Search Results

The GEM Premier ChemSTAT is a portable critical care system for use by health care professionals to rapidly analyze lithium heparinized whole blood samples at the point of health care delivery in a clinical setting and in a central laboratory. The instrument provides quantitative measurements of Creatinine (Crea), Blood Urea Nitrogen (BUN) and Total Carbon Dioxide (tCO2) from arterial and venous heparinized whole blood. These parameters, along with derived parameters, aid in the diagnosis of a patient's acid/base status and metabolite balance.

· Creatinine (Crea) measurements are used in the diagnosis and treatment of renal diseases and in monitoring renal dialysis.
· Blood Urea Nitrogen (BUN) or urea measurements are used for the diagnosis, monitoring, and treatment of certain renal and metabolic diseases.
· Total carbon dioxide/tCO2 (also referred to as bicarbonate/HCO3-) is used in the diagnosis, monitoring, and treatment of numerous potentially serious disorders associated with changes in body acid-base balance.

Device Description

The GEM Premier ChemSTAT is a portable system that analyzes arterial and venous lithium heparinized whole blood at the point of health care delivery in a clinical setting and in a central laboratory for Creatinine, BUN and tCO₂. All tests are included in a single self-contained, disposable GEM Premier ChemSTAT PAK (cartridge).

Key Components:
Analyzer: The GEM Premier ChemSTAT analyzer has the internal logic and processing power necessary to perform analysis. It employs a unique touch-sensitive color screen and a simple set of menus and buttons for user interaction. The analyzer guides operators through the sampling process with simple, clear messages and prompts.
PAK (Cartridge): The disposable, multi-use GEM Premier ChemSTAT PAK is a completely closed cartridge that houses all components necessary to operate the instrument once the GEM PAK is validated. These components include the sensors, Process Control (PC) Solutions, sampler, and waste bag. The values of all PC Solutions are read from the GEM PAK Electronically Erasable Programmable Read Only Memory (EEPROM) chip. The components and processes used to manufacture the PC Solutions in the GEM PAK are traceable to National Institute of Standards and Technology (NIST) standards, Clinical & Laboratory Standards Institute (CLSI) procedures or other internal standards, where available and appropriate. The GEM Premier ChemSTAT PAK has flexible menus to assist facilities in maximizing efficiency. As part of this program, GEM ChemSTAT CVP (Calibration Valuation Products) are external solutions intended to complete the calibration process and final accuracy assessment of the iQM cartridge calibration following warm-up.
Intelligent Quality Management (iQM): Intelligent Quality Management (iQM) is used as the quality control and assessment system for the GEM Premier ChemSTAT system. iQM is an active quality process control program designed to provide continuous monitoring of the analytical process before and after sample measurement with real-time, automatic error detection, automatic correction and automatic documentation of all corrective actions. iQM performs 4 types of continuous, quality checks to monitor the performance of the GEM PAK, sensors, and reagents throughout the cartridge use-life. These checks include System, Sensor, Pattern Recognition (PR) and Stability Checks.

AI/ML Overview

Here's a breakdown of the acceptance criteria and study information for the GEM Premier ChemSTAT device, based on the provided text:

1. Table of Acceptance Criteria and Reported Device Performance:

The document doesn't explicitly state quantitative acceptance criteria in a dedicated table format. However, it indicates that "All results were within specification" and successful performance in comparison to predicate devices. For this summary, I've inferred the performance metrics as the reported study outcomes.

