Search Results

DCM is a small worn activity monitor designed for documenting physical movement associated with applications in physiological monitoring.

The device is intended to monitor the activity associated with movement during sleep.

DCM can be used to analyze circadian rhythms and assess activity in any instance where quantifiable analysis of physical motion is desirable.

DCM is indicated for monitoring of adult patients only.

DCM is a wrist-worn wearable device intended to continuously record high resolution digital acceleration data associated with a patient's physical movement.
In practice, a healthcare professional or researcher can prescribe the device to collect physiological data from patients during sleep and in applications where quantifiable analysis of physical motion is desirable.
The device is set up to collect data by the healthcare professional then placed on the subject's wrist. The device is designed to be worn during normal activities and/or during sleep over a period of days to weeks. The patient does not need to interact with the device to control data collection.
The data stored on the device can be transmitted to the cloud for storage, and made accessible to healthcare professionals or researchers for further analysis. Downloaded data can be post-processed based on the timestamp and magnitude of acceleration along each axis.
The DCM system comprises a system of components:

• wearable biosensor (PW010)
• off the shelf mobile device (PW030) running the DCM mobile app (PW400)
• cloud-based data storage and data processing (PW100) (back-end)
• investigator dashboard (PW500) accessed through a web browser (front-end)

The provided FDA 510(k) clearance letter for the DCM (PW-DCM) device does not describe a study involving a test set, ground truth experts, or human readers for assessing device performance related to diagnostic accuracy or interpretation.

Instead, the document focuses on the technical performance of the device as a physical activity monitor, comparing it to a predicate device (Actigraph LEAP) primarily on its physical and operational characteristics. The acceptance criteria and "study" described are more akin to verification and validation (V&V) testing of hardware and software components, rather than a clinical performance study measuring accuracy against a diagnostic gold standard involving human interpretation.

Therefore, many of the requested categories (e.g., number of experts, adjudication method, MRMC study, effect size on human readers, type of ground truth for diagnostic accuracy) are not applicable or cannot be extracted from this document, as the device's function is data collection and not direct diagnostic interpretation.

However, I can extract the information that is present and explain why other information is not available from this document.

Acceptance Criteria and Reported Device Performance

The table below summarizes the technical acceptance criteria for the DCM device and the reported outcomes, as found in the "Summary of Testing" section.

Study Details (based on available information)

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

   • Test set sample size: Not explicitly stated for each test. The tests described are bench tests ("Bench testing with the biosensor in a range of orientations," "Bench testing with mobile app paired to biosensor," "manual interruption and restoration of connectivity"). This implies testing of device units, not a patient cohort.
   • Data provenance: Not explicitly stated. Given the nature of the tests (bench testing, design validation), the "data" being generated is measurement data from the device itself rather than clinical patient data. The document does not refer to geographical origin or patient type for these validation tests.
   • Retrospective or Prospective: Not applicable in the context of device design verification and validation testing. These are controlled engineering tests.

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

   • Not applicable for these types of tests. The "ground truth" for these engineering and software tests would be established by calibrated measurement equipment (e.g., accelerometers for accuracy, timing devices for accuracy) and adherence to international standards (e.g., IEC 62366-1 for usability, IEC 60601 series for safety, IEC 62304 for software). There is no mention of human experts interpreting data to establish a ground truth for diagnostic purposes because the device's function is data collection, not interpretation.

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

   • None. This concept is for clinical performance studies where multiple human readers interpret medical images or data. The described tests are technical performance evaluations.

4. 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, an MRMC comparative effectiveness study was not done. The document explicitly states: "DCM did not require clinical studies to support substantial equivalence to the predicate device." The device is a "small worn activity monitor designed for documenting physical movement," not a device that provides AI-assisted interpretations for human clinicians.

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

   • Yes, a form of standalone testing was done for the technical performance. The "Summary of Testing" section describes tests where the device's inherent capabilities (e.g., acceleration measurement, timing accuracy, data storage) were evaluated against predetermined engineering criteria. This is performance of the algorithm/device itself, without human interpretation in the loop beyond setting up the test and interpreting the test results (e.g., "PASS").

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

   • Technical/Engineering Standards and Calibrated Equipment: For accuracy measurements, the ground truth would be from highly accurate, calibrated reference instruments. For safety, EMC, and software, the ground truth is adherence to established international standards (e.g., IEC 60601-1, IEC 62304) and internal design specifications. There is no biological or clinical "ground truth" (e.g., pathology, outcomes data, expert consensus on patient diagnosis) applied here.

7. The sample size for the training set:

   • Not applicable / Not disclosed. The document does not describe a machine learning algorithm that requires a "training set" in the context of clinical AI. The device collects raw acceleration data. While there might be internal algorithms for processing this data (e.g., activity counts, sleep/wake detection, circadian rhythm analysis from raw data), the document describes validation of the data collection capability, not the performance of an AI model trained on a specific dataset for diagnostic tasks.

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

   • Not applicable. As no training set for a clinical AI algorithm is described, there's no ground truth establishment for such a set.

Ask a specific question about this device