires a prescription.
The Omnipod 5 iAGC (SmartAdjust technology) is a software-only medical device intended for the management of type 1 diabetes mellitus. The Omnipod 5 iAGC uses data from a connected iCGM along with user-defined parameters to predict future glucose trends and automatically increase, decrease, or suspend the delivery of insulin via a compatible alternate controller enabled (ACE) pump.
The Omnipod 5 iAGC software is part of the Omnipod 5 Automated Insulin Delivery System, which also includes the Omnipod 5 ACE Pump, Omnipod 5 Bolus Calculator, and an interoperable iCGM. The Omnipod 5 iAGC is intended to be digitally connected to an iCGM and an ACE Pump.
The Omnipod 5 iAGC software resides on the Omnipod 5 ACE Pump (the Omnipod 5 Pod and Omnipod 5 App). The iAGC software is responsible for controlling insulin delivery via compatible ACE Pump when the system is in Automated Mode. iCGM data is transmitted from the iCGM to the ACE Pump via Bluetooth Low Energy technology. The Omnipod 5 iAGC uses this transmitted iCGM data in its calculations. The Omnipod 5 iAGC can be turned off, and the Omnipod 5 ACE Pump will operate in Manual Mode, which delivers insulin based on HCP- or user-defined Basal Programs.
The Omnipod 5 iAGC has three states of operation: Automated Mode, Automated: Limited, and Activity. In Automated Mode, the system calculates insulin delivery every 5 minutes based on the user-customizable glucose target (110–150 mg/dL). Automated: Limited is enabled when the Omnipod 5 iAGC is not receiving data from a connected iCGM for 20 minutes or more and during sensor warm-up. While in Automated: Limited, the user will receive no more than preprogrammed basal insulin. When a valid glucose value is received from the iCGM, the Omnipod 5 iAGC will resume delivery of insulin in full Automated Mode. Activity is a user-selected temporary feature intended for use during periods when insulin sensitivity is expected to be higher, such as during exercise. With Activity, the algorithm reduces insulin delivery and sets a temporary glucose target of 150 mg/dL.
Acceptance Criteria and Device Performance for SmartAdjust Technology
The SmartAdjust technology, an interoperable Automated Glycemic Controller (iAGC) for type 1 diabetes management, was evaluated through a clinical study. The acceptance criteria and reported device performance are detailed below.
1. Table of Acceptance Criteria and Reported Device Performance
The provided document does not explicitly list pre-defined quantitative acceptance criteria for all aspects. Instead, the study aims to demonstrate improvements in glycemic control and safety compared to standard therapy. The "Change" columns in the glycemic results tables effectively represent the performance metrics, and a "P < 0.0001" indicates statistical significance for these changes, which serves as an implicit acceptance criterion for improvement.
Here's a summary of the key performance metrics from the clinical study, presented as a comparison to the standard therapy phase:
Glycemic Results Overall (24 hours)
| Characteristic | Acceptance Criteria (Improvement vs. Standard Therapy with P < 0.0001) | Reported Device Performance (Change with Omnipod 5) |
|---|---|---|
| Children (6 to 13.9 years) | ||
| Avg A1C, % | Significant decrease | -0.71%* |
| Avg sensor glucose, mg/dL | Significant decrease | -23* |
| Avg % time 70-180mg/dL | Significant increase | 15.6%* |
| Median % <54mg/dL | No significant increase or decrease (ideally stable or decreasing) | 0.04% |
| Median % <70mg/dL | No significant increase or decrease (ideally stable or decreasing) | 0.06% |
| Avg % >180mg/dL | Significant decrease | -15.1%* |
| Avg % ≥250mg/dL | Significant decrease | -9.4%* |
| Avg % ≥300mg/dL | Significant decrease | -5.1%* |
| Adolescents & Adults (14 to 70 years) | ||
| Avg A1C, % | Significant decrease | -0.38%* |
| Avg sensor glucose, mg/dL | Significant decrease | -8* |
| Avg % time 70-180mg/dL | Significant increase | 9.3%* |
| Median % <54mg/dL | Significant decrease (ideally) | -0.08%* |
| Median % <70mg/dL | Significant decrease (ideally) | -0.89%* |
| Avg % >180mg/dL | Significant decrease | -7.7%* |
| Avg % ≥250mg/dL | Significant decrease | -4.3%* |
| Avg % ≥300mg/dL | Significant decrease | -2.0%* |
Safety Results (Incidence of Adverse Events)
| Adverse Event Type | Acceptance Criteria (Low incidence, comparable to standard, or manageable) | Reported Device Performance (Events per person-month / Count) |
|---|---|---|
| Severe hypoglycemia | Low incidence, ideally no association with device malfunction | 0.004 events per person-month (3 events total, not attributable to system) |
| Diabetic ketoacidosis (DKA) | Low incidence, ideally no association with device malfunction | 0.001 events per person-month (1 event total, from suspected infusion site failure) |
| Other related adverse events | Acceptable and manageable rate | Hypoglycemia (non-severe): 1 (0.4%) / Hyperglycemia: 3 (1.3%) / Prolonged Hyperglycemia: 18 (6.7%) / Other: 16 (6.7%) |
*Change between standard therapy phase and Omnipod 5 System phase was significant (P < 0.0001).