Analyte	Performance Metric	Acceptance Criteria (Implicit)	Reported Device Performance
Creatinine (Crea)	Internal Precision (Total %CV)	Within specification (not explicitly stated, but < 3.6% for all levels is good)	2.1% - 3.6% (Levels 1-5, N=120)
	Reproducibility (CVP)	Within specification (not explicitly stated, but < 6.8% for CVP Level 1 is good)	3.2% - 10.1% (Levels 1-4, N=90)
	Reproducibility (PVP)	Within specification (not explicitly stated, but < 8.8% for PVP Level 1 is good)	1.7% - 8.8% (Levels 1-5, N=90)
	External Precision (SD)	Within specification (not explicitly stated, but < 0.014 for fixed range)	0.011 - 0.014 (POC 1-3, Pooled_N=147 for fixed range)
	External Precision (%CV)	Within specification (not explicitly stated, but < 1.4% for variable range)	0.6% - 1.4% (POC 1-3, Pooled_N=54 for variable range)
	Limit of Blank (LoB)	Established	0.04 mg/dL
	Limit of Detection (LoD)	Established	0.07 mg/dL
	Limit of Quantification (LoQ)	Established	0.10 mg/dL
	Linearity (R²)	Ideally close to 1 (0.9976 is excellent)	0.9976
	Reportable Range	To be met	0.20 to 15.00 mg/dL (after testing up to 16.40 mg/dL)
	Analytical Specificity	No observed interference (for non-interfering substances) or acceptable bias (for interfering substances)	No interference for numerous substances. Specific biases for Creatine and Hydroxyurea detailed.
	Method Comparison (Slope)	Close to 1.0 (1.009 is excellent)	1.009
	Method Comparison (Intercept)	Close to 0 (-0.027 is excellent)	-0.027
	Method Comparison (R)	Close to 1 (0.997 is excellent)	0.997
	Method Comparison (Sample Range)	To be covered by testing	0.20 to 14.18 mg/dL
BUN (mg/dL)	Internal Precision (Total %CV)	Within specification (not explicitly stated, but < 1.8% for all levels is good)	1.1% - 1.8% (Levels 1-5, N=120)
	Reproducibility (CVP)	Within specification (not explicitly stated, but < 3.1% for CVP Level 2 is good)	2.0% - 3.1% (Levels 1-2, N=90)
	Reproducibility (PVP)	Within specification (not explicitly stated, but < 3.0% for PVP Level 5 is good)	1.7% - 3.0% (Levels 1-5, N=90)
	External Precision (SD)	Within specification (not explicitly stated, but < 0.18 for fixed range)	0.12 - 0.18 (POC 1-3, Pooled_N=132 for fixed range)
	External Precision (%CV)	Within specification (not explicitly stated, but < 2.8% for variable range)	0.7% - 2.8% (POC 1-3, Pooled_N=66 for variable range)
	Limit of Blank (LoB)	Established	0.3 mg/dL
	Limit of Detection (LoD)	Established	0.3 mg/dL
	Limit of Quantification (LoQ)	Established	1.2 mg/dL
	Linearity (R²)	Ideally close to 1 (0.9997 is excellent)	0.9997
	Reportable Range	To be met	3.0 to 112.0 mg/dL (after testing up to 122.0 mg/dL)
	Analytical Specificity	No observed interference (for non-interfering substances) or acceptable bias (for interfering substances)	No interference for numerous substances. No explicit interfering substances for BUN shown in the table.
	Method Comparison (Slope)	Close to 1.0 (0.965 is good)	0.965
	Method Comparison (Intercept)	Close to 0 (0.441 is small)	0.441
	Method Comparison (R)	Close to 1 (0.997 is excellent)	0.997
	Method Comparison (Sample Range)	To be covered by testing	3.0 to 109.7 mg/dL
tCO2 (mmol/L)	Internal Precision (Total %CV)	Within specification (not explicitly stated, but < 2.0% for all levels is good)	1.4% - 2.0% (Levels 1-5, N=120)
	Reproducibility (CVP)	Within specification (not explicitly stated, but < 4.5% for CVP Level 2 is good)	2.5% - 4.5% (Levels 1-2, N=90)
	Reproducibility (PVP)	Within specification (not explicitly stated, but < 5.5% for PVP Level 5 is good)	1.1% - 5.5% (Levels 1-5, N=90)
	External Precision (SD)	Within specification (not explicitly stated, but < 0.12 for fixed range)	0.11 - 0.12 (POC 1-3, Pooled_N=81 for fixed range)
	External Precision (%CV)	Within specification (not explicitly stated, but < 1.1% for variable range)	0.5% - 1.1% (POC 1-3, Pooled_N=120 for variable range)
	Limit of Blank (LoB)	Established	0.0 mmol/L
	Limit of Detection (LoD)	Established	0.2 mmol/L
	Limit of Quantification (LoQ)	Established	2.0 mmol/L
	Linearity (R²)	Ideally close to 1 (0.9986 is excellent)	0.9986
	Reportable Range	To be met	5.0 to 50.0 mmol/L (after testing up to 51.3 mmol/L)
	Analytical Specificity	No observed interference (for non-interfering substances) or acceptable bias (for interfering substances)	No interference for numerous substances. No explicit interfering substances for tCO2 shown in the table.
	Method Comparison (Slope)	Close to 1.0 (0.979 is good)	0.979
	Method Comparison (Intercept)	Close to 0 (0.535 is small)	0.535
	Method Comparison (R)	Close to 1 (0.986 is good)	0.986
	Method Comparison (Sample Range)	To be covered by testing	5.5 to 47.2 mmol/L

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

The document describes several studies, and the test sets are detailed separately for each:

