Volume 21, Issue 74 (1-2019)
jha 2019, 21(74): 87-100 |
Back to browse issues page
Download citation:
BibTeX | RIS | EndNote | Medlars | ProCite | Reference Manager | RefWorks
Send citation to:
BibTeX | RIS | EndNote | Medlars | ProCite | Reference Manager | RefWorks
Send citation to:
Langarizade M, Owji L, Orooji A. Developing a decision support system for osteoporosis Prediction. jha 2019; 21 (74) :87-100
URL: http://jha.iums.ac.ir/article-1-2758-en.html
URL: http://jha.iums.ac.ir/article-1-2758-en.html
1- , School of Health Management and Information Sciences, Iran University of Medical Sciences, Tehran, Iran
2- School of Health Management and Information Sciences, Iran University of Medical Sciences, Tehran, Iran ,leila.owji@gmail.com
2- School of Health Management and Information Sciences, Iran University of Medical Sciences, Tehran, Iran ,
Abstract: (4894 Views)
Introduction: Osteoporosis is a common disease in women. Osteoporosis fractures may cause irreparable damages; therefore, early diagnosis and treatment before fractures is an important issue. The ojectiveof this study was to develop a decision support system for diagnosing osteoporosis using artificial neural networks.
Method: This developmental study has been done in second half of 2017 based on crossectional method. In present study initially Osteoporosis affecting clinical factors were identified and the most significant clinical factors were selected through incorporation of a questionnaire-based survey. Subsequently, information of 256 female participants and their BMD test results, five years after initial data entry were used to train neural network. The information was obtained fromwomen who refered to department of Bone Mineral Densityof oushehr university of medical sciences. In order to identify the best network, trial and error method was used and neural networks with different topologies were trained using Scaled Conjugate Gradient and Levenberg-Marquardt algorithms. Confusion matrix was used to evaluate the network’s accuracy, sensitivity and specificity.
Results: In the first stage, out of 15 essential variables, 12 variables were selected as the most important risk factors. Multilayer perceptron neural network was designed. Results showed that the best structure of network was due to Scaled Conjugate Gradient algorithm with 10 neurons and Levenberg-Marquardt algorithm with 12 neurons in hidden layer. Accuracy comparison was showed that generally Levenberg-Marquardt algorithm had better result. The best sensitivity, specificity, and accuracy was 83.1%, 89.4%, and 86.3% respectively.
Conclusion: In this study developed a diagnostic tool based on local data that could be effective in tracking osteoporosis. Utilizing such a diagnostic tool as a timely referral of individuals and initiating therapy as soon as possible may prevent fractures from occurring and help avoiding the frequent complications of osteoporosis.
Method: This developmental study has been done in second half of 2017 based on crossectional method. In present study initially Osteoporosis affecting clinical factors were identified and the most significant clinical factors were selected through incorporation of a questionnaire-based survey. Subsequently, information of 256 female participants and their BMD test results, five years after initial data entry were used to train neural network. The information was obtained fromwomen who refered to department of Bone Mineral Densityof oushehr university of medical sciences. In order to identify the best network, trial and error method was used and neural networks with different topologies were trained using Scaled Conjugate Gradient and Levenberg-Marquardt algorithms. Confusion matrix was used to evaluate the network’s accuracy, sensitivity and specificity.
Results: In the first stage, out of 15 essential variables, 12 variables were selected as the most important risk factors. Multilayer perceptron neural network was designed. Results showed that the best structure of network was due to Scaled Conjugate Gradient algorithm with 10 neurons and Levenberg-Marquardt algorithm with 12 neurons in hidden layer. Accuracy comparison was showed that generally Levenberg-Marquardt algorithm had better result. The best sensitivity, specificity, and accuracy was 83.1%, 89.4%, and 86.3% respectively.
Conclusion: In this study developed a diagnostic tool based on local data that could be effective in tracking osteoporosis. Utilizing such a diagnostic tool as a timely referral of individuals and initiating therapy as soon as possible may prevent fractures from occurring and help avoiding the frequent complications of osteoporosis.
