青少年非特异性下背痛发病风险预测模型的构建:基于logistic回归和BP神经网络模型

Development of a Prediction Model for Non-specific Low Back Pain in Adolescents: Based on Logistic Regression and BP Neural Network Models

摘要

目的:探索Logistic回归和BP神经网络(BPNN)模型在青少年非特异性下背痛(NLBP)发病风险中的预测效果。方法:于2022年9-11月,采用多阶段整群随机抽样方法选取上海市区1800名在校学生进行问卷调查。采用多因素Logistic回归分析青少年NLBP影响因素并构建Logistic回归模型,采用MATLAB R2023b软件构建BPNN。通过准确率、灵敏度、特异度、约登指数及ROC曲线下面积(AUC)评价各模型的预测效果。结果:共有1630名青少年纳入研究,NLBP者406名(流行率为24.91%)。经单因素及多因素分析确定9个变量纳入青少年NLBP发病风险预测模型。训练集显示,Logistic回归模型、BPNN的准确率分别为78.79%、83.26%,灵敏度分别为78.01%、85.18%,特异度分别为79.06%、77.66%,约登指数分别为0.571、0.628,AUC分别为0.761、0.867。测试集显示,Logistic回归模型、BPNN的准确率分别为77.10%、81.80%,灵敏度分别为75.65%、84.22%,特异度分别为77.54%、73.91%,约登指数分别为0.532、0.581,AUC分别为0.758、0.838。结论:与Logistic回归模型相比,BPNN模型的预测效果更好,可以有效预测上海市青少年非特异性下背痛发病风险。

Abstract

Objective: To explore the predictive effects of Logistic regression and Back Propagation Neural Network (BPNN) models on the risk of non-specific low back pain (NLBP) in adolescents. Methods: From September to November 2022, a multi-stage cluster random sampling method was used to select 1,800 in-school students in urban areas of Shanghai for a questionnaire survey. Multivariate Logistic regression was applied to analyze the influencing factors of NLBP in adolescents and construct a Logistic regression model, while MATLAB R2023b software was used to build the BPNN. The predictive effects of each model were evaluated using accuracy, sensitivity, specificity, Youden index, and the area under the receiver operating characteristic curve (AUC). Results: A total of 1,630 adolescents were included in the study, among whom 406 had NLBP (with a prevalence rate of 24.91%). Through univariate and multivariate analyses, 9 variables were identified to be included in the risk prediction model for NLBP in adolescents. In the training set, the accuracy rates of the Logistic regression model and BPNN were 78.79% and 83.26%; the sensitivities were 78.01% and 85.18%; the specificities were 79.06% and 77.66%; the Youden indices were 0.571 and 0.628; and the AUC values were 0.761 and 0.867, respectively. In the test set, the accuracy rates of the Logistic regression and BPNN models were 77.10% and 81.80%; the sensitivities were 75.65% and 84.22%; the specificities were 77.54% and 73.91%; the Youden indices were 0.532 and 0.581; and the AUC values were 0.758 and 0.838, respectively. Conclusion: Compared with the Logistic regression model, the BPNN model has a better predictive effect and can effectively predict the risk of NLBP in adolescents in Shanghai.

关键词

下背痛 / 青少年体质 / Logistic回归分析 / BP神经网络模型

Key words

low back pain / adolescent physical fitness / logistic regression analysis / back propagation neural network model

参考文献

[1] HWANG J, LOUIE PK, PHILLIPS FM, et al. Low Back Pain in Children: A Rising Concern[J]. Eur Spine J, 2019, 28(2): 211-213.

[2] HARTVIGSEN J, HANCOCK MJ, KONGSTED A, et al. What Low Lack Lain is and Why We Need to Pay Attention[J]. Lancet, 2018, 391(10137): 2356-2367.

[3] ROY R, GALÁN S, SÁNCHEZ-RODRÍGUEZ E, et al. Cross-National Trends of Chronic Back Pain in Adolescents: Results From the HBSC Study, 2001-2014[J]. J Pain, 2022, 23(1): 123-130.

[4] WU A, DONG W, LIU S, et al. The Prevalence and Years Lived with Disability Caused by Low Back Pain in China, 1990 To 2016: Findings from The Global Burden of Disease Study 2016[J]. Pain, 2019, 160(1): 237-245.

[5] JUNGE T, WEDDERKOPP N, BOYLE E, et al. The Natural Course of Low Back Pain from Childhood to Young Adulthood-A Systematic Review[J]. Chiropr Man Therap, 2019, 20; 27: 10.

