(264 days)
The RIA SafeGuard is intended as a room recirculating air cleaner. The system is used for filtering out airborne particles PM2.5 from the air for medical purposes, and is intended for over-the-counter use.
The RIA SafeGuard has wind speed of 1.3 m/s and airflow of 135 CMH (m3/h). Using the HEPA H13 filter, it provides a 4 log (99.99%) reduction in suspended particulate PM2.5 in 120 minutes within a 29.75 m³ room.
RIA SafeGuard (SG020MS120) is an air purifier that contains the highefficiency HEPA technology and Bi-Polar Ionization (BPI) unit to remove 99.99% of particles as small as 2.5 microns from the air stream. SafeGuard addresses suspended particles with a single wind speed of 1.3 m/s, providing a 4 log reduction in PM2.5 particulates within 120 minutes in a 29.75 m³ space. The airflow is 135 CMH (m3/h). It is intended to be placed in a room such as medical care institutions and hospitals etc. RIA SafeGuard is a product that is designed as one-touch, easy to use, and no contact with patients. RIA SafeGuard has built in a blue LED as the switch-on indicator. All the maintenance should be done by a trained technician. Changing the HEPA is the only recommended maintenance to do by a non-trained technician, recommended renewing every 12 months.
The provided text describes the mechanical and electrical performance of an air purification device, not an AI/ML-enabled medical device. Therefore, the information typically required for an AI/ML device's acceptance criteria, such as details about ground truth, expert adjudication, MRMC studies, or training set data, is not present in this document.
However, I can extract the acceptance criteria and study details relevant to this specific device based on the provided text.
Here's a breakdown of the acceptance criteria and the study that proves the device meets them, according to the document:
1. A table of acceptance criteria and the reported device performance:
| Test Name | Purpose | Acceptance Criteria | Reported Performance |
|---|---|---|---|
| Filtration of Particles | Demonstration of Device Air Cleaner Efficacy for 4 Log Reduction of PM2.5 Air Particulates in 120 minutes in a 29.75 m³ room. | Reduction rate > 99.99% (4 log reduction) | Pass |
| Ozone Emissions UL867 | Determine the ozone emissions of the device. | The maximum ozone concentration emitted is below the 0.050 ppm standard. | Pass |
| Electrical Safety Testing IEC 60601-1:2005, IEC 60601-1:2005+AMD1:2012, IEC 60601-1:2005+AMD1:2012+AMD2:2020 | Determine the general requirements for basic safety and essential performance of the device. | Meet the test specification according to the standards. | Pass |
| Electromagnetic Compatibility (EMC) Testing IEC 60601-1-2:2014 | Determine the electromagnetic compatibility, emissions and immunity of the device. | Meet the test specification according to the standards. | Pass |
2. Sample sized used for the test set and the data provenance:
- Sample Size: The document does not specify a "sample size" in the context of a test set for an AI/ML model. Instead, it refers to a single device (RIA SafeGuard, Model Number: SG020MS120) that underwent non-clinical testing.
- Data Provenance: The tests were conducted on the subject device, RIA SafeGuard. The exact location or origin of the testing data is not explicitly stated, but the manufacturer is RIA Tech Co., Ltd. located in Taipei City, Taiwan (R.O.C.). The tests were "non-clinical," implying they were conducted in a laboratory or controlled environment, rather than involving patient data (retrospective or prospective).
3. Number of experts used to establish the ground truth for the test set and the qualifications of those experts:
This information is not applicable to this device's testing. "Ground truth" in the context of an AI/ML model usually refers to labeled data verified by human experts. For a physical device like an air cleaner, the "ground truth" is established through direct physical measurements against predefined engineering and safety standards (e.g., measuring particle reduction with sensors, ozone concentration with detectors). No human experts were involved in establishing the "ground truth" for these physical performance metrics.
4. Adjudication method (e.g. 2+1, 3+1, none) for the test set:
This information is not applicable. Adjudication methods are relevant for subjective interpretations of data, typically in AI/ML model validation where multiple experts might disagree on labels. For the physical performance tests conducted, the results are objective measurements that do not require adjudication.
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:
This information is not applicable. An MRMC study is relevant for evaluating the impact of an AI/ML system on human performance (e.g., how radiologists improve diagnostic accuracy with AI assistance). The RIA SafeGuard is an air cleaning device, not an AI/ML diagnostic tool, and therefore, an MRMC study was not performed.
6. If a standalone (i.e. algorithm only without human-in-the-loop performance) was done:
This information is not applicable. This question pertains to AI/ML algorithm performance. The "Filtration of Particles" test can be considered an analogue to a standalone performance test for the device in that it measures the device's inherent ability to clean air without human interaction determining its efficacy. However, it's not an "algorithm-only" test.
7. The type of ground truth used (expert consensus, pathology, outcomes data, etc):
The "ground truth" for this device's performance was established through objective physical measurements against established standards and design specifications.
- For "Filtration of Particles": The ground truth was the measured reduction of PM2.5 particulates using scientific instruments.
- For "Ozone Emissions": The ground truth was the measured ozone concentration.
- For "Electrical Safety" and "Electromagnetic Compatibility": The ground truth was compliance with the specified IEC standards determined through testing.
8. The sample size for the training set:
This information is not applicable. This device is a physical product, not an AI/ML model that undergoes training with a dataset.
9. How the ground truth for the training set was established:
This information is not applicable, as there is no training set for this type of device.
§ 880.5045 Medical recirculating air cleaner.
(a)
Identification. A medical recirculating air cleaner is a device used to remove particles from the air for medical purposes. The device may function by electrostatic precipitation or filtration.(b)
Classification. Class II (performance standards).