- Количество слайдов: 1
Comparing the accuracy of BMI, skin fold measurements of body type, and waist circumference to bioelectrical impedance within individuals. By: Mallory Hoff Advisor: Dr. Hancock Introduction Conclusions Results This project explores four different simple measurements for determining body fat within an individual. An accurate measurement is needed to aid in lowering the overwhelming number of people who are overweight in America. The Center for Disease Control and Prevention (2011) estimates that 18. 7 -32% of America’s population is obese. Determining the most accurate measurement can help in the diagnosis, prevention, and management of obesity. BMI, or body mass index, is being used more frequently as an accurate indicator of body type. In fact BMI has been declared the principle measurement for determining obesity for many organizations such as the World Health Organization (Neovius, 2005, 164). Skin fold calipers are instruments used to measure subcutaneous tissue. Skinfold calipers are thought to be a simple but accurate method for determining body fat percentage (Ball, et. al. , 2006, 257). Waist circumference is probably the simplest type of body fat measurement. The control method in this project is bioelectrical impedance. This is done using a low-level electrical current and measuring resistance to this electrical current within the tissues (Hong et. al. , 2009, 1266). The question is, are these simple measurements capable of accurately predicting a persons body type? This project examines the ability of BMI, waist circumference, and skin fold calipers to classify a person as normal or overweight compared to what bioelectrical impedance has classified the person. My hypothesis is that BMI and waist circumference will be the worst predictors of body type because they are such simple tests. I also hypothesized that skin fold calipers will be the best predictor of body type because it is an actual test of body fat on the individual. Table 1: The number of true negatives (TN), false negatives (FN), true positives (TP), and false positives (FP). BMI WC SF F BMI WC SF M M M TN 18 20 14 20 20 18 TP 15 12 18 16 4 5 7 1 4 16 7 FP 1 0 6 0 0 and false positive rate (FPR) 13 FN Table 2: The true positive rate (TPR) BMI WC SF F BMI WC M M SF M 2 TPR 0. 75 0. 63 0. 95 0. 8 0. 11 0. 65 FPR 0. 05 0 0. 3 0. 1 In conclusion, my hypothesis was not completely supported. I had hypothesized that BMI would be the least accurate measurement. BMI turned out to be the best predictor in both sexes. I also hypothesized that skinfold would be the best predictor. It turned out to be the 2 nd best predictor instead of the best. However, I was correct in my hypothesis that waist circumference would be inaccurate. This measurement turned out to be the worst predictor. It is interesting that BMI was so accurate when it is so simple and only takes into account weight and height. My findings support why previous research stated that BMI is the principle measurement by many professional organizations such as the World Health Organization (Neovius, 2005, 164). It is also noteworthy that waist circumference in males was significantly worse than females. One would expect the opposite due to the curviness of the female body in the waist and hips. It would be interesting to further this research by comparing the different measurements to each other instead of the control of bioelectrical impedance. It would also be interesting to calculate muscle mass of the participants and see how this altered the accuracy of the measurement. It is thought that BMI is less accurate for individuals with higher muscle mass (Romero-Corral et al. , 2008, 964). BMI = body mass index; WC = waist circumference; SF = skin fold measurement; F = female; M = male Literature Cited Ball S, Swan P, Altena T. 2006. Skinfold assessment: accuracy and application. Measurement in Physical Education and Exercise Science 10(4): 255 -264. Methods & Experimental Strategy The methodology of this project consisted of 81 human subjects, both men and women. The 4 tests (BMI, waist circumference, skin fold, and bioelectrical impedance) were performed on each participant. The bioelectrical impedance was done first using an Omron HBF-510 body scale. Figure 1 shows a participant using this scale by placing his hands and feet on the appropriate electrodes and raising his arms to a 90 degree angle. The second measurement done was BMI which was automatically calculated by the scale for each participant. The third test was the skinfold caliper which was done using a Lange Skin fold caliper on 4 different test sites; biceps, triceps, subscapular, and suprailiac. Figure 2 shows a subscapular skinfold measurement. The last measurement performed was the waist circumference measurement. This was done using simple, house-hold measuring tape to measure the waist circumference as seen in figure 3. After collecting all of the raw data, participants were classified as normal or overweight using standards provided by the World Health Organization for each measurement. Next, the number of true negatives, true positives, false negatives, and false positives were found by comparing the classifications of the different measurements to the control measurement of bioelectrical impedance. The true positive rate and false positive rates were calculated to construct a ROC sensitivity and specificity graph which could graphically depict the most accurate measurements. Figure 1 Figure 2. Figure 3. Centers for Disease Control and Prevention. 2011. Behavioral Risk Factor Surveillance System. Retrieved from http: //www. cdc. gov/brfss/ Esmaillzadeh A, Mirmiran P, Azizi F. 2004. Waist-to-hip ratio is a better screening measure for cardiovascular risk factors than other anthropometric indicators in Tehranian adult men. International Journal of Obesity 28: 1325 -1332. Hong KH, Lim YG, Park KS. 2009. Effectiveness of thigh-to-thigh current path for the measurement of abdominal fat in bioelectrical impedance analysis. Medical & Biological Engineering & Computing 47: 1265 -1271. Neovius M, Linne Y, Rossner S. 2005. BMI, waist-circumference and waist-hip-ratio as diagnostic tests for fatness in adolescents. International Journal of Obesity 29: 163 -169. Figure 4. With these TPR’s and FPR’s the ROC sensitivity and specificity graph in Figure 4 was constructed to determine the most accurate measurement compared to bioelectrical impedance Interpretations These tables and graphs show that waist circumference in males was the least accurate. This measurement has 16 FN’s meaning that 16 males were misclassified as normal. This produced a very low TPR at 0. 11 which is much less then a perfect TPR which would be 1. Waist circumference in females was the 2 nd worst measurement, however, it was much better then it was for males. This measurement only had 7 FN or only 7 females were misclassified as normal. It also had a TPR much closer to 1 at 0. 63. BMI in males was the best measurement overall with BMI in females being a close second. BMI for males only had 4 FN’s and females had 5. This means only a few participants were misclassified as normal. The TPR’s were very close to 1 for both sexes with males at 0. 80 and females at 0. 75. Males had a perfect FPR at 0 and females were very close to 0 at 0. 05. Skinfold was next best to BMI and was pretty equal between the sexes. The females had more FP’s or more participants that were misclassified as overweight. However the males had more FN’s or more participants were misclassified as normal. This makes the TPR closer to 1 for females at 0. 95 and 0. 65 for males. Males had a close FPR to 0 at 0. 1 and females at 0. 3. Romero-Corral A, Somers VK, Sierra-Johnson J, Thomas RJ, Collazo-Clavell ML, Korinek J, Allison TG, Batsis JA, Sert-Kuniyoshi FH, Lopez-Jimenez F. 2008. Accuracy of body mass index in diagnosing obesity in the adult general population. International Journal of Obesity 32: 959 -966 Rothman KJ. 2008. BMI-related errors in the measurement of obesity. International Journal of Obesity (32)1: S 56 -S 59 Acknowledgements I’d like to thank Dr. Hancock, my capstone supervisor, for helping me throughout this project. I would also like to thank Dr. Brown, my capstone instructor, for looking over my Power. Point's. Finally, I’d like to thank my capstone class of 2011, my family, and my friends for providing feedback on my project.