Search Results

The intended use of the Digistat Smart Central is to provide an interface with clinical systems to forward information associated to the particular event from patient monitors, ventilators, and infusion pumps to the designated display device(s). For medical, near real time alarms, the Digistat Smart Central is intended to serve as a parallel, redundant, forwarding mechanism to inform healthcare professionals of particular medical related events. The Digistat Smart Central does not alter the behavior of the primary medical devices and associated alarm annunciations. The display device provides a visual, and/or audio and/or vibrating mechanism upon receipt of the alert.

The Digistat Smart Central is intended for use as a secondary alarm. It does not replace the primary alarm function on the medical devices.

Digistat Smart Central is an on-site integration solution which forwards medical device status and alarm information to the user via display devices provided by Ascom or third-party companies. Users receive, time-critical information from connected medical devices directly via their display devices as text, alarms or data. Digistat Smart Central allows users to be aware of their patients' status and alarm conditions when they are away from the patient and associated medical devices.

Digistat Smart Central connects to the information sources through wired ethernet connections which are part of the customer's infrastructure. Digistat Smart Central software acquires patient data and information through DIGISTAT Connect, a MDDS, from medical devices including patient monitors, ventilators, and infusion pumps. The user configures Digistat Smart Central to determine which information, including alarm notifications, is delivered to which users. Digistat Smart Central then formats the data for delivery to the display devices.

Digistat Smart Central is not in contact with the patient. It is intended to forward the information generated by the connected medical devices and systems and it does not generate patient related alarms. As such, the patient population and patient conditions are established by the medical devices and systems to which the product is connected.

This document, a 510(k) Premarket Notification for the Digistat Smart Central, primarily focuses on demonstrating substantial equivalence to predicate devices rather than proving a device meets specific performance acceptance criteria for an AI/ML algorithm. The product described is a medical device integration solution that forwards information and alarms, not an AI/ML diagnostic or predictive algorithm.

Therefore, many of the requested elements pertaining to AI/ML model validation (e.g., sample size for test/training sets, ground truth establishment by experts, MRMC studies, standalone performance) are not applicable or detailed in this document. The document outlines performance testing more generally related to software development, cybersecurity, and usability engineering, rather than clinical performance metrics typical for AI/ML algorithms.

Here's an attempt to answer the questions based on the provided text, highlighting where information is not present due to the nature of the device:

1. A table of acceptance criteria and the reported device performance

The document does not explicitly present a table of quantitative acceptance criteria for performance metrics (like sensitivity, specificity, accuracy) akin to AI/ML device validation. Instead, the "Summary of Performance Testing" section indicates that the device's software and benchmark performance were verified and validated against:

• Software Standards/Guidance:

  • FDA guidance: "The content of premarket submissions for software contained in medical devices, 11 May 05."
  • FDA guidance: "Off-the-shelf software use in medical devices, 09 Sep 99."
  • FDA guidance: "General principles of software validation; Final guidance for industry and FDA staff, 11 Jan 02."
  • FDA guidance: "Content of premarket submissions for management of cybersecurity in medical devices, 02 Oct 14."
  • "Cybersecurity for Networked Medical Devices Containing Off-The-Shelf (OTS) software, 14 Jan 05."
  • "Design Considerations and Premarket Submission Recommendations for Interoperable Medical Devices, 06 Sep 17."
  • IEC 62304: 2006, "Medical device software - Software life cycle processes."

• Performance Testing – Bench (Usability):

  • IEC 62366-1: 2015, "Medical devices – Application of usability engineering to medical devices."

Reported Device Performance:

The document states:

• "Test results indicate that the Digistat Smart Central software complies with its predetermined specifications and the applicable guidance documents."
• "Test results indicate that the Digistat Smart Central complies with its predetermined specifications and the applicable standards."
• "The results of these activities demonstrate that the Digistat Smart Central is performs as well as the predicate devices."

This indicates qualitative confirmation of meeting specifications and standards rather than specific quantitative performance numbers (e.g., latency, throughput, error rates) that would typically be associated with performance acceptance criteria for an AI/ML device. The "acceptance criteria" here are compliance with software and usability standards, and the "performance" is that it did comply.

2. Sample sizes used for the test set and the data provenance (e.g., country of origin of the data, retrospective or prospective)

This information is not provided in the document. As the device is a data forwarding system and not an AI/ML diagnostic tool, there's no mention of "test sets" in the context of patient data or clinical images/signals for AI/ML model evaluation. The "tests" mentioned are verification and validation of software and usability, which don't typically involve "sample sizes" of clinical data to the same extent an AI/ML model would.

3. Number of experts used to establish the ground truth for the test set and the qualifications of those experts (e.g. radiologist with 10 years of experience)

This information is not applicable/provided. The device does not generate diagnostic interpretations or predictions that would require expert-established ground truth for performance evaluation in the context of AI/ML.

4. Adjudication method (e.g. 2+1, 3+1, none) for the test set

This information is not applicable/provided. Adjudication methods are used in studies involving human interpretation or uncertain ground truth, which is not the primary focus for a data forwarding device like Digistat Smart Central.

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/provided. MRMC studies are typically done for AI/ML devices that assist human interpretation (e.g., radiologists reading scans with vs. without AI assistance). The Digistat Smart Central is a messaging system, not an AI-assisted diagnostic tool.

6. If a standalone (i.e. algorithm only without human-in-the-loop performance) was done

This information is not applicable/provided. The device is a "Smart Central" system that forwards information; its "performance" is its ability to reliably forward data and alarms, not to make a standalone diagnosis or prediction.

7. The type of ground truth used (expert consensus, pathology, outcomes data, etc)

This information is not applicable/provided. Ground truth, in the AI/ML sense, is not relevant for the type of device (data forwarding system) described in this document. The "truth" for this device would be whether the data was forwarded accurately and in a timely manner, which would be verified through system testing rather than clinical ground truth labels.

8. The sample size for the training set

This information is not applicable/provided. The Digistat Smart Central is not described as an AI/ML algorithm that requires a training set. Its "software" is rigorously engineered and verified/validated against specifications and standards, not "trained" on a dataset.

9. How the ground truth for the training set was established

This information is not applicable/provided. As the device is not an AI/ML algorithm requiring a training set, the concept of "ground truth for the training set" does not apply.

Ask a specific question about this device