Catalog of Regulatory Science Tools to Help Assess New Medical Devices
This tool is a method to estimate the likelihood of wireless coexistence of a wireless medical device in its intended use environment using two sources of information:
- wireless coexistence testing of the device and
- wireless spectrum surveys of the use environment.
First, the outcomes of wireless coexistence testing—as specified in ANSI C63.27:2017 standard for evaluation of wireless coexistence—are used to derive a logistic regression model describing the probability that the device successfully performs its wireless functionality under a given set of wireless coexistence testing conditions. The least absolute shrinkage and selection operator (LASSO) algorithm is used to select a reduced set of explanatory variables for the model.
The method then specifies the integration of realistic spectrum survey data to estimate the likelihood of wireless coexistence in a specific use environment. Surveying the electromagnetic spectrum used by the wireless medical device (for example, 2.4 GHz industrial, scientific, and medical band) for communication in a use environment can be done by monitoring the power flux spectral density in the environment and recording the received power to calculate the channel utilization in the spectrum band of interest.
The method is intended to model the likelihood of wireless coexistence representing the probability that a wireless medical device performs its wireless functionality under a given set of coexistence testing conditions, which can be used in simulation to facilitate the estimation of the device wireless performance in an intended use environment.
Wireless-enabled medical devices operating in shared wireless environments with other systems, for example devices using Wi-Fi, Bluetooth, ZigBee, or other technologies that operate in unlicensed spectrum bands in a hospital or home environment, are in the scope of this tool.
The method can be used by medical device developers and testing labs to complement their evaluation of wireless coexistence and inform the medical device risk management with an estimate of the likelihood of wireless coexistence.
The method was demonstrated through an experimental realization of the radiated open environment coexistence test for a device under test operating ZigBee and realistic Wi-Fi sources of unintended signals.
Al Kalaa MO, Seidman SJ, Refai HH. Estimating the Likelihood of Wireless Coexistence Using Logistic Regression: Emphasis on Medical Devices. IEEE Trans Electromagnetic Compat. 2018 Oct;60(5):1546-1554. doi: 10.1109/temc.2017.2777179. PMID: 36248761; PMCID: PMC9558297.
The method is limited by the availability of wireless coexistence test data for the subject device and spectrum measurement data of the intended use environment.
The method integrates device performance data obtained experimentally for a specific device functionality. Accordingly, changes to how the device function is specified should be addressed with new experimental data.
The method does not establish or imply acceptable medical device wireless coexistence risk.
Table of Contents/Supporting Documentation
- Full description of the Logistic regression model specification including the selection of a reduced set of explanatory variables, fitting, and testing are provided in a peer-reviewed publication. Also described is the use of Monte Carlo simulation with realistic spectrum survey data to estimate the likelihood of wireless coexistence in a hospital environment.
- Example spectrum survey to characterize the use environment in a hospital is reported in a peer-reviewed publication. This includes the detailed coefficients of a generalized extreme value (GEV) distribution that fits the channel utilization observed for Wi-Fi channels 1, 6, and 11 in the surveyed hospital.
In addition to citing relevant publications please reference the use of this tool using DOI: 10.5281/zenodo.7883662
For more information:
Note: This tool was previously listed in the catalog as “Estimating the likelihood of medical device wireless coexistence using logistic regression.”