(61 days)
The ACCU-CHEK Connect Diabetes Management App is indicated as an aid in the treatment of diabetes. The software provides for electronic download of blood glucose meters, manual data entry, storage, display, transfer, and self-managing of blood glucose and other related health indicators which can be shown in report and graphical format.
The ACCU-CHEK Bolus Advisor, as a component of the ACCU-CHEK Connect Diabetes Management App, is indicated for the management of diabetes by calculating an insulin dose or carbohydrate intake based on user-entered data. Before its use, a physician or healthcare professional must activate the bolus calculator and provide the patient-specific target blood glucose. insulin-to-carbohydrate ratio, and insulin sensitivity parameters to be programmed into the software.
The ACCU-CHEK Connect Diabetes Management App is designed to facilitate efficient collecting, transmitting, and analyzing of blood glucose results and other diabetes management data. The App helps:
Wireless transfer of data from ACCU-CHEK Aviva Connect Blood Glucose Meter. Assist in general diabetes management through logging of contextual data. ACCU-CHEK Bolus Advisor support of mealtime insulin dosing calculations. Perform structured testing. Wireless transfer of data from mobile devices to ACCU-CHEK Connect Online Diabetes Management System and optionally share this data with healthcare provider (HCP) or caregiver.
The insulin bolus calculations provided by the app are meant for patients undergoing multiple daily injection therapy. Bolus calculators, such as the ACCU-CHEK Bolus Advisor, have been demonstrated to facilitate the optimization of glycemic control in patients who are trained in multiple daily insulin injection therapy and under the supervision of healthcare professional experienced in managing insulin-treated patients. Such calculators have also been shown to reduce patient fear of hypoglycemia and improve patient confidence in diabetes management.
The ACCU-CHEK Connect Diabetes Management App is not intended to serve as an accessory to an insulin pump.
The provided document describes a 510(k) premarket notification for the "ACCU-CHEK Connect Diabetes Management App" for iOS platform. The submission aims to demonstrate substantial equivalence to an existing predicate device, the Android OS version of the same app (K141929). The core of the argument for substantial equivalence relies on the fact that the bolus calculator algorithm and intended use have not changed, and the modifications are primarily related to adapting the app to the iOS operating system.
Here's an analysis of the acceptance criteria and the study that proves the device meets them, based on the provided text:
1. Table of Acceptance Criteria and Reported Device Performance
The document does not explicitly present a table of acceptance criteria with numerical targets and corresponding performance metrics for the modified device. Instead, the acceptance is based on demonstrating that the iOS version performs equivalently to the already cleared Android version. The performance is assessed by confirming that the changes do not introduce new hazards or alter the core functionality.
| Acceptance Criterion | Reported Device Performance |
|---|---|
| Functional Equivalence | The ACCU-CHEK Connect Diabetes Management App for iOS retains all the core functionalities of the predicate Android version, including: |
- Electronic download of blood glucose meters
- Manual data entry
- Storage, display, transfer, and self-managing of blood glucose and other related health indicators
- ACCU-CHEK Bolus Advisor for insulin dose/carbohydrate intake calculation
- Structured testing
- Wireless transfer of data to ACCU-CHEK Connect Online Diabetes Management System
- Bolus calculator algorithm is unchanged from the predicate device.
- Bolus calculator activation prescription control process, activation, and patient training materials, and user interface screens (related to bolus calculator) are unchanged from the predicate device. |
| Safety - Risk Assessment | A risk analysis according to ISO 14971 was carried out. Potential faulty conditions and hazards were systematically identified and evaluated using "Failure Mode Effect and Criticality Analysis." Adequate protection measures were implemented. The risk assessment for the iPhone version "relied heavily" on the risk assessment performed for the Android OS version. Post-launch monitoring of the Android version did not identify possible faulty conditions leading to hazards for the patient. |
| Performance Requirements | "Design verification bench testing on the modification of ACCU-CHEK Connect Diabetes Management App demonstrated that the device meets the performance requirements for its intended use." (Specific metrics are not provided, as the claim is equivalence to the predicate). |
| Human Factors | An "expert evaluation" (human factors) was performed to show that the predicate design validation can be used to support the iPhone version's design validation. This evaluation used side-by-side comparison of screenshots between Android and iOS versions to review changes. - Changes reviewed were due to inherent differences between OS user interface standards.
