Search Results

The t:slim X2 Insulin Pump with Interoperable Technology (the pump) is intended for the subcutaneous delivery of insulin, at set and variable rates, for the management of diabetes melliting insulin. The 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 pump is intended for single patient, home use and requires a prescription. The Pump is indicated for use in individuals 2 years of age and greater.

The t:slim X2 Insulin Pump with Interoperable Technology is an ambulatory, battery operated, rate-programmable infusion pump designed for the subcutaneous delivery of insulin, at set and variable rates, for the management of diabetes mellitus in persons requiring insulin.

The t:slim X2 insulin pump includes a disposable cartridge with a 300 unit reservoir which is filled with insulin by the user with the use of the syringe and needle. The cartridge needs to be changed every 48-72 hours depending on the type of insulin used.

The pump is motor driven to deliver patient programmed basal rates and boluses through an infusion set into subcutaneous tissue. The desired timing and quantity of insulin delivery (bolus or basal) is programmed by the patient based on their healthcare provider's recommendations.

The pump is able to reliably and securely communicate with compatible, digitally connected devices, including automated insulin dosing software, execute, and confirm commands from these devices.

The t:connect mobile app enables a user to connect a smartphone to the pump using Bluetooth® wireless technology to display pump information and perform some pump functions on the smartphone as well as display pump notifications. The t:connect mobile app can transmit pump and therapy data from the pump to the cloud as long as the user's smartphone is connected to the internet.

The pump is designed to be able to receive and display alerts and alarms to users based on information received from other interoperable devices.

The pump is compatible with Interoperable Automated Glycemic Controllers, such as Basal-IQ Technology (K193483) and Control IQ Technology (K200467) to aid in diabetes management. The latter is being submitted concurrently.

In addition, the Subject Device is compatible with iCGM systems cleared under K223931 21 CFR 862.1355.

The provided text describes the 510(k) summary for the Tandem Diabetes Care t:slim X2 Insulin Pump with Interoperable Technology (K232380). This submission focuses on a change to the age indication for the device, lowering it to individuals 2 years of age and greater.

Here's an analysis of the acceptance criteria and study information based on the provided text:

1. Table of Acceptance Criteria and Reported Device Performance

The document does not explicitly present a table of specific quantitative acceptance criteria (e.g., minimum accuracy percentages, specific error rates) for the device's performance in relation to the expanded age indication. Instead, it makes a general statement about safety and effectiveness.

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

The document mentions "clinical testing was performed to support the expanded age indication" and "the clinical testing provided in this 510(k) demonstrate that the t:slim X2 Insulin Pump with Interoperable Technology is the same as the predicate device in safety and effectiveness."

• Sample Size for Test Set: Not explicitly stated.
• Data Provenance: Not explicitly stated as retrospective or prospective, nor the country of origin. However, the nature of "clinical testing" for an expanded age indication typically implies a prospective clinical study involving the new patient population.

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 text does not describe a process of expert review or establishing ground truth in the context of the clinical testing.

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

This information is not provided in the document. The study described is a clinical evaluation of the device itself rather than a comparison of human readers with and without AI assistance.

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

The device is an insulin pump, which is a physical medical device that delivers insulin. While it has interoperable technology and can communicate with automated insulin dosing software, the "performance data" mentioned pertains to the overall device in a clinical setting with human interaction (patient programming, healthcare provider recommendations). Therefore, a "standalone algorithm only" performance study as typically understood for AI/CAD devices is not directly applicable or explicitly described in this context. The focus is on the device's safety and effectiveness for insulin delivery in the expanded age group.

7. The Type of Ground Truth Used

The "ground truth" for the clinical testing would be the actual physiological responses and outcomes of the pediatric patients using the insulin pump, as assessed by healthcare professionals. This would include parameters like blood glucose levels, incidence of hypoglycemia/hyperglycemia, and adverse events. The document refers to "safety and effectiveness," which points to clinical outcomes.

8. The Sample Size for the Training Set

The document describes a clinical study to support an expanded age indication for an existing device. This implies that the device's core algorithm and functionality were developed and "trained" (in a non-AI sense, meaning designed and validated) prior to this specific submission. The text does not provide information regarding a training set for an AI/algorithm, as this submission is not about the initial development or a new AI component. If the "interoperable technology" involves AI, the training data for that specific AI component is not detailed here.

