(77 days)
The IntelliVue GuardianSoftware is intended for use by healthcare providers whenever there is a need for generation of a patient record.
The IntelliVue GuardianSoftware is indicated for use in the collection, storage and management of data from Philips specified measurements, Philips Patient Monitors and qualified 3rd party measurements that are connected through networks.
The IntelliVue GuardianSoftware is a stand-alone 'Clinical Information Management System (CIMS)' software application with client-server architecture and designed to be used in professional healthcare facilities (i.e. hospitals, nursing homes) and is intended to be installed on a 'customer-supplied', compatible off-the-shelf (OTS) information technology (IT) equipment.
The IntelliVue GuardianSoftware is a documentation, charting, and decision-support software that is configurable by the hospital to suit the needs of individual clinical units. The device collects data/vital signs from the following Philips compatible patient monitor/measuring devices.
Using the collected data, the device provides trending, review, reporting and notification. The 'Guardian Early Warning Score (EWS)' application is integrated into the IntelliVue GuardianSoftware to provide the healthcare professional/provider basic assessment and the ability to recognize early signs of deterioration in patients.
The IntelliVue GuardianSoftware is not an alarming device and displays alarms from patient monitors as supplemental information only.
Here's an analysis of the acceptance criteria and study information based on the provided text:
1. Table of Acceptance Criteria and Reported Device Performance
The provided text does not contain specific quantitative acceptance criteria or a side-by-side performance comparison of the new device versus these criteria. Instead, the document focuses on demonstrating substantial equivalence to a predicate device. The acceptance is implied by the successful software verification and validation, as well as human factors testing, which collectively show the device "met all safety and reliability requirements and performance claims."
The closest we get to "performance" in the context of the comparison is in the "Substantial Equivalence Determination" column, which consistently states "IDENTICAL" or provides explanations for differences that do not affect substantial equivalence.
Key areas assessed for equivalence (and thus, implicit "performance" against the predicate):
| Feature | Predicate Device (Rev. D.0) | Subject Device (Rev. E.0X) | Acceptance/Equivalence Determination |
|---|---|---|---|
| Intended Use | Identical | Identical | Substantially Equivalent (IDENTICAL) |
| Indications for Use | Collection, storage, management of data from Philips specified measurements & Patient monitors. | Collection, storage, management of data from Philips specified measurements, Patient Monitors, and qualified 3rd party measurements. | Substantially Equivalent (Difference in indications for use does not affect substantial equivalence, HL7 testing verified 3rd party measurements.) |
| System Platform | Client Server Architecture, Microsoft OS, OTS IT equipment | Identical | Substantially Equivalent (IDENTICAL) |
| Operating System(s) & Database | Windows 7/8.1/10, Win Server 2008R2/2012R2/2016, SQL 2014/2016/2017, Android 4.4+ | Windows 8.1/10, Win Server 2012R2/2016/2019, SQL 2014/2016/2017, Android 5.0+ | Substantially Equivalent (Updates to OS versions (removal of unsupported, addition of newer) ensure continued support and do not affect substantial equivalence.) |
| Programming Language | Microsoft® .NET C#, Microsoft® .NET C++, Java (mobile client) | Identical | Substantially Equivalent (IDENTICAL) |
| Maximum # of Supported Patients/Servers/Clients | Patients: 1200, Servers: 120, Clients: 240, SW Clients: 40 | Identical | Substantially Equivalent (IDENTICAL) |
| Compatible Devices | Philips IntelliVue Cableless Measurements, MP5/MP5SC, MX400/XG50, SureSigns VS3/VS4, Biosensor EarlySense Insight Device | Adds Philips EarlyVue VS30 (K190624) and Philips Biosensor BX100 (K192875) | Substantially Equivalent (Addition of new patient monitoring devices does not affect substantial equivalence.) |
| Software Functionality (General Overview) | Clinical Documentation, Patient Data Management, Reporting (SBAR), Calculations (Protocol Watch, EWS Scoring), Clinical decision support, Storage | Identical | Substantially Equivalent (IDENTICAL) |
| System Interfaces (IT Network Requirements) | Hospital IT (W)LAN infra, HL7, ADT, Labs, Paging | Adds HL7 data import extension | Substantially Equivalent (Addition of HL7 data import expands compatibility to 3rd party systems and does not affect substantial equivalence.) |
| Device Interfaces | Internal interface for connection to measuring devices via hospital LAN | Identical | Substantially Equivalent (IDENTICAL) |
| Remote Viewing/Operation | Independent display/operating interface, operations from host measuring device, PC UI (mouse/touchscreen), XDS Infrastructure Service | Identical | Substantially Equivalent (IDENTICAL) |
2. Sample size used for the test set and the data provenance:
The document does not specify a "test set" in terms of patient data. The evaluation relies on:
- Software Verification and Validation Testing: This is typically performed on software builds and simulated environments, not directly on patient data.
- Human Factors and Usability Testing: This involves human users interacting with the device. The sample size for this specific testing is not mentioned.
- Data Provenance: Not applicable in the context of patient data testing, as no patient data was used for performance evaluation of the software itself.
3. Number of experts used to establish the ground truth for the test set and the qualifications of those experts:
This information is not provided. Since the evaluation focused on software verification/validation and human factors/usability, and not on clinical performance with patient data requiring expert ground truth, this type of detail is absent.
4. Adjudication method (e.g. 2+1, 3+1, none) for the test set:
No adjudication method is mentioned, as there was no test set requiring expert adjudication for clinical ground truth.
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 MRMC comparative effectiveness study was conducted or mentioned. The device is a "Clinical Information Management System" and not an AI-assisted diagnostic tool for human readers. It collects, stores, and manages data, and provides decision support (e.g., Early Warning Score), but it doesn't appear to directly assist human readers in interpreting medical images or other complex data where "improvement" with AI would be measured.
6. If a standalone (i.e. algorithm only without human-in-the-loop performance) was done:
The device is inherently a "Clinical Information Management System (CIMS) software application" that operates in a standalone manner as software. However, its function as data collection, storage, management, and providing decision support (like EWS) means it is intended for use by healthcare providers and is integrated into clinical workflows. The performance testing focuses on its software functionality, reliability, and human factors, rather than a diagnostic algorithm's standalone performance. The "Guardian Early Warning Score (EWS)" is an algorithm, and its performance would be assessed for accuracy in calculating scores, but the document doesn't provide details on its standalone performance metrics.
7. The type of ground truth used (expert consensus, pathology, outcomes data, etc.):
For the software verification and validation, the "ground truth" would be the expected functional behavior and output of the software as defined by its requirements and specifications. For human factors testing, the ground truth would be the expected safe and effective interaction of users with the device. There isn't an external clinical ground truth (like pathology or outcomes) applied to the software's performance itself in this submission.
8. The sample size for the training set:
Not applicable. This is not an AI/machine learning model where a specific training set (of patient data) would be used. The software is developed based on engineering principles and regulatory requirements.
9. How the ground truth for the training set was established:
Not applicable, as there is no training set for an AI/machine learning model.
§ 870.2450 Medical cathode-ray tube display.
(a)
Identification. A medical cathode-ray tube display is a device designed primarily to display selected biological signals. This device often incorporates special display features unique to a specific biological signal.(b)
Classification. Class II (performance standards).