Search Results

The Plum Duo™ Precision IV Pump is intended for administration of parenteral fluids, medications, and whole blood and blood products through the following routes of administration: intravenous, intra-arterial, and subcutaneous.

The Plum Duo™ Precision IV Pump is intended for use in clinical environments in the hospital and other outpatient healthcare facilities by licensed healthcare professionals. These healthcare professionals are trained in the use of the infusion pump and the administration of therapies consistent with the intended use.

The Plum Duo™ Precision IV Pump is intended for adults, pediatric (including infants and children), and neonatal patient populations.

The Plum Duo™ Precision IV Pump is a large volume pump (LVP) with two independent pump channels that can deliver fluid to a patient from 1 to 4 lines independently. In addition, although the channels can operate independently, patient parameters can be shared across the channels to aid in ease of programming. The Plum Duo™ Precision IV Pump can only be used with dedicated PlumSet™ administration sets (not subject of this filing). Each pump channel accepts a cassette that is part of a PlumSet™ administration set and can connect to a primary and/or secondary container. The fluid is delivered from the upstream lines either serially (piggyback) or concurrently through the cassette to the downstream line. Each pump channel cassette has an independent downstream line (patient line), so the clinician can connect each downstream line to a single infusion site or two separate infusion sites. The flow rate accuracy precision has been improved (lower allowed variance) by implementing the new motor mechanism, as well as increased precision for programming concentration, flow rate and VTBI entries. The overall delivery accuracy of the system has improved to +/- 3% per TIR 101 standard condition testing and +/-5% for non-standard conditions.

The provided FDA 510(k) clearance letter and summary for the Plum Duo™ Precision IV Pump discuss the device's technical specifications and how it meets regulatory requirements for significant equivalence to a predicate device. However, this document does not describe the acceptance criteria and a study proving the device meets those criteria in the context of an AI/ML-driven medical device, as implied by the user's detailed request.

The Plum Duo™ Precision IV Pump is an infusion pump, a hardware medical device with embedded software, not an AI/ML diagnostic or prognostic tool. Therefore, the requested information (such as sample size for test sets, data provenance, number of experts for ground truth, adjudication methods, MRMC studies, standalone performance, training set details, etc.) does not apply to this specific device or the information provided in the 510(k) summary.

The 510(k) summary focuses on "Non-Clinical Testing" which includes:

• Verification testing of product requirements
• Human factors validation testing
• Reliability goals testing
• Safety assurance case following FDA Guidance for Infusion Pumps
• Flow rate and bolus accuracy testing per AAMI TIR101
• Software verification and validation per relevant FDA guidance documents (2005 and 2021 draft for software functions), which are focused on traditional software engineering principles, not AI/ML model performance.
• Human factors evaluations per FDA guidance and IEC 62366-1
• Electrical and Electromagnetic Compatibility testing per IEC 60601-1 and IEC 60601-1-2
• Cybersecurity testing per FDA guidance documents (2014 and 2016)
• Risk management activities per ISO 14971:2019 +A11 2021

Conclusion based on the provided document:

The provided document does not contain the information required to answer the specific questions about acceptance criteria and study proving an AI/ML device meets them. The device described, the Plum Duo™ Precision IV Pump, is an infusion pump, and its clearance relies on non-clinical performance and safety data relevant to its mechanical and software functions, not AI/ML-driven insights or diagnostics.

Therefore, I cannot populate the table or answer the specific questions regarding AI/ML study design and ground truth establishment for this device based on the given text. The text explicitly states: "Clinical evaluation is not required for this submission to support substantial equivalence." This further indicates that the type of studies and data provenance you are asking about (which are typical for AI/ML diagnostic devices) were not part of this 510(k) submission.

Ask a specific question about this device

The Cogent™ Hemodynamic Monitoring System (HMS) is intended for patients for whom the monitoring of continuous cardiac output and calculated hemodynamic parameters is indicated for diagnostic and prognostic evaluation by a clinician. Suitability for use on a patient is up to the physician's judgment and the diameter to be used. The target population includes patients for whom hemodynamic monitoring will improve clinical care. The target populations are identical to those for the predicate devices and include: Critical Care Patients, Trauma Patients, Cardiac Surgery Patients. The Cogent™ HMS is intended for use with ICU Medical pulmonary artery catheters and central venous oximetry catheters, and with ICU Medical Cogent™ sensors. The Cogent™ HMS is intended to measure and calculate venous oxygen saturation in patients. PulseCO functionality is limited to adult patients. The intended environment for use is the hospital including Critical Care Units (such as Medical, Surgical, and Coronary), Trauma and Accident Emergency Units, Post Anesthesia Care Units, Operating Rooms, and Cardiac Catheterization labs. The Cogent™ HMS is intended to be used by trained and qualified individuals in medical and surgical intensive care units, operating rooms, trauma and accident emergency units, coronary and intensive care units and cardiac catheterization laboratories. The CogentTM HMS is restricted to one patient at a time.