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

Given that this submission is for an expanded age indication of an existing device, and not a new AI model, the concept of a "training set" and its ground truth establishment, as typically understood for AI/ML, is not applicable or described in this document. The device's fundamental insulin delivery mechanism and safety profile would have been established through prior design, testing, and regulatory submissions. The current clinical testing validates the device's safety and effectiveness for a younger population.

Ask a specific question about this device

The t:slim X2 insulin pump with interoperable technology (the Pump) is intended for the subcutaneous delivery of insulin. at set and variable rates, for the management of diabetes mellitus in persons requiring insulin. The 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 Pump is intended for single patient, home use and requires a prescription. The Pump is indicated for use with NovoLog or Humalog U-100 insulin. The Pump is indicated for use in individuals 6 years of age and greater.

The t:connect mobile app can be wirelessly paired to the t:slim X2 Insulin Pump with Interoperable Technology, for the purposes of allowing limited control of bolus insulin therapy. The t:connect mobile app is available on iOS and Android compatible smartphones via the Apple Store® and Google Play Store®. When successfully paired with The Pump, users have the ability to perform the following via the t:connect mobile app:

• View The Pump therapy data, trends, alerts, alarms, notifications, and reminders as a secondary display.
• Program Correction Boluses, Bolus Override, and Food (Standard) Boluses.
• Terminate (Cancel or stop) all bolus types regardless of origin of bolus request being made on the t:slim X2 Insulin Pump or the t:connect mobile app.
  Outside of programming and terminating boluses, the t:connect mobile app will have no other controlling action on The Pump.

The provided document describes the FDA 510(k) clearance for the t:slim X2 Insulin Pump with Interoperable Technology (with t:connect mobile app). It focuses on the addition of the mobile app to allow limited control of bolus insulin therapy.

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

No specific quantitative acceptance criteria or detailed study results (like sensitivity, specificity, or specific error rates) are explicitly reported in this 510(k) summary for the t:connect mobile app. The submission primarily relies on non-clinical testing (human factors, software verification and validation, cybersecurity) and demonstrates substantial equivalence to a predicate device.

The clearance is for a medical device (insulin pump with a mobile app for control), not an AI/ML algorithm that predicts or diagnoses based on data. Therefore, many of the typical AI/ML study components (like expert ground truth establishment, MRMC studies, standalone algorithm performance, or specific performance metrics like AUC, sensitivity, specificity, etc.) are not applicable or not explicitly detailed in this type of regulatory submission for this device.

The study aims to demonstrate that the mobile app addition is safe and effective and does not raise new questions of safety or effectiveness compared to the predicate device.

Acceptance Criteria and Reported Device Performance

Given the nature of this device (insulin pump with a mobile app for control), the "acceptance criteria" are more about functionality, safety, and human usability rather than diagnostic accuracy metrics.

• Viewing Pump therapy data, trends, alerts, alarms, notifications, and reminders as a secondary display.
• Programming Correction Boluses, Bolus Override, and Food (Standard) Boluses.
• Terminating (Cancelling or stopping) all bolus types regardless of origin of bolus request.
  Pump functions independently from the mobile app.
  Mobile app does not control or impact automated dosing algorithms (Basal-IQ, Control-IQ).
  Bolus Calculator: Mobile app includes a built-in Bolus Calculator with the same purpose and function as the Pump's. Pump sends info to app for bolus calculation.
  Display of Alerts/Notifications: Mobile app displays Pump Notifications, Alerts, Alarms, and Reminders. Specific alerts for "Pump Connection Lost" and "t:connect Mobile App Incomplete Bolus Alert." Specific notification for "Bolus in Progress on Pump." (Mobile app cannot clear/alter; requires Pump for that.)
  Logging Critical Events: Pump logs bolus requests and terminations from the app. Mobile app logs actions of bolus initiation/termination. | The device (Pump with t:connect mobile app) met specified requirements and performed as intended.
• The t:connect mobile app enables the listed functionalities for viewing data and controlling bolus insulin.
• The mobile app and pump retain independent functionality where appropriate.
• The bolus calculator functions as described.
• Alerts, alarms, and notifications are displayed on the app as intended.
• Critical events, including actions from the mobile app, are logged by the system. |
  | Safety | Human Factors: Demonstrate intended users can effectively use the Subject Device for its intended purpose in expected use environments, ensuring safe use.
  Software Verification & Validation (V&V): Ensure software conforms to patient needs and intended uses, implying absence of critical errors, accurate execution of commands, and data integrity.
  Cybersecurity: Evaluate and ensure the device's resilience against cyber threats to prevent unauthorized access or manipulation.
  Risk Management: Compliance with ISO 14971 implies a robust risk management process. | Human Factors: Results from the human factors study demonstrate users can safely and effectively use the features of the Subject Device.
  Software V&V: Carried out activities (requirements linking, code inspection, static analysis, unit/system testing) to ensure specified conformity.
  Cybersecurity: Evaluations were carried out.
  Risk Management: Compliance with relevant standards (ISO 14971, IEC 62304) for risk management and software lifecycle processes implicitly ensures safety.
  Overall: The device does not raise any new or different questions of safety or effectiveness. |
  | Effectiveness | The device, with the mobile app, should continue to effectively manage diabetes mellitus in persons requiring insulin. The mobile app's role is to facilitate limited control of bolus insulin therapy. | The device functions as an alternate controller enabled infusion pump, providing the intended functionality for insulin delivery and bolus control via the app. |

