(189 days)
The GE Blood Glucose Monitoring System 333 is intended to be used for the quantitative measurement of glucose (sugar) in fresh capillary whole blood samples drawn from the fingertips, forearm or palm. The GE Blood Glucose Monitoring System 333 is intended to be used by a single person and should not be shared.
The GE Blood Glucose Monitoring System 333 is intended for self testing outside the body (in vitro diagnostic use) by people with diabetes at home as an aid to monitor the effectiveness of diabetes control. The GE Blood Glucose Monitoring System 333 should not be used for the diagnosis of, or screening for diabetes or for neonatal use. Alternative site testing should be done only during steady - state times (when glucose is not changing rapidly).
The GE Blood Glucose Test Strips 33 are for use with the GE Blood Glucose Meter 333 to quantitatively measure glucose (sugar) in fresh capillary whole blood samples drawn from the fingertips, forearm or palm.
The GE Blood Glucose Monitoring System 333 consists of the following devices: GE Blood Glucose Meter 333, GE Blood Glucose Test Strip 333, and Rightest Control Solution GC550. The GE333 Blood Glucose Meter and GE Blood Glucose Test Strips 333 are manufactured by BIONIME Corporation. The GE Blood Glucose Meter 333, when used with the GE Blood Glucose Test Strips 333, quantitatively measures glucose in fresh capillary whole blood. The performance of the GE Blood Glucose Monitoring System 333 is verified by the Rightest Control Solution GC550.
The Rightest Control Solution GC550 has been previously been cleared (K092052) and is being repackage for use with the GE Blood Glucose Monitoring System 333.
GE Blood Glucose Test Strip 333 paired with GE Blood Glucose Meter 333 that utilizes electrical characteristic technology for measuring the glucose level in human blood. A relatively small drop of blood (minimum requirement of only 0.75 µL) is placed on the disposable GE Test Strip 333 coated with Glucose Oxidase which interacts with the GE Blood Glucose Meter 333. Within 5 seconds after testing, the blood glucose level is indicated on the meter's digital display screen.
This document describes K143387, for the GE Blood Glucose Monitoring System 333, which is substantially equivalent to the Rightest Blood Glucose Monitoring System GM550.
Here's an analysis of the acceptance criteria and the studies performed:
1. Table of Acceptance Criteria and Reported Device Performance
The document frequently refers to "acceptance criteria" but does not explicitly list them in a single table with associated performance statistics. Instead, the performance is reported against implicit or general criteria (e.g., "falls within the acceptance criteria," "within the criteria of ±15%," "met the acceptance criteria").
Below is a table summarizing the performance evaluation where acceptance criteria are discernable from the text:
| Study/Test | Acceptance Criteria (Implied) | Reported Device Performance |
|---|---|---|
| Linearity | Linear regression mean slopes above 0.99, R2 values above 0.99 | Linear regression between device and reference method shows mean slopes above 0.9912 and R2 values above 0.999. Glucose assay is linear from 20-600 mg/dL. |
| Precision (Repeatability) | Falling within defined acceptance criteria (not explicitly stated numerically) | Repeatability evaluation results demonstrate that repeatability falls within the acceptance criteria. |
| Precision (Intermediate) | Falling within defined acceptance criteria (not explicitly stated numerically) | Intermediate precision evaluation results demonstrate that intermediate precision falls within the acceptance criteria. |
| Sample Volume (Minimum) | Less than 0.60 µL should display an error message (Er4); other volumes within ±10% bias compared to reference | When sample volumes were less than 0.60 µL, meter displayed error message (Er4). At all other sample volumes (greater than 0.65 µL), the results were within the accepted criteria of ±10% bias compared to reference. Minimum sample volume determined to be 0.75 µL. |
| Interference | Percentage of interference within ±10% bias (as implied by the report) | Ascorbic acid (≥ 5 mg/dL), cholesterol (600 mg/dL), and uric acid (≥ 10 mg/dL) indicated percentage of interference over ±10% bias in high glucose concentration (leading to limitations statement). The remaining 16 interference substances were within the criteria for normal and high glucose concentrations. |
| Hematocrit | Stability compared to reference method within ±15% of reference value | Stability compared to the reference method during various HCT ranges were all within the criteria of ±15% of reference value. The acceptable HCT range is 20-60%. |
| System Accuracy | Meets acceptance criteria and EGA (Error Grid Analysis) for clinical performance, no significant difference in slope, intercept, and correlation coefficient between subject device and comparative meter/reference lab instrument. | The accuracy of subject device met the acceptance criteria, and there was no significant difference in slope, intercept, and correlation coefficient between the comparative meter or reference lab instrument. |
| User Performance | Achieves accurate results such that it is substantially equivalent to predicate device | The study result shows substantial equivalence to predicate device used in finger, palm and forearm position for accurate results by lay users. |
2. Sample Size Used for the Test Set and Data Provenance
- Linearity: 15 levels of glucose concentrations. (Specific number of samples per level not stated). Data provenance is likely internal lab testing ("Venous blood samples spiked").
- Precision (Repeatability): 1500 test strips used (over 10 days, 5 glucose concentrations, 10 meters, 3 lots). Data provenance is internal lab testing ("venous blood sample results").
- Precision (Intermediate): 900 test strips used (over 10 days, 3 glucose concentrations, 10 meters, 3 lots). Data provenance is internal lab testing ("venous blood sample results").