Internal Precision Study – Whole Blood:
- Sample Size: 5 different concentrations of whole blood per analyte, with 8 replicates per run, 1 run per day for 5 days on 3 analyzers. Total N = 120 (5 concentrations * 8 replicates * 5 days * 3 analyzers).
- Data Provenance: Not explicitly stated, implied as internal lab data ("internal precision study"). No country of origin mentioned. Likely prospective, as it's a controlled study.
Reproducibility Study with Aqueous Controls – Point-of-Care (POC) Setting:
- Sample Size: 9 levels for Creatinine (4 CVP, 5 PVP) and 7 levels for BUN and tCO2 (2 CVP, 5 PVP). Each control level run in triplicate, twice a day for 5 days. Total N = 90 (3 replicates * 2 times/day * 5 days * 3 sites) pooled across 3 sites per level for each analyte.
- Data Provenance: Conducted at "control clinical point-of-care (POC) sites." No specific country of origin mentioned. Likely prospective, as it's a controlled study.
External Precision – Whole Blood:
- Sample Size: Patients samples (whole blood). N = 147 for Creatinine (fixed), 54 for Creatinine (variable), 132 for BUN (fixed), 66 for BUN (variable), 81 for tCO2 (fixed), 120 for tCO2 (variable). Less than 10% of samples were "contrived."
- Data Provenance: Conducted at "three (3) external clinical point-of-care (POC) sites." No specific country of origin mentioned. Likely a mix of prospective (as part of the study) and retrospective (existing patient samples) given the mention of "patient samples" and "contrived" samples.
LoB, LoD, and LoQ:
- Sample Size: 3 lots of GEM Premier ChemSTAT PAKs (cartridges). Specific number of measurements not detailed but implied to be sufficient for CLSI EP17-A2.
- Data Provenance: Internal laboratory study. Likely prospective.
Linearity:
- Sample Size: 9 levels per analyte, prepared by spiking or diluting whole blood. Each blood level analyzed in triplicate on 6 GEM Premier ChemSTAT test analyzers. Total N per level = 18 (3 replicates * 6 analyzers).
- Data Provenance: "whole blood" samples, but could be internal or simulated. Likely prospective.
Analytical Specificity (Interference Study):
- Sample Size: Various test substances at specified concentrations mentioned in tables. Specific number of replications or blood samples not detailed.
- Data Provenance: Not explicitly stated, implied as internal lab study. Likely prospective.
Clinical Testing (Method Comparison Study):
- Sample Size: N = 405 for Creatinine and BUN, N = 416 for tCO2. These were "lithium heparinized whole blood patient samples." Less than 10% of samples were "contrived."
- Data Provenance: Conducted at "three (3) point-of-care (POC) sites." No specific country of origin mentioned. This is a clinical study, likely a mix of prospective and retrospective patient samples.

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

The device is an in-vitro diagnostic (IVD) for quantitative measurements of analytes (Creatinine, BUN, tCO2) using established laboratory methods. For such devices, "ground truth" is typically established by comparative analysis against a reference method or predicate device, which are themselves considered gold standards in laboratory medicine. There is no mention of human experts establishing a "ground truth" through interpretation or consensus for these quantitative measurements, as would be the case for image-based diagnostic AI.

The method comparison study compares the candidate device to legally marketed predicate devices:

Cobas Integra Creatinine Plus Ver. 2 (K024098)
Cobas Integra Urea/BUN (K954000)
Cobas Integra Bicarbonate Liquid (K031879)

These predicate devices serve as the reference standard for the "ground truth" in the clinical method comparison.

4. Adjudication Method for the Test Set:

Not applicable. For this type of quantitative IVD device, ground truth for patient samples is typically established by measurement on a reference method/predicate device, not through expert adjudication in the way it's done for qualitative interpretations (e.g., radiology reads). The comparison involves statistical analyses (slope, intercept, R-value) rather than an adjudication process.

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:

Not applicable. This device is an automated in-vitro diagnostic (IVD) system that provides direct quantitative measurements. It does not involve human "readers" interpreting cases with or without AI assistance, nor is it an AI-powered diagnostic aid to human interpretation. Therefore, an MRMC study is not relevant here.

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

Yes, the studies presented (internal precision, external precision, LoB/LoD/LoQ, linearity, analytical specificity) all demonstrate the standalone performance of the GEM Premier ChemSTAT device. The device operates automatically to provide quantitative measurements without human intervention in the measurement process itself, beyond sample loading and initiation. The "clinical testing" (method comparison) also evaluates the device's performance against predicate devices as a standalone instrument.

7. The Type of Ground Truth Used (expert consensus, pathology, outcomes data, etc.):

The ground truth for the performance studies of the GEM Premier ChemSTAT is established by:

Reference Methods/Predicate Devices: For the method comparison study, the measurements obtained from the predicate devices (Cobas Integra Creatinine Plus Ver. 2, Cobas Integra Urea/BUN, Cobas Integra Bicarbonate Liquid) served as the reference standard for patient samples.
Prepared Standards/Controls: For precision, linearity, LoB/LoD/LoQ, and analytical specificity studies, the ground truth is established by precisely prepared control materials, aqueous solutions, or spiked/diluted whole blood samples with known concentrations of the analytes. These are traceable to national or international standards (e.g., NIST, CLSI).

8. The Sample Size for the Training Set:

This document describes a pre-market notification (510(k)) for a medical device (IVD analyzer). It does not explicitly mention a "training set" in the context of machine learning or AI models. The performance studies detailed are for verification and validation of the device's analytical and clinical performance against predicate devices and controlled materials, not for training a model. Therefore, this information is not provided because the device is not an AI/ML-driven product in the described context.

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

Not applicable, as a "training set" in the context of AI/ML is not described or implied for this device. The development process for such IVD devices typically involves rigorous analytical characterization and clinical validation against reference methods, rather than machine learning training.

Ask a Question

Ask a specific question about this device

Page 1 of 1