Study Details

1. Sample sizes used for the test set and data provenance:
   This document focuses on non-clinical testing (human factors, software V&V, cybersecurity) and does not specify a "test set" in the context of a clinical trial or algorithm performance evaluation on a patient dataset.

   • Human Factors Validation Testing: While the document states "Human factors validation testing was conducted," it does not specify the sample size (number of participants) or data provenance (e.g., country of origin, retrospective/prospective).
   • Software Verification and Validation: This testing involves internal software evaluations, code reviews, unit testing, and system-level testing. There isn't typically a "sample size" in the same way as a clinical dataset for these activities. The provenance relates to the software development process itself.
   • No new clinical testing was required. This indicates no patient data was used for a new clinical study to support this specific 510(k) for the mobile app.

2. Number of experts used to establish the ground truth for the test set and the qualifications of those experts:

   • Not Applicable/Not Specified in this context. For human factors, "ground truth" would be related to user performance outcomes and safety, assessed by usability engineers or similar experts rather than medical specialists establishing diagnostic ground truth from images or clinical data. The document does not specify the number or qualifications of experts for these assessments.

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

   • Not Applicable/Not Specified. This type of adjudication is typically for establishing medical "ground truth" in diagnostic studies, which was not performed here.

4. 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:

   • No. This type of study (MRMC for AI assistance) is not applicable to an insulin pump and its control app. This device does not involve "human readers" interpreting medical images or data with or without AI assistance.

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

   • Not Applicable in the traditional AI/ML sense. The software itself undergoes comprehensive standalone verification and validation to ensure its functional correctness and safety, but this is distinct from "standalone performance" of a diagnostic AI algorithm. The device, including the app, is designed for human-in-the-loop use for insulin delivery.

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

   • Not Applicable. For a device like an insulin pump and its control app, "ground truth" isn't established in the same way as for diagnostic AI. Instead, the "ground truth" for safety and effectiveness is established through:
     • Functional Requirements: The expected, correct operation of the software and hardware.
     • Usability Objectives: The desired safe and effective user interaction, verified through human factors testing.
     • Regulatory Standards & Guidance: Adherence to established medical device software, cybersecurity, and risk management standards.

7. The sample size for the training set:

   • Not Applicable. This device is not described as using machine learning that has a "training set" in the context of data-driven model training. Its functionality is based on pre-defined algorithms for insulin delivery control.

8. How the ground truth for the training set was established:

   • Not Applicable. No training set for an AI/ML algorithm is mentioned.

Ask a specific question about this device

The t:slim X2 insulin pump with interoperable technology (the Pump) is intended for the subcutaneous delivery of insulin, at set and variable rates, for the management of diabetes mellitus in persons requiring insulin. The 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 Pump is intended for single patient, home use and requires a prescription. The Pump is indicated for use with NovoLog or Humalog U-100 insulin. The Pump is indicated for use in individuals 6 years of age and greater.

The Subject Device is an ambulatory, battery operated, rate-programmable infusion pump designed for the subcutaneous delivery of insulin, at set and variable rates, for the management of diabetes mellitus in persons requiring insulin. The device includes a disposable cartridge which is motor driven to deliver patient programmed basal rates and boluses through an infusion set into subcutaneous tissue. The desired timing and quantity of insulin delivery (bolus or basal) is programmed by the patient based on their healthcare provider's recommendations. The Subject Device can send and receive data to and from other interoperable devices and is designed to act on commands from other authorized digital pump controller devices to adjust insulin dosing. The Subject Device is designed to be able to receive and display alerts and alarms to users based on information received from other interoperable devices. The Subject Device is compatible with Interoperable Automated Glycemic Controllers, such as Basal-IQ Technology (K193483) and Control IQ Technology (DEN190034) to aid in diabetes management. In addition, the Subject Device is compatible with iCGM systems cleared under 21 CFR 862.1355 and marketed separately from the ACE Pump and Interoperable Automated Glycemic Controllers.

