Search Results

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

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

Tidepool Loop is intended for single patient use.

Tidepool Loop is Rx - For Prescription Use Only.

Tidepool Loop is a mobile application with algorithm technology that works to control an ACE (Alternate Controller Enabled) insulin pump to automatically increase, and suspend delivery of basal insulin based on readings from an iCGM (integrated continuous glucose monitor) and glucose values predicted by Tidepool Loop. Tidepool Loop can also recommend, and with the user's confirmation, control the delivery of correction boluses when glucose values are predicted to exceed user configurable thresholds.

Tidepool Loop predicts glucose levels up to 6 hours in the approximate duration of insulin action for U-100 rapid-acting insulin) based on prior iCGM readings, insulin delivery history, and user input (e.g., carbohydrate intake and exercise, and in some cases fingerstick glucose) and uses that prediction to adjust insulin delivery. Tidepool Loop can be used to adjust or suspend basal insulin delivery every 5 minutes based on actual CGM sensor and predicted glucose readings. iCGM values are automatically used by the Tidepool Loop Bolus Recommendation Tool (TLBRT) when the Tidepool Loop Algorithm technology is active, i.e. when the device is operating in closed-loop mode with an active iCGM sensor session. When closed-loop mode is off, such as when it is manually disabled or when there is no active iCGM sensor session, the Tidepool Loop Bolus Recommendation Tool (TLBRT) is disabled. The user will use Tidepool Loop's simple bolus calculator. into which iCGM values are not automatically populated into the glucose field.

Users must manually enter information about carbohydrates to initiate a meal bolus. When closed-loop mode is on, recommended bolus delivery is calculated using the Tidepool Loop Bolus Recommendation Tool (TLBRT) and can be manually adjusted.

The Tidepool Loop app requires that specific, initial therapy settings are established by a health care provider as part of creating the prescription order. These settings include:

• Target Correction Ranges for normal operation, Pre-Meal and Workout Presets .
• Carb to Insulin Ratios .
• Insulin Sensitivity Factors ●
• Basal Rates ●
• Max Basal Rate ●
• Max Bolus .

Tidepool Loop uses two glucose-specific settings that may be different from the user's experience with traditional glucose monitoring or CGM therapy. These are Correction Range and Glucose Safety Limit.

Correction Range is the range of glucose values that the user wants Tidepool Loop to work to bring their glucose to. Correction Range can be set as low as 87 mg/dL and as high as 180 mg/dL. Tidepool Loop will warn the user if values outside the recommended bounds of 100-115 mg/dL are selected. The user can add different Correction Ranges for different times of day. Tidepool Loop supports up to 48 Correction Range segments in a 24-hour period.

Tidepool Loop allows these user-customizable target Correction Ranges:

• . Normal operation
• Pre-meal Preset
• . Workout Preset

Glucose Safety Limit (mg/dL) is a safety feature of the Tidepool Loop algorithm. If the current CGM value or any future predicted glucose value is below this safety limit. Tidepool Loop will suspend insulin delivery in an effort to prevent low glucose. The algorithm will also not recommend a bolus.

Glucose Safety Limit can be set as low as 67 mg/dL. It can be set as high as 110 mg/dL or to the Correction Range minimum, whichever qlucose value is lower. Tidepool Loop will warn the user if values outside Tidepool's recommended bounds of 74 to 80 mg/dL are selected.

The Glucose Safety Limit is also part of the Dosing Safety Threshold, which is part of the Tidepool Loop insulin delivery algorithm. The Dosing Safety Threshold is a period of time that has the same duration as the insulin activity duration (i.e., 6 hours). The Dosing Safety Threshold is equal to the user's Glucose Safety Limit for the first half of the insulin activity duration (i.e., 3 hours), and then increases until it is at the midpoint of the Correction Range at the end of the insulin activity duration (i.e., 6 hours).

Tidepool Loop is designed to be installed on an iPhone running iOS operating systems (version 15 or higher). The Tidepool Loop application includes an optional extension for Apple Watch devices running watchOS operating system (version 6.1 or higher).

The provided text describes the acceptance criteria and the study that proves the Tidepool Loop device meets these criteria. The device is an Interoperable Automated Glycemic Controller (iAGC) intended for the management of type 1 diabetes mellitus.

Here's a breakdown of the requested information:

1. Table of Acceptance Criteria and Reported Device Performance

The document doesn't explicitly present a formal "acceptance criteria" table with corresponding "reported device performance" values in a single, consolidated format. However, the outcomes of the clinical study, particularly for the 'intended user population', implicitly serve as the performance demonstration against the safety and effectiveness expectations.

Based on the "Summary Diabetes Outcomes" and "Summary Safety Outcomes" for the "Study population limited to intended user population" (Page 19), and further elaborated in the "Results (Intended User Population)" section (Page 22), we can construct the following:

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

The "test set" for the clinical performance evaluation was derived from an observational study of the "do-it-yourself (DIY) Loop system," which Tidepool Loop is based on.

