Validation
1) The first requirement for this device was validated by testing the device on human skins of varying tones and determining that the device correctly identified when the skin was inflamed at least 85% of the time. The final device prototype was placed on 6 individuals sampled from undergraduates in the TCNJ BME department. These individuals had varying ethnic backgrounds and skin colors so this validation testing did meet the criteria for varying skin tones. The device prototype was then used to take 10 images of the participants leg which were labelled as healthy images and then 10 images after treatment; a pressure cuff placed above the knee at 120 mmHg for 2 minutes, which were labelled as inflamed. The final result of our validation testing on this human data showed that the device was able to identify most of the inflamed skin images with a sensitivity of 90% and a AUC ROC score of 0.875 a large improvement relative to the AUC ROC score of 0.500 for a random guessing algorithm. These results validate that this device does pass the first requirement.
ROC curve for requirement 1 validation testing on human data.
Confusion matrix of validation testing results for human device trials.
2) Can take % O2 readings from the hospital’s oximeter and transmit through the bluetooth module.
3) Test it on a range of different sized people. This was passed as we tested the device on 6 individuals of different sizes. The device worked successfully and was able to change length accordingly.
4) The device will be powered on by the chosen battery pack for 12 continuous hours and all parts are functioning as intended. This was not tested but a calculation shows that the device would work for approximately 19 hours with all parts continuing to function as intended. This was calculated using the total power usage of the device and based on the power the battery supplied we calculated the duration of how long the device can be powered.
5) Evaluate redness after using the prototype for 12 hours. We did not test for 12 hours, but we believe the light would not induce inflammation or cause any redness to the skin. The device and its components should be able to work for 19 hours continuously while maintaining functionality.
6) The validation testing of this requirement was conducted at the same time as the validation testing for requirement 1. The SpO2 and erythema values for each device trial on a participant were recorded by the device and reviewed after testing to determine whether alarms were triggered correctly for detected threshold levels. After a review of the recorded threshold values the device was determined to have passed this validation test for requirement 6 as it only sounded an alarm on trials where the calculated DVT risk was above the set threshold values.