- Sample Volume: 9 sample volumes (ranging from 0.60 to 3.0 µL), 3 lots of strips, 3 glucose concentration levels. Data provenance is internal lab testing ("Venous blood").
- Interference Study: 3 lots of test strips, 2 glucose concentration levels, 19 interference substances. Data provenance is internal lab testing ("venous blood samples").
- Hematocrit Study: 3 lots of blood glucose test strips, 6 glucose concentration levels, tested in 5 replicates. Data provenance is internal lab testing ("venous blood samples").
- System Accuracy Study: 103 patients participated. Data provenance is clinical study ("capillary whole blood obtained from fingertip, palm, and/or forearm"). The location (country) is not specified, but typically these studies are either internal to the company's location (Taiwan) or a contract research organization. It does not state if it was retrospective or prospective, but clinical accuracy studies are generally prospective.
- User Performance Study: 160 laypersons in multiple sites. Data provenance is a clinical study ("capillary whole blood from fingertip, palm and forearm sample sites"). The location (country) is not specified. It does not state if it was retrospective or prospective, but user performance studies are generally prospective.
3. Number of Experts Used to Establish the Ground Truth for the Test Set and Qualifications of Those Experts
For the non-clinical studies (Linearity, Precision, Sample Volume, Interference, Hematocrit), the ground truth was established by the YSI 2300 Plus Glucose Analyzer (or YSI 2300 Analyzer), which is a laboratory reference method. Therefore, human experts were not directly establishing the ground truth for these measurements; rather, it was established by a recognized gold standard instrument. Expert qualifications for operating and maintaining such instruments are implied but not stated.
For the System Accuracy Study, the ground truth was established by the YSI 2300 Plus Glucose Analyzer. A comparative effectiveness was also done against a reference device, the Roche ACCU-CHEK Performa. Human experts were involved in collecting samples and running the analyzers but not in establishing the ground truth value itself, which came from the YSI machine.
For the User Performance Study, the primary ground truth reference is the subject device's own performance against its predicate (implying the predicate device's established accuracy as a baseline for "accurate results"). However, in clinical studies like this, a lab reference method (like YSI) would typically be used to compare both the user's results and potentially results from trained personnel. The document does not explicitly state how ground truth was established for the user performance aspect, beyond showing substantial equivalence to the predicate.
4. Adjudication Method for the Test Set
Not applicable. For the analytical studies, the ground truth was established by an automated laboratory reference instrument (YSI 2300 Plus Glucose Analyzer). For the clinical studies, no "adjudication" between multiple human readers is mentioned, as the primary ground truth was instrumental or comparison to a predicate.
5. If a Multi Reader Multi Case (MRMC) Comparative Effectiveness Study was Done
No, an MRMC comparative effectiveness study in the context of human readers improving with AI vs. without AI assistance was not performed. This device is a blood glucose monitoring system, not an AI-powered diagnostic imaging tool requiring human interpretation.
A System Accuracy Study was performed comparing the device to a reference method (YSI 2300 Plus Glucose Analyzer) and a comparative meter (Roche ACCU-CHEK Performa). This is a comparative study but not an MRMC study involving human readers and AI assistance.
A User Performance Study involved lay users, effectively acting as "multiple readers" for the device, but it was to demonstrate the device's usability and accuracy by laypersons, not to assess an AI's impact on human reader performance.
6. If a Standalone (i.e. algorithm only without human-in-the loop performance) was Done
Yes, the analytical performance studies (Linearity, Precision, Sample Volume, Interference, Hematocrit) and the System Accuracy Study are essentially standalone performance evaluations of the device (meter and test strips) against a laboratory reference standard. These studies assess the algorithm's ability to accurately measure glucose without direct human interpretive intervention beyond operating the device and reference method. The device itself is an algorithm-driven measurement tool.
7. The Type of Ground Truth Used
For non-clinical analytical studies (Linearity, Precision, Sample Volume, Interference, Hematocrit) and the System Accuracy Study, the ground truth was established by a laboratory reference method: YSI 2300 Plus Glucose Analyzer.
For the User Performance Study, the ground truth was implied by the established accuracy of the predicate device and the reference method used in other studies, to determine if lay users could obtain "accurate results."
8. The Sample Size for the Training Set
The document does not specify a training set size. Blood glucose monitoring systems typically do not involve machine learning models that require distinct "training sets" in the same way AI/ML algorithms for image analysis do. Instead, the device's design, calibration, and internal algorithms are developed and validated using extensive lab testing and engineering principles. The studies described are primarily for verification and validation of the final product's performance and accuracy.
9. How the Ground Truth for the Training Set Was Established
As no specific training set for a machine learning model is mentioned, this question is not directly applicable. The device's underlying technology relies on electrochemical reactions and calibration curves established through rigorous laboratory testing using reference methods like the YSI 2300 Plus Glucose Analyzer, rather than a "training set" in the context of AI.
§ 862.1345 Glucose test system.
(a)
Identification. A glucose test system is a device intended to measure glucose quantitatively in blood and other body fluids. Glucose measurements are used in the diagnosis and treatment of carbohydrate metabolism disorders including diabetes mellitus, neonatal hypoglycemia, and idiopathic hypoglycemia, and of pancreatic islet cell carcinoma.(b)
Classification. Class II (special controls). The device, when it is solely intended for use as a drink to test glucose tolerance, is exempt from the premarket notification procedures in subpart E of part 807 of this chapter subject to the limitations in § 862.9.