- Enhancements based on results of Android version summative and iPhone version formative human factors study.
- Changes to validation study tasks.
"No new use-related hazard was identified during the expert evaluation." |
2. Sample Size Used for the Test Set and Data Provenance
- Test Set Sample Size: The document does not specify a numerical sample size for a "test set" in the traditional sense of a performance study with patient data.
- For functional and performance requirements: "Design verification bench testing" was conducted, but no sample size for this testing is provided. The testing aimed to confirm that the device meets performance requirements, likely through technical validation.
- For Human Factors: The human factors evaluation was an "expert evaluation" involving a comparison of screenshots and review of changes. It does not appear to be a study with user participants from a specific test set.
- Data Provenance: Not applicable in the context of this submission, as it largely focuses on the technical equivalence between two versions of the same software for different operating systems. The core bolus algorithm, if it was validated with patient data, would have been done for the predicate device.
3. Number of Experts Used to Establish the Ground Truth for the Test Set and Their Qualifications
- Human Factors: An "Human Factors expert" performed the evaluation. The exact number of experts (singular or plural often used generically) is not explicitly stated if it was more than one, nor are their specific qualifications (e.g., years of experience, specific certifications) detailed, beyond them being "Human Factors expert".
- Other Testing: The document does not describe the establishment of a "ground truth" for a test set in the context of clinical outcomes or diagnostic accuracy, as the device is a diabetes management app with an unchanged bolus calculator algorithm. The "ground truth" for the bolus calculation would have been established during the development and validation of the predicate device's algorithm, adhering to medical and physiological principles of insulin dosing.
4. Adjudication Method for the Test Set
Not applicable. There is no mention of an adjudication method, as the studies described (bench testing, expert human factors evaluation) do not involve subjective interpretation or a need for external consensus on a "ground truth" for clinical cases in this specific 510(k) submission.
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
No. An MRMC comparative effectiveness study was not done. This type of study is typically used for diagnostic imaging devices where human readers interpret medical images with and without AI assistance. The ACCU-CHEK Connect Diabetes Management App is a diabetes management software, not an imaging diagnostic device, and thus this methodology is not relevant to its validation as described here.
6. If a Standalone (i.e., algorithm only without human-in-the-loop performance) was done
Yes, in essence, the predicate device's bolus calculator algorithm would have undergone standalone validation. The current submission explicitly states: "The insulin bolus calculator algorithm is unchanged as compared to the predicate device." This implies that the standalone performance of the algorithm itself was established during the predicate's clearance. The "Design verification bench testing" for the modified app would primarily verify that the implementation of this unchanged algorithm on the new platform correctly computes the same results.
7. The Type of Ground Truth Used
For the bolus calculator functionality (which is based on the unchanged algorithm from the predicate):
- The ground truth would be based on established medical and physiological principles for insulin dosing, as programmed into the algorithm's parameters (e.g., target blood glucose, insulin-to-carbohydrate ratio, insulin sensitivity, insulin action profiles). The accuracy of its calculations would be verified against these predefined parameters and typically, in a predicate device, against known physiological models or expert-derived reference calculations.
For the human factors evaluation:
- The "ground truth" was implicitly the established safety and usability of the predicate (Android) version, and the goal was to ensure the iOS version maintained this without introducing new hazards. The "expert evaluation" served as the primary method to assess this.
8. The Sample Size for the Training Set
The document does not mention a training set or its sample size. This submission is for a software modification (porting to a new OS) rather than the development of a new AI/ML algorithm that typically requires a large training set. The bolus calculator algorithm is based on predefined physiological equations and parameters, not on machine learning models trained on vast datasets.
9. How the Ground Truth for the Training Set Was Established
Not applicable, as there is no mention of a training set for an AI/ML model in this submission. The bolus calculator algorithm's "ground truth" (i.e., its correctness) would have been established at the time of the predicate device's development through clinical and algorithmic validation against medical standards and physiological models, as it is a rule-based system, not a data-driven learning system.
§ 868.1890 Predictive pulmonary-function value calculator.
(a)
Identification. A predictive pulmonary-function value calculator is a device used to calculate normal pulmonary-function values based on empirical equations.(b)
Classification. Class II (performance standards).