The Tandem Diabetes Care t:slim X2 Insulin Pump with Interoperable Technology underwent non-clinical testing to confirm that its User Settable Max Basal Limit software update feature met specified requirements and performed as intended.

1. Table of Acceptance Criteria and Reported Device Performance:

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

The document mentions "human factors validation testing" but does not specify the sample size of participants in this study. The data provenance (e.g., country of origin, retrospective or prospective) is not explicitly stated. It is implied to be prospective and conducted by Tandem Diabetes Care Inc., as it's part of their submission for their device.

3. Number of Experts and Qualifications:

The document does not provide details on the number of experts used to establish ground truth or their specific qualifications for the human factors study. It refers to "intended users" as participants. For software verification and validation, it implies internal experts from Tandem Diabetes Care were involved in activities like code inspection and walkthroughs, but their specific number and qualifications are not mentioned.

4. Adjudication Method:

The document does not describe an adjudication method for the test set.

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

No MRMC comparative effectiveness study was mentioned. The device is an insulin pump, and the evaluation focused on the safety and effectiveness of a new software feature, not on the interpretation of medical images or data by multiple readers.

6. Standalone Performance:

Yes, a standalone performance of the algorithm (or the software feature in this case) was done through "software verification and validation testing." This testing ensured the software conformed to patient needs and intended uses independently. The human factors study also evaluated the device's performance when used by intended users, but the software's foundational performance was assessed in isolation.

7. Type of Ground Truth Used:

For the human factors study, the ground truth was based on the ability of "intended users" to "effectively use the Subject Device for its' intended purpose in expected use environments" as evaluated against established usability standards (ANSI AAMI HE 75:2009, ANSI/AAMI/IEC 62366-1:2015, and FDA guidance).

For software verification and validation, the ground truth was conformity to defined "detailed software requirements," "patient needs," and "intended uses."

8. Sample Size for the Training Set:

The document does not explicitly mention a separate "training set" or its sample size. Software development and verification and validation processes were used, which would involve iterative development and testing, but a distinct training set in the context of machine learning is not applicable here as the description is of a feature update to an existing software, not a new AI algorithm that learns from data.

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

Not applicable, as a distinct training set for a machine learning algorithm is not described. The ground truth for the overall software development and testing was established through detailed software requirements, patient needs, intended uses, and adherence to established software development and risk management standards (ISO 14971-2012, ANSI AAMI IEC 62304:2006/A1:2016, and FDA guidance).

Ask a specific question about this device

The t:slim X2 insulin pump with interoperable technology (the Pump) is intended for the subcutaneous delivery of insulin, at set and variable rates, for the management of diabetes mellitus in persons requiring insulin. The 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 Pump is intended for single patient, home use and requires a prescription. The Pump is indicated for use with NovoLog/NovoRapid or Humalog U-100 insulin.

The t:slim X2 insulin pump with interoperable technology (as shown in Figure 1) is an ambulatory, battery operated, rate-programmable infusion pump designed for the subcutaneous delivery of insulin, at set and variable rates, for the management of diabetes mellitus in persons requiring insulin. The device includes a custom disposable cartridge which is motor-driven to deliver patient programmed basal rates and boluses through an infusion set into subcutaneous tissue.

The t:slim X2 insulin pump with interoperable technology consists of:

• a user-operated interface display; ●
• an electronic microprocessor software control system, including Bluetooth radio ● module and signal processing algorithms allowing the pump to communicate with digitally connected devices:
• motor and encoder; ●
• rack drive mechanism; and ●
• an audible speaker and a vibrator to provide alarms, alerts and reminders to the user. ●

Here's a breakdown of the acceptance criteria and the studies that prove the t:slim X2 insulin pump with interoperable technology meets those criteria, based on the provided document.

1. Table of Acceptance Criteria and Reported Device Performance

The document outlines special controls that act as acceptance criteria for this device type. The reported device performance is extracted directly from the analytical performance studies.

