Search Results

Surgical gown are devices that are intended to be worn by operating room personnel during surgical procedures to protect both the surgical patient and the operating room personnel from transfer of microorganisms, body fluids, and particulate material.

The surgical gown is intended to be worn by operating room personnel during surgical procedures to protect both the surgical patient and the operating room personnel from transfer of microorganisms, body fluids, and particulate material. It is made of soft, air permeable SMS non-woven fabric. The Jianerkang Surgical Gown is made of a laminate with adhesive taped seams and have a hook and loop closure at the back of the neck and a waist tie feature to secure the gown to the body of the user. The sleeves of the gown have knit cuffs sewn onto the end of the sleeve at the user's wrists to keep the sleeves in place on the wearer. The entire gown including the gown sleeves are made of the same material and utilize the same manufacturing processes.

The provided text is a 510(k) summary for a medical device called "Surgical Gown." It details the product, its intended use, and its comparison to a predicate device to demonstrate substantial equivalence for FDA clearance.

However, the questions you've asked are typically relevant to the performance evaluation of AI/ML-enabled medical devices, specifically regarding their accuracy and how they improve human performance. The document describes a traditional medical device (a surgical gown) and focuses on its physical properties, material performance, and biocompatibility, not on AI/ML algorithm performance.

Therefore, I cannot provide answers to your specific questions based on the provided text, as it does not contain information about:

• Acceptance criteria for an AI/ML algorithm's performance
• Study data for an AI/ML algorithm
• Sample sizes for test or training sets for AI/ML
• Ground truth establishment by experts for AI/ML
• MRMC studies for AI/ML
• Standalone AI algorithm performance

The document states that "Performance testing was conducted on the Surgical Gown. All of the tested parameters met the predefined acceptance criteria," and then lists those criteria in a table for properties like water resistance, tearing strength, tensile strength, and seam strength. This is a physical product, not a diagnostic or prognostic AI/ML device.

To answer your request, if this were an AI/ML device, the structure of the answer below demonstrates what would be needed, but the actual values and details cannot be extracted from the provided text.

Assuming the Request Pertains to a Hypothetical AI/ML-enabled Medical Device based on the Questions Provided:

Given the nature of the provided document, which describes a surgical gown (a physical medical device), it does not contain the kind of information requested regarding acceptance criteria and studies for an AI/ML-enabled software device. The acceptance criteria and performance studies described in the document relate to the physical properties and barrier performance of the gown itself, not to an algorithm's diagnostic or analytical capabilities.

Therefore, for the type of device described in the document, none of the following information is applicable or available:

• A table of acceptance criteria and reported device performance for an AI/ML algorithm
• Sample sizes used for AI/ML test sets or training sets
• Data provenance for AI/ML
• Number of experts used to establish ground truth for AI/ML
• Adjudication method for AI/ML
• MRMC comparative effectiveness study results for AI/ML
• Standalone AI algorithm performance
• Type of ground truth used for AI/ML (e.g., expert consensus, pathology, outcomes data)
• How ground truth for AI/ML training was established

However, if we were to reinterpret your request as an example of what such a description would entail for an AI/ML device, even though not present in this document, here's a hypothetical structure that addresses your points:

Hypothetical Description for an AI/ML Medical Device (Not Applicable to the Provided Document)

This section describes the hypothetical acceptance criteria and the study conducted to prove the AI-enabled device meets these criteria, as if the "Surgical Gown" were an AI/ML diagnostic tool.

1. Table of Acceptance Criteria and Reported Device Performance (Hypothetical AI/ML Device)

2. Sample Sizes and Data Provenance (Hypothetical AI/ML Device)

• Test Set Sample Size: N = 1,000 cases (e.g., medical images, patient records).
• Data Provenance: Retrospective and prospective data collected from multiple sites across the United States, Europe (e.g., Germany, UK), and Asia (e.g., Japan, South Korea). Data was collected over a period of 5 years (2018-2023).

3. Number of Experts and Qualifications for Ground Truth (Hypothetical AI/ML Device)

• Number of Experts: A panel of 5 board-certified medical specialists (e.g., Radiologists, Pathologists) with varying levels of experience.
• Qualifications:
  • 3 Senior Experts: Each with >10 years of experience in the relevant subspecialty (e.g., thoracic radiology, dermatopathology).
  • 2 Junior Experts: Each with 3-5 years of experience in the relevant subspecialty.
  • All experts were blinded to the device's output during ground truth establishment.

4. Adjudication Method for the Test Set (Hypothetical AI/ML Device)

• Method: A "3+1" consensus method was used.
  • Each case was independently reviewed by three out of the five experts.
  • If at least two out of the three experts agreed on a label, that was considered the preliminary ground truth.
  • In cases of disagreement (i.e., no two out of three experts agreed, or a 1-1-1 split), a fourth senior expert (the "plus 1") was brought in to review the case and resolve the discrepancy, establishing the final ground truth.

5. Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study (Hypothetical AI/ML Device)

• Was an MRMC study done? Yes.
• Effect Size of Human Reader Improvement: The MRMC study demonstrated a statistically significant improvement in human reader performance when assisted by the AI device.
  • Effect Size: Human readers improved their mean sensitivity by 7.2% (absolute change) and mean specificity by 4.5% (absolute change) when using the AI assistance compared to reading without AI assistance.
  • This translated to a statistically significant increase in AUC for human readers with AI assistance (e.g., AUC increased from 0.85 to 0.91, p < 0.001).

6. Standalone Performance (i.e., algorithm only without human-in-the-loop) (Hypothetical AI/ML Device)

• Was standalone performance done? Yes.
• Performance: The standalone algorithm achieved a sensitivity of 92.5%, specificity of 83.2%, and an AUC of 0.93 on the independent test set. These metrics met the predefined acceptance criteria for standalone performance.

7. Type of Ground Truth Used (Hypothetical AI/ML Device)

• Type: Expert consensus (as detailed in point 3 and 4) was the primary method for establishing ground truth.
• For certain conditions, (e.g., malignancy), confirmation was sought from pathology reports or long-term clinical outcomes data where available, serving as the gold standard or confirmatory reference. All ground truth labels were meticulously documented and linked to the source.

8. Sample Size for the Training Set (Hypothetical AI/ML Device)

• Sample Size: N = 10,000 cases (e.g., medical images). This large dataset allowed for robust model training and generalization.

9. How Ground Truth for the Training Set Was Established (Hypothetical AI/ML Device)

• Ground truth for the training set was established through a combination of methods:
  • Automated Labeling: For a large portion of the data, ground truth was derived from structured clinical reports, electronic health records (EHR), and existing pathology results. Natural Language Processing (NLP) models were used to extract relevant information and assign preliminary labels.
  • Expert Review (Subset): A statistically significant subset (e.g., 20%) of the automatically labeled training data, particularly complex or uncertain cases, underwent manual review by experienced clinicians (e.g., 2 specialists per case) to validate and correct labels. Discrepancies were resolved through a consensus meeting.
  • Active Learning: An active learning strategy was employed to prioritize cases for expert review that the model found most challenging or ambiguous, ensuring that expert annotation effort was focused on maximizing model learning.

Ask a specific question about this device

The medical face mask is intended to be worn to protect both the patient and healthcare personnel from transfer of microorganisms, body fluids, and particulate materials in infection control practices to reduce the potential exposure to blood and body fluids. It is for single-use and provided non-sterile.

The medical face mask is designed and manufactured by Jiangsu Province Jianerkang Medical Dressing Co., Ltd. It is non-sterile and for single use. The medical face mask is manufactured with three-layers, the inner and outer layers are made of polypropylene, and the middle layer is made of melt blown polypropylene. The elastic ear loop of proposed device is made of spandex and nylon, not made with natural rubber latex. The nose piece contained in the proposed device allows the user to fit the face mask around their nose, which is made from steel wire. It is a self-inhalation filter mask, which works by filtering the air containing microorganisms, body fluids, and particulate matters through the filter material of the mask before being inhaled or exhaled. The product is level 1 according to ASTM F2100-19.

The document describes the acceptance criteria and performance data for a Medical Face Mask (K202061).

1. Table of Acceptance Criteria and Reported Device Performance:

Note: While the device is classified as Level 1 according to ASTM F2100-19 in the device description, the performance data often reports results that meet Level 3 criteria, indicating a higher performance than the minimum required for Level 1 for those specific tests.

2. Sample size used for the test set and the data provenance:

• The document does not explicitly state the sample size used for each performance or biocompatibility test.
• The data provenance is from Jiangsu Province Jianerkang Medical Dressing Co., Ltd. in China. The data is retrospective, collected during the device's development and testing to support the 510(k) submission.

3. Number of experts used to establish the ground truth for the test set and the qualifications of those experts:

• This information is not applicable as the document describes performance testing for a medical face mask, not an AI/ML-based diagnostic device requiring expert interpretation for ground truth. The acceptance criteria are based on established international and national standards (ASTM, ISO, CFR).

4. Adjudication method for the test set:

• This information is not applicable for the same reason as point 3. Testing is based on objective, standardized laboratory methods rather than subjective 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:

• This information is not applicable. The device is a medical face mask, 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. The device is a medical face mask, not an AI algorithm.

7. The type of ground truth used:

• The "ground truth" for the device's performance is established by adherence to internationally recognized standards and test methods. For example, Flammability is tested per ASTM F2100-19 and 16 CFR Part 1610-2008. BFE is tested per ASTM F2101-2019. Biocompatibility is tested per ISO 10993 series. These standards define the objective measurements and criteria for performance.

8. The sample size for the training set:

• This information is not applicable. The device is a physical medical face mask, not a machine learning model that requires a training set.

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

• This information is not applicable for the same reason as point 8.

Ask a specific question about this device