[6] CIEZA A, CAUSEY K, KAMENOV K, et al. Global Estimates of The Need for Rehabilitation Based on The Global Burden of Disease Study 2019: A Systematic Analysis for The Global Burden of Disease Study 2019[J]. Lancet, 2021, 396(10267): 2006-2017.

[7] 何秉涛, 黄美欢, 贠国俊, 等 . 2022年德国儿童和青少年医学会《儿童和青少年腰背痛病因、危险因素和诊断以及非特异性腰背痛治疗和预防的循证指南》解读[J]. 中华实用儿科临床杂志, 2023, 38(5): 321-326.

[8] WALL J, MEEHAN WP 3RD, TROMPETER K, et al. Incidence, Prevalence and Risk Factors for Low Back Pain in Adolescent Athletes: A Systematic Review and Meta-Analysis[J]. Br J Sports Med, 2022, 56(22): 1299-1306.

[9] VITTA A, BENTO TPF, CORNELIO GP, et al. Incidence and Factors Associated with Low Back Pain in Adolescents: A Prospective Study[J]. Braz J Phys Ther, 2021, 25(6): 864-873.

[10] ANDAUR NAVARRO CL, DAMEN JAA, VAN SMEDEN M, et al. Systematic Review Identifies The Design and Methodological Conduct of Studies on Machine Learning-Based Prediction Models[J]. J Clin Epidemiol, 2023, 154: 8-22.

[11] SHEHAB M, ABUALIGAH L, SHAMBOUR Q, et al. Machine Learning in Medical Applications: A Review of State-of-the-Art Methods[J]. Comput Biol Med, 2022, 145: 105458.

[12] ELOQAYLI H, AL-YOUSEF A, JARADAT R. Vitamin D and Ferritin Correlation with Chronic Neck Pain Using Standard Statistics and A Novel Artificial Neural Network Prediction Model[J]. Br J Neurosurg, 2018, 32(2): 172-176.

[13] 王新亭, 王琪, 徐聃弟, 等 . 基于BP神经网络的高足弓异常程度评价模型的构建[J]. 中国生物医学工程学报, 2021, 40(2): 252-256.

[14] 姚卫光. 在校儿童和青少年非特异性下腰痛的危险因素研究[D]. 广州: 南方医科大学, 2011.

[15] 陈秋, 于伟平, 陈瑞美, 等 . 学龄前儿童生活方式对情绪与行为问题影响的研究[J]. 现代预防医学, 2021, 48(1): 82-85.

[16] LIU Q, ZHOU Y, XIE X, et al. The Prevalence of Behavioral Problems Among School-Aged Children in Home Quarantine During The COVID-19 Pandemic in China[J]. J Affect Disord, 2021, 15: 412-416.

[17] 吴正慧, 赵占锋, 谭咏梅, 等 . 小学生校园欺凌及其与问题行为和生活满意度的关系[J]. 中国心理卫生杂志, 2023, 37(3): 213-218.

[18] CORP N, MANSELL G, STYNES S, et al. Evidence-Based Treatment Recommendations for Neck and Low Back Pain Across Europe: A Systematic Review of Guidelines[J]. Eur J Pain, 2021, 25(2): 275-295.

[19] 余敏, 朱敏, 何海燕, 等 . 母亲情绪症状及童年期不良经历与儿童情绪行为问题的关联[J]. 中国学校卫生, 2023, 44(3): 411-414+ 418.

[20] 中国肥胖问题工作组, 季成叶. 中国学龄儿童青少年超重、肥胖筛查体重指数值分类标准[J]. 中华流行病学杂志, 2004(2): 10-15.

[21] 曹镓玺, 王鑫, 雷光春. 基于遗传算法优化BP神经网络的青藏高原海北高寒湿地CO₂通量模拟及其影响因子 [J]. 山东大学学报(理学版), 2021, 56(5): 33-41+50.

[22] ALQAISSI EY, ALOTAIBI FS, RAMZAN MS. Modern Machine-Learning Predictive Models for Diagnosing Infectious Diseases[J]. Comput Math Methods Med, 2022: 6902321.

[23] 曹瑞晴, 杨继春, 于淼. 佛山市高中生腰痛流行病学调查及影响因素分析[J]. 中国校医, 2020, 34(8): 576-579+ 591.

[24] YANG G, CAO J, LI Y, et al. Association Between Internet Addiction and The Risk of Musculoskeletal Pain in Chinese College Freshmen - A Cross-Sectional Study[J]. Front Psychol, 2019, 10: 1959.

