(84 days)
The Willow Wearable Breast Pump 2.0 is intended to express milk from lactating women in order to collect milk from their breasts. The device is intended for a single user.
The Willow Wearable Breast Pump 2.0 (Willow) is a small electric breast pump that is intended for lactating women to express and collect breast milk. Willow may be operated as a single or double pumping system. For a user to pump both breasts simultaneously, she would need to use two Willow devices at the same time, one on each breast. Willow is for use by a single user only.
Willow is a battery-powered electro-mechanical device that contains software. The device includes the following main components: Pump, milk collection component (milk bag or container), and a charger.
All milk contacting components of the device are compliant with 21 CFR 174-179.
This submission was for clearance of modifications made to the predicate device, including addition of a milk container for milk collection, updates to device firmware, inclusion of Bluetooth for connection to an optional mobile app, and a change in the flange attachment method.
The provided text is a 510(k) summary for the Willow Wearable Breast Pump 2.0. It describes the device, its intended use, and compares it to a predicate device to demonstrate substantial equivalence. However, it does not contain the specific information requested in your prompt regarding acceptance criteria, a study proving the device meets those criteria, or details about test set size, ground truth establishment, or training set size typically found in AI/ML device submissions.
The breast pump is a physical medical device, not an AI/ML-driven diagnostic or assistive device that would typically have the kind of performance metrics and study details you've asked for (e.g., effect size of human readers with AI assistance, standalone algorithm performance, expert consensus for ground truth). The document focuses on showing that the modifications to the device (e.g., addition of a milk container, firmware updates, Bluetooth connectivity) do not raise new questions of safety or effectiveness compared to the predicate device.
Here's a breakdown of why I cannot answer your specific questions with the provided text:
-
Acceptance Criteria and Reported Device Performance (Table): The document lists various bench testing (vacuum profile, leakage, flow rate, mechanical stress), electrical safety, EMC, biocompatibility, software, and cybersecurity testing performed. While these tests have internal acceptance criteria (e.g., device operation maintained, within specified ranges for vacuum), the specific numerical acceptance criteria and reported performance values are not explicitly stated in this summary. It only states that these tests were performed to assess modifications and support substantial equivalence.
-
Sample size for the test set and data provenance: Not applicable in the context of a breast pump's 510(k) submission as described. There isn't a "test set" in the sense of a dataset for an AI/ML algorithm. The performance data refers to physical device testing.
-
Number of experts and qualifications for ground truth: Not applicable. Ground truth for a breast pump is functionality, safety, and performance as a physical device, based on engineering standards and regulatory requirements, not expert interpretation of outputs.
-
Adjudication method: Not applicable.
-
Multi-reader multi-case (MRMC) comparative effectiveness study: Not applicable. This type of study is for evaluating observer performance with and without an AI/ML aid, which is not relevant for a breast pump.
-
Standalone (algorithm only without human-in-the-loop performance): Not applicable. The device is a physical breast pump; it does not have a standalone algorithm in the sense of a diagnostic or assistive AI.
-
Type of ground truth used: For the physical device, it's based on engineering specifications, physical measurements, and compliance with standards (e.g., IEC 60601 for electrical safety, ISO 10993 for biocompatibility), not expert consensus, pathology, or outcomes data in the context of disease diagnosis.
-
Sample size for the training set: Not applicable. This refers to AI/ML model training data.
-
How the ground truth for the training set was established: Not applicable.
In summary, the provided document is a regulatory submission for a physical medical device (breast pump) and therefore does not contain the information requested about AI/ML algorithm performance, ground truth establishment for datasets, or comparative effectiveness studies involving human readers, which are typically found in submissions for AI/ML-driven software as a medical device (SaMD) or AI-assisted devices.
§ 884.5160 Powered breast pump.
(a)
Identification. A powered breast pump in an electrically powered suction device used to express milk from the breast.(b)
Classification. Class II (performance standards).