Search Results

The YJ-010B series is a mechanical wheelchair which is a manually operated device with wheels that is intended for medical purposes to provide mobility to persons restricted to a sitting position. It can be folded for transport by bringing the two sides together. The manual wheelchair incorporates a main frame, a seat, two adjustable footplates and four wheels. The larger rear wheels have hand rims of slightly smaller diameter projecting just beyond the tire. These allows the user to maneuver the chair by pushing them on without requiring them to grasp the tires. The manual wheelchairs have brakes that bear on the tires of the rear wheels and two push handles at the upper rear of the frame to allow for manual propulsion by an assistant.

AI/ML Overview

The provided document is a 510(k) premarket notification for a Bariatric Heavy Duty Wheelchair. It does not describe an AI medical device or a study involving AI assistance or human readers. Therefore, I cannot extract information related to acceptance criteria and study proving a device meets those criteria for an AI-powered device, as the document pertains to a mechanical wheelchair.

The document discusses:

Acceptance criteria (implicitly): The device (a manual wheelchair) needs to demonstrate substantial equivalence to a predicate device, primarily through non-clinical performance testing and biocompatibility assessments, adhering to various ISO standards.
Device performance: The document lists various specifications of the proposed wheelchair (e.g., dimensions, weight capacity, stability angles, brake effectiveness) and states that these comply with relevant ISO standards.
Ground Truth: For a mechanical device like a wheelchair, "ground truth" would relate to its physical properties and performance metrics as measured according to established engineering and safety standards (e.g., ISO 7176 series).
Sample Size: Not applicable in the context of an AI test set. This refers to physical samples of the wheelchair models being tested.
Experts/Adjudication/MRMC/Standalone: These concepts are not applicable to the evaluation of a mechanical wheelchair. There is no AI algorithm being evaluated.
There is no clinical study included in this submission.

To answer your request, here's a blank template for an AI medical device study, illustrating the elements you've asked for, as this document does not contain the necessary information.

Acceptance Criteria and Study for an AI Medical Device

Study Description:
(This section would be filled if the document were about an AI device. As it is not, this remains a template.)

1. Table of Acceptance Criteria and Reported Device Performance

Acceptance Criteria Category	Specific Metric	Threshold for Acceptance	Reported Device Performance	Meets Criteria?
Clinical Performance	Sensitivity	≥ 90%	[Reported Sensitivity]%	[Yes/No]
	Specificity	≥ 80%	[Reported Specificity]%	[Yes/No]
	AUC (ROC)	≥ 0.85	[Reported AUC]	[Yes/No]
Technical Performance	Inference Time	≤ 2 seconds/image	[Reported Time]	[Yes/No]
	Usability Score	> 70 (SUS)	[Reported SUS Score]	[Yes/No]
Safety	False Positive Rate	As low as reasonably achievable, not impacting patient management negatively	[Reported FPR]%	[Yes/No]
	False Negative Rate	As low as reasonably achievable, not impacting patient management negatively	[Reported FNR]%	[Yes/No]

Note: The above table is a placeholder for a hypothetical AI device study and cannot be populated from the provided document, which describes a mechanical wheelchair.

2. Sample Size Used for the Test Set and Data Provenance

Test Set Sample Size: [Number of cases/patients/images]
Data Provenance: [e.g., Retrospective or Prospective; Country of origin (e.g., Multi-site from US, Europe, Asia)]

3. Number of Experts Used to Establish Ground Truth for the Test Set and Qualifications

Number of Experts: [ e.g., 3]
Qualifications of Experts: [e.g., Board-certified Radiologists with >10 years of experience in [specific domain, e.g., chest imaging, mammography]; Certified Pathologists with >5 years of experience]

4. Adjudication Method for the Test Set

Adjudication Method: [e.g., 2+1 (two initial readers, third independent reader for discordance); 3+1 (three initial readers, fourth independent reader for discordance if majority not reached); Consensus meeting; None (if single reader ground truth)]

5. If a Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study Was Done, What Was the Effect Size of How Much Human Readers Improve with AI vs. Without AI Assistance?

MRMC Study Conducted?: [Yes/No]
Effect Size (e.g., improvement in AUC or F-measure for human readers with AI assistance vs. without AI): [e.g., Human readers' AUC improved from X to Y (an absolute increase of Z); F-measure increased by Z%]

6. If a Standalone (i.e., algorithm only without human-in-the-loop performance) Was Done

Standalone Performance Study Conducted?: [Yes/No]
Key Performance Metrics (if Yes): [e.g., Sensitivity: X%, Specificity: Y%, AUC: Z]

7. The Type of Ground Truth Used

Type of Ground Truth: [e.g., Expert consensus (radiologist readings); Pathological diagnosis (biopsy/histology); Clinical outcomes data (e.g., hospital discharge codes, follow-up imaging); Longitudinal follow-up data]

8. The Sample Size for the Training Set

Training Set Sample Size: [Number of cases/patients/images]

9. How the Ground Truth for the Training Set Was Established

Ground Truth Establishment for Training Set: [e.g., Clinical reports by attending physicians; Single expert review; Automated label extraction; A subset might be expert-reviewed, while the majority are from routine clinical practice.]

Ask a Question

Ask a specific question about this device

Page 1 of 1