[25] LEMES ÍR, OLIVEIRA CB, SILVA GCR, et al. Association of Sedentary Behavior and Early Engagement in Physical Activity with Low Back Pain in Adolescents: A Cross-Sectional Epidemiological Study[J]. Eur Spine J, 2022, 31(1): 152-158.

[26] BROOKS SJ, FELDMAN I, SCHIÖTH HB, et al. Important Gender Differences in Psychosomatic and School-Related Complaints Iin Relation to Adolescent Weight Status[J]. Sci Rep, 2021, 11(1): 14147.

[27] SANTOS ES, BERNARDES JM, NOLL M, et al. Prevalence of Low Back Pain and Associated Risks in School-Age Children[J]. Pain Manag Nurs, 2021, 22(4): 459-464.

[28] AZEVEDO N, RIBEIRO JC, MACHADO L. Back Pain in Children and Adolescents: A Cross-Sectional Study[J]. Eur Spine J, 2023, 32(9): 80-89.

[29] KĘDRA A, PLANDOWSKA M, KĘDRA P, et al. Non-Specific Low Back Pain: Cross-Sectional Study of 11,423 Children and Youth and The Association with the Perception of Heaviness in Carrying of Schoolbags[J]. PeerJ, 2021, 9: e11220.

[30] ALHOWIMEL AS, ALFAIFI RM, ALLUHA YBI AA, et al. Prevalence of Low Back Pain and Associated Risk Factors Among Saudi Arabian Adolescents: A Cross-Sectional Study[J]. Int J Environ Res Public Health, 2022, 19(18): 11217.

[31] NOLL M, CANDOTTI CT, DA ROSA BN, et al. Back Pain and its Risk Factors in Brazilian Adolescents: A Longitudinal Study[J]. Br J Pain, 2021, 15(1): 16-25.

[32] DARIO AB, KAMPER SJ, O'KEEFFE M, et al. Family History of Pain and Risk of Musculoskeletal Pain in Children and Adolescents: A Systematic Review and Meta-Analysis[J]. Pain, 2019, 160(11): 2430-2439.

[33] 成湘, 王亮. 体卫融合视域下中小学学校体育的发展特征、挑战及策略[J]. 广州体育学院学报, 2023, 43(3): 102-108.

[34] MASIERO S, SARTO F, CATTELAN M, et al. Lifetime Prevalence of Nonspecific Low Back Pain in Adolescents: A Cross-Sectional Epidemiologic Survey[J]. Am J Phys Med Rehabil, 2021, 100(12): 1170-1175.

[35] SHIRI R, FALAH-HASSANI K. Does Leisure Time Physical Activity Protect Against Low Back Pain? Systematic Review and Meta-Analysis of 36 Prospective Cohort Studies[J]. Br J Sports Med, 2017, 51(19): 1410-1418.

[36] LI FD, KONG QJ, WANG YX, et al. Predicting the Risk of Non-Specific Low Back Pain in The Young Population: Development and Assessment of A New Predictive Nomogram[J]. Eur Rev Med Pharmacol Sci, 2022, 26(23): 8795-8807.

[37] LINDELL M, GRIMBY-EKMAN A. Stress, Non-Restorative Sleep, And Physical Inactivity as Risk Factors for Chronic Pain in Young Adults: A Cohort Study[J]. PLoS One, 2022, 17(1): e0262601.

[38] ROMAN-JUAN J, ROY R, JENSEN MP, et al. The Explanatory Role of Sedentary Screen Time and Obesity in the Increase of Chronic Back Pain Amongst European Adolescents: The HBSC Study 2002-2014[J]. Eur J Pain, 2022, 26(8): 1781-1789.

[39] FOSTER NE, ANEMA JR, CHERKIN D, et al. Prevention and Treatment of Low Back Pain: Evidence, Challenges, and Promising Directions[J]. Lancet, 2018, 391(10137): 2368-2383.

[40] VAN LOOVEREN E, BILTERYS T, MUNNEKE W, et al. The Association Between Sleep and Chronic Spinal Pain: A Systematic Review from The Last Decade[J]. J Clin Med, 2021, 10(17): 3836.

[41] YUE C, WENYAO G, XUDONG Y, et al. Dose-Response Relationship Between Daily Screen Time and The Risk of Low Back Pain Among Children and Adolescents: A Meta-Analysis of 57831 Participants[J]. Environ Health Prev Med, 2023(28): 64.

