BME100 s2017:Group5 W1030AM L3

BME 100 Spring 2017	Home People Lab Write-Up 1 | Lab Write-Up 2 | Lab Write-Up 3 Lab Write-Up 4 | Lab Write-Up 5 | Lab Write-Up 6 Course Logistics For Instructors Photos Wiki Editing Help
OUR TEAM Name: Esther Sim Name: Marci Bandala Name: Emina Causevic Name: Neil Collins Name: Chase Frailey Name: Cooper Bertke LAB 3 WRITE-UP Descriptive Stats and Graph Mean Value Standard Deviation Error of mean Heart Rate (Gold Standard) 98.08977 23.03054 1.323069 Heart Rate (Spree) 98.9538 24.87754 1.371728 Temperature (Gold Standard) 96.64716 1.922602 0.1 Spree Temperature 95.53086 0.870378 0.05 Inferential Stats Pulse Ox/Spree Band (Pearson’s Coefficient (r)): 0.690806 Oral Thermometer/Spree Band (Pearson’s Coefficient (r)): .192798 p-value (Paired t-test): Pulse Ox vs Spree Band: .427116 - Conclusion: Not statistically relevant p-value (Paired t-test): Oral Thermometer vs Spree Band: 1.09676E-21 - Conclusion: Statistically relevant Design Flaws and Recommendations The standard deviation for both the gold standard for heart rate and Spree Band have extremely high deviations. This means that the accuracy of the readings are very low, whereas not much variance exists for the Spree Band’s body temperature monitor and the temperature monitor gold standard. Therefore, its accuracy is high. Since the R coefficient between the Pulse Ox and Spree Band was approximately 0.7, it shows that the Spree band has fairly accurate heart rate monitoring because the value is close to positive 1. However, the monitoring could be more related and still needs further testing. The R coefficient between the Oral Thermometer and Spree Band is extremely low, 0.2, and cannot be relied on to provide accurate readings because the correlation between the gold standard for temperature and the device is significantly low. A high p value means that the difference between the Spree Band and Pulse Ox is not significant, whereas the low p value between the Spree Band and Oral Thermometer means the difference significant. This means the Spree Band is very accurate for taking the heart rate. On the other hand there are significant differences between the oral thermometer and the Spree Band’s measurements meaning it is not particularly accurate for temperature. Overall, the Heart Rate monitor feature on the Spree Band functions well because it has a high positive correlation with no significant difference. The temperature monitor on the other hand is not reliable at all because it has a very low positive correlation and very significant difference. There are some noticeable design flaws in their experiment. First, there isn’t a large range of temperatures and heart rates since walking is a very mild exercise. The experiment should be run with walking as well as running and other more intense forms of exercise to ensure that the device works for relatively higher and lower body temperatures. Additionally, since the prototype is worn outside the body, tests should be run to ensure that the temperature and weather of the environment does not alter the efficacy of the prototype when worn while exercising outside, for example. Experimental Design of Own Device An experiment to test our own prototype would be to use the ear lobe glucose monitor that we have designed and compare it to an already verified Medtronic continuous glucose monitor (CGM), using a paired T-test. Because our device is mainly centered around the idea that we are able to measure glucose levels non-invasively, we will need experimentation to show that we are able to accurately measure glucose levels through the earlobe sensor on our prototype. Since both devices continuously monitor glucose levels, each participant will need to wear both monitors for the duration of the experiment (24 hours). We could then record data before and after each meal, before and after exercise, and when the participants wake up, go to sleep, and during sleep. This would demonstrate that our monitor is able to show differences in a body’s glucose levels after meals and after physical activity; moreover, this experiment would show that our prototype is capable of measuring glucose levels when they’re both high or low. Our sample size would be around 50 people, and we would make sure to limit our sampling to individuals who need constant glucose monitoring and have diabetes. The sampling would need to be diverse with respect to age and race so that we ensure our device is consistent across different demographics. We would also run inferential statistics to make inferences about our population sample. We would use our smaller sample size and use it to make generalizations about the population of people with diabetes that need constant glucose monitoring. We would graph the results of the glucose levels measured by the Medtronic CGM against the results of our prototype. If the graphs show similar results and show a positive p-value, then we can conclude that our prototype will yield accurate results. We would similarly calculate the error and statistical significance of our results to also show the margin of accuracy, which should also yield close to a positive R value of 1.