Type of Study: Research |
Received: 2018/07/10 | Accepted: 2019/01/23 | Published: 2019/01/23
Received: 2018/07/10 | Accepted: 2019/01/23 | Published: 2019/01/23
References
1. Becker KL. Principles and Practice of Endocrinology and Metabolism: Lippincott Williams and Wilkins; 2001.
2. Matin N, Tabatabaie O, Keshtkar A, Yazdani K, Asadi M. Development and validation of osteoporosis prescreening model for Iranian postmenopausal women. JDMDC. 2015; 14(1): 12-21. [DOI:10.1186/s40200-015-0140-7]
3. Curtis EM, Moon RJ, Harvey NC, Cooper C. The impact of fragility fracture and approaches to osteoporosis risk assessment worldwide. Bone. 2017;104:29-38. [DOI:10.1016/j.bone.2017.01.024]
4. Golob AL, Laya MB. Osteoporosis: Screening, Prevention, and Management. Med Clin North Am. 2015;99(3):587-606. [DOI:10.1016/j.mcna.2015.01.010]
5. Yu X, Ye C, Xiang L. Application of artificial neural network in the diagnostic system of osteoporosis. Neurocomputing. 2016;214:376-81. [DOI:10.1016/j.neucom.2016.06.023]
6. Iliou T, Anagnostopoulos C-N, Anastassopoulos G. Osteoporosis detection using machine learning techniques and feature selection. Int J Artif Intell T. 2014;23(05):145-59. [DOI:10.1142/S0218213014500146]
7. Keshtkar A, Khashayar P, Mohammadi Z, Etemad K, Dini M, Meybodi HA, et al. A suggested prototype for assessing bone health. Archives of Iranian Medicine. 2015;18(7):411-5.
8. Pazokian M, Yaghmaie F. Development and psychometric evaluation of "Assessment form of Osteoporosis" Urmia Medical Journal. 2013;24(3):176-83. [In Persian]
9. Saxon SV, Mary Jean Etten EDGNPFT, Dr. Elizabeth A. Perkins PDR. Physical Change and Aging, Sixth Edition: A Guide for the Helping Professions: Springer Publishing Company; 2014.
10. McCloskey EV, Johansson H, Oden A, Kanis JA. From relative risk to absolute fracture risk calculation: the FRAX algorithm. Curr Osteoporos Rep. 2009;7(3):77-83. [DOI:10.1007/s11914-009-0013-4]
11. Chang H-W, Chiu Y-H, Kao H-Y, Yang C-H, Ho W-H. Comparison of classification algorithms with wrapper-based feature selection for predicting osteoporosis outcome based on genetic factors in a taiwanese women population. Int J Endocrinol. 2013;2013:1-8.
12. Unni S, Yao Y, Milne N, Gunning K, Curtis J, LaFleur J. An evaluation of clinical risk factors for estimating fracture risk in postmenopausal osteoporosis using an electronic medical record database. Int J Osteoporos. 2015; 26(2): 581-7. [DOI:10.1007/s00198-014-2899-7]
13. Fracture Risk Assessment Tool [Internet]. 2016. [cited: 15 Dec 2017]. Available from: https://www.shef.ac.uk/FRAX/faq.aspx.
14. Amarnath ALD, Franks P, Robbins JA, Xing G, Fenton JJ. Underuse and overuse of osteoporosis screening in a regional health system: a retrospective cohort study. J Gen Intern Med. 2015;30(12):1733-40. [DOI:10.1007/s11606-015-3349-8]
15. Lasser E, Pfoh E, Chang H, Chan K, Bailey J, Kharrazi H, et al. Has Choosing Wisely affected rates of dual-energy X-ray absorptiometry use? Osteoporosis Int. 2016;27(7):2311-2316. [DOI:10.1007/s00198-016-3511-0]
16. Pinheiro M, Neto ER, Machado F, Omura F, Szejnfeld J, Szejnfeld V. Development and validation of a tool for identifying women with low bone mineral density and low-impact fractures: the São Paulo Osteoporosis Risk Index (SAPORI). Osteoporos Int 2012;23(4):1371-9. [DOI:10.1007/s00198-011-1722-y]
17. Lisboa PJG, Ifeachor EC, Szczepaniak PS. Artificial Neural Networks in Biomedicine: Springer London; 2012.
18. Du Q, Nie K, Wang Z. Application of Entropy-Based Attribute Reduction and an Artificial Neural Network in Medicine: A Case Study of Estimating Medical Care Costs Associated with Myocardial Infarction. Entropy. 2014;16(9):4788-800. [DOI:10.3390/e16094788]
19. Al-Shayea QK. Artificial neural networks in medical diagnosis. International Journal of Computer Science Issues. 2011;8(2):150-4.
