(104 days)
Hypotension Decision Assist is indicated to acquire, process and display arterial pressure and other key cardiovascular characteristics of adult patients who are at least eighteen years of age that are undergoing surgery where their arterial pressure is being continuously monitored by a vital-signs monitor. It is indicated for use to assist anesthesia healthcare professionals manage the blood pressure, hemodynamic stability and the cardiovascular system during such surgery.
Hypotension Decision Assist (HDA) is a clinical decision support Software as a Medical Device (SaMD) that is installed upon a medically-rated touch-screen computer. HDA connects to a multi-parameter patient monitor supplied by other manufacturers, from which it acquires vital signs data continuously including the arterial blood pressure waveform and cardiovascular-related numeric parameters.
HDA continually processes this data to display, in graphical charts and numeric format, vital signs data and derived variables including mean arterial pressure (MAP), heart rate, systolic and diastolic blood pressure, cardiac output and systemic vascular resistance. HDA compares MAP to user set targets to indicate when MAP is above or below the target range. It allows the user to mark the administration of vasopressors and volume challenges to the MAP trend.
The Hypotension Decision Assist (HDA-OR2) device, as described in the provided FDA 510(k) summary, is a clinical decision support software intended to assist healthcare professionals in managing blood pressure and cardiovascular stability during surgery.
Here's an analysis of the acceptance criteria and the study that proves the device meets them, based on the provided text:
1. Table of Acceptance Criteria and Reported Device Performance
The document does not explicitly present acceptance criteria in a quantitative table format with corresponding performance metrics for the HDA-OR2 device's intended use (i.e., assisting anesthesia healthcare professionals manage blood pressure, hemodynamic stability, and the cardiovascular system). Instead, the performance data focuses on system verification, measurement accuracy, artifact detection, and software validation.
The "Performance Data" section primarily details verification tests rather than clinical performance acceptance criteria directly related to the device's indications for use. The overall conclusion states that "the accuracy of its measurements is substantially equivalent to its predicate device and that HDA-OR2 performs as well as its predicate device."
Here's an interpretation of the performance data as it relates to implicit acceptance criteria:
| Acceptance Criterion (Implicit) | Reported Device Performance |
|---|---|
| Measurement Accuracy | Verified: Bench testing following IEC 60601-2-34 Edition 3.0 2011-05 demonstrated measurement accuracy across the intended use measuring range for each physiologic parameter (systolic blood pressure, diastolic blood pressure, MAP, heart rate, cardiac output, systemic vascular resistance) over serial and network connections. The device's measurement accuracy is substantially equivalent to its predicate device. |
| Artefact Detection | Verified: Bench testing over a network connection confirmed the device's capability to detect each signal artifact and anomaly that has the potential to impact its performance. |
| Power Interruption Tolerance | Verified: Bench testing in accordance with IEC 60601-2-34 Edition 3.0 2011-05 demonstrated that HDA can tolerate a sudden power interruption without loss of user-input or patient data, remaining in the correct operating mode, including when the battery is disconnected. |
| Software Verification & Validation (Moderate Level of Concern) | Completed: Performed and documented in accordance with FDA guidance for "Software Contained in Medical Devices" requirements, indicating that potential malfunction or latent design flaws would likely lead to 'Minor Injury'. No specific performance metrics (e.g., uptime, error rate) are provided, but the V&V process itself is the compliance metric. |
| Electrical Safety and EMC Compliance | Compliant: The supplied touch screen computer complies with FDA recognized standards ES60601-1 2005/(R) 2012 and A1:2012 for safety, and IEC60601-1-2:2014 for EMC. |
| Remote Update Reliability in Noisy Environments | Verified: Testing confirmed HDA's ability to receive remote updates in electromagnetically noisy environments (hospital installation site), responding as designed, not installing interrupted updates, and detecting/rejecting malware masquerading as legitimate updates. This implies a successful update rate or error handling robustness, though specific numbers are not given. |
| Functional Equivalence to Predicate | Affirmed: The device has the "same intended use and indications for use" and "same technological characteristics" as the predicate (HDA-OR1), with minor differences (battery, connectivity options, internet features) that were also subject to verification testing. The conclusion explicitly states it "performs as well as its predicate device." This is the core "acceptance" for 510(k) cleared devices - substantial equivalence. |
2. Sample Size Used for the Test Set and Data Provenance
The document describes bench testing for verification of the device's technical performance (measurement accuracy, artefact detection, power interruption, remote updates, software V&V, electrical safety/EMC).
