Using neural network as a screening and educational tool for abnormal glucose tolerance in the community

Archives of Public Health

L Shi¹  S Nie¹  F Dong²  W Gao³ 

[1]King's College London, School of Medicine, Department of Palliative Care, Policy and Rehabilitation, Cicely Saunders Institute, London, UK

[2]General Hospital of Pingdingshan Coal Mining District, Department of Internal Medicine, Pingdingshan, China

[3]Huazhong University of Science & Technology, Tongji Medical College, Department of Epidemiology and Health Statistics, Wuhan, China

关键词: neural networks;    risk factors;    Abnormal Glucose Tolerance;    diabetes mellitus;    Screening;   

Others  :  1083896
DOI  :  10.1186/0778-7367-68-4-143

【 摘 要 】

Background

Accurate, simple and non-invasive tools are needed for efficient screening of abnormal glu-cose tolerance (AGT) and educating the general public.

Aim

To develop a neural network-based initial screening and educational model for AGT.

Data and methods

230 subjects with AGT and 3,243 subjects with normal glucose tolerance (NGT) were allocated into training, validation and test sets using stratified randomization. The ratios of AGT versus NGT in three groups were 150:50, 30:570 and 50:950, respectively. A feed-forward neural network (FFNN) was trained to predict 2-hour plasma glucose of 75 g Oral Glucose Tolerance Test (OGTT) using age, family history of diabetes, weight, height, waist and hip circumference. The screening performance was evaluated by the area under the receiver operating characteristic (ROC) curve (AUC) and the partial AUC (in the range of false positive rates between 35 and 65%) and compared to those from logistic regression, linear regression and ADA Risk Test.

Results

Sensitivity, specificity, accuracy and percentage that needed further testing at 7.2 mmol/L in test group were 90.0%(95%CI: 78.6 to 95.7%), 47.7% (95%CI: 44.5 to 50.9%), 49.8% (95%CI: 46.7 to 52.9%) and 54.2% (95%CI: 51.1 to 57.3%) respectively. The entire and partial AUCs were 0.70 (95%CI: 0.62 to 0.78) and 0.26 (95%CI: 0.22 to 0.30). The partial AUC of the NN was higher than those of logistic regression (p = 0.06), linear regression (p = 0.06) and ADA Risk Test (P = 0.006).

Conclusion

NN can be used as a high-sensitive and non-invasive initial screening and educational tool for AGT.

【 授权许可】

   
2011 Gao et al.

【 图 表 】

Figure 1.