Table 1: Acceptance Criteria and Device Performance for t.slim X2 Insulin Pump

• Bolus (3 units): Average time to detection 0:01:00, max 0:01:35
• Bolus (25 units): Average time to detection 0:01:05, max 0:01:25
• Basal (2.0 units/hour): Average time to detection 1:01:39, max 1:29 |
  | 1.c. Validation testing results demonstrating compatibility with drugs which may be used with the pump based on its labeling. Testing must include assessment of drug stability under reasonably foreseeable use conditions which may affect drug stability (e.g., temperature, light exposure, or other factors as needed). | Insulin Compatibility and Stability: In vitro testing for extractables/leachables and insulin compatibility with Humalog and Novolog. Stability evaluated for 6 days at 25°C and 3 days at 37°C (stressed). Acceptable results for degradation products, extractables, and leachables. |
  | 1.d. The device parts that directly or indirectly contact the patient must be demonstrated to be biocompatible. This shall include chemical and particulate characterization on the final, finished, fluid contacting device components demonstrating that risk of harm from device-related residues is reasonably low. | Biocompatibility: Insulin cartridge tested per ISO-10993-1. Results: non-toxic (cytotoxicity, systemic toxicity), no sensitization, non-irritant (irritation/intracutaneous, implantation), non-hemolytic, negative subacute/subchronic toxicity, non-mutagenic/non-genotoxic (Ames, Chromosome Aberration, Mouse Micronucleus). |
  | 1.e. Evidence verifying and validating that the device is reliable over the ACE pump use life, as specified in the design file, in terms of all device functions and in terms of pump performance. | Mechanical Engineering: Referenced previous submissions (P140015, P180008) for mechanical reliability, stress, performance, packaging, shelf life, and storage testing. New hazard analysis accounted for interoperability risks. Specific tests mentioned include: Environmental Storage, Drop Resistance, Fluid Ingress (IPX7), Battery Verification, Random Vibration, Mechanical Shock, etc. |
  | 1.f. The device must be designed and tested for electrical safety, electromagnetic compatibility, and radio frequency wireless safety and availability consistent with patient safety requirements in the intended use environment. | EMC and Wireless Coexistence: Passed IEC 60601-1-2 (EMC, EMI, wireless coexistence). Passed FCC Part 15 (Radiated Emissions, Occupied Bandwidth, Band-edge Measurement). Demonstrated communication with digitally connected device at max distance of 20 feet.
  Electrical Safety and Essential Performance: Passed IEC 61010-1 (ed. 3) and collateral standards (IEC 60601-1-8, 1-11, 2-24). |
  | 1.g. For any device that is capable of delivering more than one drug, the risk of cross-channeling drugs must be adequately mitigated. | Not explicitly addressed as a separate study point, but "Design mitigations to prevent cross-channeling" is listed as a mitigation measure for "Patient harm due to inadequate drug delivery accuracy" in Section K. |
  | 1.h. For any devices intended for multiple patient use, testing must demonstrate validation of reprocessing procedures and include verification that the device meets all functional and performance requirements after reprocessing. | Not applicable; device is for single patient, home use (Section G.1). |
  | 2.a. Risk control measures shall be implemented to address device system hazards and the design decisions related to how the risk control measures impact essential performance shall be documented. | Hazard Analysis: Revised hazard analysis to account for interoperability risks, unique design elements, and intended use. Traced all identified risks to adequate design controls and demonstrated appropriate implementation and validation. |
  | 2.b. A traceability analysis demonstrating that all hazards are adequately controlled and that all controls have been validated in the final device design. | Hazard Analysis: "traced all identified risks to adequate design controls, and demonstrated that design features were appropriately implemented and validated." |
  | 3. The device shall include validated interface specifications for digitally connected devices... | Interoperability: Plan and approach for interoperability provided per FDA Guidance "Design Considerations and Pre-market Submission Recommendations for Interoperable Medical Devices." Included secure, accurate, reliable communication, sharing state information, handling out-of-bounds data, and a process for sharing specs with connected devices. |
  | 4. The device must include appropriate measures to ensure that safe therapy is maintained when communications with digitally connected alternate controller devices is interrupted, lost, or reestablished after an interruption... Validation testing results must demonstrate that critical events that occur during a loss of communications... are handled appropriately... | Interoperability: Validated software protocols for failsafe design features to mitigate risks of communication interruption. The hazard analysis also covered scenarios where compatible/incompatible digital devices attempted communication and delivered commands. |
  | 5. The device design must ensure that a record of critical events is stored and accessible for an adequate period to allow for auditing of communications between digitally connected devices... | Data Logging: Validated software protocols enable recording critical events: insulin delivery, pump commands/confirmations, connectivity states, malfunctions, alarms. These were reviewed and found adequate. |
  | 6. Design verification and validation must include results obtained through a human factors study that demonstrates that an intended user can safely use the device for its intended use. | Human Factors: Non-randomized, multicenter study with 30 representative participants (plus separate study with 36 participants aged 6-11) in a simulated use environment. Assessed comprehension and usability for critical tasks, alone and with a digitally connected device. Demonstrated safe use by intended users. |