- Test Set Sample Size: No specific "sample size" of patients or cases is mentioned for the performance data section, as the testing described is primarily technical and bench-level, not clinical. For example, for measurement accuracy, it mentions verification "across the intended use measuring range" and "over the serial and network connections available," implying a range of test conditions and inputs rather than a patient count.
- Data Provenance: The data provenance is not specified as clinical patient data (e.g., country of origin, retrospective/prospective). The described tests are laboratory/bench tests, not studies on patient data. The device acquires data from "multi-parameter patient monitor supplied by other manufacturers," but the testing described here relates to the device's capability to process that signal and interact with its environment, not its performance in specific patient scenarios or outcomes.
3. Number of Experts Used to Establish Ground Truth for the Test Set and Qualifications
- Not Applicable: The studies described are technical verification tests (measurement accuracy, artefact detection, electrical safety, software validation), not clinical studies requiring expert ground truth establishment from patient data for diagnostic or prognostic performance. The "ground truth" for these technical tests would be derived from calibrated instruments, known signal inputs, and established engineering standards.
4. Adjudication Method for the Test Set
- Not Applicable: As the described tests are technical verification rather than clinical performance studies using human experts evaluating patient cases, an adjudication method for a test set of clinical data is not mentioned or required.
5. Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study
- No MRMC Study was done: The document does not mention any MRMC comparative effectiveness study, or any study involving human readers/users comparing performance with and without AI assistance. The device is a "clinical decision support" tool, implying assistance, but no study is presented to quantify this assistance's effect on human performance. The 510(k) pathway for this device did not require such a study, as it demonstrated substantial equivalence primarily through technical performance and predicate comparison.
6. Standalone (i.e., algorithm only without human-in-the-loop performance) Study
- Standalone Performance Studied (Technically): The performance data presented (measurement accuracy, artifact detection, power interruption, remote updates) represents the standalone technical performance of the algorithm and hardware. The device "continually processes this data to display... vital signs data and derived variables." The verification tests confirm the accuracy and robustness of these processing and display functions.
- However, no clinical outcome study or diagnostic accuracy study (e.g., predicting hypotension with a specific accuracy) was performed in a standalone context. The device's role is to "assist" healthcare professionals, not autonomously diagnose or treat.
7. Type of Ground Truth Used
- Technical/Engineering Standards and Calibrated Inputs: The ground truth for the verification tests was established based on:
- IEC 60601-2-34 Edition 3.0 2011-05: For measurement accuracy verification using bench testing. This implies using known, precisely controlled electrical or physiological signals as inputs and comparing the device's output to these known inputs.
- Known Artefacts: For artifact detection, specific types of known signal aberrations were introduced to test the device's ability to detect them.
- Controlled Power Situations: For power interruption testing, the power supply was intentionally cut.
- Controlled Electromagnetic Environments: For remote update testing in noisy environments.
- Software Design Specifications and Requirements: For software verification and validation.
8. Sample Size for the Training Set
- Not applicable / Not disclosed: The document describes a "Software as a Medical Device (SaMD)" but does not specify if it employs machine learning or AI models that require a training set in the conventional sense. The device "aquires, processes and displays" data and "derives and displays" variables. This sounds more like rule-based or signal-processing software rather than a learned AI model, at least not one that would typically require a large training dataset for its core function of calculating and displaying vital signs. If there are any internal predictive or pattern-recognition algorithms, their training data and sample size are not mentioned.
9. How the Ground Truth for the Training Set Was Established
- Not applicable / Not disclosed: Since the existence and nature of a training set (in the context of machine learning) are not discussed, the method for establishing its ground truth is also not mentioned.
§ 870.1435 Single-function, preprogrammed diagnostic computer.
(a)
Identification. A single-function, preprogrammed diagnostic computer is a hard-wired computer that calculates a specific physiological or blood-flow parameter based on information obtained from one or more electrodes, transducers, or measuring devices.(b)
Classification. Class II (performance standards).