• Total Participants Enrolled: 1,127
• Participants Initiating Study: 872 (606 in Cohort A - new users, 266 in Cohort B - existing users)
• Sample Size for Intended User Population (Primary Analysis Set): 175 participants (This subset met the criteria of being ≥6 years old, using Humalog or Novolog insulin only, and making use of the Tidepool Loop guardrails at least 90% of the time).
• Data Provenance:
  • Country of Origin: Not explicitly stated, but the study was conducted by the Jaeb Center for Health Research (a US-based entity) and involved the "DIY Loop system" which has a global user base. Given the context of FDA submission, the data is likely predominantly from the United States or a multinational study supporting US regulatory needs.
  • Retrospective or Prospective: The study was an observational, single-arm study that collected data prospectively after enrollment. Weekly web-based surveys were used for Adverse Events, and CGM/insulin dosing data was collected continuously and uploaded automatically. Historical data (e.g., pre-Loop SH events) were also collected as baseline.

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

The document does not describe the establishment of a "ground truth" for the test set in the traditional sense of expert radiologists annotating images. Instead, for clinical outcomes:

• Safety Adjudication: Severe Hypoglycemia (SH) events were adjudicated.
  • Number of Experts: Events were reviewed "both by the Jaeb Center for Health Research and by Tidepool staff, including review with Tidepool's Chief Medical Advisor." This implies at least two entities/individuals, with the Chief Medical Advisor being a qualified medical expert. Specific number beyond this is not provided.
  • Qualifications of Experts: The Chief Medical Advisor is a medical expert. The Jaeb Center for Health Research is a reputable clinical research organization focused on diabetes, implying their staff involved in adjudication would have relevant clinical expertise. Specific individual qualifications (e.g., "radiologist with 10 years of experience") are not provided as this is not an imaging study.

4. Adjudication Method for the Test Set

• Method: For Severe Hypoglycemia events, the review was performed "both by the Jaeb Center for Health Research and by Tidepool staff, including review with Tidepool's Chief Medical Advisor." This indicates a consensus-based review. The phrasing "adjudicated as 'Related to Loop'" suggests a systematic process for attributing events. It's not a typical "2+1" or "3+1" for initial findings, but rather an expert review of reported events for causality.

5. If a Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study Was Done

• No, an MRMC comparative effectiveness study was not done. This describes a clinical trial design relevant to diagnostic imaging interpretation studies, where different readers interpret cases with and without AI assistance. The Tidepool Loop study is a clinical observational study of an automated insulin delivery system, not an imaging device.
• Effect size of human readers improving with AI vs without AI assistance: Not applicable as this was not an MRMC study and the device does not assist human interpretation of diagnostic images. The study rather focused on how the automated system (Tidepool Loop / DIY Loop) directly impacts glycemic control and safety in real patients. The improved glycemic outcomes (TIR, HbA1c) and reduced incidence of severe hypoglycemia demonstrate the direct therapeutic effect of the AI algorithm.

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

• Yes, in essence. The "Tidepool Loop" product is an "Interoperable Automated Glycemic Controller (iAGC)" which runs an algorithm to automatically adjust basal insulin delivery and recommend boluses. While the user confirms boluses, the core automated function (basal adjustment) is algorithm-driven.
• Clinical Study Design: The clinical study evaluated the "DIY Loop" system, which is based on the same algorithm as Tidepool Loop, in "real-world, unsupervised, patient-driven use." This implies the study assessed the performance of the algorithm in situ as it was used by patients, with its automated functions and interactions. The primary glycemic and safety outcomes (TIR, HbA1c, SH, DKA) were derived directly from the overall system's performance, which is predominantly driven by the algorithm.

7. The Type of Ground Truth Used

• Clinical Outcomes / Patient Reported Data / Expert Adjudication: The "ground truth" for this device's performance was established through:
  • Clinical Measurements: HbA1c (blood samples) and Continuous Glucose Monitoring (CGM) data.
  • Patient-Reported Outcomes: Weekly web-based surveys for adverse events (severe hypoglycemia, DKA, hospitalization) and device issues. Psychosocial/quality of life aspects were collected via validated questionnaires (e.g., INSPIRE).
  • Expert Consensus/Review: Adjudication of Severe Hypoglycemia events as described in point 4.

8. The Sample Size for the Training Set

The document does not provide information on the training set or its size for the Tidepool Loop algorithm. The clinical study described is a validation study for regulatory submission, assessing the performance of an already developed algorithm (modeled on DIY Loop) in a real-world setting. Training data for such algorithms are typically much larger and often proprietary to the developer.

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

Since information on the training set is not provided, how its ground truth was established is also not detailed in this document. Given it's a medical device algorithm, training data would likely involve large datasets of CGM readings, insulin delivery, meal data, and corresponding verified glucose outcomes, potentially derived from clinical studies or real-world usage data with rigorous data curation and annotation.

Ask a specific question about this device