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

• Basal Delivery Accuracy: 32 pumps (16 new, 16 aged to simulate 4 years use). For both aged and unaged, 8 tested with new cartridges, 8 with 2-year aged cartridges.
• Bolus Delivery Accuracy: 32 pumps (16 new, 16 aged to simulate 4 years use). For both aged and unaged, 8 tested with new cartridges, 8 with 2-year aged cartridges.
• Occlusion Detection: 29 pumps.
• Human Factors: 30 representative participants for general use, plus an additional study with 36 representative participants aged 6-11.
• Data Provenance: Not explicitly stated regarding country of origin, but the testing appears to be retrospective analytical and simulated use studies conducted by the manufacturer as part of the De Novo submission. "Simulated 4 years of typical regular use" and "2 years of real time aging" for cartridges indicate specific aging protocols rather than field data.

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

• Not Applicable: The studies described for analytical performance (delivery accuracy, occlusion detection) and biocompatibility use physical measurements and laboratory standards (e.g., scales for fluid delivery, time measurement for occlusion, ISO standards for biocompatibility) to establish ground truth, not expert consensus.
• For Human Factors studies, "ground truth" is typically defined by successful completion of critical tasks or identification of use errors based on pre-defined criteria, which don't require external experts to "establish ground truth" in the same way clinical image interpretation might. The design and validation of the Human Factors study would be overseen by human factors engineering specialists.

4. Adjudication Method for the Test Set

• Not Applicable: The analytical performance and biocompatibility studies rely on direct measurement and established laboratory standards, not human interpretation requiring adjudication. Human Factors studies would identify use errors, but adjudication (e.g., 2+1, 3+1) is typically for disagreements in subjective interpretation, which isn't the primary method for these engineering and usability tests.

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

• Not Applicable: This device is an insulin pump and not an AI-powered diagnostic imaging device or a system where "human readers" interpret data in the context of improving with or without AI assistance. The "AI" component mentioned is an "automated insulin dosing software" which is an external, interoperable device. This document focuses on the pump's ability to reliably act on commands from such devices, not on the performance of human readers assisted by AI.

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

• Yes, for the pump itself: The detailed analytical performance testing for basal delivery accuracy, bolus delivery accuracy, and occlusion detection represents the standalone performance of the t:slim X2 pump without human intervention during the measurement phase. Water was pumped, and weights/times were recorded. The pump's ability to receive, execute, and confirm commands from other devices (as per the interoperability section) also indicates standalone testing of specific functionalities.

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

• Objective Measurements and Standardized Testing:
  • Basal/Bolus Delivery Accuracy: Measured weight of water delivered vs. expected volume. This is an objective, quantitative measurement.
  • Occlusion Detection: Time elapsed between occlusion initiation and pump detection, an objective time measurement.
  • Biocompatibility: Results of standardized lab tests according to ISO 10993-1 (e.g., "Non-toxic", "No evidence of sensitization").
  • Electrical Safety, EMC, Wireless: Compliance with recognized international standards (IEC, FCC).
  • Human Factors: Identification of critical use errors and successful completion rates based on predefined safety criteria during simulated use.

8. The Sample Size for the Training Set

• Not Applicable for AI/Machine Learning Model Training: This document describes the evaluation of a medical device (insulin pump) through engineering, analytical, and human factors testing, not the training of an AI or machine learning algorithm. The pump's software includes "signal processing algorithms" but there's no indication that these algorithms are "trained" in the typical machine learning sense with a distinct training dataset. The development relied on hazard analyses and software development processes.

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

• Not Applicable: As there's no indication of a machine learning training set in the conventional sense within this document, the method for establishing its ground truth is not applicable.

Ask a specific question about this device