[42] IN TS, JUNG JH, JUNG KS, et al. Spinal and Pelvic Alignment of Sitting Posture Associated with Smartphone Use in Adolescents with Low Back Pain[J]. Int J Environ Res Public Health, 2021, 18(16): 8369.

[43] WONG JJ, TRICCO AC, CÔTE P, et al. Association Between Depressive Symptoms or Depression and Health Outcomes for Low Back Pain: A Systematic Review and Meta-analysis[J]. J Gen Intern Med, 2022, 37(5): 1233-1246.

[44] 孙军, 罗秋兰, 罗小兵. “双减”背景下我国中小学体育高质量发展: 要点、难点与路向[J]. 广州体育学院学报, 2023, 43(2): 44-51.

[45] YANG H, HURWITZ EL, LI J, et al. Bidirectional Comorbid Associations between Back Pain and Major Depression in US Adults[J]. Int J Environ Res Public Health, 2023, 20(5): 4217.

[46] ANDREUCCI A, CAMPBELL P, RICHARDSON E, et al. Sleep Problems and Psychological Symptoms as Predictors of Musculoskeletal Conditions in Children and Adolescents[J]. Eur J Pain, 2020, 24(2): 354-363.

[47] 徐皓铭, 宋君毅, 李国兴, 等 . 青少年学生锻炼行为的心理适应: 基于扎根理论的探索性研究[J]. 广州体育学院学报, 2022, 42(3): 77-85.

[48] BENTO TPF, CORNELIO GP, PERRUCI NI PO, et al. Low Back Pain in Adolescents and Association with Sociodemographic Factors, Electronic Devices, Physical Activity and Mental Health[J]. J Pediatr(Rio J), 2020, 96(6): 717-724.

[49] STAPP EK, PAKSARIAN D, HE JP, et al. Mood and Anxiety Profiles Differentially Associate with Physical Conditions in US Adolescents[J]. J Affect Disord, 2022, 299: 22-30.

[50] THOMAS PA, GOODIN BR, MEINTS SM, et al. Adverse Childhood Experiences and Chronic Low Back Pain in Adulthood: The Role of Emotion Regulation. J Pain, 2024, 25(9): 104551.

[51] 王凯, 闵少雄, 徐新毅, 等 . 社区非特异性下腰痛流行病学调查及预测模型的构建[J]. 实用医学杂志, 2015, 31(2): 306-308.

[52] HEIJNEN B J, BOHRINGER S, SPEYER R. Prediction of Aspiration in Dysphagia Using Logistic Regression: Oral Intake and Self-Evaluation[J]. Eur Arch Otorhinolaryngol, 2020, 277(1): 197-205.

[53] 闫瑞平, 王习亮, 姚粉霞, 等 . 决策树模型与Logistic回归分析模型识别高血压危险因素的效果比较[J]. 中华疾病控制杂志, 2022, 26(2): 218-222.

[54] WANG X Y, JIANG Z Z, JI R D, et al. Detection of Ningnanmycin Using Fluorescence Spectroscopy Combined with BP Neural Network[J]. Comb Chem High Throughput Screen, 2023, 26(7): 1414-1423.

[55] OWARI Y, MIYATAKE N. Prediction Of Chronic Low Back Pain Using the Hierarchical Neural Network: Comparison with Logistic Regression - A Pilot Study[J]. Medicina(Kaunas), 2019, 55(6): 259.

[56] SHIM JG, RYU KH, CHO EA, et al. Machine Learning Approaches to Predict Chronic Low Back Pain in People Aged over 50 Years[J]. Medicina(Kaunas), 2021, 57(11): 1230.

[57] 朱宵彤, 庞春颖, 朱涵. 基于深度学习的心血管疾病预测模型[J]. 计算机应用, 2021, 41(S2): 346-350.

[58] 陈珂锐, 孟小峰. 机器学习的可解释性[J]. 计算机研究与发展, 2020, 57(9): 1971-1986.

[59] LYU J, SHI H, ZHANG J, et al. Prediction Model for Suicide Based on Back Propagation Neural Network and Multilayer Perceptron[J]. Front Neuroinform, 2022(16): 961588.

[60] 曾慧娟, 田波, 袁红伶, 等 . 机器学习算法构建慢性肾脏病伴高血压或糖尿病的预测模型[J]. 昆明医科大学学报, 2024, 45(3): 99-105.

基金资助

上海市哲学社会科学规划项目(2022BTY003)

教育部人文社会科学规划基金(24YJAZH207)

上海市教育科学研究项目(C2021037)