【 参考文献 】

• [1]Aroda VR, Ratner R: Approach to the patient with prediabetes. J Clin Endocrinol Metab 2008, 93(9):3259-65.
• [2]Millett C, Car J, Eldred D, Khunti K, Mainous AG, Majeed A: Diabetes prevalence, process of care and outcomes in relation to practice size, caseload and deprivation: national cross-sectional study in primary care. J R Soc Med 2007, 100(6):275-83.
• [3]Waugh N, Scotland G, McNamee P, Gillett M, Brennan A, Goyder E, et al.: Screening for type 2 diabetes: literature review and economic modelling. Health Technol Assess 2007, 11(17):iii-xi. 1
• [4]Saudek CD, Herman WH, Sacks DB, Bergenstal RM, Edelman D, Davidson MB: A new look at screening and diagnosing diabetes mellitus. J Clin Endocrinol Metab 2008, 93(7):2447-53.
• [5]Hoerger TJ, Harris R, Hicks KA, Donahue K, Sorensen S, Engelgau M: Screening for type 2 diabetes mellitus: a cost-effectiveness analysis. Ann Intern Med 2004, 140(9):689-99.
• [6]Icks A, Rathmann W, Haastert B, Gandjour A, Holle R, John J, et al.: Clinical and cost-effectiveness of primary prevention of Type 2 diabetes in a 'real world' routine healthcare setting: model based on the KORA Survey 2000. Diabet Med 2007, 24(5):473-80.
• [7]Griffin SJ, Little PS, Hales CN, Kinmonth AL, Wareham NJ: Diabetes risk score: towards earlier detection of type 2 diabetes in general practice. Diabetes Metab Res Rev 2000, 16(3):164-71.
• [8]Simoens S, Foulon E, Dethier M, Mathieu C, Laekeman G: Promoting targeted screening for Type 2 diabetes mellitus: the contribution of community pharmacists. Diabet Med 2005, 22(6):812-3.
• [9]Schneider H, Shaw J, Zimmet P: Guidelines for the detection of diabetes mellitus - diagnostic criteria and rationale for screening. Clin Biochem Rev 2003, 24(3):77-80.
• [10]Krogh A: What are artificial neural networks? Nat Biotechnol 2008, 26(2):195-7.
• [11]Ji W, Naguib RN, Ghoneim MA: Neural network-based assessment of prognostic markers and outcome prediction in bilharziasis-associated bladder cancer. IEEE Trans Inf Technol Biomed 2003, 7(3):218-24.
• [12]Linder R, Albers AE, Hess M, Poppl SJ, Schonweiler R: Artificial neural network-based classification to screen for dysphonia using psychoacoustic scaling of acoustic voice features. J Voice 2008, 22(2):155-63.
• [13]Modai I, Kurs R, Ritsner M, Oklander S, Silver H, Segal A, et al.: Neural network identifi-cation of high-risk suicide patients. Med Inform Internet Med 2002, 27(1):39-47.
• [14]Mofidi R, Duff MD, Madhavan KK, Garden OJ, Parks RW: Identification of severe acute pancreatitis using an artificial neural network. Surgery 2007, 141(1):59-66.
• [15]Nieminen P, Hakama M, Viikki M, Tarkkanen J, Anttila A: Prospective and randomised public-health trial on neural network-assisted screening for cervical cancer in Finland: results of the first year. Int J Cancer 2003, 103(3):422-6.
• [16]Duh MS, Walker AM, Ayanian JZ: Epidemiologic interpretation of artificial neural networks. Am J Epidemiol 1998, 147(12):1112-22.
• [17]Colagiuri S, Hussain Z, Zimmet P, Cameron A, Shaw J: Screening for type 2 diabetes and impaired glucose metabolism: the Australian experience. Diabetes Care 2004, 27(2):367-71.
• [18]Lindahl B, Weinehall L, Asplund K, Hallmans G: Screening for impaired glucose tolerance. Results from a population-based study in 21,057 individuals. Diabetes Care 1999, 22(12):1988-92.
• [19]Qiao Q, Hu G, Tuomilehto J, Nakagami T, Balkau B, Borch-Johnsen K, et al.: Age- and sex-specific prevalence of diabetes and impaired glucose regulation in 11 Asian cohorts. Diabetes Care 2003, 26(6):1770-80.
• [20]Schmidt MI, Duncan BB, Vigo A, Pankow J, Ballantyne CM, Couper D, et al.: Detection of undiagnosed diabetes and other hyperglycemia states: the Atherosclerosis Risk in Communities Study. Diabetes Care 2003, 26(5):1338-43.
• [21]The DECODE Study Group: Age- and sex-specific prevalences of diabetes and impaired glucose regulation in 13 European cohorts. Diabetes Care 2003, 26(1):61-9.
• [22]Wang KA, Li TL, Bai J, Feng JG, Fu ZY: [Study on the epidemiological characteristics of diabetes mellitus and IGT in China]. Zhonghua Liu Xing. Bing Xue Za Zhi 1998, 19:282-5.
• [23]Chester DL: Why two hidden layers are better than one. 1990, 265-8.
• [24]Sontag ED: Feedback stabilization using two-hidden layer nets. IEEE Trans Neural Netw 1992, 3:981-90.
• [25]Xiang HD, Wu W, Liu CQ, Li K, Feng JG, Zhang YT: [An epidemiological study on DM 1995-1996, in China]. Zhongguo Tang Niao Bing Za Zhi 1998, 6:131-3.
• [26]Hagan MT, Menhaj MB: Training feedforward networks with the Marquardt algorithm. IEEE Trans Neural Netw 1994, 5(6):989-93.
• [27]Tu JV: Advantages and disadvantages of using artificial neural networks versus logistic regression for predicting medical outcomes. J Clin Epidemiol 1996, 49(11):1225-31.
• [28]Metz CE: ROC methodology in radiologic imaging. Invest Radiol 1986, 21(9):720-33.
• [29]Walter SD: The partial area under the summary ROC curve. Stat Med 2005, 24(13):2025-40.
• [30]Zhang DD, Zhou XH, Freeman DH Jr, Freeman JL: A non-parametric method for the comparison of partial areas under ROC curves and its application to large health care data sets. Stat Med 2002, 21(5):701-15.
• [31]Risk Test - Text Version http://www diabetes org/risk-test/text-version jsp webcite [ 2008 [cited 2008 Nov. 12]; Overweight & Obesity (BMI). http://apps who int/bmi/index jsp?introPage=intro_3 html webcite [ 2008 [cited 2008 Nov. 12]
• [32]Herdzik E, Safranow K, Ciechanowski K: Diagnostic value of fasting capillary glucose, fructosamine and glycosylated haemoglobin in detecting diabetes and other glucose tolerance abnormalities compared to oral glucose tolerance test. Acta Diabetol 2002, 39(1):15-22.
• [33]Glumer C, Jorgensen T, Borch-Johnsen K: Targeted screening for undiagnosed diabetes reduces the number of diagnostic tests. Inter99(8). Diabet Med 2004, 21(8):874-80.
• [34]Mannucci E, Ognibene A, Sposato I, Brogi M, Gallori G, Bardini G, et al.: Fasting plasma glucose and glycated haemoglobin in the screening of diabetes and impaired glucose tolerance. Acta Diabetol 2003, 40(4):181-6.
• [35]Kahn LB, Marshall JA, Baxter J, Shetterly SM, Hamman RF: Accuracy of reported family history of diabetes mellitus. Results from San Luis Valley Diabetes Study. Diabetes Care 1990, 13(7):796-8.
• [36]Greenland S: Modeling and variable selection in epidemiologic analysis. Am J Public Health 1989, 79(3):340-9.