BME100 f2017:Group1 W1030 L3
|BME 100 Fall 2017|| Home |
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
Wiki Editing Help
LAB 3 WRITE-UP
Descriptive Stats and Graph
Given the readings for the Spree Band, pulse ox and oral thermometer, the following measures of central tendency and variability were calculated:
Spree Band - Heart Rate
Mean: 98.95379538 BPM
Standard Deviation: 24.87753802 BPM
Spree Band - Temperature
Mean: 95.53250774 degrees
Standard Deviation: 0.8712250758 degrees
Gold Standard - Pulse ox
Mean: 98.08976898 BPM
Standard Deviation: 23.03054395 BPM
Gold Standard - Oral Thermometer
Mean: 96.63770898 degrees
Standard Deviation: 1.920911488 degrees
Pearson's Correlation Coefficient (r) = 0.6908064894
Pearson's Correlation Coefficient (r) = 0.1745355645
Using the means and standard deviations for 1) the Spree Sensor for both heart rate and temperature, 2) Pulse ox (gold standard) and 3) oral thermometer (gold standard), a paired t-test to compare both the Spree Sensor to the pulse ox and the Spree Sensor to the oral thermometer was run and interpreted below.
Spree Sensor vs. Pulse Ox
paired t-test value: p = 0.427116
Spree Sensor vs. Oral Thermometer
paired t-test value: p = 4.98*10^-21 (<<<0)
When comparing the Spree Sensor to the pulse ox, it can be stated that there is not enough evidence to reject the null hypothesis, where the difference of the means of the Spree Sensor and pulse ox is equal to zero. This is because the p-value calculated by the paired t-test is much greater than 0.05, and that indicates that there is a high probability of no difference between the two devices. Therefore, the Spree Sensor adequately measures heart rate with very similar accuracy to that of the pulse ox (gold standard). This strong relationship between the two devices and their measurements at the same time intervals is verified with the correlation coefficient r=0.6908064894, which indicates a fairly strong, positive correlation between the measurements for the Spree Sensor and the measurements for the oral thermometer (gold standard).
When comparing the Spree Sensor to the oral thermometer, the p-value obtained from the paired t-test was 4.98*10^-21, which is very very small and less than 0.05. Therefore, there is an extremely low probability that the difference between the means of temperature data calculated from both the Spree Band and the oral thermometer is equal to zero; there is sufficient evidence to reject the null hypothesis that this difference is equal to zero. From this paired t-test, it was determined that there is a significant difference in temperature readings from the two devices. This inconsistency between the two devices is also seen by a very small correlation coefficient r= 0.1745355645, indicating a weak relationship between the measurements of both devices at the same time intervals.
Design Flaws and Recommendations
Every design has flaws. Our design has several flaws that may impact its success in the market. First, due to the advanced nature of the sensing devices such as ours, we decided to make the sensors irremovable from the bands and insole. This leads to the customer having to make the decision on what color they want their device to be before purchase. Once the purchase is made, the color cannot be changed without purchasing new sensors. Basically leading the customer into buying a whole new device. We understand that this is not ideal, but feel it is necessary to make sure that the quality of our device holds firm throughout its use. Another design flaw that can impede mass production is the fact that these devices are tailored to the customer's foot. Feet tend to come in all sorts of shapes and sizes, so each device’s insole will have to be customized to fit each user. This shouldn’t be too much of a problem though, as shoe companies face this same problem and easily overcome it. The manufacturer of this device will just have to adopt the standard way shoe companies measure a person's foot. Lastly, it was difficult finding the perfect force sensor to go into the devices insole. The sensors would either be too small and couldn’t handle a human's typical weight, or the sensors would be a tad to big and be able to measure upwards of 1000 lbs. To fix this we would need to contact a company that specializes in making custom force sensors such as Flintec. They would make our ideal sensor and then give us an estimate of price per unit. Other than these flaws, we see no other aspect that is holding back our device from its full potential.
Experimental Design of Own Device
The goal of performing analysis on inferential statistics with our device is to determine its accuracy by comparing it to the gold standard. In our case, the gold standard for heart rate calculations, tools to measure weight, calorie counters and step counters are the pulse ox, a standing scale, a calorie watch and a perambulator, respectively.
All obese patients in the U.S., all ages and ethnicities
We will be randomly selecting 500 obese patients from across the country; random selection will take place from the top 100 obesity clinics, with five patients selected from each. Age, ethnicity, location, weight, and any other factors will not be considered in selection of patients to avoid selection bias.
5 heart rate readings taken in the lab (one every two minutes) using our device and the pulse ox (gold standard) at the same time. Walked around campus and take one reading every five minutes for an hour using both our device and the pulse ox (gold standard). Returned to lab and took 5 heart rate readings, one every two minutes using both the pulse ox (gold standard) and our device.
Take weight measurements once in lab, once around campus, and again in the lab with both the device and a standing scale (gold standard).
Steps taken and distance:
Use click wheel/surveyor's wheel/perambulator (gold standard) in comparison with the device both in the lab and around campus.
In the lab, the study participant uses a treadmill for 20 minutes (10 minutes walking, 10 minutes running) taking 10 measurements (one every two minutes) while wearing the device and the calorie watch (gold standard). Walk around campus for an hour and take measurements every five minutes while wearing the calorie watch (gold standard) and our own device to determine the accuracy of our device.. Return to the lab and run treadmill test again while walking 10 min and running 10 min, wearing device on a treadmill to determine if there was any type of difference.
For all four different types of measurements- heart rate, weight, steps taken/ distance and calorie counting- measurements will be taken and recorded prior to the campus walk in order to establish a baseline value for each. Then, measurements are to be taken during the campus walk at various intervals (noted above) in order to compile a set of data that can be used to compare our device to the gold standard. Measurements taken after the walk can verify the capability of our device and will reflect this when the customer is not in motion. Thus, all types of motion are represented in this procedure, with measurements at the correct interval size to accurately compare the device to the gold standard. A paired t-test will be used to make this comparison and verify whether there is a significant difference in measurements between our device and the gold standard.