20. Amato F, López A, Pe-a-Méndez EM, Vaňhara P, Hampl A, Havel J. Artificial neural networks in medical diagnosis. J Appl Biomed. 2013; 11(2): 47-58. [DOI:10.2478/v10136-012-0031-x]
21. Kriesel D. A Brief Introduction to Neural Networks [Internet]. 2007. [cited: 17 Dec 2017]. Available from: http://www.dkriesel.com.
22. Birch K, George K, McLaren D. BIOS Instant Notes in Sport and Exercise Physiology: Taylor & Francis; 2004.
23. Giudici P. Applied Data Mining: Statistical Methods for Business and Industry: Wiley; 2005.
24. Sivanandam SN, Deepa SN. Introduction to Neural Networks Using Matlab 6.0: Tata McGraw-Hill; 2006.
25. Langarizadeh M, Saeedi M, Far M, Hoseinpour M. Predicting premature birth in pregnant women via assisted reproductive technologies using neural network. Journal of Health Administration (JHA). 2016;18(62): 42-51. [In Persian]
26. Sharifkhani M, Alizadeh S, Abbasi M, Ameri H. Providing a model for predicting the risk of osteoporosis using decision tree algorithms. J Mazandaran Univ Med Sci. 2014;24(116):110-18. [In Persian]
27. Ghafoori S, Keshtkar A, Khashayar P, Ebrahimi M, Ramezani M, Mohammadi Z, et al. The risk of osteoporotic fractures and its associating risk factors according to the FRAX model in the Iranian patients: a follow-up cohort. JDMDC. 2014;13(1):93-103. [DOI:10.1186/s40200-014-0093-2]
28. Harrar K, Hamami L, Akkoul S, Lespessailles E, Jennane R, editors. Osteoporosis assessment using Multilayer Perceptron neural networks. Image Processing Theory, Tools and Applications (IPTA), 2012 3rd International Conference on; 2012: IEEE. [DOI:10.1109/IPTA.2012.6469528]
29. Halldorsson B, Bjornsson A, Gudmundsson H, Birgisson E, Ludviksson B, Gudbjornsson B. A clinical decision support system for the diagnosis, fracture risks and treatment of osteoporosis. Comput Math Method M. 2014;2015:1-7.
30. Oh SM, Nam B-H, Rhee Y, Moon S-H, Kim DY, Kang DR, et al. Development and validation of osteoporosis risk-assessment model for Korean postmenopausal women. J Bone Miner Metab. 2013;31(4):423-32. [DOI:10.1007/s00774-013-0426-0]
31. Edwards M, Jameson K, Denison H, Harvey N, Sayer AA, Dennison E, et al. Clinical risk factors, bone density and fall history in the prediction of incident fracture among men and women. Bone. 2013;52(2):541-7. [DOI:10.1016/j.bone.2012.11.006]
32. Reid D. Handbook of Osteoporosis: Springer Healthcare Limited; 2011.
33. Bartl R, Frisch B. OSTEOPOROSIS: Diagnosis, Prevention, Therapy : a Practical Guide for All Physicians--from Pediatrics to Geriatrics: Springer; 2004.
34. Shafiee G, Ostovar A, Heshmat R, Darabi H, Sharifi F, Raeisi A, et al. Bushehr Elderly Health (BEH) programme: study protocol and design of musculoskeletal system and cognitive function (stage II). BMJ open. 2017;7(8):e013606. [DOI:10.1136/bmjopen-2016-013606]
35. Liu Q, Cui X, Chou Y-C, Abbod MF, Lin J, Shieh J-S. Ensemble artificial neural networks applied to predict the key risk factors of hip bone fracture for elders. Biomed Signal Process Contr. 2015;21:146-56. [DOI:10.1016/j.bspc.2015.06.002]
Send email to the article author
| Rights and permissions | |
 |
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License. |
Site Keywords
Health, Administration, Scientific Article, Iran University of Medical Sciences,