The Cogent™ HMS system is designed to compute and display cardiac and oximetry parameters relevant to patient care in the hospital acute care areas including Intensive Care Units and the Operating Room. Parameters include cardiac output and blood oxygen saturation levels, as well as other derived hemodynamic parameters. Measurements are obtained through ICU Medical pulmonary artery and central venous oximetry catheters, and ICU Medical CardioFloTM sensors. Input data for derived parameters may be keyed in by a clinician or may be obtained from a bedside monitor. The Cogent™ HMS provides the following functions: monitors patient cardiac output continuously, using continuous thermodilution, and intermittently, using bolus thermodilution; monitors cardiac output continuously using Pulse Power analysis on an arterial pressure waveform; monitors venous oxygen saturation by measuring the reflectance spectrum of the blood; and provides a general-purpose interface to the analog input/output channels of other monitoring devices. The Cogent" HMS consists of a base unit (patient interface module, PIM), a dedicated touch-screen display unit (user interface module, UIM) which allows for patient monitoring remotely (up to 50 feet), and the associated cables. The modules communicate with each other in docked, tethered (wired) or wireless mode. A physically separate optical module (OpMod) connects with an oximetry catheter. The Cogent™ HMS is designed for compatibility with PA catheters via connection to existing patient cables, i.e. unchanged cables as supplied with the primary predicate Q2 Plus. For the purpose of PulseCO™ data acquisition, the Cogent™ HMS is designed for compatibility with the CardioFloTM sensor and the new CardioFloTM reusable cable. In order to calculate blood oxygen saturation, the Cogent™ HMS is designed for compatibility with the existing optical module, its existing integrated cable and its associated compatible PA and oximetry catheters.

Here's an analysis of the provided text regarding the acceptance criteria and the study that proves the device meets those criteria:

The document primarily focuses on demonstrating substantial equivalence to predicate devices rather than defining specific acceptance criteria for a novel device performance. However, based on the content, we can infer some "acceptance criteria" through the comparison to predicate devices.

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

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

• Bench Tests (CCO, TdCO, SO2 algorithms): The sample size or specific dataset characteristics for the simulated physiological data are not explicitly provided. It simply states "bench simulation."
• Bench Test (PulseCO algorithm): The sample size or specific dataset characteristics are not explicitly provided. It states "the same simulated physiological data set as was used for the secondary predicate device."
• Animal Study: The sample size was 5 pigs. The provenance is not specified (e.g., country of origin, prospective/retrospective), but it describes an intervention-based study, suggesting it was prospective in design within the animal model.

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)

• For the bench tests, the ground truth was established by the simulated physiological data set. There is no mention of human experts establishing ground truth for these tests.
• For the animal study, the ground truth was established by "accepted reference devices." The number, qualifications, or adjudication method of human experts in establishing this "ground truth" or comparing to reference devices is not specified.

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

• None explicitly mentioned. The ground truth for bench tests was simulated data, and for the animal study, it was "accepted reference devices." There's no indication of human expert 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

• No MRMC study was done. The device (Cogent™ Hemodynamic Monitoring System) is a diagnostic computer for physiological parameters, not an imaging AI device where human readers interact directly with AI output. The document explicitly states: "No clinical performance testing was required to demonstrate device safety and effectiveness."

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

• Yes, standalone performance was assessed. The bench studies and the animal study evaluated the algorithms and hardware of the Cogent™ HMS independently against simulated data or accepted reference devices. The system is designed to compute and display parameters, not to provide an initial AI interpretation for a human to then review or modify.

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

• Bench Studies: Simulated physiological data.
• Animal Study: Measurements from "accepted reference devices" following pharmacological and ventilator interventions.

8. The sample size for the training set

• Since this is a 510(k) submission for a device primarily relying on established algorithms (continuous thermodilution, bolus thermodilution, oximetry, and PulseCO, which is also from a predicate), and not explicitly an AI/machine learning model that undergoes a distinct "training" phase with a large dataset, the concept of a "training set sample size" as typically understood for deep learning is not applicable or mentioned. The algorithms are described as being "measurement algorithms" or "analysis" algorithms, implying deterministic or well-defined computational processes rather than adaptive learning from a large training dataset.

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

• As noted above, a distinct "training set" with ground truth in the context of machine learning is not applicable to the description of this device. The algorithms are based on established physiological measurement principles. For the PulseCO algorithm, it is stated to be "the PulseCO algorithm which is used in the secondary predicate device," suggesting its design and validation are tied to previous work on that predicate, not a new training process described here.

Ask a specific question about this device