2. Sample Size Used for the Test Set and Data Provenance
- Sample Size for Test Set: 240 subjects. This included 112 children (6 to 13.9 years) and 128 adolescents and adults (14 to 70 years).
- Data Provenance: The data was collected from a multicenter, prospective clinical study conducted across 17 investigational sites. The document does not explicitly state the country of origin, but given the FDA submission, it is highly likely the study was conducted in the United States or in accordance with international good clinical practice standards acceptable to the FDA.
3. Number of Experts Used to Establish the Ground Truth for the Test Set and Their Qualifications
The study involved a prospective clinical trial in human subjects with Type 1 Diabetes. In this context, the "ground truth" for glycemic control and safety are direct measures from the subjects (A1C, sensor glucose readings, adverse event reporting) rather than expert consensus on images or interpretations. Therefore, there were no "experts" establishing a ground truth for a test set in the traditional sense of medical image analysis. Clinical endpoints were measured directly from the study participants.
4. Adjudication Method for the Test Set
The document does not describe an explicit adjudication method for a "test set" in the context of expert review, as the ground truth was derived from direct clinical measurements. However, adverse events and clinical outcomes would have been reviewed and documented by clinical investigators and potentially an independent medical monitor or data safety monitoring board, which is standard for clinical trials.
5. Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study
No, an MRMC comparative effectiveness study was not done. This type of study is typically relevant for interpretative medical devices (e.g., imaging AI) where human readers interpret cases with and without AI assistance. The SmartAdjust technology is a therapeutic device that directly controls insulin delivery, and its effectiveness is measured by clinical outcomes (e.g., A1C, time in range) in patients.
6. Standalone Performance Study
Yes, a standalone study was conducted. The clinical study described is a standalone performance study of the SmartAdjust technology (Omnipod 5 iAGC) in automated mode, meaning the algorithm was operating without human intervention for insulin delivery adjustments. The study compared the device's performance against a "standard therapy phase" which served as a baseline. The algorithm made automatic adjustments based on iCGM data, and the outcomes were observed directly in patients.
7. Type of Ground Truth Used
The ground truth used was outcomes data from human subjects. Specifically:
- Glycemic Control: Measured by laboratory-assessed A1C, and continuous glucose monitor (CGM) data (e.g., average sensor glucose, percentage of time in various glucose ranges: 70-180 mg/dL, <54 mg/dL, <70 mg/dL, >180 mg/dL, ≥250 mg/dL, ≥300 mg/dL).
- Safety: Measured by the incidence of severe hypoglycemia and diabetic ketoacidosis (DKA), as well as other adverse events reported during the clinical trial.
8. Sample Size for the Training Set
The document does not provide information on the sample size for the training set. This is typical for device submissions where the algorithm's development (training) might involve proprietary data and methods not usually detailed in the 510(k) summary. The clinical study focuses on the validation and performance of the fully developed device.
9. How the Ground Truth for the Training Set Was Established
The document does not specify how the ground truth for any potential training set was established. Given the nature of an automated glycemic controller, it's plausible that the training could involve:
- Simulated physiological models of diabetes.
- Retrospective data from patients with T1D (e.g., CGM data, insulin delivery logs, meal information).
- Rule-based logic derived from clinical guidelines and expert knowledge, rather than a "trained" machine learning model in the conventional sense.
Without further information in the submission, the exact method for establishing ground truth for training cannot be determined.
Ask a specific question about this device
Page 1 of 3